111Un1ive1rs~ity~F~re~e Srtaute~~OO~I~~
34300004297002
Universiteit Vrystaat
~------'----=~· ft._.==. ~
2 1 SEP 20n9
WATER AND NITROGEN MANAGEMENT
FOR rusx M][TIGATION
IN SEMI-ARID CROPPING SYSTEMS
by
WALTER T. MUPANGWA
A dissertation submitted
in accordance with
the requirement for the degree of
Doctor of Philosophy in Agrometeorology/Soil Science
In the Faculty of Natural and Agricultural Sciences
Department of Soil, Crop and Climate Sciences
University of the Free State
Supervisor: Prof. Sue Walker
Co-Supervisor: Dr. S.J. Twomlow
Bloemfontein
November 2008
DECLARA TION
I declare that the dissertation hereby submitted by me for the PhD degree in
Agrometeorology/Soil Science at the Faculty of Natural and Agricultural Sciences,
University of the Free State, is my own independent work and has not been submitted by
me at another university. I cede the copyright of the dissertation in favour of the
University of the Free State.
Signature: ~ . Date: "1,}~/ó/ .
II
ACKNOWLEDGEMENTS
'For with God nothing shall be impossible - Luke J: 37'. I thank the Living God for
taking me this far. I am grateful to WaterNet for funding my study through the CGIAR
Challenge Programme on Water and Food Project (PN 17) 'Integrated Water Resources
Management for Improved Rural Livelihoods: Managing Risk, Mitigating Drought and
Improving Water Productivity in the Water Scarce Limpopo Basin'.
To Professor Sue Walker and Or. Stephen J. Twomlow who guided me in this study, may
the Living God richly bless you. You endured the heat and rough roads of Gwanda and
Insiza districts so that I can have a better future. I would also like to thank the fatherly Dr.
John P. Dimes for assisting me during the modeling exercise. To my father, Peter Claver,
thank you for riding the bicycle for all those years so that I could go to school. To my
mother, Lilian Otilia, thank you for waking up early every morning and preparing
breakfast and packed lunch during my school days. To my wife Terry and the girls,
Mufaro and Munashe, thank you for enduring the long days when I was in the field, in
the office at Matopos and in South Africa.
I owe special thank you to Sifiso'Gogo' Ncube, Ronelle and Rida for putting together all
the logistics in order, you made my life easy. I am grateful to Gertrude Mpofu,
Beckimpilo Ncube and Thulani Ndlovu for assisting me during the years of carrying out
my experiments. To Dr. Andre van Rooyen and Prof. C.C. du Preez, I say thank you for
the words of encouragement during the process of me getting this far.
III
Abstract
WATER AND NITROGEN MANAGEMENT FOR RISK MITIGATION
IN SEMI-AruD CROPPING SYSTEMS
By
WALTER T. MUPANGWA
PhD Agrometeorology, University of the Free State, Bloemfontein
November 2008
This study was conducted with three main objectives which were firstly to characterize
the smallholder farming system of semi-arid southern Zimbabwe and its rainfall pattern.
The second objective was to quantify the crop yield and soil water benefits derived from
in situ (single and double ploughing, ripping and planting basins) and inter-field (dead
level contours and infiltration pits) soil water management techniques in southern
Zimbabwe using field trials established on the farmers' fields. The on-farm study also
explored the effect of combining in situ soil water management technologies (single and
double ploughing, ripping and planting basins) and nitrogen fertilizer (0, 10 and 20
kgblha") application on crop yield under semi-arid conditions. An on-station experiment
was established to assess the effect of combining in situ soil water management
techniques (single ploughing, ripping and planting basins) and mulching (0, 0.5, 1, 2, 4, 8
and 10 tha') on maize, cowpea and sorghum yields, and soil water dynamics. The third
objective was the evaluation of tillage systems on the farmers' fields over three growing
seasons (2005/06, 2006/07 and 2007/08). The smallholder farmers appraised the tillage
systems at the end of the last growing season of the study. Simulation modeling was then
used to assess the long term effect of using the basin tillage system over a 69 year period.
IV
The daily rainfall data was collected from five meteorological stations located in the
Mzingwane catchment of the Limpopo basin. The analysis revealed that neither the total
annual rainfall, based on the July-June calendar, nor the start nor end of growing season
has changed significantly over the past 50-74 year period in southern Zimbabwe. The
analysis indicated that the length of the growing season decreases along the Bulawayo to
Beitbridge transect. The growing season starts during the first eight days of December at
all stations except at Filabusi where the season starts during the last week of November.
The number of wet days per growing season has also not changed along the Bulawayo to
Beitbridge transect. There are better chances of getting rainfall during the January-March
period compared to the first half of the growing season.
Our study revealed that there were no in situ soil water management techniques practiced
by smallholder farmers in either Gwanda or Insiza districts during 2006. In Insiza district,
the graded contours were the only structures constructed between fields while in Gwanda
the dead level contours and infiltration pits were found on most farms particularly in
wards 17 and 18. The dead level contours were being promoted by a Non-Governmental
Organization (NGO) called Practical Action. Smallholder farmers in Insiza district used
both manure and inorganic fertilizer as soil fertility amendments. However, in Gwanda
district the majority of smallholder farmers used neither manure nor inorganic fertilizer
for fear of crop burn. Farmers who used manure and fertilizer in Gwanda district had
been exposed to how much manure and fertilizer is applied in semi-arid areas through
interaction with agriculture extension officers and researchers. However, there is need for
v
wider promotion of training and demonstrations on soil water and fertility management in
the semi-arid smallholder farming areas.
The on-farm experimentation assessed the effect of integrating soil water and nitrogen
management under smallholder farming conditions. The study assessed the effect of
single and double conventional ploughing, ripping and planting basins combined with
nitrogen ferti Iizer on maize yields, surface runoff and soi I water dynam ics. Resu Its of the
on-farm experimentation showed that the double conventional ploughing combined with
nitrogen fertilizer outperformed the other three tillage systems regardless of the rainfall
pattern in Insiza and Gwanda districts. Nitrogen fertilizer increased maize yields and
water use efficiency in each season regardless of the tillage system used under
smallholder farming conditions. The planting basin system had higher maize crop
establishment at most farms during the period of experimentation.
The on-station experiment showed that mulching had a significant influence on maize
grain production across the three tillage systems in a season with below average rainfall.
There were no significant maize yield differences across the three tillage systems tested
at the on-station experimental site. Delayed planting in the conventional system resulted
in reduced cowpea yields in a season with below average rainfall. Planting basin system
gave lower sorghum yield as a result of reduced plant stand which was caused by rodent
attack that was experienced at crop establishment stage. The on-station experiment
indicated that sorghum and cowpea can be grown at the 0.9 m x 0.6 m spacing of the
basin system without significantly reducing yield compared to the conventional system.
VI
The soil water dynamics were similar under single and double ploughing, ripper and
basin tillage systems. The on-station experiment also showed similar soil water dynamics
in the conventional, ripper and basin systems under mulched conditions at Matopos (clay
soil) and Lucydale (sandy soil). The basin system had more soil water during the
November-December period when the growing seasons started. Surface runoff
measurements indicated that planting basins significantly reduce surface runoff water
losses from cropped or uncropped field. However, the reduced surface runoff and higher
initial soil water content was not translated into higher yields under the basin tillage
system compared to the other tillage systems.
The study on the soil water contribution of dead level contours and infiltration pits
indicated that these inter-field structures have no significant effect on soil profile water
content in seasons with below average rainfall. During seasons that receive daily rainfall
events of more than 40 mm, the dead level contours and infiltration pits collect more
r
rainwater than the dead level contour only. Lateral soil water movement occurred after
rainfall events of 60-70 mm particularly downslope of the contour with significant
changes in soil water being observed at 3 m from the contour. The dead level contours
and infiltration pits supplied soil water to the 0.25-0.45 m layers of the 0.6 m profile
measured in this study.
The evaluation of the in situ tillage systems by smallholder farmers revealed that labour
demand and crop yields are major factors considered by smallholder farmers in semi-arid
southern Zimbabwe when selecting a technology for adoption. The majority of the
VII
farmers achieved the highest yields under the double ploughing system, hence they
ranked it as the most appropriate tillage system to use under their conditions. Availability
of cereal and legume seed is one of the major challenges being faced by all households
sampled in Gwanda and Insiza districts during the period of our study. The long term
assessment of the basin system through simulation modeling revealed that basins give
only marginal maize yield benefits over the conventional system regardless of the
nitrogen level used. The long term simulation also indicated that crop failures can be
experienced in both conventional and basin systems due largely to uneven distribution of
rain events through the growing season.
VIII
Uittreksel
WATlER- EN STIKSTOFBESTUUR VIR RISIKO- VERMINDlERING IN SEMI-
ARIEDE GlEWASSISTEME
deur
WALTER T. MUPANGWA
PhD Landbouweerkunde/Grondkunde, Universiteit van die Vrystaat, Bloemfontein
November 2008
Hierdie studie is uitgevoer met drie doelwitte in gedagte. Die eerste daarvan is die
karakterisering van die kleinhoewe boerderysisteem en reënvalpatroon van semi-ariede
suidelike Zimbabwe. Die tweede doelstelling is om die gewasopbrengs- en
grondwatervoordele wat ontstaan het uit in situ (enkel en dubbel ploeg, skeurploeg en
plantbakkies) en tussenlandgrondwaterbestuur (waterpasgelyke kontoere en
infiltrasiepitte) in suidelike Zimbabwe, te kwantifiseer, deur gebruik van veldproewe
gevestig op plaaslanderye. Verder is die effek van samevatting van in situ
grondwaterbestuurtegnologieë (enkel en dubbel ploeg, skeurploeg en plantbakkies) en
stikstofbemestingtoediening (0, 10 en 20 kglvha") op gewasopbrengs onder semi-ariede
toestande, ondersoek. 'n Eksperiment by die stasie is opgestel om die effek van
kombinering van in situ grondwaterbestuurtegnieke (enkel ploeg, skeurploeg en
plantbakkies) en deklaag (0, 0.5, 1,2,4, 8 en 10 tha") op mielies, akkerboon en sorghum
opbrengste, asook grondwaterdinamika, te assesseer. Die derde doelstelling is die
evaluering van tegnologieë getoets op boerelanderye oor drie groeiseisoene (2005/06,
2006/07 en 2007/2008). Die grondbewerkingsisteme is deur die kleinhoeweboere teen die
eiende van die laaste groeiseisoen van die studie ge-evalueer. Simulasiemodelle is daarna
gebruik om die langtermyneffekte van gebruik van bakkiegrondbewerkingsisteem oor 'n
69 jaar periode te evalueer.
IX
Die daaglikse reënvaldata is vanaf meteorologiese stasies geleë in die Mzingwane
opvanggebied van die groter Limpopo-opvanggebied, versamel. Die analises het getoon
dat beide die totale jaarlikse reënval gebasseer op die Julie-Junie almanak, en die begin of
einde van die groeiseisoen ewemin noemenswaardig verander het oor die 50-74 jarige
periode in suidelike Zimbabwe. Die analise het wel getoon dat die lengte van die
groeiseisoen verminder het langs die Bulawayo tot Beitbrugdeursnit. Die groeiseisoen
begin gedurende die eerste agt dae van Desember by alle stasies behalwe by Filabusi
waar die seisoen gedurende die laaste week van November begin. Die aantal nat dae per
groeiseisoen het ook nie verander langs die Bulawayo tot Beitbrugdeursnit nie. Die kanse
is goed dat reënval gedurende Januarie-Maart periodes ontvang sal word vergeleke met
die eerste helfte van die groeiseisoen.
Ons bevindinge het gewys dat daar geen in situ grondwaterbestuurstegnieke beoefen is
deur kleinhoeweboere in beide die Gwanda en lnsiza distrikte gedurende 2006 nie. In die
Insiza distrik, is gegradeerde kontoere die enigste stukture opgerig tuseen landerye terwyl
in Gwanda is waterpasgelyke kontoere en infiltrasiepitte op meeste plase, veral in wyke
17 en 18 gevind. Waterpasgelyke kontoere word tans bevorder deur 'n Nie-
Regeringsorganisasie of "Non-Governmental Organization" (NGO) genoem "Practical
Action". Kleinhoeweboere in Insiza distrik gebruik beide mis asook anorganiese
bemesting as grondbemestingswysigings. Die meerderheid kleinhoeweboere in Gwanda
distrik gebruik egter nóg mis nóg anorganiese bemesting omrede gewasbrand
("cropburn") gevrees word. Boere wat wel mis en bemesting in Gwanda distrik gebruik,
x
is deur interaksie met landboukundige offisiere en navorsers touwys gemaak aangaande
toedieningshoeveelhede van mis en bemesting in semi-ariede gebiede. Daar is nogtans 'n
behoefte aan nog meer bevordering van opleiding en demonstrasies van grondwater en
fertiliteitsbestuur in die serni-ariede kleinhoeweboerdery gebiede.
Die proewe uitgevoer op die plase het die effek van grondwater en stikstofbestuur onder
kleinhoewe plaastoestande ge-assesseer. Die studie het die effek van enkel en dubbel
konvensionele ploeg, skeurploeg en plantbakkies gekombineer met stikstokbemesting op
mielieopbrengs, oppervlakafloop en grondwaterdinamika, ge-assesseer. Die resultate het
gewys dat die dubbel konvensionele ploeg gekombineer met stikstofbemesting die ander
drie grondbewerkingsisteme oortref, ongeag die reënvalpatroon in die Insiza en Gwanda
distrikte. Die studie het ook gewys dat die stikstofbemesting mielieopbrengste in elke
seisoen verhoog het, ongeag die grondbewerkingsisteem gebruik onder
kleinhoeweboerdery toestande. Die plantbakkiesisteem het hoër mieliegewasvestiging op
meeste plase tydens die proeftydperk opgelewer.
Die proef by die stasie het getoon dat die deklaag 'n noemenswaardige invloed op
mieliegraanproduksie vir al drie grondbewerkingsisteme In 'n seisoen met onder
gemiddelde reënval getoets, gehad het. Daar is geen noemenswaardige
mielieopbrengsverskille tussen die drie grondbewerkingsisteme getoets rue. Laat
aanplanting by die konvensionele sisteem het gelei tot afname in akkerboon opbrengste
binne 'n seisoen met onder gemiddelde reënval. Plantbakkiesisteem het laer sorghum
opbrengs as gevolg van afname in plantstand wat deur knaagdiere geteister is by
XI
gewasvestigingsstadium. Dieselfde proef het getoon dat die sorghum en akkerboon by die
0.9 m X 0.6 m spasiëring van die bakkiesisteem sal groei sonder noemenswaardige
afname in opbrengs vergeleke met die konvensionele sisteem.
Die grondwaterdinamika van die enkel konvensionele en dubbelploeg, skeurploeg en
bakkiegrondbewerkingsisteme is gevind as redelik ooreenstemmend. Die stasieproef het
ook soortgelyke grondwaterdinamika in konvensionele, skeurploeg en bakkiesisteme
onder deklaetoestande by Matopos (kleigrond) en Lucydale (sandgrond), blootgelê. Die
bakkiesisteem het meer grondwater gedurende die begin van die November-Desember
groeiseisoen bevat. Oppervlakaflooplesings het aangedui dat bakkies waterverliese vanaf
beide verboude en onverboude landerye verminder. Die afname in oppervlakafloop en
hoër aanvanklike grondwaterinhoud het egter nie daartoe bygedra tot hoër opbrengstes
onder die bakkiegrondbewerkingssisteem In vergelyking met ander
grondbewerkingsisteme nie.
Die studie van grondwaterbydrae tot waterpasgelyke kontoere en infiltrasiepitte het aan
die lig gebring dat hierdie tussen-landerye stukture geen effek gehad het op
grondprofielwaterinhoud binne seisoene met ondernormale reënval nie. Gedurende
seisoene met daaglikse reënvalgevalle van meer as 30 mm, het die waterpasgelyke
kontoere en infiltrasiepitte meer reënwater versamel as slegs die waterpas kontoere.
Laterale grondwaterbewegings het plaasgevind na reënvalgevalle van 60-70 mm, veral
teen die skuinste van die kontoer, met noemenswaardige veranderinge in grondwater
waargeneem op 'n afstand 3 m van die kontoer. Die waterpasgelyke kontoere en
XII
infiltrasiepitte het grondwater verskaf aan die 0.25-0.45 m lae van die 0.6 m profiel
gebruik by hierdie studie.
Die evaluasie van die in situ grondbewerkingsisteme deur kleinhoeweboere het getoon
dat arbeidsvereistes en gewasopbrengstes as belangrike faktore deur kleinhoeweboere in
semi-ariede suidelike Zimbabwe oorweeg word, by tegnologiese keuses. Die groot
meerderheid boere het die hoogste opbrengstes onder dubbelploegsisteem behaal, en
daarom het hulle dit verkies as die mees geskikte grondbewerkingsisteem vir verbruik
onder die gegewe toestande. Beskikbaarheid van graan en peulsade is een van die grootse
probleme ondervind by alle steekproef huishoudings in die Gwanda en Insiza disktrikte
gedurende die studieperiode. Die langtermyn assessering van die bakkiesisteem deur
middel van simulasie modelering het getoon dat bakkies slegs marginale opbrengs
voordele bo die konvensionele sisteem behaal het, ongeag die stikstofvlak gebruik. Die
langtermyn simulasie het ook aangedui dat gewasmislukkings by beide konvensionele
asook bakkiesisteme voorkom, hoofsaaklik as gevolg van oneweredige verspreiding van
reënvalgevalle deur die groeiseisoen.
XIII
TABLE OF CONTENTS
Title i
Declaration ii
Acknowledgements 111
Abstract Iv
Uitttreksel IX
Table of Contents xiv
List of Tables XVI
List of Figures XXII
List of symbols and abbreviations xxxi
CHAPTER 1: General Introduction 1
CHAPTER 2: Literature Review 6
CHAPTER 3: General Materials and Methods 29
CHAPTER 4: Characterisation of Rainfall Pattern for Improved Crop Production
in Semi-Arid Cropping Systems of Southern Zimbabwe .46
CHAPTER 5: Soil Water and Fertility Management Practices on Smallholder
Farms in Insiza and Gwanda Districts of Semi-Arid Southern Zimbabwe 81
CHAPTER 6: Farm Characteristics and Evaluation of Soil Water and Fertility
Management Options for Smallholder Semi-Arid Cropping Systems 110
CHAPTER 7: Integrated Tillage and Nitrogen Management for Improved Soil and
Water Productivity on Smallholder Farms in Semi-Arid Zimbabwe 137
CHAPTER 8: Effect of Dead Level Contours and Infiltration Pits on Soil
Water Content and Crop Yields on Smallholder Farms in Gwanda District,
XIV
Southern Zimbabwe 174
CHA]>TER 9a: Soil Water and Maize Yield Responses to Minimum Tillage
and Mulching on Clayey and Sandy Soils in Semi-Arid Southern Zimbabwe 202
CHAPTER 9b: Minimum Tillage and Mulching Effects on Soil Water Regimes,
and Cowpea and Sorghum Yields on a Red Clay Soil in Semi-Arid
Southern Zimbabwe 238
CHAPTER 9c: Cumulative Effects of Minimum Tillage, Mulching and Rotation on
Selected Soil Properties and Maize Yield on a Clayey Soil in Semi-Arid Southern
Zimbabwe 262
CHAPTER 10: Productivity of Planting Basin Tillage System and Nitrogen under
Highly Variable Rainfall Regimes of Semi-Arid southern Zimbabwe: A Modelling
Assessment 292
CHAPTER 11: Summary and Recommendations 316
REFERENCE . 330
APPENDIX 1 355
APPENDIX 2 357
xv
.List of Tables
Table 2.1. Effect of double spring ploughing on sorghum grain yield for seven soil
types in Botswana in 1988/89 15
Table 3.1. Geographical description of meteorological stations and rainfall database
of the five stations used in the analyses 34
Table 3.2. Experimental fields used and crops grown in each field from 2004/05 to
2007/08 seasons at Matopos Research Station .42
Table 4.1. Drought classification indices adapted from Hayes et al. (1999) .49
Table 4.2. Homogeneity test for annual total rainfall for the five meteorological
stations in southern Zimbabwe 51
Table 4.3. Characteristics of the total annual rainfall (based on July-June calendar)
recorded at five meteorological stations in semi-arid southern Zimbabwe
....................................................................................... 52
Table 4.4. Median dates for the start and end of the growing season based on daily
rainfall data obtained from five meteorological stations in southern
Zimbabwe 63
Table 4.5. Rainfall characteristics for first half of the growing season at five
stations in semi-arid southern Zimbabwe 76
Table 4.6 Rainfall characteristics for second half of the growing season at five
stations in semi-arid southern Zimbabwe 77
Table 5.1. Effect of combining manure or compound D (8: 14:7 - N:P20S:K20) with
modified tied ridges on crop yield in Gwanda district (after Rusike and
Heinrich, 2002) 86
Table 5.2. Rainfall patterns and maize grain yield (kg ha") responses to farmer
practice and planting basins for eight districts in southern Zimbabwe in
2004/2005, (after Twomlowel al. 2006a) 88
Table 5.3. Major soil types, crop areas and possible yields in a good growing season
in ward I of Insiza district 91
Table 5.4. Characteristics of different winter and growing seasons and crops grown
as reported by farmers in ward I of Insiza district 94
XVI
Table 5.5. Soil fertility amendments applied during bad and good seasons by farmers
in ward I, Jnsiza district 96
Table 5.6. Risk factors and their severity on household food security as identified by
farmers in ward 1 of Insiza district 97
Table 5.7. Major soil types, crop areas and possible grain production in a good
growing season in Gwanda district 97
Table 5.8. Crop production in good and bad farming seasons in Gwanda district ... 99
Table 5.9. Characteristics of different winter and growing seasons and crops grown
as observed by farmers in Gwanda district I05
Table 5.10. Characteristics of different winter and growing seasons and crops
grown as observed by farmers in Gwanda district 106
Table 5.1 I. Risk factors and their severity on household food security as identified by
farmers in Gwanda district 107
Table 6.1. Farmer resource status as an average of Gwanda and Insiza districts of
southern Zimbabwe 114
Table 6.2. Cropping calendar for the farmer practice fields in Gwanda and Insiza
districts during 2005/06,2006/07 and 2007/08 growing seasons 122
Table 6.3. Cropping calendar for the single conventional ploughing in Gwanda and
Insiza districts during 2005/06,2006/07 and 2007/08 growing seasons
..................................................................................... 122
Table 6.4. Cropping calendar for the double conventional ploughing in Gwanda and
Insiza districts during 2005/06,2006/07 and 2007/08 growing seasons
...................................................................................... 122
Table 6.5. Cropping calendar for the ripper system in Gwanda and Insiza districts
during 2005/06, 2006/07 and 2007/08 growing seasons 123
Table 6.6. Cropping calendar for the basin system in Gwanda and Insiza districts
during 2005/06,2006/07 and 2007/08 growing seasons 123
Table 6.7. Farmer assessment of labour demands for land preparation on a 0.25 ha
plot of different tillage systems tested on their farms averaged for three
growing seasons 126
XVII
Table 6.8. Farmer assessment of weeding methods, frequency and labour demands
during seasons with different rainfall pattern in Gwanda and Insiza
districts 127
Table 6.9. Pest species in different tillage system affected as observed by farmers in
Gwanda and Insiza districts during 2005/06, 2006/07 and 2007/08
growing seasons 131
Table 6.10. Farmer ranking of organisations working in Gwanda and Insiza, and
support rendered and expected by farmers in Gwanda and Insiza districts.
(1 = lot of support; 5 = little support) 134
Table 7.1. Selected initial soil chemical and physical properties (0 - 0.6 m) at farms
used from 2005/06 to 2007/08, Insiza and Gwanda districts 146
Table 7.2. Average maize plant stands under four tillage systems (CP, OP, ripper and
basin) measured two weeks after crop emergence during the three seasons
of experimentation in Insiza and Gwanda districts 159
Table 7.3. Maize responses to four tillage systems (CP, OP, ripper and basin) and
nitrogen fertilizer (0 and lO kgNha·l) averaged across five farms in
2005/06 season, Insiza and Gwanda districts 160
Table 7.4. Maize responses to four tillage systems (CP, OP, ripper and basin) and
nitrogen fertilizer averaged across three farms in 2006/07 season, Insiza
and Gwanda districts 161
Table 7.5. Maize responses to four tillage systems (CP, OP, ripper and basin) and
nitrogen fertilizer (0, lO and 20 kgNha·l) averaged across seven farms in
2007/08 season, Insiza and Gwanda districts 162
Table 7.6. Effects of season and tillage system on maize crop performance during
three seasons of experimentation in Insiza and Gwanda districts 163
Table 7.7. Effects of season and nitrogen fertilizer on maize crop performance during
three seasons of experimentation in Insiza and Gwanda districts 163
Table 7.8. Agronomic nitrogen use efficiency of maize in four tillage systems (CP,
OP, ripper and basin) averaged across farms for each season during
2005/06, 2006/07 and 2007/08 growing seasons Insiza and Gwanda
districts 166
XVIII
Table 8.1. Soi I textural variation with soi I depth at the four farms used for
quantifying soil water supply from dead level contours and infiltration pits
in Gwanda district 181
Table 8.2. Components of water balance averaged across two farms (Moyo and
Ncube) at each distance along unploughed and ploughed transects the
during 2007/08 eason 198
Table 8.3. Components of the soil water balance averaged across two farms (Dube
and Siziba) at each distance along unploughed and ploughed transects
across dead level contours and infiltration pits during 2007/08 season
...................................................................................... 198
Table 8.4. Pearl millet and maize yields and water use efficiency between different
positions from dead level contour averaged across two farms (Moyo and
Ncube) along the ploughed transect at the end of the 2007/08 growing
season 199
Table 8.5. Pearl millet and maize yields and water use efficiency between different
positions from dead level contour averaged across two farms (Dube and
Siziba) with dead level contours and infiltration pits along the ploughed
transect at the end of 2007/08 season 200
Table 9a.l. Experimental fields used and crops grown in each field from 2004/05. to
2007/08 seasons at Matopos Research Station 205
Table 9a.2. Physical and chemical characteristics of Matopos and Lucydale soils, after
Moyo (2001) 208
Table 9a.3. Characteristics of the 2004/05 and 2005/06 growing seasons at Lucydale
experimental site 210
Table 9a.4. Effects of tillage and mulch treatments on maize crop performance at
Lucydale experimental site in 2004/05 growing season 214
Table 9a.5. Effects of tillage and residual mulch treatments on maize crop
performance at Lucydale experimental site in 2005/06 growing season
...................................................................................... 215
Table 9a.6. Characteristics of the 2004/05, 2005/06, 2006/07 and 2007/08 growing
seasons at Matopos experimental site 217
XIX
Table 9a.7. Effects of tillage and mulch treatments on maize crop performance at
Matopos experimental site in 2004/05 growing season 227
Table 9a.8. Effects of tillage and mulch treatments on maize crop performance at
Matopos experimental site in 2005/06 growing season 227
Table 9a.9. Effects of tillage and mulch treatments on maize crop performance at
Matopos experimental site in 2006/07 growing season 229
Table 9a.10. Effects of tillage and mulch treatments on maize crop performance at
Matopos experimental site in 2007/08 growing season 231
Table 9a.ll. Effects of tillage and mulching on maize performance averaged across
four growing seasons at Matopos experimental site 233
Table 9a.12. Effect of the growing season on maize crop performance at Matopos
experimental site 234
Table 9a.13. Effects of tillage and mulching on maize performance averaged across four
growing seasons at Matopos experimental site 235
Table 9b.l. Effect of tillage method and mulch cover on cowpea plant stand two
weeks after planting (plants m") on a red clay soil at Matopos Research
Station 256
Table 9b.2. Effect of tillage system and mulch treatment on sorghum stand two weeks
after planting (plants m-2) on a red clay soil at Matepos Research Station
..................................................................................... 257
Table 9b.3. Cow pea grain yield responses (kgha") averaged across three tillage
systems and seven mulch levels at Matopos Research Station during
2005/06 and 2006/07 growing seasons 258
Table 9b.4. Sorghum yield responses (kgha') averaged across three tillage systems
and seven mulch treatments at Matopos Research Station during 2006/07
growing season 260
Table 9c.l. Experimental fields used and crops grown in each field from 2004/05 to
2007/08 seasons at Matopos Research Station 264
Table 9c.2. van Genuchten parameters for 12 textural classes and A values for 2.2 cm
disk radius and suction values from 0.5 to 6.0 cm (Adapted from Carsel
and Parrish, 1988) : 268
xx
Table 9c. 3. Effect of tillage and period of exposure to CA practices averaged across
three mulch levels (0, 4 and 10 tha') on soil organic carbon content (%) of
a clay soi I at Matopos Research Station 270
Table 9c.4. Effect of tillage, mulching and year interaction on soil bulk density of a
red clay soil at Matepos Research Station 272
Table 9c.5. Effect of minimum tillage, mulching and period of exposure to CA
practices on soil hydraulic conductivity (K) (1O-3mms-') and capillary
sorptivity (S) (mm/..Js) of a clay soil at Matopos Research Station 285
Table 9c.6. Volumetric water content (%) measured by capacitance probe in 0 - 0.10
m soil layer before infiltration runs at Matopos Research Station 287
Table 9c.7. Volumetric soil water content (%) in 0 - 0.10 m layer measured after
infiltration runs in the four fields at Matopos Research Station 288
Table 9c.8. Maize crop performance measured at the end of the 2007/08 season as
influenced by different periods of exposure to CA practices on a clay soil
at Matopos Research Station................................................. 289
Table 9c.9. Maize yield (kgha'), harvest index and plant stand (m") responses to three
tillage methods at Matopos Research Station in 2007/08 season 290
Table 10.1. Soil chemical and physical properties of the clay soil used for Matopos
Research Station experimental site (from ICRISA T unpublished data)
...................................................................................... 296
Table 10.2. Soil chemical and physical properties of the sand soil used for Lucydale
experimental site (from Masikati, 2006) 296
Table 10.3. Dates for field activities carried out at Matopos Research Station during
the four seasons of experimentation (Chapter 9a) 297
Table 10.4. Dates for field activities carried out at Lucydale experimental site during
the two seasons of experimentation (Chapter 9a) 298
Table I l.I. Typical cropping calendar for smallholder cropping systems of semi-arid
southern Zimbabwe 318
XXI
List of Figures
Figure 2.1. Agro-ecological regions of Zimbabwe (Source: ICRISAT GIS office,
2008) 13
Figure 3.1 Agro ecological regions of Zimbabwe and location of experimental sites in
Insiza and Gwanda districts, Matebeleland South province 34
Figure 3.2. Schematic diagram of the set up of equipment for collection of runoff
water from the plots under single and double conventional ploughing,
ripper and basin systems 38
Figure 3.3. Schematic diagram of the set up of access tubes across dead level contours
in Gwanda district .40
Figure 4.1. Total annual rainfall (July to next June) for Bulawayo (a), Matopos (b)
and Mbalabala (c) meteorological stations of semi-arid southern
Zimbabwe 53
Figure 4.2. Total annual rainfall for Filabusi (d) and Beitbridge (e) meteorological
stations of semi-arid southern Zimbabwe 54
Figure 4.3. Standardized precipitation indices derived from total annual rainfall data
for Bulawayo (a), Matopos (b) and Mbalabala (c) meteorological stations
....................................................................................... 57
Figure 4.4. Standardized precipitation indices derived from total annual rainfall data
for Filabusi (d) and Beitbridge (e) meteorological stations 58
Figure 4.5. Number of wet days per season based on daily rainfall data obtained from
Bulawayo (a), Matopos (b) and Mbalabala (c) meteorological stations ...60
Figure 4.6. Number of wet days per season based on daily rainfall data obtained from
Filabusi (d) and Beitbridge (e) meteorological stations 61
Figure 4.7. Start and end of growing seasons derived from daily rainfall data for
Bulawayo (a), Matopos (b) and Mbalabala (c) Meteorological stations
.................................................................................................................... 64
Figure 4.8. Start and end of growing seasons derived from daily rainfall data for
Filabusi (d) and Beitbridge (e) meteorological stations 65
Figure 4.9. Length of the growing season based on daily rainfall data obtained from
Bulawayo (a), Matopos (b) and Mbalabala (c) meteorological stations ....68
XXII
Figure 4.10. Length of the growing season based on daily rainfall data obtained from
Filabusi (d) and Beitbridge (e) meteorological stations 69
Figure 4.11. Relationship between length and start of growing season at Bulawayo (a)
meteorological station in southern Zimbabwe 70
Figure 4.12. Relationship between length and start of growing season at Matopos (b),
Mbalabala (c) and Filabusi (d) meteorological stations in southern
Zimbabwe 71
Figure 4.13. Relationship between length and start of growing season at Filabusi (d)
and Beitbridge (e) meteorological stations in southern Zimbabwe 72
Figure 4.14. Probability of getting 14 and 21 day spells within 30 days from a wet day
based on the fitted first order Markov chain probability values for
Bulawayo (a) and Matopos (b) meteorological stations in southern
Zimbabwe 73
Figure 4.15. Probability of getting 14 and 21 day spells within 30 days from a wet day
based on the fitted first order Markov chain probability values Mbalabala
(c), Filabusi (d) and Beitbridge (e) stations in southern Zimbabwe 74
Figure 4.16. Cumulative distribution functions for the first and second halves of the
growing season based on three monthly rainfall totals from Bulawayo (a)
meteorological station 77
Figure 4.17. Cumulative distribution functions for the first and second halves of the
growing season based on three monthly rainfall totals from Mbalabala,
Filabusi (d) and Beitbridge (e) meteorological stations 78
Figure 6.1. Resource flow map for Mr. John Ncube of Gwanda district. The map
represents flow of resources at a farm that interacted with researchers for
three cropping seasons 117
Figure 6.2. Resource flow map for Mrs. Malotha of Humbane village, Gwanda
district. The map represents flow of resources at a farm that had no
interaction with researchers from 2005/06 to 2007/08 seasons 119
Figure 6.3. Average sources of sorghum seed planted by 14 households in 2005/06,
2006/07 and 2007/08 seasons in ward I of Insiza and ward 17 of Gwanda
district 120
xxiii
Figure 6.4. A verage sources of maize seed planted by 14 households in 2005/06,
2006/07 and 2007/08 seasons in ward 1 of Insiza and ward 17 of Gwanda
district 121
Figure 6.5. Farmer ranking of basin, double ploughing (OP) and ripper tillage systems
tested on their fields for three growing seasons in Gwanda and Insiza
districts. Ranking was based on maize grain yields achieved over three
growing seasons 130
Figure 6.6. Tillage systems chosen for adoption by female and male farmers during
focus group discussions in Gwanda and Insiza districts 132
Figure 7.1. Cumulative rainfall distribution at Mpofu and Moyo farms of ward I in
Insiza district during 2005/06 growing season 147
Figure 7.2. Profile soil water changes (0 - 0.50 m) averaged across two farms (Moyo
and Mpofu) in ward 1 of Insiza district during 2005/06 growing season .
..................................................................................... 148
Figure 7.3. Cumulative rainfall distribution at Sibanda, Siziba and J. Ncube farms of
ward 17 in Gwanda district during 2005/06 growing season 149
Figure 7.4. Profile soil water changes (0 - 0.60 m) averaged across three farms in
ward 17 ofGwanda district during 2005/06 growing season 150
Figure 7.5. Cumulative rainfall distribution at Mpofu, Nyathi, Mguni and Mlalazi
farms of ward 1 in Insiza district during 2006/07 growing season 151
Figure 7.6. Profile soil water changes (0 - 0.60 m) averaged across two farms (Mpofu
and Nyathi) in ward 1 of Insiza district during 2006/07 growing season .
..................................................................................... 152
Figure 7.7. Cumulative rainfall distribution at Sibanda, Siziba and J. Ncube farms of
ward 17 in Gwanda district during 2006/07 growing season 153
Figure 7.8. Profile soil water changes (0 - 0.60 m) averaged across three farms
(Sibanda, Siziba and J. Ncube) in ward 17 of Gwanda district during
2006/07 growing season 153
Figure 7.9. Cumulative rainfall distribution at Mpofu, N. Ncube, Nkomo, Mguni and
Mlalazi farms of ward I in Insiza district during 2007/08 growing season
...................................................................................... 154
XXIV
Figure 7.10. Profile soil water changes (0 - 0.60 m) averaged across three farms
(Mpofu, N. Ncube and Nkomo) in ward I of Insiza district during 2007/08
growing season 155
Figure 7.11. Cumulative rainfall distribution at Sibanda, Siziba, J. Ncube and Tlou
farms of ward 17 in Gwanda district during 2007/08 growing season ... 156
Figure 7.12. Profile soil water changes (0 - 0.60 m) averaged across three farms (J.
Ncube, Sibanda and Siziba) in ward 17 of Gwanda district during 2007/08
growing season 157
Figure 7.13. Seasonal runoff losses from plots under four tillage systems in 2006/07
growing season, Insiza (Mpofu and Nyathi) and Gwanda (J. Ncube and
Sibanda) districts 158
Figure 7.14. Seasonal runoff losses from plots under four ti Ilage systems in 2007/08
growing season, Insiza (Mpofu and N. Ncube) and Gwanda (J. Ncube and
Sibanda) districts 159
Figure 7.15. Water use efficiency as affected by four tillage systems (CP, OP, ripper
and basin) and nitrogen fertilizer (0 and 10 kgNha-l) across five farms
(Moyo, Mpofu, J. Ncube, Sibanda and Siziba) in 2005/06 season, Insiza
and Gwanda districts 164
Figure 7.16. Water use efficiency as affected by four tillage systems (CP, OP, ripper
and basin) and nitrogen fertilizer (0, 10 and 20 kglvha') across six farms
in 2007/08 season, Insiza and Gwanda districts 165
Figure 8.1. Drained upper limit (OUL) and lower limit (LL) derived from 2007/08
soil water data for Moyo, Ncube and Oube farms in Gwanda district ... 180
Figure 8.2. Drained upper limit (DUL) and lower limit (LL) derived from 2007/08
soil water data for Siziba farm in Gwanda district 181
Figure 8.3. Cumulative rainfall distribution at Moyo, Ncube, Oube and Siziba farms
of ward 17 in Gwanda district during 2006/07 growing season 183
Figure 8.4. Cumulative rainfall distribution at Moyo, Ncube, Oube and Siziba farms
of ward 17 in Gwanda district during 2006/07 growing season 184
xxv
Figure 8.5. Profile soil water changes along ploughed transects averaged across two
farms with dead level contours only (Mayo and Ncube) during the
2006/07 growing season 185
Figure 8.6. Soil water distribution with respect to depth at different distances from the
dead level contour only averaged across two farms (Mayo and Ncube) on
the driest day along unploughed (a) and ploughed (b) transects during
2006/07 season 186
Figure 8.7. Profile soil water changes averaged across two farms (Mayo and Ncube)
along un ploughed (a) and ploughed (b) transects across farms with only
dead level contours during 2007/08 growing season 187
Figure 8.8.. Soil water distribution with respect to depth at different distances from the
dead level contour only averaged across two farms (Mayo and Ncube) on
the driest day along unploughed transect during 2007/08 season 188
Figure 8.9. Soil water distribution with respect to depth at different distances from the
dead level contour only averaged across two farms (Mayo and Ncube) on
the driest day along ploughed transect during 2007/08 season 189
Figure 8.10 .. Soil water distribution with respect to depth at different distances from the
dead level contour only averaged across two farms (Mayo and Ncube) on
wettest day along unploughed (a) and ploughed (b) transects during
2007108 season 190
Figure 8.11. Profile soil water changes at different distances from the dead level
contour with infiltration pit averaged across two farms (Dube and Siziba)
during 2006/07 growing season 192
Figure 8.12. Soil water distribution with respect to depth at different distances from
the dead level contour and infiltration pits averaged across two farms
(Dube and Siziba) on driest and wettest days during 2006/07 season .... 193
Figure 8.13. Profile soil water changes along unploughed (a) and ploughed (b)
transects at farms with dead level contours and infiltration pits averaged
across two farms (Du be and Siziba) during 2007/08 growing season ... 194
Figure 8.14. Soi I water distribution with respect to depth at different distances from the
dead level contour and infi Itration pits averaged across two farms (Dube
XXVI
and Siziba) on driest day along unploughed transect during 2007/08
season 195
Figure 8.15. Soil water distribution with respect to depth at different distances from the
dead level contour and infiltration pits averaged across two farms (Dube
and Siziba) on driest day along ploughed transect during 2007/08 season
..................................................................................... 196
Figure 8.16. Soil water distribution with respect to depth at different distances from the
dead level contour and infiltration pits averaged across two farms (Dube
and Siziba) on wettest day along unploughed (a) and ploughed (b)
transects during 2007/08 season 197
Figure 9a.l. Daily rainfall distribution at Lucydale experimental site during 2004/05
and 2005/06 growing seasons 21 0
Figure 9a.2. Average seasonal soil water content in the 0-0.30 m profile at Lucydale
during 2004/05 cropping season 211
Figure 9a.3. Soil water content in the 0-0.60 m profile at Lucydale during 2005/06
cropping season 212
Figure 9a.4. Cumulative rainfall distribution at Matopos experimental site during
004/05,2005/06,2006/07 and 2007/08 growing seasons 217
Figure 9a.5. Average seasonal soil water content in the 0 - 0.30 m profile under three
tillage systems and mulching treatments at Matopos during 2004/05
season 221
Figure 9a.6. Soil water content in the 0 - 0.30 m profile at Matopos during 2005/06
cropping season 221
Figure 9a.7. Profile soil water changes in the conventional ploughing, ripper and basin
system and four mulch treatments (0, 2, 4 and 10 tha") at Matopos
experimental site during 2006/07 growing season 223
Figure 9a.8. Soil water changes in four different layers of the soil profile under the
conventional ploughing tillage system during 2006/07 growing season at
Matopos site 224
xxvii
Figure 9a.9. Profile soil water changes in the conventional, ripper and basin tillage
systems and four mulch treatments (0, 2, 4 and 10 tha') at Matopos
experimental site during 2007/08 growing season 225
Figure 9a.l O. Soil water changes in four different layers of the soil profile during 2007/08
growing season at Matopos experimental site 226
Figure 9a.ll. Maize grain yield responses to mulching on a red clay soil under
conventional, ripper and basin tillage system during the 2006/07 growing
season 232
Figure 9b.l. Daily rainfall distribution at Matopos site during the 2005/06 growing
season 243
Figure 9b.2. Soil water changes in 0 - 0.30 m profile in the cowpea field under
conventional ploughing tillage system and four mulch treatments (0, 2, 4
and lOt ha-I) on a clay soil during 2005/06 growing season
.................................................................................................. 245
Figure 9b.3. Soil water changes in 0 - 0.30 m profile in the cowpea field under two
tillage systems (ripper and basin) and four mulch treatments (0, 2, 4 and
10 t ha") on a clay soil during 2005/06 growing season 246
Figure 9b.4. Daily rainfall distribution at Matopos site during the 2006/07 growing
season 248
Figure 9b.5. Profile soil water changes in 0 - 0.55 m profile in the cowpea field under
two tillage systems (ripper and basin) and four mulch treatments (0, 2, 4
and lOt ha") on a clay soil during 2006/07 growing season 249
Figure 9b.6. Average changes in equivalent soil water depth in different layers of a clay
soil under conventional ploughing tillage system in the cowpea field at
Matopos Research Station during 2006/07 growing season 250
Figure 9b.7. Average changes in equivalent soil water depth in different layers of a
clay soil under ripper and basin tillage systems in the cowpea field at
Matopos Research Station during 2006/07 growing season 252
Figure 9b.8. Soil water changes (0 - 0.65 m profile) in response to conventional, ripper
and basin tillage systems and four mulching treatments (0, 2, 4 and 10 tha'
xxviii
I) in the sorghum field during 2006/07 growing season at Matopos
Research Station 254
Figure 9c.l. Effect of ti lIage and soil depth on organic carbon content (%) of a red
clay soil at Matopos Research Station 271
Figure 9c.2. Soil bulk density distribution with respect to soil depth in different tillage
treatments in field I (first year of CA treatments) at Matopos Research
Station ' 276
Figure 9c.3. Soil bulk density distribution with respect to soil depth in different tillage
treatments and fields at Matopos Research Station 277
Figure 9c.4. Effect of conventional ploughing and ripper systems, 0 tha·1 mulch cover
and period of exposure to CA practices on cumulative infiltration of a clay
soil , 279
Figure 9c.5. Effect of basin system, 0 tha" mulch cover and period of exposure to CA
practices on cumulative infiltration ofa clay soil 280
Figure 9c.6. Effect of tillage, 4 tha' mulch cover and period of exposure to CA
practices on cumulative infiltration of a clay soil 281
Figure 9c.7. Effect of tillage, 10 tha' mulch cover and period of exposure to CA
practices on cumulative infiltration of a clay soil 283
Figure 10.1. Observed and predicted grain yield from different mulch levels over four
growing seasons at Matopos Research Station 30 I
Figure 10.2. Observed and predicted total biomass yields from different mulch levels
over four growing seasons at Matopos Research Station 302
Figure 10.3. Observed and predicted grain and total biomass yields for different mulch
levels on a sand soil over two growing seasons at Lucydale experimental
site _ 304
Figure 10.4. Observed and predicted soil water content in a clay soil (0 - 0.25 m layer)
under plough, ripper and basin tillage systems at Matopos Research
Station during 2006/07 growing season .306
Figure 10.5. Observed and predicted soil water in a clay soil (0 - 0.25 m layer) under
plough, ripper and basin tillage systems at Matopos Research Station
during 2007/08 growing season 307
XXIX
Figure 10.6. Observed and predicted soil water in a clay soil (0 - 0.25 m layer) under
plough (a), ripper (b) and basin (c) tillage systems at Matopos Research
Station during 2007/08 growing season 308
Figure 10.7. Comparison of grain yield achieved in the conventional and basin tillage
systems at four N application rates (0, 10,20 and 52 kglvha") on a granitic
sandy soil under semi-arid conditions 310
Figure 10.8. Cumulative distribution function for maize grain yield response to four N
application rates (0, 10,20 and 52 kgNha-l) on a granitic sandy soil in the
conventional tillage system under semi-arid conditions 311
Figure 10.9. Cumulative distribution function for maize grain yield response to four N
application rates (0. 10, 20 and 52 kgNha-l) on a granitic sandy soil in the
basin tillage system under semi-arid conditions 312
Figure 10.10. Surface runoff water losses from the conventional ploughing and basin
tillage systems on a granitic sandy soil under semi-arid conditions ...... 313
Figure 10.11. Deep drainage soil water losses from conventional ploughing and basin
tillage systems during the crop growing period under semi-arid conditions
.......................................... , 314
xxx
List of Symbols and Abbreviations
AGRITEX Zimbabwe Department of Agricultural technical and extension
ANOVA Analysis of variance
ANUE Agronomic nitrogen use efficiency
APSIM Agricultural Production Simulator
AREX Agricultural Extension Services
C Capillary rise
CA Conservation agriculture
CADEC Catholic Development Commission
CIMMYT International Maize and Wheat Improvement Centre
CP Conventional ploughing
CV Coefficient of variation
D Deep drainage
DFID United Kingdom Department for International Development
DLC Dead level contour
DP Double ploughing
DRSS Zimbabwe Department of Research and Specialist Services
DUL Drained upper limit of soil water content
ECAF European Conservation Agriculture Federation
ENSO El Nifio-Southern Oscillation
ET Evapotranspiration rate
FAO United Nations Food and Agricultural Organization
GCM Global Circulation Model or Global Climate Model
GMB Zimbabwe Grain marketing board
GTZ Germany Agency for Technical Cooperation
lAE Zimbabwe Institute of agricultural engineering
ICRISAT International Crops Research Institute for the Semi-Arid Tropics
ITCZ Inter-tropical convergence zone
LL Lower limit of soil water content
Lsd Least significant difference
ME Modeling efficiency
XXXI
NGO Non-Governmental Organization
NR Zimbabwe Agroecological region
Oi Observed values
ORAP Organization of Rural Associations for Progress
P Precipitation (mm)
PAWC Plant available water capacity
Pi Model predicted values
PRA Participatory Rural Appraisal
REML Restricted maximum likelihood model
RMSD Root mean square deviation
Roff Surface runoff out of the field (mm)
Ron Surface runon into the field (mm)
SAFIRE Southern Alliance for Indigenous Resources
se Standard error of means
SEDAP South eastern dry areas project
sed standard error of difference
SPI Standardized precipitation index
SST Seas Surface Temperature
UNDP United Nations Development Programme
USAID United States Agency for International Development
WUE Water use efficiency
6SW Change in soil water (mm)
6Y Change in grain yield (kgha')
ZFC Zimbabwe Fertilizer Company
xxxii
CHAPTER 1
1.1 Background
Every year an estimated six million hectares of productive land are being lost through
desertification, land degradation and declining agricultural productivity (UNDP,
2002). The population of Africa is expected to reach 1.2 billion by the year 2020
(Love et al., 2006). This means more 'mouths' have to be fed from a diminishing
resource base (Diagana, 2003), and leads to overgrazing and cultivation of marginal
lands (Ngwenya, 2006; Twomlowet al., 2006a). Agricultural production of major
cereals and legumes has remained stagnant or even declined in most African countries
(Henao and Baanante, 2006). In Zimbabwe, more than 75% of smallholder farming
families, who depend on rainfed agriculture, live in the semi-arid areas of
agroecological regions (NR) IV and V (Chuma and Haggmann, 1998) where rainfall
is erratic and soils are inherently infertile.
Smallholder agriculture in semi-arid sub-Saharan Africa is heavily dependant on the
seasonal characteristics of the rainfall. Total annual rainfall in semi-arid southern
Zimbabwe is less than 500 mm (against a national average of 657 mm) and exhibits
extreme spatial and temporal variations (Unganai, 1996; Phillips et al., 1998).
According to the agro-ecological classification of Zimbabwe based on climate and
soils, NRs IV and V are not suitable for intensive rainfed crop production (Vincent
and Thomas, 1960) due to low and highly variable annual rainfall. Rainfall occurs as
short duration, high intensity convective storms (Rockstrëm and Falkenmark, 2002).
Complete crop failure and reduced crop yields due to frequent intra-seasonal dry
spells and drought make rainfed crop production risky. Soil water availability remains
1
one of the major challenges to crop production in NRs IV and V given the erratic
rainfall pattern.
Despite the use of drought-tolerant maize, sorghum and pearl millet varieties, fanners
in semi-arid southern Zimbabwe still achieve small yield gains. The potential yields
for sorghum and millet range between 1.7 and 4.8 tha-1 but average yields in
Zimbabwe are currently less than 0.6 tha-1 (Ahmed et al., 1997). Most soils in the
smallholder farming areas are sands derived from granitic parent material (Grant,
1981) with poor water holding capacity, low inherent soil fertility and organic matter
(Anderson et al., 1993) making them marginal for crop production (Twomlow, 1994).
Lack of inorganic' and/or organic fertilizer use is undermining household food security
in the semi-arid areas. In semi-arid southern Zimbabwe fertilizer and livestock
manure usage is also low and the free handouts of fertilizer have failed to stimulate
farmers to use more soil fertility amendments (Ahmed et al., 1997). Farmer
perception of fertilizer usage is that it is too risky given the low and erratic nature of
rainfall in the region. Many farmers in the semi-arid areas have had few positive
experiences with organic and inorganic fertilizers. However, studies conducted by the
International Crops Research Institute for Semi-Arid Tropics (ICRISAT) across 17
districts in southern Zimbabwe have shown that cereal yield gains of 30-50% are
achievable using small doses (10 kgNha-1) of nitrogen in a normal season (Twomlow
et al., 2008b). Supplementing livestock manure with small doses of nitrogen fertilizer
gives maize yield gains even in seasons with severe plant water stress (Ncube et al.,
2007).
2
Low crop productivity in smallholder farming systems is further exacerbated by poor
management. Appropriate timing of field operations such as planting and weeding is a
potential key to successful cropping in semi-arid areas. Late planting in most
instances is due to lack of draught power resulting in farmers failing to utilize the
effective planting rains. Nyagumbo (2007) reports that delaying planting by more than
3 weeks can reduce maize yield by 32 % in sub humid northern Zimbabwe. Excessive
weed growth is one of the major factors limiting crop production in smallholder
farming systems (Dhliwayo et al., 1995). Field studies have shown that weed
transpiration has a significant impact on soil water regimes in semi-arid environments
(Vander Meer, 2000). Besides reducing unproductive water flows through weed
transpiration, effective weed control also reduces competition for nutrients and
radiation between crops and weeds.
A range of interventions that collect and conserve rainwater and prolong the time of
soil water availability to crops are currently being developed and promoted in the
semi-arid cropping systems of southern Zimbabwe. Such in situ techniques include
planting basin and ripper tillage systems which are part of the recently introduced
Conservation Agriculture (CA) program (Twomlowand Hove, 2006). By using a
planting basin system, smallholder farmers with limited access to animal draught
power can plant on time in terms of a few days after an effective rainfall event
(Twomlowet al., 2008a). Dead level contours and infiltration pits are between field
structures designed to collect runoff water from the field and supply soil water to
crops through lateral movement from the contours. However, the contribution of both
in situ tillage systems and between field rainwater harvesting structures to soil water
3
dynamics and crop yields under semi arid cropping systems is not well understood in
terms of soil water storage and use.
1.2 Hypotheses
In order to understand the effect of rainwater management techniques and nitrogen on
soil water dynamics and crop yields in semi-arid cropping systems, the following
hypotheses were tested:
(1) Smallholder fanners in Gwanda and Insiza districts are currently using available
soil water and fertility management practices.
(2) Rainfall patterns and growing seasons in semi-arid southern Zimbabwe have been
changing over the past 50 to 74 years.
(3) The use of minimum tillage techniques such as planting basins and ripping, and
nitrogen fertilizer with or without mulching improves soil water supply and crop
yields in semi-arid cropping systems.
(4) Dead level contours and infiltration pits supply soil water laterally to crops
resulting in increased yields.
(5) The use of planting basin tillage system and nitrogen fertilizer improves maize
yields and reduces risk in semi-arid smallholder cropping systems.
1.3 Objectives
The current study was designed to quantify soil water benefits derived from in situ
and inter-field water management techniques in semi-arid Gwanda and Insiza districts
of southern Zimbabwe. The study also explored synergies between soil water and
nitrogen management for improved water and crop productivity in semi-arid cropping
systems. Simulation modelling was applied to assess the long-term effect of planting
4
basin tillage system and nitrogen on maize productivity and the soil water balance in
smallholder cropping systems. The specific objectives were:
(1) To identify current soil water and fertility management practices and farmers'
perceptions of risk in Gwanda and Insiza districts;
(2) To characterise rainfall and growing season patterns of the semi-arid southern
Zimbabwe;
(3) To quantify the effects of in situ minimum tillage techniques and N fertilizer on
soil water dynamics and maize yields;
(4) To quantify the effects of in situ minimum tillage techniques and mulching on soil
water dynamics, and cowpea, maize and sorghum yields;
(5) To evaluate the effect of dead level contours and infiltration pits on in-field soil
water dynamics in Gwanda district;
(6) To conduct a farmer evaluation of in situ soil water and fertility management
systems being promoted in Gwanda and Insiza districts; and
(7) To assess the long-term effect of planting basin tillage system and nitrogen on
maize productivity and soil water balance in smallholder cropping systems using
APSIM systems model.
5
CHAPTER2
Literature Review
2.1 Conservation agriculture
The introduction of the conventional plough in the 19th century enabled farmers to till
more land and control problematic weeds (Stocking, 1989). Conventional ploughing
became the 'civilized' way ofland preparation resulting in expansion of cropped land
area in the smallholder farming sector (Alvord, 1936). Continuous cultivation has led
to the loss of physical; chemical and· biological soil properties responsible for
maintaining soil health (Xiano-Bin et al., 2006; Nhamo, 2007). Cereal yields from the
smallholder farming sector of sub-Saharan Africa continue to decline at a time when
Africa's population is expected to double the 1995 figures by the year 2020 (Love et
al.,2006).
Conservation agriculture (CA) has the potential to address some of the challenges
being faced in the smallholder rainfed agriculture (Giller et al., 2008). Conservation
agriculture is a farming system that encompasses the use of minimum tillage, animal
and tractor drawn implements, or hand powered methods, together with integrated
pest and disease management (Twomlowet al., 2008a). The CA system also
emphasizes the integrated management of soil and water resources. Conservation
agriculture systems hinge on three cornerstones: (1) minimum disturbance of the soil,
(2) keeping the soil covered as much as possible with a minimum cover of 30 %, and
(3) mixing and rotating crops (ECAF, 1999; Baudeon et al., 2007).
In southern Africa CA has been practiced in the large-scale commercial farming
sector (Oldrieve, 1993; Haggblade and Tembo, 2003; Nyagumbo, 2007). The
6
commercial farmers have access to appropriate equipment and crop residues for
mulching, and realised profits due to reduced land preparation costs and less wear on
implements (Nyagumbo, 2007). After attaining independence Zimbabwe evaluated
and promoted CA techniques such as no-till tied ridges, clean and mulch ripping, no-
till strip cropping and tied furrows (Vogel, 1992; Twomlowand Dhliwayo, 1999;
Twomlowand Bruneau, 2000; Nyagumbo, 2002). The research results indicated crop
yield, soil and water conservation benefits derived from using the developed CA
techniques (Smith, 1988; Vogel, 1992; Twomlowand Bruneau, 2000; Nyagumbo,
2002).
Despite the efforts made in developing and promoting CA techniques and the benefits
shown by research results, adoption in the smallholder farming sector has remained
low for various reasons that include the top-down approach used during the
development of the technologies as well as the lack of appropriate equipment for CA
practices (Nyagumbo, 2007). For example specialized equipment is required for
planting and weeding in unploughed and mulched fields (Friedrich and Kienzle,
2007). One of the major challenges to adoption of CA technologies developed for the
smallholder sector is the availability of adequate labour on the farm as well as farming
inputs. In semi-arid areas where livestock plays a more critical role than crops in the
livelihood of households, competition for crop residues between livestock and crop
sub-systems poses a threat to adoption of mulching (Palm et al., 1997). Jf higher
levels of production are reached then livestock would also benefit from the
improvements in crop production. The use of cereal residues and livestock manure
with high C:N ratio can promote temporary immobilization of soil nitrogen (Giller et
7
al., 1997; Palm et al., 1997; Abiven and Recous, 2007) resulting in depressed crop
yields.
2.2 Semi-arid smallholder cropping system
Crop production in semi-arid smallholder farming sector is strongly integrated with
livestock activities. The crop enterprise depends on livestock mainly for draught
power and manure. Draught power is required for land preparation and transport, and
donkeys are the major source of draught power in semi-arid southern Zimbabwe.
Cattle manure is available in relatively larger quantities than goat or sheep manure in
most parts of Zimbabwe (Mapfumo and Giller, 2001). However, in semi-arid districts
goat manure is predominant in some communities such as Kezi in Matebeleland South
(Aluned et al., 1997). Crops are grown for dual purposes namely household food and
as cash crops (Ncube, 2007). The cash is spent on paying school fees, buying
agricultural inputs (seed, fertilizer, farm equipment etc.) and food among other
household requirements.
The major crops grown include cereals such as maize (Zea mays L.), sorghum
(Sorghum bicolour (L.) Moeneh) and pearl millet (Pennisetum glaucum (L.) R.Br.)
and legumes namely bambara nut (Vigna subterranean (L) Verde), groundnut
(Arachis hypogaea) and cowpea (Vigna unguiculata (L.) Walp). Cereals are grown on
more than 70 % of cultivated area while legumes take up the remaining portion
(Twomlowet al., 2006b). Intereropping is a common practice with full crop rotation
rarely practised by smallholder farmers. Lack of adequate seed hampers the
production of legumes at a large scale in semi-arid areas (Ncube, 2007). Minor crops
usually grown together with cereals include sunflower (Helianthus annus L.),
8
watermelons (Citrullus lanatus (Thunb), sweet potatoes (Ipomoea batatas), sweet
sorghum (Sorghum vulgare Pers.) and pumpkins (Cucurbita maxima L.).
Land preparation often starts in October-November but this activity can be delayed
because draught animals are in poor condition at that time of the year (Shumba et al.,
1992). The recommended ploughing depth of 0.23 - 0.25 m is rarely achieved on
smallholder farms because of poor plough settings and other plough conditions
(Smith, 1988). Tsimba et al. (1999) observed that smallholder farmers managed a
0.11 - 0.15 m ploughing depth in a study conducted in Mutoko and Chinyika
communal areas. Small grains are planted first with maize being planted as late as
February. Legumes are normally planted during December and January. Planting date
is season dependant as the onset of the first effective rains signalling the start of
cropping period varies from year to year. Most weeding takes place within the first
two months after crop emergence (Van der Meer, 2000). Hand weeding or ox-drawn
cultivation followed by hand hoeing are common practices for weed control.
Frequency of weeding operations depends on the rainfall pattern for each season with
a single weeding being common in a dry cropping season and three or more weeding
in a wet year. Problem weeds include striga (Striga asiatica (L.) Kuntze) and couch
grass (Cynodon guyana L.).
In the low input crop production systems of the smallholder sector, maize and
sorghum yields are often below 1.0 and 0.2 tha" respectively (Twomlowet al.,
2006b). In the south western Tsholotsho district of semi-arid Zimbabwe yields of 0.4
tha" for cowpea, 0.5 tha" for pearl millet, 0.7 tha" for sorghum and 0.8 tha-1 maize
have been reported by the Department of Research and Extension (AREX) (Ncube,
9
2007). The Zimbabwe national yield averages for groundnuts and cowpea are 0.3 and
cereals is 0.6 tha-I (Nhamo et a/., 2003).
2.3 Climatic characteristics of semi-arid areas
Rainfall over southern Africa is associated with the movement of the Inter-Tropical
Convergence Zone (ITCZ) (Dennett, 1987). Following the movement of the sun, the
ITCZ and associated rainfall moves towards southern African countries during the
October to November period. The intensity of the !TCZ and the resultant rainfall
amounts over southern Africa are also influenced by topography and the El Nifio-
Southern Oscillation (ENSO) phenomenon (Hulme and Sheard, 1999; Clay et al.,
2003; Trenberth et al., 2007). There is a significant relationship between the ENSO
phenomenon and the inter-annual rainfall variability in some parts of southern Africa
(Hulme and Sheard, 1999; Clay et al., 2003). Regional sea surface temperatures (SST)
and latitude also have an influence on rainfall patterns in southern Africa (Clay et al.,
2003). Under normal circumstances the ITCZ often moves between Tanzania and
Zimbabwe and rarely goes beyond the Limpopo river in the south (Love et al., 2008).
Rainfall, occurring as short duration, high intensity convective storms is confined to
one distinct rainy season stretching from October to April (Unganai, 1996; Nonner,
1997). Generally the rainy season ends in March-April following the northward
retreat of ITCZ and reduced heating of the continent (Tadross et al., 2007). The high
spatial variability of rainfall in semi-arid regions arises from its convective nature
(Dennett, 1987). Typical coefficients of variation range from about 20 to 40 % in East
.Africa (Rockstrëm et al., 2003) and 34 to 44 '% in semi-arid areas of Zimbabwe
(Oosterhout, 1996; Chapter 4). Climatic variability is enhanced by climate change and
10
climatic phenomenon such as ENSO which increases the potential for drought
occurrence (Dennett, 1987; Phillips et al., 1998).
Statistically, complete crop failure in semi-arid areas due to droughts occurs almost
once every 10 years (Rockstrórn et aI., 2002). Analysis of rainfall data from stations
in semi-arid NR V of Zimbabwe showed a high probability (once every five years) of
crop failure for sorghum and pearl millet (Nyamudeza, 1998). A climatic study
conducted by Unganai (1996) using Global Climate Models (GCM) revealed that,
between 1900 and 1993, total precipitation over Zimbabwe declined by up to 10 % on
average.
Analysis of rainfall data has shown that water-related problems in semi-arid tropics
are often associated with intra-seasonal dry spells during critical stages of crop growth
rather than cumulative rainfall (Rockstrêm et al., 2003; Oosterhout, 1996). For
example, a study of five semi-arid districts in Zimbabwe by Oosterhout (1996)
showed that years with the highest total rainfall did not coincide with years of highest
crop yields. An earlier study by Unganai (1990) had indicated that approximately 480
mm of rainfall, well distributed throughout the growing season was sufficient for
successful production of sorghum and maize crops.
Research in east Africa showed that the probability of occurrence of a dry spell of two
to four weeks during the growing season far exceeds that of droughts (Rockstrêm et
al., 2002). In Machakos (Kenya), where seasonal rainfall is bimodal findings by
Barron (2004) indicated that the probability of occurrence of dry spells exceeding five
days was higher for the long rains (90 %) than for the short rains (78 %). These dry
11
spells are detrimental to crop yields if their occurrence coincides with critical phases
of crop development. In semi-arid regions, severe crop yield reductions due to dry
spells occur once or twice in every five years (Rockstrëm et al., 2003).
The uncertainties and variability of rainfall in semi-arid tropics formed a basis for
numerous studies on analysing the cropping seasons. In Zimbabwe, the anomalies and
generalizations associated with the pioneering work on agro-ecological classification
conducted by Vincent and Thomas (1960) has led to several reassessments. The
Department of Agricultural Technical and Extension Services (AGRITEX, 1990)
refined the classification by defining and mapping five agro-ecological regions based
on rainfall amount, distribution and altitude (Fig. 2.1). In northern Zimbabwe agro-
hydrological analysis of daily rainfall data from four stations (Binga, Kamativi, Bumi
Hills and Siabuwa) by Chiduza (1995) revealed that within the same NR, some sites
had a better cropping potential than others.
Despite advances in agroclimatology, forecast of seasonal rainfall distribution from
early events such as onset of the rainy season remains uncertain. Dennett (1987)
concluded that there is little evidence to suggest such relationships. Analysis of
rainfall data provides important information for agricultural management in rainfed
agriculture. The determination of start, end and length of a growing period, together
with the frequency and duration of dry spells is useful to farmers and land use
planners in crop selection, timing of farming activities and variety selection. By
incorporating soil water characteristics and yield reduction factors, yield estimates
also can be obtained (Barron, 2004).
12
~ Cities
Agro.regions
.Gt~ I above 10QQmm Very High
:::,1:·1;j:li.e;' 750-1 OOOmm Moderelely high
lib 750-1 OOOmm Moderetely high
III 65D.800mm Moderale
IV 45D.B50mm Feirly low
V below 450mm Too low end erreue
50 0 50 100 150 200 Kilometers
F+3 F+3 F+3
Figure 2.1. Agro-ecological regions of Zimbabwe (Source: ICRISAT GIS office,
2008)
An analysis of the amount and distribution of rainfall in NRs. II, IV and V of
Zimbabwe showed that the growing season averages 96 days (Hussein 1987 in Morse
1996). In NR V the season starts in early December and finishes by 24 March (Morse,
1996). False start of growing season is a common feature in semi-arid areas as early
rains are often followed by long dry spells before crop establishment (Dennett, 1987).
Regression analysis of rainfall data by Hussein (1987 in Morse 1996) showed that the
length of the growing season increased with earlier onset of the rainy season. At three
semi-arid sites in Botswana, Kanemasu et al. (1990) observed a similar relationship
(R2 = 0.76) between onset and duration of rainfall. These climatic features of the area
need to be considered when formulating recommendations for smallholder farming
systems.
13
2.4 Tillage techniques and other in situ strategies
Good soil water management in rainfed agriculture can be achieved through various
. tillage methods, both conventional and reduced, and rainwater harvesting teclmiques
or structures. Various researchers and development agencies have explored in situ
rainwater harvesting. For upgrading rainfed agriculture in semi-arid tropics
Rockstrom (2002) classified rainwater management methods into the following
categories:
(1) Systems that prolong the duration of soil moisture availability in the soil for
example mulching practices;
(2) Systems that promote infiltration of rainwater into the soil. These teclmiques
include pitting, ridging or furrowing and terracing;
(3) Systems that store surface and sub-surface runoff water for later use e.g. rainwater
harvesting systems with storage for supplementary irrigation.
Soil tillage, to be defined as 'The manipulation, generally mechanical, of soil properties
to modify soil conditions for crop production' (SSSA, 1987) has, through the ages been
applied in farming systems where natural vegetation is replaced by arable crops. The
majority of the soils (sands, loamy sands and sandy loam) in the drier regions of
southern Africa are poorly structured, self compacting and have a tendency to crust.
The primary functions of tillage in these areas are to prepare a seed bed, control weeds,
incorporate crop residues and manures, enable water infiltration into the soil and permit
the growth of roots down the profile, where soil water may be stored.
Throughout southern Africa the timing of primary tillage operations such that
cultivation is done during the post-harvest autumn period has been advocated as a
14
sound practice, as weeds are destroyed (Willat, 1967), draught animals are still in
good condition at the end of the growing season, and the soil surface is in a state
which permits infiltration of the early spring rains. However, few smallholder
households manage to achieve this post harvest ploughing, despite the obvious
advantages on some soil types (Grant et al., 1979; DLFRS, 1985). The reasons for this
are many and varied, and include the fact that the soils are too hard to plough at the
end of the season when dry, traditional practices in many areas is to allow livestock to
graze crop residues in situ over the winter period, other household activities such as
marketing of produce take priority in allocation of labour. To overcome the inability
to carry out an autumn ploughing, work in Botswana clearly demonstrated the
advantages of a double spring ploughing, increasing yields by as much as 71 % across
a range of soil types (Table 2.1). Similar results have been observed in Zimbabwe for
a range of soil types (Twomlowand Bruneau, 2000).
Table 2.1. Effect of double spring ploughing on sorghum grain yield for seven soil
types in Botswana in 1988/89, after Heinrich 1989
Adjusted grain yields kg ha-i at 10% moisture content
Soil Type shallow shallow calci ortbic deep cambic h.nvic
ferric ferrallic cambisol luvisol ferric arenosol arenosol
luvisol arenosol luvisol
Conventional 790 1996 1255 1059 606 1351 1365
ploughing
Double 1020 2674 1369 1273 682 2812 1470
spring
ploughing
From recent studies of smallholder farmers in southern Africa it is clear that farmers
already employ tillage practices with the aim of accomplishing several short-term
goals including seedbed preparation, weed control and rainwater retention (Morse,
1996; Twomlowand Bruneau, 2000). In Zimbabwe, as in Tanzania, it has been
reported that farmers recognise that timely inter-row cultivation is important for both
15
weed control and for maintaining a rough soil surface which can retain subsequent
rainfall.
2.5 Soil water management in semi-arid smallholder farming systems
The water balances can be formulated at different spatial scales (soil, field, farm,
catchment, basin) and temporal scales (days, weeks, decade, season, year). As shown
by field studies conducted by Klaij and Vachaud (1992) and Verplancke (1994) in the
. Sahel region (Niger), water balance analysis is important for rainwater management in
rainfed cropping systems. It can be used to determine the effect of management
systems on rainfall partitioning. The results are useful for adapting crop growth to
water availability in the soil. In rainfed systems, the water balance can be formulated
thus:
P + Ron = ET + Rof! - (C-D) +/- L1SW Equation 2.1
where
P = Precipitation (nun)
Ron = Surface runon into field (mm)
ET = Evapotranspiration (mm)
Roff = Surface runoff out of the field (nun)
C = Capillary rise (nun)
D = Deep percolation (drainage) (nun)
!1SW = Change in soil water over the considered time (mm).
16
Techniques that capture rainwater, reduce surface runoff and promote infiltration have
traditionally been used in Africa (Reij et al., 1996). In sub-Saharan Africa these
techniques include planting pits, infiltration pits, terracing, mulching, stone bunds and
ridges/furrows (Hudson, 1987). Elsewhere in Africa other systems exist such as the
Trus system in Sudan, the Zai system in Mali and Burkina Faso, and the Tassa system
in Niger (Mandiringana et al., 2003). In South Africa extensive research work has
been conducted on runoff in-field rainwater harvesting using basins and mulch under
semi-arid conditions (Botha et aI., 2003) and used by smallholder farmers in marginal
areas with maize and a variety of vegetables.
In Zimbabwe, water management under rainfed cropping systems has been the focus
of several field studies, encompassing the effects of cereal-legume rotations on soil
water (Ncube, 2007), weeding effects on soil water regimes (Twomlowand Bruneau,
1998), tillage effects on soil water dynamics and crop yields (Nyamudeza, 1993;
Nyakatawa et al., 1996; Nyagumbo, 2002), tillage effects on weeds (Dhliwayo et al.,
1995, and soil erosion control (Chuma, 1993; Chuma and Haggmann, 1998;
CONTILL, 1998). AGRITEX through the Lowveld Research Stations, Cotton
Research Institute, Makoholi Experiment Station and Agronomy Institute conducted
numerous studies on soil and water conservation systems from the mid-1980s.
At Matopos Research Station, Ncube (2007) monitored the residual soil water benefits
of four grain legumes (bambara nuts, cowpea, groundnut and pigeon pea) to sorghum
in a legume-sorghum rotation over three cropping seasons. Results from that study
showed that soil water dynamics in the 0 - 0.25 m soil layer during the sorghum
growing phase of the rotation depended on rainfall pattern and not the previous
17
legume crop. However, soil water content differences between the grain legumes were
observed in the 0 - 0.55 m profile of the sandy soil at the experimental site.
Improved soil water supply, rooting depth and crop yields through use of minimum
tillage techniques have been reported in semi-arid regions of Zimbabwe (Nyamudeza,
1993, Vogel, 1993). On sandy soils at Chivi (NR V) and Matibi I (NR IV) Nyakatawa
et al. (1996) observed a 22 - 85 % increase in maize yields due to tied furrows
compared to the farmers' practice of planting on the flat. The response was low for
sorghum with an average yield increase of 18 %. The yield increase rose to 35 - 115
% and 59 - 200 % for maize and sorghum respectively by adding inorganic fertilizer
although the interaction was not significant. On a sandy clay loam soil at Chiredzi
Research Station, tied furrows increased sorghum yield by 4 - 62 % in 4 out of 5
seasons compared to the farmer's practice of sowing on the flat (Nyamudeza, 1993).
In a very wet year, Nyamudeza et al. {l992), working on vertisols in south eastern
Zimbabwe, reported that tied furrows did not significantly affect sorghum yields in
years with adequate rainfall. Crop response to tied furrows and fertilizers was affected
by rainfall distribution within the season rather than total annual rainfall. In all cases,
significant increases in grain yields were observed in seasons with mediocre or good
rainfall distribution. Droughts and poorly distributed rainfall resulted in no positive
response to either tied furrows or fertilizers, forcing farmers to revert to their old
production practices.
The presence of crop residue mulch at the soil-atmosphere interface has a direct
influence on infiltration of rainwater into the soil and evaporation from the soil. Trials
conducted in the higher potential areas of Zimbabwe between 1988 and 1995
18
indicated that mulching significantly reduced surface runoff and hence soil loss
(Erenstein, 2002). Mulch cover shields the soil from solar radiation thereby reducing
evaporation losses from the soil surface. Soil biota increase in a mulched soil
environment thereby improving nutrient cycling and organic matter builds up over a
period of several years (Holland, 2004).
On a sandy loam soil in semi-arid Kenya, studies by Gicheru et al. (2003) revealed
that a combination of 3 tha' mulch cover and 3 tha-I manure promoted more
recharging of soil profiles with water under minimum tillage than conventional
ploughing. Nyagumbo (2002) monitored the effect of mulch ripping, no-till tied
ridging and conventional tillage on recharge of groundwater on a red clay soil. Mulch
ripping reduced surface runoff and resulted in higher soil water content in the top 0.45
m of the soil profile.
Bescansa et al. (2006) in semi-arid parts of Spain compared the effect of combining
surface mulch with zero tillage and conventional ploughing on soil water retention.
Results from that study showed development of more soil pores (6 - 9 urn pore
diameter) for water storage in the top 0.15 m of the soil profile under no-till plus
mulching. Conventional ploughing had pores predominantly greater than 9 urn which
are good for soil water drainage under gravity. Under any given suction, soil under
no-till with mulch cover retains more soil water than under conventional tillage
suggesting soil structural improvement (Bhattacharyya et al., 2006)
Excessive weed growth has been identified as one of the major factors limiting crop
production in smallholder farming systems of Zimbabwe (Dhliwayo et al., 1995).
19
Field studies have shown that weed transpiration has a profound impact on soil water
regimes in semi-arid environments (Van der Meer, 2000). Thus apart from in situ
rainwater. harvesting through tillage practices, weed control is an important water
conservation practice. Besides reducing unproductive water flows through weed
transpiration, weed control also reduces competition for nutrients and radiation
between crops and weeds. In an assessment of eight hand weeding treatments at
Makoholi Experiment Station (Masvingo in Zimbabwe), Twomlowet al. (1997)
reported that the control treatment (no weeding) had the driest soil profiles and lowest
maize yields. The weed free treatment had a significantly wetter soil profile than the
other 7 treatments. Studies conducted by Shumba et al. (1992) at Matopos and
Makoholi also showed that early weeding resulted in 40 % more maize grain than the
farmers' practice (control). The study also showed significant weeding x nitrogen
interaction effects on maize yields. Twomlowand Bruneau (1998) reported that crop
yields would be reduced if weed control is inadequate, even when water conservation
measures such as tied ridges are used. They therefore, concluded that weed
management is critical for maximizing soil water availability for crop growth.
Despite the role of in situ rainwater harvesting systems in improving water
availability to crops, their potential to mitigate dry spells and other problems of yield
reduction appears limited, especially on sandy soils characterized by low inherent
water holding capacity. In semi-arid Machakos district of Kenya smallholder farmers
cited insufficient rainfall as the major 'reason for low crop yields (Barron and
Rockstrërn, 2003), suggesting that water deficits are still being experienced. In
addition, most studies on in situ rainwater harvesting systems overlooked the problem
of soil fertility, which can curtail the benefits associated with improved soil water
20
status. Exceptions are studies by Nyakatawa et al. (l996) which investigated the
interaction of tied ridge/furrow system and soil fertility.
2.6 Soil fertility management in semi-arid smallholder farming systems
Continuous cropping without addition of nutrients and organic matter is a major threat
to sustainable crop production in sub-Saharan Africa (Waddington et al., 1994).
Unlike the effects of water stress due to droughts and dry spells, the effects of low and
declining soil fertility on crop yields are gradual. In Zimbabwe, most soils found in
the communal areas are sands, derived from granitic parent material (Grant, 1981).
The soils have low inherent fertility, poor water holding capacity and low organic
matter (Grant, 1981). In Africa it is reported that nitrogen (N) is lost at a rate of more
than 30 kgNha-1year-1 from the smallholder cropping systems (Hikwa et al., 2001).
Over a period of five years, Plucknett (l994) reported a 21 and 14 % decline in soil
organic carbon and phosphorus. Due to rapid population growth and limited arable
land, traditional soil fertility management practices such as shifting cultivation are no
longer feasible (Muller-Samann and Kotschi, 1994).
A soil fertility mapping study involving 3 500 smallholder farmers in Zimbabwe
during 1994/5 and 1999/2000 seasons showed that N and P were limiting in 70 % of
the soils (Hikwa et al., 2001). An assessment of nutrient status of 250 soil samples
submitted to Soil Analysis Laboratory of the Chemistry and Soil Research Institute
revealed serious N deficiencies in 82 % of the soil samples (Nyamangara et al., 2000).
Mapfumo and Mtambanengwe (l998) reported similar results in a survey of ten
smallholder farm sites dominated by continuous maize cropping systems. In the same
21
study by Mapfumo and Mtambanengwe (1998), nutrient losses through removal of
crop residues accounted for 44 % of the N losses from the cropping system.
Despite the low and declining soil fertility, inorganic fertilizer use in smallholder
farming systems is low. A survey conducted in semi-arid Chivi district (NR V) of
southern Zimbabwe showed that 59 % of smallholder farmers did not apply fertilize~
or manure (CIMMYT, 1993). In 1983, it was estimated that inorganic fertilizer use
under smallholder cropping systems in Sub-Saharan Africa was about 5 kgha"
(Plucknett, 1994), while in Zimbabwe, a country with a high fertilizer use by 1983,
average rates monitored by CIMMYT (1993) during the 1992/93 season were 30
kgNha-I and 10 kgP205ha-I in NR N and 12 kg Nha-I and 3 kgP205ha-I inNR V. The
decrease in rates from NR N to the drier NR V was attributed to variability of
rainfall, which makes investment in fertilizers risky. Fertilizer application was higher
when government or Non-Governmental Organizations (NGOs) provided free
fertilizer. At Beitbridge, a participatory rural appraisal (PRA) conducted by the South
Eastern Dry Areas Project (SEDAP) (2001) highlighted that farmers did not consider
soil fertility a problem in their farming system. Accordingly, a follow-up focused
diagnostic survey showed that only 5 and 7 %, of smallholder farmers in Gwanda and
Beitbridge districts respectively used fertilizers (Nyamudeza, pers. com.). A similar
trend was observed by Nyamudeza et al. (pers. cam.) in a survey of five districts in
NRs N and V. A survey conducted by ICRISAT in southern Zimbabwe revealed that
40 % of the interviewed farmers acknowledged declining soil fertility as a problem
(Ahmed et al., 1997).
22
As soil fertility in smallholder farming systems continues to decline the problem is
further aggravated by inherent low fertility of some of the soils and low fertilizer
inputs. Nutrient balance analysis provides a basis for maintaining soil fertility
(Archer, 1988) and is crucial for identifying unsustainable land use systems
(Bindraban et al., 2000). A positive nutrient balance indicates an improvement in soil
fertility. Sahel studies have shown that smallholder farming systems lack the internal
capacity to replenish nutrient losses due to crop biomass off take for use by the
livestock part of the farming system (Powell et al., 1996). As a result nutrient
balances with regards to N are negative in most cases (Bosch et al., 1998). However,
few field studies have been conducted to determine the nutrient balances and fluxes in
smallholder farming systems. Pilbeam et al. (2000) attributed this lack of studies on
nutrient balances to difficulties in quantifying some of the components (for example
leaching and gaseous losses of nitrogen). Attempts to quantify all nutrient inflows and
outflows include lysimeter studies (Archer, 1988; Haggmann, 1994) and lSN isotope
analysis (Archer, 1988).
Studies have shown that in drought-prone environments, crop response to fertilizer is
highly dependent on seasonal distribution of rainfall (Shumba et al., 1992; Nyakatawa
et al., 1996; SEDAP, 2001). Fertilizer recommendations for semi-arid areas of
.Zimbabwe have been based on research conducted in low-risk high rainfall areas. In
some cases, farmers' experiences with inorganic fertilisers have not been encouraging
due to lack of knowledge and skills in using inorganic fertilisers (Lee, 1993) resulting
in crop bum, lack of response and high labour costs. In some remote semi-arid areas,
farming inputs including fertilisers are difficult to access when required. Investment
of scarce resources into inorganic fertilisers is therefore risky particularly at the
23
current recommended rates of 300 kgha' ammonium nitrate (34.5 % N) and 300
kgha' compound D (8 % N, 14 % P205, 7 % K20).
2.7 Water use efficiency in semi-arid smallholder systems
Agricultural productivity in arid (precipitation/potential evapotranspiration < 0.2) and
semi-arid (0.2 < precipitation/potential evapotranspiration < 0.5) environments is
limited by water availability (Debaeke and Aboudrare, 2004). Greater water use
efficiency can be achieved through elimination of biophysical and socio-economic
constraints facing smallholder farmers. In smallholder agriculture these constraints
include timely planting and weed control, labour availability, balanced soil nutrient
management, use of recommended crop type and varieties, and availability and
accessibility of fanning inputs to farmers.
In rainfed smallholder agriculture timely planting is crucial in order to make full use
of early rains in the season. Late planting at times leads to no crop harvests as seasons
end abruptly. Weed control is probably the greatest challenge in improving water use
efficiency in semi-arid cropping systems. Weed transpiration leads to more rapid
depletion of soil water thereby exposing crops to severe soil water stress during dry
spells or drought (Van der Meer, 2000). Studies conducted at Makoholi Research
Station (NR IV) of Zimbabwe revealed that weeding at 2 and 6 weeks after maize
crop emergence significantly increased water use efficiency compared to 4 weeks
after emergence (Twomlowand Dhliwayo, 1999). Maize (Zea mays L.) water use
efficiency increased from 3.9 to 5.0 kgha'lmm" with two weedings (at 2 and 6 weeks
after crop emergence) compared to one weeding at 4 weeks after emergence.
24
Balanced soil nutrient management ensures efficient use of soil water at each stage of
crop growth (Bouman, 2007). Adequate soil nutrient supply promotes early
development of crop canopy leading to better interception of solar radiation which in
turn increases root development and allocation of root -extracted water to transpiration
(Sivakumar and Salaam, 1999). In semi-arid West Africa application of 30 kgNha-1
and 45 kgP20sha-1 increased both yield and water use efficiency of pearl millet
(Sivakumar and Salaam, 1999). Studies by Ncube (2007) showed that maize yields
and water use efficiency can be significantly increased by application of small rates of
manure (3 tha") and N (lO kgNha-l) in a semi-arid environment where farmers do not
traditionally use the two nutrient inputs.
2.8 Current risk mitigation strategies in semi-arid southern Zimbabwe
Efforts to reduce the impacts of drought and dry spells are underway in the semi-arid
regions of sub-Saharan Africa. The current initiatives aim at harvesting rainwater, and
managing soil water and fertility on smallholder farms. This will lead, hopefully, to a
reduction in the risk of total crop failure, improve water productivity and crop yields.
Several international and national organisations are spearheading the spread of
technologies that have shown potential for semi-arid smallholder agriculture.
International organisations such as ICRISAT and CIMMYT, NGOs such as World
Vision, Catholic Development Commission (CADEC) and Practical Action are all
taking part in the development activities sponsored by Department for International
Development (DFID) and European Commission Humanitarian Organisation (ECHO)
(Twomlowet al., 2006b). The development work is taking a participatory approach
giving the smallholder farmers and extension officers an equal-partner-status at all
stages of the activities in their communities.
25
Water management techniques at a field scale in semi-arid southern Zimbabwe
include ordinary contours, dead level contours with underground storage tanks and
infiltration pits (Mwenge Kahinda, 2004; Mupangwa et al., 2006). In situ water
management techniques include planting basin tillage and deep winter ploughing soon
after crop harvest. The microcatchment structures include dead level contours with or
without infiltration pits (Plate 2.1) and ordinary contour ridges « 5 % slope).
Between the fields dead level contours harness water originating from the area
upslope. The dead level contours vary in size; the minimum having a cross-sectional
dimension of 1.5 m width and 0.5 m depth (Mwenge Kahinda, 2004). The length of
each contour varies according to the length of the field. Currently soil water and crop
yield benefits derived from dead level contours with or without infiltration pits have
not been quantified.
Plate 2.1 Dead level contours with storage tanks and infiltration pits in ward 17 of
Gwanda district
The promotion of in situ water management techniques such as planting basins and
ripper tillage systems began during the 2004/05 cropping season in eight districts of
semi-arid Zimbabwe. The program was built upon the seed and fertilizer relief
26
program sponsored by DFID (Twomlowet al., 2006b). Officials from AGRITEX and
NGOs received training on concepts of conservation agriculture. The initial training
workshop. had a strong practical focus.with participants being trained in site selection,
laying out demonstration plots with emphasis being placed on spacing, digging the
basins, fertilization, planting and weeding. Results from paired plot demonstrations
(planting basins against conventional ploughing) showed an increase (P < 0.001) in
cereal yields through the use of planting basin tillage across five of the eight districts
(Twornlow et aI., 2006a). The significant cereal yield increase was attributed to
timeliness of planting and better management of planting basin system compared to
conventional practice. For the semi-arid southern Zimbabwe environments, simulation
modelling using Agricultural Production Simulator (APSIM) showed that combining
rainwater harvesting technologies with organic and inorganic fertilizers could reduce
the risk of crop failure from 20 to 7 % (Mwenge Kahinda, 2004).
The drip irrigation technology was introduced to improve household food security and
incomes in semi-arid communities (Moyo et al., 2006). The United States Agency for
International Development (USAID) funded program was introduced with the.aim of
reducing the negative impact of water shortages and increase productivity of the
available limited water resources. Moyo et al. (2006) assessed the sustainability of
drip irrigation in home gardens in Gwanda and Beitbridge districts of southern
Zimbabwe. The study revealed that beneficiary farmers encountered challenges that
included availability of adequate water during the dry season, availability of drip kit
spare parts and lack of technical back up support during the use of the drip kits. Only
. 5 % of the beneficiary farmers in Beitbridge district had used their kits at the time of
study by Mayo et al. (2006). In Beitbridge 49 % of the water sources available for
27
drip irrigated farming had dried up when the survey team visited the district in
February 2004 while Gwanda district had lost 5 % of its water sources (Moyo et al.,
2006). Lack of co-ordinated approach by relevant stakeholders negatively impacted
on the sustainability of the drip irrigation programme.
A review of literature from Zimbabwe and other southern African countries shows
that technologies have been developed for improving crop and water productivity
under rainfed smallholder farming conditions. Currently the planting basin and ripper
tillage systems, and dead level contours are being promoted in semi-arid southern
Zimbabwe. However, technical information on the contribution of these soil water
management technologies is still lacking. Integration of the soil water and fertility
management is a potential key to improved soil and crop productivity in the
smallholder farming system. This study focuses on quantifying soil water derived
from in situ and between-field rainwater management practices for drought mitigation
in water-scarce environments such as the Mzingwane catchment of the Limpopo
Basin. The study goes further and explores the effect of integrating soil water and
fertility management practices on crop yield under semi-arid smallholder farming
conditions as well as documenting the farmers' opinion of the interventions.
28
CHAPTER3
General Materials and Methods
. 3.1 Site descriptions
3.1.1 On-farm sites
Insiza and Gwanda districts lie in the Mzingwane catchment which is part of the
Limpopo river basin. Insiza district lies in NR IV which is characterized by semi-arid
climatic conditions with total annual rainfall ranging between 450 and 650 mm (FAO,
. 2006). Gwanda district lies in NR V which receives annual rainfall of less than 450
mm. In both districts, the rainfall season is unimodal and begins in
November/December and ends in March/April. The cropping season experiences
periodic dry spells particularly in January and is followed by a cool to warm dry
season from May to September.
The predominant soils in Insiza and Gwanda districts are coarse-grained sands to
loamy sands and clay loams to clay with minor occurrences of vertisols (Anderson et
al., 1993). The soils are classified as 2, 4P and 5G according to the Zimbabwe Soil
Classification system, Eutric/Dystric Regosols and Chromic Luvisols according to
FAOIUNESCO classification, and as Ustalfic Haplargid and Lithic/Ustic Torriorthent
according to Soil Taxonomy (Nyamapfene, 1991; FAO, 2006). Landform is almost
flat to undulating pediplain with some local hills and rock outcrops. Vegetation
consists of Colophospermum mopane as the dominant tree species with scattered
associated tree species of Commiphora spp, Combretum apiculatum and Adonsonia
digitata.
29
3.1.2 On-station sites
Matopos Research Station lies at 28°30.9iE, 20023.3is and 1 344 m above sea level.
Lucydale experimental site lies at 28°24.46'E, 20025.64'S and 1 378 m above sea level.
Matopos Research Station is located in NR IV, which is characterized by semi-arid
climatic conditions and rainfall season is unimodal and begins In
November/December and ends in March/April. The long-term average rainfall for
Matopos and Lucydale is 590 mm (Ncube, 2007). As in Insiza and Gwanda districts,
the cropping season at Matopos and Lucydale also experiences periodic dry spells
particularly in January. The cool period stretches from May to August with a rise in
temperatures experienced as from September.
The red clay soil at Matopos Research Station is classified as a siallitic soil (4E.l) and
Chromic-Leptic Cambisol according to the Zimbabwean and FAO systems (Moyo,
2001). The internal drainage of Matopos clay soil shows signs of saturation for short
periods during the rainy season and external drainage is characterised by slow surface
runoff (Moyo, 2001). The granitic sand at Lucydale is classified in the Zimbabwean
system as moderately deep to deep well-drained fersiallitic soil (5G.2) (Nyamapfene,
1991) and is classified as Eutric Arenosol by FAO (1998). The chemical and physical
properties of the two soil types are described in Table 9a.2.
3.2 Focus group discussions and resource flow mapping
3.2.1 Focus group discussions
The first focus group discussions were conducted in Gwanda and Insiza districts of
Matebeleland South province of Zimbabwe in March 2006. The researchers first
sought permission from the District Administrators of Gwanda and Insiza districts so
30
that meetings could be held with farmers. After permission was granted, dates for the
resource mapping were agreed upon with farmer representatives. In Insiza district
farmers who participated in the focus group discussions were drawn from Mpumelelo
and Masiyepambili villages (Plate 3.1). In Gwanda district participants came from
Fumukwe, Humbane, Magaya, Mnyabezi D and Mnyabetsi S villages of ward 17
(known as Manama). Experimental details and results for soil water and fertility
management assessment are presented in Chapter 5.
The focus group discussions held at the end of 2007/08 growing season aimed at
evaluating the performance of single and double ploughing, ripping and planting basin
tillage systems with farmers who were hosting the experiments. Two focus group
meetings were convened, one in ward one of Insiza district and another in ward 17 of
Gwanda district (Plate 3.1). The initial intention was to convene a third focus group
meeting with farmers who had been praeticing single conventional ploughing and
planting basin tillage systems in Matobo district. However, this could not take place
because of the prevailing political situation in Zimbabwe between March and July
2008. The experimental details and results for the evaluation of technologies are
presented in Chapter 6. A checklist of guiding questions used during the focus group
discussions is given in Appendix 6.1.
31
V\'·'I;~\''.(' ·\;."~.·. . '~" :(~..~...
: .. ~~t -~.-~,,~.r~~~'" 'ti.' ' 1~'Q
, <;, f 'Wl,'f" J~.~ J_:
Plate 3.1. Focus group discussions in progress at (a) Humbane and (b) Mpumelelo
villages of Gwanda and Insiza districts
3.2.2 Resource now mapping
At the end of the 2005/06 growing season six farms hosting field based research trials
were visited and the flow of agricultural resources was mapped on each farm. Four
farms that were not hosting research trials were also visited and mapped during the
same week. At the end of 2007/08 nine farms hosting research trials were visited and
flow of agricultural resources was mapped (Plate 3.2). Out of the nine farms hosting
trials three farms had not been visited at the end of 2005/06 growing season. Two of
the four farms without research trials were revisited also at the end of 2007/08 season,
the other two could not be visited because family heads were not available at the time
of the resource mapping exercise. The experimental details and results are presented
in Chapter 6 and names of participants are given in Appendix 6.2.
32
Plate 3.2 (a) Mrs. Sibanda of Mnyabezi D village of Gwanda district drawing her field
map on the ground and (b) ICRISAT researchers interviewing Mr. Siza Mguni of
Insiza district in April 2008
3.3 Rainfall analysis
3.3.1 Meteorological stations
Five meteorological stations that lie within the semi-arid southern Zimbabwe were
selected for analysis of daily rainfall data collected over varying periods. The stations
were considered because they lie within the Mzingwane river subcatchment and
districts with sites for agronomic trials conducted in our study (Fig. 3.1). Daily
rainfall data were obtained from the Zimbabwean Department of Meteorological
Services. The different locations of the meteorological stations and their geographical
descriptions are shown in Table 3.1. The rainfall analysis details and results are
presented in Chapter 4.
33
N
A
tare
D Cities
• Farmer elte s
Agroecologleal roglons
t;nm I above 1000mm
~ 1I.750-1000mm
lib 750-1000mm
III 650·800mm
IV 450-600mm
V below 450mm
50 0 50 lOO 150 200 Kilomeiers
E=3 E3 E3
Sou",e:ICRISATGIS Let; 200B
Figure 3.1 Agro ecological regions of Zimbabwe and location of experimental sites in
Insiza and Gwanda districts, Matebeleland South province
Table 3.1. Geographical description of meteorological stations and rainfall database of
the five stations used in the analyses
Station Latitude Longitude Data period Duration of
data set
Bulawayo -20.22 28.62 1930-2001 71
Matepos -20.38 28.50 1939-2008 69
Mbalabala -20.35 28.95 1931-1994 64
Filabusi -20.55 29.28 1921-1995 74
Beitbridge -22.22 30.00 1951-2001 50
3.4 On-farm experiments
3.4.1 Seasonal rainfall
For the on-farm experiments, each farmer was provided with a plastic raingauge for
measuring daily rainfall (Plate 3.6). Farmers were first trained by ICRISAT
researchers on how to use the plastic raingauge correctly and keep the rainfall records.
Each farmer was provided with a book for keeping records of rainfall and agronomic
34
activities on the experimental fields and the record books were collected by ICRISAT
researchers at the end of each season.
Plate 3.6. An ordinary raingauge installed at Mr. John Ncube's homestead in
Fumukwe village of ward 17 (Manama), Gwanda district
3.4.2 Farmer selection
3.4.1.1 Insiza district
The host farmers for the research plots were chosen from a group of farmers that had
been formed by World Vision for the Conservation Agriculture programme. Each
farmer had a field that was fenced, accessible, had uniform soil type and adequate
land to host the research trials. Three farmers namely Mpofu, Moyo and Nkomo
hosted the trials in the first season. The number was increased from three to seven in
the 2006/07 growing season. The seven farmers were drawn from Masiyepambili,
Mpumelelo and Thandanani villages of ward 1 and the same farmers were maintained
35
in 2007/08 growing season. The names of the farmers who hosted the experiments are
given in Appendix 7.1.
3.4.1.2 Gwanda district
Two groups of farmers were selected in ward 17 (also known as Manama) to host the
two experiments that were run in the district. The farmers who hosted the tillage and
nitrogen management experiment were selected based on land availability, soil type
.uniformity, accessibility and protection of field site. Three farmers namely Sibanda,
Siziba and Ncube hosted the trials in the 2005/06 season and the number was
increased to seven in 4006/07 growing season and then retained in 2007/08. Farmers
were spread over four villages namely Mnyabezi D, Mnyabetsi S, Fumukwe and
Magaya of ward 17. The farmers who hosted the second experiment, which was
quantifying the soil water supply from dead level contours, were chosen by village
representatives of the rainwater harvesting project that is being promoted in ward 17
by Practical Action.
3.4.2 Experimental layout
3.4.2.1_Experiment 1
The experiment consisted of four tillage treatments that were combined factorially
with three nitrogen application rates. The four tillage treatments were hand dug
planting basins (Basins), tine ripping (Ripper), single (CP) and double (DP) ploughing
(Plate 3.4). The three nitrogen rates were 0, 10 and 20 kgblha' applied as ammonium
nitrate (34.5% N). The experiment was established at five farms in the 2005/06
growing and the number of farms was increased to 14 farms in 2006/07 season. The
36
14 farms used in the 2006/07 growing season were maintained in the 2007/08 season.
The experimental details and results are given in Chapter 7.
Plate 3.4. (a) Conventional ploughing (b) hand dug planting basins (c) ZimPlow
ripper tine used for ripping tillage treatment at on-farm and on-station experimental
sites and (d) Rip furrows opened by donkey/ox drawn ripper tine
Surface runoff was measured at four farms during 2006/07 and 2007/08 growing
seasons. Runoff plots measuring 10 m x 10 m were established for each tillage
treatment at each farm (Fig. 3.2). Each farm had a total of four runoff plots
corresponding to each of the four tillage treatments. The boundaries of the runoff
plots were constructed using a heat resistant black polythene plastic that prevented
water flow from other parts of the field onto the runoff plot. At the down slope side of
37
each runoff plot a triangular surface of dimensions 10m x 1.7 m was used to receive
runoff water from the plot and lead it into 210 litre drum through a polythene plastic
gutter.
10 m
Runoff plot lOm
1.7m
Plastic gutter
210 litre drum
Figure 3.2. Schematic diagram of the set up of equipment for collection of runoff
water from the plots under single and double conventional ploughing, ripper and basin
systems
3.4.2.2 Experiment 2
This experiment consisted of a transect of access tubes across dead level contours
with and without infiltration pits in Gwanda district (Plate 3.5 and Fig. 3.3). In the
2006/07 season one transect of six access tubes, two upslope and four downslope of
the dead level contour were installed for soil water monitoring. In the 2007/08 season,
38
a second transect was established in order to separate the effect of conventional tillage
from that of dead level contours on soil water content measured in the field. The area
around access tubes of the second transect was left undisturbed and no crop was
planted along this unploughed transect throughout the growing season. On the
downslope side of contour access tubes were installed at 3, 8, 13 and 18 m from the
centre of the dead level contour. On the upslope side of the contour access tubes were
installed at 2 and 7 m from the centre of the contour. Weeds around access tubes of
the second transect were controlled by spraying gramoxone herbicide. Further details
and the results of this experiment are presented in Chapter 8.
Plate 3.5. Dead level contour with (a)
Gwanda district
39
I IAccess tubes tt UpslopeJ<. ~·7m
Dead level contour
~
1 ~
)
~
1Sm Downslope
I~
6
Figure 3.3. Schematic diagram of the set up of access tubes across dead level contours
in Gwanda district
3.6 On-station experiment
3.6.1 Experimental layout
The experiment was run for two growing seasons at the Lucydale sandy soil site and
four seasons at the Matopos clay soil site. At the Lucydale site, the experiment was
run during 2004/05 season and the 2005/06 growing season. At the Matopos site, the
experiment was run from the 2004/05 season to 2007/08 growing season. The
experiment consisted of three tillage treatments, namely conventional ploughing,
ripper and planting basins, combined factorially with seven levels of mulch cover (0,
0.5, 1,2,4, 8 and 10 tha") (Plate 3.7). As part of a planned rotation, maize, cowpea
40
and sorghum were grown at the Matopos site between 2004/05 and 2007/08 seasons
and a new field was opened in each season (Table 3.2). A sole maize crop was grown
at the Lucydale site during the 2004/05 to 2005/06 growing seasons. Crop yields, soil
water content and soil physical and chemical properties were measured in the
experiment. The experimental details and results are presented in Chapters 9a, 9b and
9c.
Plate 3.7.Maize growing in the basin tillage system and under 4 tha-1 mulch cover at
the Matopos clay soil site in March 2008 during the 2007/08 season
41
Table 3.2. Experimental fields used and crops grown in each field from 2004/05 to
2007/08 seasons at Matopos Research Station
Field number Period under CA Season(s) field was Crop sequence
(seasons) used
1 .. . 1... . . 2007/08 Maize (M)
~[~~i~~~J~~~$~~:h~~f~J~m~~'~]iiL:1~~;S~i11i~.~$rJ~.~~ï:~i~adTL1f~~~Z~UY{~l~!~f~~~1t~?~f~~~~j~likl~l~$~\~i~'i':'Yt'~l~~~~~~1\1ff~áJ~~~~'it~~~'1~j~f;~~~\~U~i~:~~'f~~~~~1f~li~~;';~ffl~r!,á~l~~ff.~~t@~
2 2 2006/07 and Maize - maize
2007/08 (MM)
lIill.'1.111Itlllli~EtBltiliDiEi.i.~i':I'iIlli:'Il_.1
3 3 2005/06, 2006/07 Maize - cowpea-
and 2007/08
4 4 2004/05, 2005/06, Maize - cowpea -
2006/07 and sorghum - maize
2007/08 (MCSM)
3.6.2 Crop yield, soil water and infiltration measurements
Crop yield from each treatment was determined from a netplot consisting of five
middle rows with a running length of 6 m. In the 2004/05 and 2005/06 seasons, soil
water was measured using the gravimetric method. Soil samples were collected from
on-farm and on-station experimental sites using equipment shown in Plate 3.8. During
the 2006/07 and 2007/08 seasons, soil water measurements were taken using the
microgopher capacitance probe (Plate 3.9). During April 2008, infiltration
measurements were taken from selected tillage and mulch cover combinations at
Matopos in all the four fields used in the experiment. A minidisk infiltrometer was
used for measuring infiltration (Plate 3.10). The details of soil water and infiltration
measurements and results are presented in Chapters 7, 8, 9a, 9b and 9c.
42
Plate 3.8. Soil sampling equipment used for collecting soil samples for gravimetric
soil water determination in 2005/06 growing season
Data
Plate 3.9. Capacitance probe used for soil water measurement during 2006/07 and
2007/08 seasons
43
Plate 3.10. Minidisk infiltrometer used for measuring infiltration in fields exposed to
conservation agriculture practices for one, two, three and four growing seasons at
Matopos site in April 2008
3.7 Simulation modelling
On-farm crop and soil data showed a lot of variability across the farms used during.
the three seasons of experimentation. These data sets were therefore considered not
suitable for calibrating APSIM. Crop and soil data collected from on-station
experiments were used for evaluating the performance of APSIM model as there was
less variability in the datasets. The APSIM model was tested for two soil types, clay
and sand, to check how close the model could predict the observed soil and crop data.
The light textured soil at Lucydale has properties similar to those found in granite-
derived sandy soils of the smallholder farming areas in Gwanda and Insiza districts of
southern Zimbabwe and was therefore used for the long term simulation in the present
study. Daily rainfall, temperature and radiation data were collected from Matopos
Research Station weather station which is located 3 km from Matopos experimental
44
site and up to 10 km from the Lucydale site. The climate record used for APSIM
calibration stretched from 1 October 2004 to 30 June 2008. The details of the setting
up of the APSIM model.and results from the simulations are given in Chapter 10.
45
CHAPTER4
Characterisation of Rainfall Pattern for Improved Crop Production
in Semi-Arid Cropping Systems of Southern Zimbabwe
4.1 Introduction
Total annual rainfall in semi-arid southern Zimbabwe is less than 500 mm (against
national average of 657 mm) and exhibits spatial and temporal variation (Unganai,
1996; Phillips et al., 1998). Total annual rainfall decreases from 800 mm in the north-
east of Zimbabwe to less than 400 mm in the south (Unganai, 1996). Zimbabwe has a
unimodal rainfall pattern falling between October and April (Unganai, 1996; Phillips
et al., 1998). The rainy season reaches the peak during January and February, and
distribution during the season depends on the interplay between tropical and mid-
latitude weather systems (Clay et al., 2003). In semi-arid southern Zimbabwe the
rains are poorly distributed during the growing season and often occur as high
intensity convective storms (Rockstrëm and Falkenmark, 2000). Typical coefficients
of variation range between 20 and 40 % in semi-arid NRs IV and V of Zimbabwe
increasing with decreasing seasonal rainfall averages (Oosterhout, 1996).
Smallholder agriculture in semi-arid southern Zimbabwe is heavily dependant on the
seasonal characteristics of the rainfall. An analysis of the amount and distribution of
rainfall in NR II, IV and V of Zimbabwe showed that the cropping season averages 96
days (Morse, 1996). InNR V the season starts early December and finishes in March
(Morse, 1996). Analysis of rainfall data from northern Zimbabwe (Sebungwe)
indicated that the length of the growing season increased with earlier onset of the
rainy season (Chiduza, 1995). A strong relationship (R2 = 0.76) between onset and
46
duration of rainfall has been reported in some parts of semi-arid southern Africa
(Kanemasu et al. (1990).
In semi-arid southern Zimbabwe a false start of rain during the November and
December period may be followed by a long dry spell and this can be fatal to crop
establishment (Stern et al., 1981; Dennett, 1987). Water-related problems in semi-arid
smallholder agriculture are often complicated by occurrence of intra-seasonal dry
spells during critical stages of crop growth (Rockstrëm et aI., 2003; Oosterhout,
1996). Intra-seasonal dry spells are recurrent in southern Zimbabwe, often coinciding
with anthesis stage of commonly grown, cereals (maize, sorghum, pearl millet).
Approximately 480 mm of rainfall, well distributed within the growing season, is
sufficient for successful production of sorghum and maize crops (Unganai, 1990).
Analysis of rainfall data provides important information for agricultural management
in semi-arid cropping systems. The determination of start and end of summer rainfall
season, and length of growing season, together with the frequency and duration of dry
spells is useful to farmers in selecting suitable crop types and varieties, and timing of
farming activities.
4.2 Objectives
This study was designed to determine
(1) trends oftotal annual rainfall;
(2) the start and end of growing season;
(3) trend of wet days per growing season;
(4) probability of dry spells during growing season and
47
(5) behaviour of first and second halves of the growing season in semi-arid southern
Zimbabwe.
4.3 Materials and Methods
4.3.1 Data analyses
4.3.1.1 Testfor homogeneity of total annual rainfall data
Rainfall data, based on the July to June calendar, were subjected to a test of
homogeneity using the Run test (Freund, 1979; Dixon and Massey, 1983). A Run is
defined as a sequence of like events or symbols that are preceded and followed by an
event or symbol of a different type or by none at all (Freund, 1979). The Run test was
conducted on total annual rainfall for each station based on the July to June calendar.
The median of the total annual rainfall data was calculated at each station. The
median was then subtracted from each total annual rainfall value in the data series.
The number of times the data would make a run above (positive) or below (negative)
the median was counted. This gave the persistence of positive and negative values
from the annual rainfall data series. Significance tables were used based on the
number of runs (denoted by 'u') and number of positive values (NA) to determine
thresholds for homogeneity.
4.3.1.2 Standardized precipitation index (SPI)
To assess the trends in drought throughout the periods under review at each station,
standardized precipitation indices for each station were calculated using the following
procedure outlined by Le Barbe et al. (2002):
SPI = (P-M) / S Equation 4.1
48
where P is total annual rainfall for each year in the data series,M is mean rainfall for
the data series and S is the standard deviation for the data series. The categories in
Table 4.1 were adopted for classifying the data series according to wetness and
dryness.
Table 4.1. Drought classification indices adapted from Hayes et al. (1999)
SPI value Drought category
2.00 and above Extremely wet
1.50 to 1.99 Very wet
1.00 to 1.49 Moderately wet
-0.99 to 0.99 Near normal
-1.00 to -1.49 Moderately dry
-1.50 to -1.99 Severely dry'
-2.00 and below Extremely dry
4.3.1.3 Dry and wet days
Meteorologically a wet day definition of 'any day with more than 0.85 mm
accumulated in 1or 2 days' might suffice. However, for crop production purposes in
a region experiencing daily pan evaporation of 5 - 8 nun (Woltering 2005) rainfall of
0.85 nun has very limited influence on crop growth. Stem et al. (2003) suggest that
for other purposes a threshold of 4.95 nun can be used to define a wet day. For the
purposes of our study in semi-arid southern Zimbabwe the following definitions were
used in the analyses of dry and wet days;
o Dry day: Any day that accumulates less than 5 nun of rain in 1 day
o Wet day: Any day that accumulates 5 nun or more in 1 day
4.3.1,4 Start and end of growing season
The start and end of the growing seasonwere defined as;
Cl Start: the first day after 1 October when the rainfall accumulated over 1 or 2
days is at least 20 mm (Stem et al., 2003)
49
o End: the last day before 30 June that accumulates 10 mm or more rainfall in 1
day
The cut off point for end of season catered for late maturing or late planted crops.
After 1 June temperature normally drops quite significantly and crop growth rate is
slowed down. Given the above definitions Instat Statistical programme (Version 3.33)
(Stem et al., 2003) was used to analyse the rainfall data for start and end of season,
and length of the growing season. The F- and t-tests, to confirm significance of
change in total annual rainfall, number of wet days per growing season, start and end
of growing season, were conducted using Genstat Discovery Edition 3
(}j,ww. vsni. co. uk). Regression analysis was conducted to determine the relationship
between start and length of growing season.
4.3.1.5 Dry spell analysis
Daily rainfall data for each station was fitted to the simple Markov chain model as
outlined by Stem et al. (1982; 2003). The Markov chain model was run so that it will
give the probability of getting 14 and 21 day dry spells using the July to June
calendar. The analyses were performed using Instat Statistical Programme (Version
3.33, http://www.reading.ac.uk/ssc/software/instatlclimatic.pdf.) (Stem et al., 2003).
4.3.1.6 Cumulative distribution functions for first and second halves of rainy season
In-season rainfall patterns normally vary between the first and second halves of the
growing season. To assess the rainfall pattern in the two halves of the growing season,
the rainy period was divided into two, first half covering the October to December
period and second half starting in January and ending in March. Three monthly
50
rainfall totals were used to determine the below-normal, near-normal and above
normal rainfall ranges during each half of the growing season.
4.4 Results and Discussion
4.4.1 Testfor homogeneity of total annual rainfall data
The Run test showed a fairly equal persistence of counts below and above the median
rainfall amount for each station (Table 4.2). At Mbalabala and Beitbridge there were
marginally more total annual rainfall values greater than the long term median rainfall
for the data series reviewed. At Bulawayo, Matopos and Filabusi there was an equal
number of total annual rainfall values greater and less than the long term median
rainfall. The distribution of the total annual rainfall values around the median rainfall
values suggests that data for each station was drawn from a homogenous sample.
Table 4.2. Homogeneity test for annual total rainfall for the five meteorological
stations in southern Zimbabwe
Station Median Number of Total count Positive Negative
(mm) runs counts (NA) counts (NB)
Bulawayo 561 28 71 34 38
Matopos 537 39 69 35 34
Mbalabala 617 42 64 21 21
Filabusi 538 37 74 36 36
Beitbridge 362 20 50 26 24
4.4.2 Total annual rainfall
The gradient of rainfall from Bulawayo (NR N) southwards towards Beitbridge (NR
V) can clearly been seen on the mean and median values as well as the long term
graph of annual values (Figs. 4.1 and 4.2). However, variability is also seen at each
station with extremes of over 1 200 mm at Bulawayo and lowest of less than 100 mm
at Beitbridge with an extremely high coefficient of variation of 44 %. The total annual
rainfall variations for the Bulawayo, Matopos and Mbalabala stations in semi-arid
lJV., r-
51
southern Zimbabwe are summarized in Figure 4.1. There was no significant (P =
0.996) change in total annual rainfall at Bulawayo over the 71 year period. The lowest
annual rainfall recorded at Bulawayo was 199 mm in 1947, and the highest rainfall
was 1 259 mm received in 1978. The long term average annual rainfall for Bulawayo
based on these data was 584 mm with standard deviation of 206 mm and coefficient
of variation of 35 % (Table 4.3). At Matopos total annual rainfall decreased (P =
0.506) slightly between 1960 and 2000, and the lowest annual rainfall (263 mm) was
recorded in 1991. The 69 year average rainfall for Matopos was 573 mm with
standard deviation 203 mm and coefficient of variation of 35 % (Table 4.3). At both
Bulawayo and Matopos the driest periods were 1960-1972 and 1980 to early 1990
(Fig.4.1).
Table 4.3. Characteristics of the total annual rainfall (based on July-June calendar)
recorded at five meteorological stations in semi-arid southern Zimbabwe
Station Mean Median Standard Coefficient NR
(mm) (mm) deviation of
(mm) Variation
0/0
Bulawayo 584 561 206 35 IV
Matopos 573 537 203 35 IV
Mbalabala 626 617 213 34 IV
Filabusi 546 538 185 34 IV
Beitbridge 376 362 166 44 V
Total annual rainfall decreased (P > 0.05) in Filabusi and Mbalabala stations between
1980 and 2000 with 546 and 626 mm being average rainfall recorded for the two
stations (Table 4.3 and, Figs. 4.1 and 4.2). At Mbalabala the lowest rainfall was 195
mm recorded in 1991. The lowest annual rainfall was 250 mm recorded in 1945 at
Filabusi. The 1959 to 1973 was the driest period at Mbalabala while 1980 to 1990 was
the driest decade at Filabusi since record keeping began in 1921. Beitbridge recorded
the lowest rainfall amount of 83 mm in 1982. The highest rainfall at Beitbridge was
52
1177 mm recorded during the 1999-2000 rainy season and this coincided with the
period when cyclone Eline was experienced in the southern parts of Zimbabwe. The
driest period at Beitbridge was between 1960 and 1970 (Fig. 4.2).
1400 Bulawayo I 0 Annual rainfull -- 5 year moving averag~
1200 <>
,-,1.0.00
1800
J 600
400 o 0 0
0000 0
200
o +-----,-----,-----,-----,-----,-----,-----,-----,-----,
1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Year
1400 Matopos e Annual rainfall -- 5 year moving average
1200 e
11000 o 0'-'800
~0.2 600 00
400 <> <>
e
200 e o <>
o +-----,------.-----r-----.-----.------.-----,-----.-----~
1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Year
1400 Mbalabala <> Annual rainfull -- 5 year moving average I
1200
<>
~1000
e <> <>
.a
'-'800
§ 600
~ 400 e <>
200
o 4-----,-----,-----,----,-----,-----,-----,-----,----,
1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Year
Figure 4.1. Total annual rainfall (July to next June) for Bulawayo (a), Matopos (b)
and Mbalabala (c) meteorological stations of semi-arid southern Zimbabwe
53
1400
1200 Filabusi o Annual rainfull -- 5 year moving average I0
,-.1.0.00
], 800 o oe o
0
I 0600 0400 00
200
0
1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Year
1400 Beitbridge I <> Annual rainfall -- 5 year moving average I
1200 e
I 1000
.._,800
I 600400
200
o +-----,----,-----,-----,-----,----,-----,-----,-----,
1920 1930 1940 1950 1960 1970 1980 1990 2000 201
Year
Figure 4.2. Total annual rainfall for Filabusi (d) and Beitbridge (e) meteorological
stations of semi-arid southern Zimbabwe
The slight decline in total annual rainfall has also been observed in studies conducted
by Unganai (1996) using Global Climate Models (GCM). The study by Unganai
(1996) showed aIO % decline in total rainfall over Zimbabwe between 1900 and
1993. From the same study by Unganai (1996) from the late 1950s to about 1972 and
1980 to 1993 were the driest periods in Zimbabwe. This compares well with the
current analysis as three of the stations had 1960 to' early 1970s as the driest period on
record, whereas two stations had 1980 to early 1990s as the driest decade. Globally
Trenberth et al. (2007) also reported a similar trend of declining annual rainfall and
this has been attributed to a number of reasons including a decrease in the number of
54
wet days (Christensen et al., 2007). Other possible explanations for this decreasing
trend in rainfall include the effect of land use changes on moisture cycling, and the
increasing frequency and severity of El Nifio events (Trenberth et al., 2007).
The coefficient of variation for total annual rainfall increased with decreasing mean
annual rainfall (Table 4.3). This suggests that the drier areas experience higher rainfall
variability than wetter areas in southern Zimbabwe. The coefficient of variation range
of 34 to 44 % is higher than 28 to 36 % reported by a previous study by Oosterhout
(1996) in Zimbabwe's NRs IV and V. Despite the lack of significant changes in total
annual rainfall, farmers in semi-arid southern Zimbabwe still experience crop yield
reductions due to soil water stress. This seems to suggest that in-season rainfall
distribution plays a more significant role in semi-arid crop production than total
annual rainfall. Clay et al. (2003) reported that dry spells that occur at reproductive
stages of crops reduce yields substantially even if total annual rainfall is at or near
normal. Unganai (1990) reported that approximately 480 mm of rainfall, well
distributed within the growing season was sufficient for successful production of
sorghum and maize crops. Tadross et al. (2007) further reported that there has been a
shift towards more sporadic rainfall events during the rainy seasons.
Beitbridge represents the southern part of Gwanda district where the on-farm
experiments were conducted during 2005/06, 2006/07 and 2007/08 growing seasons.
Seasonal rainfall distribution was erratic during two (2006/07 and 2007/08) of the
three seasons of experimentation and dry spells were experienced in January and
February in each of the seasons (Chapters 7 and 8). In Insiza district, represented by
Filabusi station, erratic rainfall distribution and dry spells during the 2006/07 and
55
2007/08 seasons were also experienced. In both districts 60 to 75 % of the farmers
hosting research trials experienced total crop failure on the research plots as well as
their other fields. This is consistent with simulation modelling results (Chapter 10)
which showed that total crop failure can be experienced in both the conventional and
planting basin systems regardless of the amount of nitrogen fertilizer used. During the
four years of experimentation at Matopos site, represented by Matopos meteorological
station, three (2004/05, 2006/07 and 2007/08) out of the four growing seasons
experienced dry spells in January and February. However, no total crop failure was
experienced at Matopos site probably due to the fact that the clay soil at the site
managed to supply soil water until the maize crop matured (Chapter 9a and 9b).
4.4.3 Standardized precipitation index
The year 1977 was the wettest at Bulawayo while the most severe drought was
recorded in 1946 (Fig. 4.3). At Matopos 1954 was the wettest year and 1946, 1968
and 1991 had the most severe drought at Matopos (Fig. 4.3). The wettest year at
Mbalabala was 1977 and the most severe drought at Mbalabala was recorded in 1991.
The wettest year at Filabusi was 1924 and driest year was 1946. No year fell into
extremely or very wet at Beitbridge. Only four years namely 1995, 1996, 1999 and
2000 were moderately wet. The most severe drought was recorded in 1998 at
Beitbridge (Fig. 4.4). The relatively wet period stretching between 1995 and 2000
coincide with the El Nifio/La Nina events of 1997 to 1999 (Rosenzweig, 2001). The
1991/92 drought was one of the worst throughout southern Africa (Rosenzweig,
2001).
56
3 BuJawayo
2
s:: 0 -t-----..
r/)
1 20 201
-1
-2
-3
Year
3 Matopos
2 G
esn:: 0
1 20 19301~'~~~1~:\0~! ¥}bil,P'1!~¥llo~~'fo1
-1
-2
-3
Year
3 Mbalabala
2
. ~ 0 +-------rY
1 20 193· 9 0 1 5 1 7 9 : .. 2000 201
- 1 -
-2
-3 Year
Figure 4.3. Standardized precipitation indices derived from total annual rainfall data
for Bulawayo (a), Matopos (b) and Mbalabala (c) meteorological stations
57
3 Filabusi
2
2000 201
-2
-3
Year
3 Beitbridge
2 G
-0..
ir: 0
IC 20 1930 1940 19 0 1 • 197 1 0 000 201~
-1 ,. ,
-2
-3 Year
Figure 4.4. Standardized precipitation indices derived from total annual rainfall data
for Filabusi (d) and Beitbridge (e) meteorological stations
4.4.4 Wet days per growing season
The changes in number of wet days per growing season based on daily rainfall data
derived from Bulawayo and Matopos meteorological stations in southern Zimbabwe
are shown in Figure 4.5. There was a negligible difference in the number of wet days
per season as one moves from Bulawayo to Filabusi station through Matopos and
Mbalabala (Figs. 4.5 and 4.6). All four stations are located in NR IV. Our findings are
inconsistent with results reported by Hudson and Jones (2002) who stated that there
has been a general decrease in the number of wet days per year in southern Africa.
58
Hudson and Jones (2002) defined a wet day as a day receiving more than 0.2 mm of
rain whereas in our study we defined a day with 5 mm or more rainfall as wet. In
addition our study focussed on number of wet days per growing season and not the
whole year. As expected Beitbridge which lies in NR V, had the least (P = 0.01)
number of wet days per growing season.
The average number of wet days per growing season recorded at Bulawayo was 21,
Matopos was 22, Mbalabala was 24, Filabusi was 22 and Beitbridge was 12. At
Bulawayo station the lowest number of wet days per season recorded was four in
1946/47 and 1981/82. Forty-two wet days were recorded in 1977/78 was the highest
number observed at Bulawayo. Matopos recorded seven and 45 wet days in 1990/91
and 1954/55 seasons. Mbalabala recorded eight and 48 wet days in 1983/84 and
1954/55 seasons. Filabusi station recorded eight and 46 wet days per growing season
in 1946/47 and 1954/55 seasons. Three and 31 wet days per growing season were
recorded at Beitbridge station in 1964/65 and 1999/00 seasons. Using the October-
September hydrological year and definitions of a wet day having more than 10, 20
and 30 mm rainfall, Love et al. (2008) assessed the trend in number of wet days per
year in southern Zimbabwe. Findings from their study revealed a decline in number of
days with more than 10, 20 and 30 mm of rainfall per year in southern Zimbabwe.
However, this is not apparent from the current analysis with a wet day defined as
above 5 mm.
59
50 Bulawayo
cIÏ
~ 40
"..0...
Q)
~ 30
4-<
0
I-<
Q) 20
,.D
§
Z 10
0
= 5season moving average + Wetdavs
50 Matopos
45
cIÏ + +
~ 40 + +
".'.C..j. 35 -+++
Q)
~ 30
4-<
0 25
I-<
Q) 20
.0
E 15::s
Z 10
5 + +
0
-= 5season moving average + Wet days
50 + Mbalabala
cIÏ 45
;;:.-. + + + + G
t1:S 40
".'.C..j. 35 +
Q) +
~ 30 +
4-<
0 25....
Q) 20
.0 +
E 15 + ++::s + + + ++ + + ++
+
+ +
Z 10 ++ +5
0
--==> 5 season movingaverage + Wet days
Figure 4.5. Number of wet days per season based on daily rainfall data obtained from
Bulawayo (a), Matopos (b) and Mbalabala (c) meteorological stations
60
50 Filabusi
..n 45
+
:>. + + G
C':I 40 +
'"0
...... 35 ++
Q)
30
...~....
0 25
I-<
Q) 20 +
.0
§ 15 +
Z 10 + + +
5
0
= 5season movingaverage + Wet days
50 Beitbridge
..n 45
:>.
C':I 40
'"...0... 35
Q)
~ 30 +.......
0 25
I-<
Q) 20
.0
S 15
:::l
Z 10
5
0
= 5seasonmovingaverage + Wet days
Figure 4.6. Number of wet days per season based on daily rainfall data obtained from
Filabusi (d) and Beitbridge (e) meteorological stations
At Bulawayo and Matopos the trend of wet days per season follows a similar pattern
to that of total annual rainfall. The driest decade (1980-1990) had the lowest number
of wet days per season. At Bulawayo and Matopos the return period of seasons with
high rainfall is 18 to 22 years. The 1961 to 1978 period had the least number of wet
days per season at Mbalabala (Fig. 4.5) which coincides with the period of lowest
annual rainfall. Return period for high number of wet days per season was 21-26
years at Mbalabala and Filabusi. Further south, Beitbridge appears to receive
61
relatively wet growing seasons after more than 20 years although the record is only 50
years long.
4.4.5 Start and end of growing season
There has been no significant (P > 0.05) change in the start of growing seasons at
each station across the Bulawayo to Beitbridge transect (Table 4.4, Figs. 4.7 and 4.8).
The variability of start and end of the growing season is quite similar at all stations
along the Bulawayo to Beitbridge transect. The growing season generally starts earlier
at Filabusi and ends earlier at Beitbridge (Table 4.4). At Bulawayo the growing
season starts on 8 December and ends on 4 April (Table 4.5) giving an average season
length of 117 days. The earliest start of season was 26 October (117 days from 1 July)
and this occurred in 1988/89. The most delayed start of growing season was recorded
in 1981/82 when the season started as late as 3 March (day 246). At Matopos the
growing season long term start on 2 December and ends on 29 March. The earliest
start of growing season was recorded on 21 October (day 112) in 1978/79 and
2001/02. The most delayed start of growing season at Matopos occurred on 23
February (day 237) in 1981/82. At Mbalabala growing season starts on 3 December
and ends on 1 April. The earliest start of growing season was on 30 October (day 122)
and recorded in 1980/81 and 1985/86. The most delayed start of season occurred on
day 261 (18 March) in 1976/77. Growing season starting as late as 18 March can be
attributed to the criteria of start of season of 20 mm of rain in one or two days used in
-.
our study. At Filabusi the growing season starts on 28 November and ends on 3 April
station. The earliest start of season was observed in 1931/32 and occurred on day 102
(100ctober). The most delayed start of season was in 1988/89 and occurred on day
229 (14 February). At Beitbridge which was the driest station studied, the season
62
starts on 7 December. The earliest start of season was recorded on day 102 (10
October) in 1931/32 and the most delayed start occurred in 1988/89 on 14 February
(day 229).
Table 4.4. Median dates for the start and end of the growing season based on daily
rainfall data obtained from five meteorological stations in southern Zimbabwe
Station Median start Standard! Median end Standard!
date deviation date deviation
(days) (days)
Bulawayo 8 December 31 4 April 32
Matopos 2 December 30 29 March 31
Mbalabala 3 December 26 1April 27
Filabusi 28 November 26 3 April 30
Beitbridge 7 December 31 25 March 29
The most abrupt end of season at Bulawayo was recorded in 1964/65 when the season
ended on 15 January (day 200). At Matopos the earliest end of growing season was
recorded on 13 December (day 166) in 1979/80. At Mbalabala the earliest end of
season was on 21 January (day 206) in 1932/33. The most abrupt end of growing
season at Filabusi occurred on 20 January (day 205) in 1932/33. The end of 1958/59
growing season was the most delayed and occurred on 13 June (day 349) at Filabusi.
At Beitbridge the season ends on 25 March. The most abrupt end of season at
Beitbridge was recorded on 20 January. The most delayed end of season was in
1958/59 and occurred on 13 June. The abrupt end of the growing seasons observed in
the long term analysis agrees with observations made during the three to four years
(Chapters 7, 8, 9a and 9b). The 2007/08 growing season ended on 24 January 2008 in
all semi-arid districts of southern Zimbabwe.
63
400 Bulawayo
~ 350 t. t.t.
~- 300I-. 250
~ 200 e6
CIl
;;>...
~ 150
'-'
.~ 100
f-o 50
0
-- 5 season moving average = 5 season moving average <> Start
400 Matopos
,...., 350
;>.,
~ 300 lJ.
lJ.llt,.
] 250
<Il 200
t......., 150 -
100 <>
~
50
0 ,----,-----.----~----.----.----~----.---
-- 5 season moving average - 5 season moving average <> Start lJ. End
400 Mbalabala
,......_
..0 350
..~.... 300 t:.
I-.
~ 250 t:.~
CIl 200 e
~
~ 150
Q)
.§ 100
f-o 50
0
~##~~$####~~~~~~$$#$#$#
~ ~ ~ ~ ~ ~) ~ ~ ~ ~ ~) ~ ~ ~) ~J %) % ~ ~) ~J ~) ~
Season
-- 5 season moving average -===> 5 season moving average ~ Start t:. End
Figure 4.7. Start and end of growing seasons derived from daily rainfall data for
Bulawayo (a), Matopos (b) and Mbalabala (c) Meteorological stations
64
400 Filabusi
350 6
~ [j...... 6
.-
250
~
en 200 6 e <> •
~ 150<;»
.§ 100 <> <>
f- 50
0
,~"v",~fo~'?~",0fo?J\'>.~':h\(~\.\~ ~",",v;;{.,fo?!\fo~~rJ<_(~\.(\\f.o",r"vt,f:fon.~~~~ (\.~rt~, o,"<v-.f>Jfon.~~~t~\~rt,
"v "v '? '? ~ '>. '>. '" ~ ~ fo b (\. ~ ~ rt,Jrt, 0, ~J ~~)~
Season
-- 5 season moving average = 5 season moving average <> Start 6 End
400 Beitbridge
,-..
..0
- 350..:..:.l. 300 6l:I
~ 250
Aa
ca 200
<Jl
~ 150
~ 100 <> 0<>Cl)
.5
f- 50o -
- 5 season moving average = 5 season moving average e Start 6 End
Figure 4.8. Start and end of growing seasons derived from daily rainfall data for
Filabusi (d) and Beitbridge (e) meteorological stations
Aviad et al. (2004) from the Mediterranean region reported that as aridity increases
the rainy season normally begins late and ends early, and subsequently the length of
growing season is short. In our study smallholder farmers in Beitbridge face the
prospects of a shorter growing season compared to those in NR IV as it only began on
7 December and ended on 25 March giving a length of only 108 days. This is further
complicated by the fact that start of growing season is also more variable in
Beitbridge than its end (Table 4.5). Farming practices that allow timely preparation of
the land and planting are more critical in the Beitbridge area so that farmers can make
effective use of rainfall received during the November to December period.
65
The lack of significant long term changes suggests that the characteristics of the
growing season are influenced by other factors in addition to total rainfall and date of
start of the rains. In Zimbabwe, Oosterhout (1996) reported that years that had the
highest rainfall did not correspond with years having the highest crop yields.
Distribution and reliability of rainfall during the growing season has a stronger
influence on the characteristics of the growing period than total rainfall (Twomlowet
al., 2006b). High spatial and temporal rainfall variability during the season increases
the risk of intra seasonal dry spells and droughts in semi-arid areas (Rockstrëm and
Falkenmark, 2000). Long term simulation results (Chapter 10) indicated that there is
more than a 20 % chance of getting no yield in the conventional and CA systems
under semi-arid conditions of southern Zimbabwe. Therefore other indices to be able
to represent this distribution and variability need to be developed.
4.4.6 Length of growing season
The length of growing season decreased as one moves from Bulawayo to Beitbridge
through Matopos, Mbalabala and Filabusi (Figs. 4.9 and 4.10). For the period
reviewed at each station, the growing season averaged 111d at Bulawayo, 110d at
Matopos, 112d at Mbalabala, 122d at Filabusi and 100 days at Beitbridge. The longest
growing season recorded across the five stations was 224 days recorded at Matopos in
2001/2002 which is more than double the average length. The shortest growing
season observed across the Bulawayo to Beitbridge transect was 38 days recorded at
Filabusi and Beitbridge. The longest growing season at Bulawayo had 215 days and
was recorded in 1939/40. The 1981/82 season was the shortest growing season with
only 41 days which coincided with the lowest number of wet days (4). At Matopos the
shortest growing season had 39 days recorded in 1979/80. In 1954/55 Mbalabala
66
recorded the longest season which had 173 days which was also the season with the
highest number (48) of wet days. The shortest growing season at the same station was
46 days recorded in 1976/77. Filabusi recorded the longest growing season of 197
days in 1958/59 while the shortest season was 38 days recorded in 1946/47 which was
the year with annual rainfall of 250 mm. At Beitbridge the longest growing season
was 162 days recorded in 1999/00 which also recorded the highest number of wet
days (31) and the highest annual rainfall (1177 mm) due to the cyclone Eline pushing
in over the African landmass and causing severe flooding. The shortest season of 38
days was recorded in 1972/73 and 1973/74.
Results from our study showed an average length of the growing season of 111 days
across the 5 stations. The 111 days length of growing season is much higher than 96
days reported by Morse (1996) for Zimbabwe's NRs II, IV and V. The differences in
length of growing season could be ascribed to differences in the criteria used in
defining the start and end of growing season. Tadross et al. (2007) reported that a
growing season of90-120 days can be experienced in southern African countries such
as Malawi and Zambia. The above analysis confirms that there is no general rule of
thumb that the season length is related to the amount of rainfall received but can have
a relationship with the number of wet days or effective rainfall events. So if this
southern part of Zimbabwe has an average season length of only 111 days,
agronomists need to use this information in selecting suitable cultivars for smallholder
farmers. The ultra short cultivars could be an option and this will help farmers to be
able to plant early and produce a crop with less exposure to dry spells.
67
250 Bulawayo
,-..
(Jl n
$ 200 o
'-~'
o c
0 150
~ c
4-<
0 100
t c
50 o c~ c c c c
0
c Growing season = 5 season movin
250 Matopos
'Vi'
.§ 200 cG
'"'" c cISO n o
~ o
4-<
0 lOO o0
t 0 c co n
S 50 c c o
0
c Growing season ==== 5 season moving average
250 Mbalabala
'Vi'
t......, 200
150
~ c
~ 100 c
J 50 c
o +--,-,--,-,,-,--,-,--,-,,-,--,-,--,-,,-,--,-,,-,--,-,--,-
~#~~#~ $~ ~~~~~~#~$~##~ #b~~~~~~n~~~~v~~o~- /q~q#~nv~ ~~
Season
c Growin season
Figure 4.9. Length of the growing season based on daily rainfall data obtained from
Bulawayo (a), Matopos (b) and Mbalabala (c) meteorological stations
68
250 Filabusi
c
c c
150
100 -
cP
50 c oc
o _.~'--'-'--'-'-'--'-'--'-'--r-'--'-'-,,-.-'--'-'--'--r
~~~ ,,~(o ~,,~ <;;0(0n\r:;.."~'\~ A\.~ ~,,~ <;;\,,(0n\@ ,::"}iJ<~ <o~~A.~ t..,0(on~~ "'~ A.~~ ~o,~ t..,f1(no\r;§J",\cr--C\f>~
~ ~ " " ~ 'r)..J r:;.. " " 9 (0 (0 A. ~)~ ~J~ 0, ~)~ ~J ~
Season
I c Growing season = 5 season moving average
250 Beitbridge
c c
o +--,-,-,,-,-,--,-,--,-,--,-,--,-,-,,-,-,--,-,--,-,--,
~~~ ,,~(o ~,,~ <;;0(0n\r:;..~~ (I.\~ ~,,~ <;;\,,(0n\(o~,::"}iJ<~ \(o~~A.~ "O(0 n~~ "'~ A.~~ ~o,~"f1(onf>~ "'~ C\f>~
~ ~ " " ~ r:;.. r:;.. " " 9 (0 (0 A. ~ ~ ~J ~ 0, 0, ~ ~J~
Season
c Growing season =-== 5 season moving average
Figure 4.10. Length of the growing season based on daily rainfall data obtained from
Filabusi (d) and Beitbridge (e) meteorological stations
Plotting length against start of growing season showed a negative inverse relationship
for all five stations (Figs. 4.11, 4.12 and 4.13). Therefore a delay to the start of
growing season results in a shorter season as it is often related with the end of season.
Based on the R2-values and gradients of the slopes, the inverse relationship between
length and start of season was strongest in the wetter part at Bulawayo (NR IV) and
weakest at Beitbridge (NR V) being the driest part of the catchment. The strength of
the relationship decreases as one move from Bulawayo to Beitbridge through
Matopos, Mbalabala and Filabusi stations. The length of the growing season at
Bulawayo is more influenced by the time when the growing season starts compared to
69
other stations. The time when growing season starts has the least influence on the
subsequent length of growing season at Beitbridge compared to stations in NR IV. At
Filabusi, which lies between Mbalabala and Beitbridge, the relationship between start
and length of season is stronger than at Matopos and Mbalabala stations. A stronger
relationship (R2 = 0.76) between the start and length of the season was reported in
Botswana and other parts of southern Africa (Kanemasu, 1990) while the highest R2
in our analysis was 0.41 at Bulawayo (Fig. 4.13). Differences in the criteria for start
and end of the season can probably explain the difference between values obtained in
our analysis and results from Kanemasu (1990).
250 Bulawayo
,......,
e;n
)K
;..
<Il 200 GI)K
"..0_, )K
c:: )KA:: )K )K
0 ~)I( ;)1()1(en 150 y = -0.8404x + 250.99
<Il .....)I( )1..(.' )1(,,:;-
(!) ';')1( ~: )I( )I(
)I( 2
..e.n.. )I( )I( R = 0.4065
0 100 )KA:: )I()I( )I(
"cto
)I(~)I( ;t~
::
.3 50 ::.c )I(
0
100 120 140 160 180 200 220 240 260
Start of season (days from 1 July)
Figure 4.11. Relationship between length and start of growing season at Bulawayo (a)
meteorological station in southern Zimbabwe
70
250 - Matopos
"("Jl' 200 y = -0_7022x+ 22336
.~._, R2 = 03236
c
0 150
~
(Jl
4-<
0 100
ts
....l 50
0
100 150 200 250
Start of season (days from 1 July)
250 Mbalabala
,-._
til
~ 200
"0
'-"
C
0 150 y = -0_6544x + 216_61~
~ 2
....... R = 0-3002
0 100
ic
Q)
....l 50
0
100 150 200 250 300
Start ofseason(days from 1 July)
1 FiJabusi,-., 250
~200
~
..! )1(",_, )t:: .... y = -0.7955x + 244.15150 2-)t:: R = 0.3345
4-<
o 100
t>-l 50
o +-------------,-------------,-------------,------------,
100 150 200 250 300
Start of season (days from 1 July)
Figure 4.12. Relationship between length and start of growing season at Matopos (b),
Mbalabala (c) and Filabusi (d) meteorological stations in southern Zimbabwe
71
250 Beitbridge
'V
~;;.
i..' 200
.".C_, y = -0.5287x+ 189.59
t:
0 150 R2 = 0.2892
~ ~~
4'-<"
0 100 ::t
t
t:
Q.) 50
.....l ::t
0
100 150 200 250
Start of season (days from 1 July)
Figure 4.13. Relationship between length and start of growing season at Beitbridge (e)
meteorological station in southern Zimbabwe
4.4.7 Dry spells
The peak rainfall period during the growing season occurs between December and
February in southern Zimbabwe. The likelihood of getting 14 and 21 day dry spells
during any time of the year is given in Figure 4.14 and Figure 4.15. Based on the data
set reviewed, Filabusi has the highest chance of getting 14 and 21 day dry spells
compared to the other four stations. Mbalabala has the least chance of getting a 21 day
dry spell during the peak rainfall period from 120 to 240 days after 1 July. The
probability of getting a 14 or 21 day dry spell decreases as one moves from Bulawayo
to Mbalabala through Matopos. There is a 27 to 40 % chance of experiencing a 21 day
dry spell at Bulawayo between la November (day 133) and 8 February (day 223).
During the same period there is a 66 to 79 % chance of getting a 14 day dry spell at
Bulawayo. At Matopos there is a 18 to 33 % chance of a 14 day dry spell occurring
between la November and 8 February. During the same period 21 day dry spells have
a 3 to 7 % chance of occurring at Matopos.
72
The probability of getting a 14 day dry spell at Mbalabala between 10 November and
8 February is 14 to 27 %. There is a 2 to 5 % chance of getting 21 continuous dry
days during the same period. Fourteen day dry spells are a common feature at Filabusi
with a 91 % chance of occurring between 10 November and 8 February. During the
same period there is a 60 to 80 % probability of a 21 day dry spell occurring. Further
down the transect Beitbridge has a 48 to 69 % chance of experiencing a two week dry
spell. A 14 to 30 % chance exists for a 21 day dry spell to occur between 10
November and 8 February at Beitbridge.
1.0 Bulawayo
0.8
~ 0.6
J o 14 0210.4
0.2
0.0
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
Matopos
o 14 021 I
.£'
:;::l 0.6
~
o£D 0.4
0.2
0.0 _j__--.------r~~~~~~--.--__r--,.__-__,
o 50 100 150 200 250 300 350 400
Time (days after 1 July)
Figure 4.14. Probability of getting 14 and 21 day spells within 30 days from a wet day
based on the fitted first order Markov chain probability values for Bulawayo (a) and
Matopos (b) meteorological stations in southern Zimbabwe
73
1.0 Mbalabala
0.8 o 14 Cl!> 21
g ;?G0.6
oo
o •
.C,l!>
~ 0.4
o .,
<\0
•
0.2 '-.2•
0.0 ----,
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
1.0
0.8 [3
$ 0.6
J Filabusi0.4 o 14 ~
0.2
0.0
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
1.0
Beitbridge
0.8
g
:0 0.6 G
Cl:!
..0
ce, 0.4 014 021
0.2
0.0
0 50 100 150 200 250 300 350 400
Time (days after 1 July)
Figure 4.15. Probability of getting 14 and 21 day spells within 30 days from a wet day
based on the fitted first order Markov chain probability values Mbalabala (c), Filabusi
(d) and Beitbridge (e) stations in southern Zimbabwe
74
The decrease in probability of getting 14 and 21 'day dry spells coincides with the
peak rainfall period. The peak rainfall period occurs from December to February
(Unganai, 1996). Figures 4.16 and 4.17 reveal that rainfall is more reliable at Matopos
and Mbalabala during the December to February peak rainfall period. Rainfall is least
reliable at Filabusi where there are 91 and 60 to 80 % chances of getting 14 and 21
day dry spells. At Beitbridge there is a rapid increase in chances for experiencing 14
and 21 day dry spells after receiving rain. This implies that the probability of reduced
crop yields or complete crop failure due to soil water deficits is also high at
Beitbridge. The probability of dry spells also increases at Bulawayo, Matopos and
Mbalabala after 8 February. This coincides with the flowering and grain filling stages
of most cereals grown in semi-arid smallholder systems. This poses a major challenge
in soil water management in semi-arid rainfed cropping systems.
4.4.8 Cumulative distribution functions for the two halves of cropping season
At all stations there are better prospects of receiving rain in the January-March than
October-December period (Tables 4.5 and 4.6; Figs. 4.16 and 4.17). However, three
monthly rainfall totals indicate more variation in the second half of the season than
the October-December period at all stations (Tables 4.5 and 4.6). This observation
suggests that the timing of planting is critical if crops are to reach maturity in semi-
arid southern Zimbabwe. There could be seasons with early cessation of rains as
observed in 2007/08 (Chapters 7, 9a and 9b) and crops may fail to reach maturity.
At Bulawayo there is a 44 % chance of getting more than 242 mm which is the long
term average for the October-December period. There is a 46 % probability of getting
more than the long term average rainfall (296 mm) for the January-March half (Table
75
4.5). There is a 19 % chance of getting 157 mm of rain in both halves of the growing
season at Bulawayo. At Matopos there is a 39 % chance of getting more than 248 mm
during October-December. The January-March period has a 51 % probability of
getting more than 291 mm. There is a 34 % chance of receiving 205 mm in both
halves of the growing season (Fig. 4.19). At Mbalabala there is a 48 % chance of
getting more than 242 mm during October-December. The January-March period has
a 49 % probability of receiving more than 340 mm at Mbalabala. Filabusi has 46 and
41 % chances of receiving more than 214 and 290 mm during October-December and
January-March periods respectively. At Beitbridge there is a 25 % probability of
getting 97 mm of rainfall during both halves of the growing season. There is 47 %
probability of receiving 140 and 176 mm of rain during October-December and
January-March periods. In general, the higher probability of getting rainfall during the
January-March period give better prospects of crops reaching maturity if planting was
done on time with respect to effective planting rains. The timing of farming
operations in the first half of season could be critical in order to fully utilize the
favourable conditions in the January-March period. These values are useful when the
three month seasonal probability forecasts are used by the Department of
Meteorology each year.
Table 4.5. Rainfall characteristics for first half (October-December) of the growing
season at five stations in semi-arid southern Zimbabwe
Station Medialll Mean Standard 33 % 66 %
deviation chance chance
(mm) (mm) (mm) (mm) (mm)
Bulawayo 230 242 91 185 280
Matopos 229 248 94 202 268
Mbalabala 238 242 104 186 272
Filabusi 200 214 94 168 241
Beitbridge 131 140 61 109 156
76
Table 4.6. Rainfall characteristics for second half (January-March) of the growing
season at five stations in semi-arid southern Zimbabwe
Station Median Mean Standard 33 % 66 %
deviation chance chance
(mm) (mm) (mm) (mm) (mm)
Bulawayo 275 296 141 215 337
Matopos 300 291 155 187 355
Mbalabala 341 340 157 245 414
Filabusi 266 290 152 220 322
Beitbridge 163 176 147 116 198
1.0 Bulawayo _-"
.e _ _,. ;or0.8~g: .J
..... 0.6
.~~ 0.4
-====Jan-Mar. - - Oot-Deel
~ 0.2
U
0.0 -I-~~~---r-----'-----r---"-----'
o 100 200 300 400 500 600
3 monthly rainfall total (mm)
1.0 Matopos _J
,-
g 0.8 (
,/
~ r: [3
oeD 0.6
0-
Q)
.:~> 0.4
~
U 0.2 Jan-Mar Oct-Dec I
0.0
0 100 200 300 400 500 600
3 monthly rainfall total (mm)
Figure 4.16. Cumulative distribution functions for the first and second halves of the
growing season based on three monthly rainfall totals from Bulawayo (a)
meteorological station
77
1.0 Mb alab ala
;~........::::0::.8
~
..D G
0
'"' 0.60..
dl
:>
-•.;:::l~ 0.4;::l
E
;::l 0.2 Jan-Mar Oct-Dec I
U
0.0
0 100 200 300 400 500 600
3 monthly rainfall total (mm)
1.0 Filabusi _ , _",... - ---
g 0.8 ,- ./
~ ..,..
.CeJ G0.6 /
0..
dl -
•.:p> 0.4
:~; Jan-Mar - - Oct-Decl
E
;::l 0.2
U
0.0
0 100 200 300 400 500 600
3 monthly rainfall total (mm)
1.0 Beitbridge - _.Ig _."0.8
: )~0
.CeJ 0.6 G
0..
dl
.:~> 0.4
:;
E::s 0.2 Jan-Mar - - Oct-Deq
U
0.0
0 100 200 300 400 500 600
3-monthly rainfall total (mm)
Figure 4.17. Cumulative distribution functions for the first and second halves of the
growing season based on three monthly rainfall totals from Mbalabala, Filabusi (d)
and Beitbridge (e) meteorological stations
78
At Bulawayo near-normal (33-66 %) 3-month rainfall totals are 185-280 and 215-337
mm for October-December and January-March periods. The near-normal 3-month
rainfall ranges are 202-268 and 187-355 mm for the first and second halves of the
season for Matopos. At Mbalabala the below-normal rainfall amounts are 186 and 245
mm for October-December and January-March periods. Data from Mbalabala station
showed near-normal rainfall ranges of 186-272 and 245-414 mm for first and second
halves of the growing season. The near-normal 3-month rainfall total ranges from 168
to 241 and 220 to 322 mm for October-December and January-March periods at
Filabusi. At Beitbridge below-normal 3-month rainfall totals are 109 and 116 mm for
October-December and January-March periods.
4.5 Conclusion
There has been no significant change in total annual rainfall over the past 50 to 74
years at five meteorological stations located in the Mzingwane sub-catchment,
southern Zimbabwe. The trend in the number of wet days per growing season is
similar to the trend in total annual rainfall at all stations reviewed. However, despite
the lack of major changes in total annual rainfall significant crop yields reductions
and total crop failures are a common feature in semi-arid districts of Mzingwane sub-
catchment in southern Zimbabwe. This suggests that the in-season rainfall distribution
is playing a stronger part in influencing crop yields than total annual rainfall. This
poses a major challenge to farmers, extension agents and researchers to improve
rainwater management and storage in semi-arid cropping systems to help through the
dry spells.
79
The start and end, and the subsequent length of the growing season have not changed
substantially over the past 50 to 74 years in southern Zimbabwe. The start of growing
season was slightly more variable than end of season at Beitbridge (NR V). In both
agro-ecological regions timeliness of farming operations could be a potential key to
full exploitation of favourable conditions. This is particularly critical given the high
chances of 14 and 21 day dry spells from February to the end of season. Rainfed
cropping gets more risky as one moves from Bulawayo to Beitbridge based on dry
spell analysis conducted in this study. It would be useful to assess the probability of
dry spells for the October-December and January-March periods separately as on-
farm and on-station experimentation (Chapters 7 and 9) showed that the second half
of the season was drier than the first half. The high variation in the three-month
rainfall totals implies that there is a risk of crops failing to reach maturity if the timing
of the planting date was not good and rainfall ends early as experienced in the
2007/08 growing season.
80
CHAPTERS
Soil Water and Fertility Management Practices on Smallholder Farms
in Insiza and Gwanda Districts of Semi-Arid Southern Zimbabwe
5.1 Introduction
Smallholder agriculture in the semi-arid areas of southern Africa is largely rainfed, and
thus risky, due to high interannual variability and the occurrences of dry spells during the
rainy season (Chapter 4; Rockstrom et al., 2003; Cooper et al., 2008). Potential
evapotranspiration exceeds rainfall for more than six months of the year. Rainfall is
seasonal and highly variable both within space and time. Annual rainfall for a single site
(e.g. Bulawayo) can vary by up to 1000 mm from year to year as a drought year may
easily record less than 250 mm, such as the 1990/91 and 2004/2005 season in southern
Zimbabwe and Mozambique (Chapter 4; unpublished data, ICRISAT).
In Zimbabwe, water management under rainfed cropping systems has been the focus of
many research studies, outside the Limpopo Basin. However, few such investigations
have been carried out in the Mzingwane Catchment, which contains the driest areas of
Zimbabwe. A major feature of the aridity of the catchment is that the mean annual
potential evapotranspiration rates (approximately 1 800 mm) are three to five times the
annual rainfall (Chapter 4) suggesting a high potential of soil water loss through
evaporation.
81
In the years with above average rainfall pattern, successful rainfed cropping in the
Mzingwane catchment is hampered by the low soil fertility status of the predominantly
granite derived sandy soils of the smallholder farming sector. These granitic sands are
fragile and inherently low in organic matter, nitrogen and phosphorus (Grant, 1981;
Twomlowet al, 2006b). The low water holding capacity and low inherent fertility of
granite derived sandy soils make them marginal for crop production (Twomlow, 1994).
The purpose of this chapter was to review soil water and fertility management techniques
that have been developed for the semi-arid Zimbabwe. The review then identifies soil and
water management technologies that have been taken up by the farming communities and
those that have been recently introduced in the Mzingwane Catchment of the Limpopo
Basin.
5.2 Objectives
The objectives of the study were:
(1) to review literature on soil and water management technologies developed for the
smallholder farming sector;
(2) to determine the current soil water and fertility management practices which are being
used by smallholder farmers in Insiza and Gwanda districts; and
(3) to identify farmers' perceptions of risk in rainfed agriculture.
82
5.3 Materials and Methods
5.3.1 What has agricultural research developed?
The desk study involved a review of soil water and fertility management technologies
developed for the smallholder farming sector. Research work has been conducted on-
station and on-farm by the Department of Research and Specialist Services (DRSS) in
partnership with other institutions such as University of Zimbabwe and the German
Agency for Technical Cooperation (GTZ). Currently soil water and fertility management
technologies are being promoted in semi-arid districts of Zimbabwe by Non-
Governmental Organisations (NGOs), research institutions and AGRITEX.
5.3.2 What dofarmers say?
During each focus group meeting held in March 2005, the number of participants, mostly
women, averaged between 20 and 40. In Mpumelelo and Masiyepambili villages of
Insiza district, 20 and 27 people participated in the group discussions. In Humbane
village of ward 17, 23 people took part in the meeting. In Buvuma and Meja villages 31
and 26 people participated in the group discussions while 34 and 40 people were present
at Mhalipe and Ngoma meetings. A checklist of guiding questions was developed by
researchers before focus group discussions were held. One researcher from ICRISAT
facilitated the group discussions while an AGRITEX officer was on hand as a eo-
facilitator and interpreter. During group discussions notes were recorded on flipcharts by
the facilitator while two other researchers recorded the proceedings in notebooks. Each
group discussion lasted two to two and half hours in each district.
83
5.4 Results and Discussion
5.4.1 What agricultural research has developed
Good soil water management in rainfed agriculture can be achieved through various
tillage, both conventional and reduced, and rainwater harvesting techniques. Various
researchers and development agencies have explored in situ soil water management
technologies for smallholder rainfed agriculture. This section summarizes the findings
from previous research studies conducted in Zimbabwe and other southern African
countries.
5.4.1.1 In situ soil water management strategies
Studies conducted in Zimbabwe and Botswana demonstrated maize and sorghum yield
benefits of ploughing to a depth of 0.20-0.25 m compared to the typical depth of 0.10 m
(Grant et al., 1979; DLFRS, 1985). Maize and sorghum grain yields increased by 25 to
100 % with ploughing to 0.2-0.25 m depth depending on the season. In these studies the
residual effects of the deeper tillage (0.25 m) in terms of reduced bulk densities, and
improved root development, crop performance and reduced weed pressures were still
apparent in the second season. Similar responses have been observed by other researchers
in Tanzania (Northwood and McCartney, 1971). Studies conducted in Botswana clearly
demonstrated that double spring ploughing increase crop yields by as much as 71 %
across a range of soil types. Similar results have been observed in Zimbabwe also for a
range of soil types (Twomlowand Bruneau, 2000).
84
Minimum tillage has been explored as a soil and water conservation strategy for the serni-
arid areas of Zimbabwe. At Makoholi (NR IV) and Domboshawa (NR II), Vogel (1992)
evaluated the effect of conventional ploughing, mulch ripping and no-till tied ridging on
soil and water loss from fields. The findings from the study by Vogel (1992) showed that
mulch ripping retained higher soil water in the topsoil of the sandy soil especially at the
beginning of the cropping season. Nyagugumbo (2002) observed similar soil water
responses to mulch ripping versus conventional tillage systems on a clay soil at the
Institute of Agricultural Engineering (lAE). Mulching protected the soil from erosion and
promoted infiltration. Regrettably smallholder farmers have not taken up this technique
as mulch material is not readily available. The technology was developed and tested in a
non-participatory, top-down approach. The development of the technology did not
address the competition for crop residue between crop and livestock enterprises common
in the smallholder farming system (Twomlowet al., 2006b).
In the Chiredzi district of south eastern Zimbabwe, tied furrows increased sorghum yields
by 4-62 % on a sandy clay loam soil in four out of five seasons (Nyamudeza, 1993). In
NRs IV and V of southern Zimbabwe, N yakatawa et al. (1996) observed 22-85 % maize
yield increase as a result of using tied furrows compared with planting on the flat. Maize
yield response to tied furrows was increased by adding inorganic fertilizer. Addition of
inorganic fertilizer increased maize and sorghum yields by 35-115 % and 59-200 %
respectively. Widescale promotion of use of low rates of nitrogen fertilizer indicated that
cereal yields increase substantially by applying 10 kgNha-1 under the conventional
system in semi-arid districts of Zimbabwe (Twomlowet al., 2006a). Trials were
85
conducted in Gwanda district to determine the effect of livestock manure or basal
fertilizer and modified tied ridging on sunflower and sorghum yields (Rusike and
Heinrich, 2002). Crop yield results showed that combining manure and inorganic basal
fertilizer with tied ridges gives the highest yield benefits (Table 5.1).
Twomlowand Dhilwayo (1999) conducted a study on the effect of conventional
ploughing and improved tied ridges and weeding on hydrological properties of three
different soil types. Tied ridges created after crop emergence had more soil water than the
conventionally ploughed plots. Good weed management ensured maximum utilization of
soil water for crop productivity. Weeds compete for the scarce water with crops
especially at the beginning of the season. Timeliness of weeding also increase crop yield
and water use efficiency irrespective of the tillage practice (Twomlowand Dhliwayo,
1999).
Table 5.1. Effect of combining manure or compound D (8:14:7 - N:P20S:K20) with
modified tied ridges on crop yield in Gwanda district, after Rusike and Heinrich (2002)
Treatment Yield (kgha-i)
Sorghum + manure + ridges 400
Sorghum + 0 manure + 0 ridges 300
Sunflower + compound D + ridges 550
Sunflower + 0 compound D + 0 ridges 495
Groundnut + ridges 0
Groundnut + 0 ridges 0
Winter ploughing is a technique that has been recommended for decades. It is now the
standard recommendation by AGRITEX for the smallholder farming sector. However,
deep winter ploughing is still not being practiced in the Insiza and Gwanda districts of the
lower Mzingwane catchment. The technique is supposed to involve ploughing to a depth
86
of 0.2 m soon after harvesting. However, the majority of smallholder farmers plough to a
depth of 0.1 - 0.12 m, resulting in the formation of a hardpan within the soil profile. The
plough pan restricts root penetration and rainwater infiltration. Crops grown under
shallow ploughing cannot withstand extended periods of soil water stress during mid-
season dry spells.
5.4.2 The current soil water and fertility management techniques
5.4.2.1 In situ soil water management techniques
A planting basin tillage system has been promoted since 2004/05 growing season across
several semi-arid districts of Zimbabwe. During the 2004/2005 season World Vision
established trials on planting basins in Gwanda and Beitbridge districts which lie within
the Mzingwane catchment. However, these initial trials yielded nothing because of severe
drought. However, significant maize yield benefits derived from the use of planting basin
system were observed in other districts that lie outside the Mzingwane catchment (Table
5.2). The basin system enabled the farmers to prepare land before the onset of the rains
and plant on time following the first effective rains. The basin system also enables
resoureed constrained smallholder farmers to make more efficient use of limited
agricultural resource such as manure and inorganic fertilizers (Twomlowet al., 2008a).
87
Table 5.2. Rainfall patterns and maize grain yield (kgha') responses to farmer practice
and planting basins for 8 districts in southern Zimbabwe in 2004/2005, after Twornlow et
al. (2006a)
Province District Rainfall Farmer Basins
(mm) practice (kg ha-I)
(kg ha-I}
Matabeleland North Hwange 350 - 600 781 1 086
Lupane 350 - 450 524 796
Nkayi 530 702 1 134
Midlands Chirumhanzu 380 - 450 533 1 017
Zvishavane 380 531 336
Masvingo Gutu 458 301 395
Masvingo 260-410 603 1 010
Mwenezi 314 114 120
5.4.2.2 Inter field water management practices
In the Mzingwane catchment of southern Zimbabwe, national and international
organizations are promoting the use of inter field rainwater harvesting structures on
smallholder farms. The organizations involved include NGOs such as World Vision and
Practical Action, AGRITEX, and international research institutes such as ICRlSAT. The
techniques that are being promoted between fields include dead level contours with
underground storage tanks and infiltration pits, and planting basins (Plates 5.1 and 5.2). A
dead level contour is pegged at zero gradient and a standard graded contour is pegged at
1:250 gradient (AGRITEX, undated; Hughes and Venerna, 2005). The dead level
contours harness water originating from the area upslope. The dead level contours vary
in size, the minimum having a cross-sectional dimension of 1.5 m width and 0.5 m depth
(Mwenge Kahinda, 2004). The length of each contour varies according to the length of
the field. Currently there is no quantitative data on crop yield response to soil water
contributed by dead level contours with or without infiltration pits (see Chapter 8). Some
underground tanks are constructed from bricks (Plate 5.2) and often plastered. They have
a capacity of 2 m x 1m x 4 m and then the water collected from these underground tanks
88
is used by the household mainly for laundry or watering small gardens (Mwenge
Kahinda,2004).
Plate 5.1. Graded and dead level contours found in Insiza and Gwanda districts of
Matebeleland South
Plate 5.2. Dead level contour with storage tank and planting basins collecting rainwater
89
5.5 What the farmers said about soil water and fertility management
5.5.1Insiza district (Agroecological region IV)
5.5.1.1 Livestock system
Each household in ward 1 of Insiza district owns at least one of the following animal
species: cattle (Bos indicus), goats (Capra hircus), donkeys (Equus asinus), chickens
(Gal/us domesticus) and turkeys (Mel/eagris gal/opavo). The livestock numbers vary
greatly from household to household with the poorest households owning only chickens.
This is consistent with observations made in Insiza district during the 2005/06 and
2007/08 seasons (Chapter 6). Cattle, goats and donkeys are grazed on communal owned
rangeland which is highly degraded in ward 1. During the dry season cattle and donkeys
graze mainly along the Mzingwane river. Crop residues are either left in the field for
livestock to graze or are stacked at the homestead for dry season feeding. World Vision
started a programme of stover treatment with urea for dry season feeding. The
organization is encouraging farmers to make use of stover that comes from dry land and
irrigated crops.
5.5.1.2 Cropping system
5.5.1.2.1 Soil types and crops grown
The total land area owned by households ranges between l.2 and 3.2 ha. The
predominant soil type in the ward is granitic sandy soil with pockets of red clay soil in
some cropping lands. Maize is the predominant crop with cowpea, bambara nuts and
groundnuts being the common grain legumes. Sorghum is grown on a very limited scale
and pearl millet is no longer grown in the area. Participants in the focus group discussions
90
singled out bird damage as the main factor driving farmers away from growing pearl
millet in this area. All crops found in ward 1 of Insiza district can be grown on sand or
red clay soils (Table 5.3).
Table 5.3. Major soil types, crop areas and possible yields in a good growing season in
ward 1 of Insiza district
Crop Soils Area (ha) Yield (kg)
Maize grain Red clay, sand 1.2-1.6 1 250-1 500
Sorghum grain Red clay, sand 0.4-0.6 100-150
Cowpea grain Red clay, sand 0.1-0.2 15-30
Groundnut pods Red clay, sand 0.1-0.2 100-250
Bambara nuts pods Red clay, sand 0.1 25-200
5.5.1.2.2 Land preparation and planting
Less than 5 % of participants in the focus group discussions practiced winter ploughing
between cropping periods. The majority of participants only plough their fields after the
first effective rains in NovemberlDecember. The planting period stretches from October
to February depending on the rainfall pattern. The growing season in Insiza district
(Filabusi) normally starts between 2 November and 24 December (Chapter 4). Planting of
maize starts after the first effective rains usually late November to December. However
the cereal can be planted as late as mid-February if the rainfall pattern is good. Planting
of sorghum starts in October and ends in December. Legumes are planted during
November and December.
Maize is either dribbled behind the plough in the third furrow spaced at 0.75-0.9 m or
planted in shallow basins opened by a hand hoe. Third furrow planting is a common
practice in the smallholder farming system (Twomlowet al., 2006b). Sorghum is either
broadcast or dribbled behind the plough. The planting rows of spreading cowpea varieties
91
are spaced by two furrows apart while the erect varieties are planted in every furrow.
Groundnuts and bambara nuts are planted in the second furrow. In-row spacing for maize
and sorghum is usually 30 - 40 cm and 20 - 30 cm respectively. In-row spacing for
cowpea depends on the growth habit of the variety. Spreading varieties are spaced at 60
cm while erect varieties are spaced at 20 - 30cm. Groundnut and bambara nut are spaced
at 10 -15 cm in the row.
5.5.1.2.3 Weeding
Hand hoe weeding is the common practice although draught animal power owners use
cultivators followed by hand weeding to remove weeds close to the crops. All legumes
are weeded by the hand hoe. The weeding frequency for all crops depends on the rainfall
pattern. In a season with poor rains both cereals and legumes are only weeded once. A
season with above average rainfall pattern can force farmers to weed three times
especially in cereal fields. Legume fields are weeded twice at the most. The weeding of
maize and sorghum stretches from December to March while legumes are weeded
between late December and February.
5.5.1.2.4 Soil fertility management
Soil fertility management involves the use of both inorganic fertilizers and manure. The
. soil fertility amendments are applied to fields where maize is grown, all the other crops
are grown on residual fertility. Farmers apply up to 1 600 kgha' cattle manure in a
farming season following one with good rainfall events. Manure is available in small
quantities in ward 1 of Insiza district because households have small herds of cattle. This
92
is consistent with reports from other districts of Zimbabwe (Ncube, 2007; Mapfumo and
Giller, 2001). Cattle manure is applied to the fields between September and October and
is left in small heaps which are later spread at ploughing. Goat manure is commonly used
in vegetable gardens. Farmers who can afford inorganic fertilizers apply compound D at
20-30 kgha' and 10 to 20 kgha" ammonium nitrate. Basal fertilizer is applied at planting
while top dressing is done before tasseling of maize. Top dressing can be delayed
because of lack of adequate soil water.
5.5.1.2.5 Harvesting
The harvesting period stretches from March to May with cowpea, which normally
matures early, being the first crop to be harvested. Maize and sorghum are harvested
during April and May. Farmers are forced to finish harvesting by the end of May because
livestock especially cattle are difficult to keep away from the crop residue after this.
5.5.1.2.6 Crop rotation and intereropping
Complete crop rotations are rarely practiced in ward 1 of Insiza district. Participants of
the focus group discussions cited lack of adequate legume seed as the main constraint to
crop rotations. Maize is only shifted from a particular field if that field becomes infested
with the parasitic striga (Striga asiatica Kuntze) weed. Striga infested fields are allocated
to legumes especially cowpea and groundnuts to try and rid them of the parasitic weed.
Pumpkins and watermelons are intercropped with any cereal or legume. Cereal-legume
intercropping, if practiced, consists of a combination of maize and cowpea.
93
5.5.1.2.7 Soil water management
There are no in situ soil water management techniques being used by farmers except for a
few farmers who have just started participating in the conservation agriculture under the
DFID Protracted Relief Programme. Farmers who use the ox/donkey drawn cultivators
during weeding pointed out that furrows created during weeding collect rainwater, which
later infiltrates into the soil. All participants of the focus group discussions had graded
contours as runoff control structures in their fields. Construction of graded contours was
compulsory in the pre-independence period. Only a few farmers with draught power are
praeticing winter ploughing and they pointed out that winter ploughing helps in
conserving soil water.
5.5.1.3 Insiza district farmer risk perceptions
During the focus group discussions, prior to the experiments described in Chapter 7,
participants were comfortable with classifying farming seasons based on crop production
achieved in years with different rainfall patterns. The farming seasons were divided into
two categories, good and bad (Table 5.4). The average production is from the total area
put under each crop in a particular season per household in each district.
Table 5.4. Characteristics of bad and good farming seasons as observed by farmers in
ward 1 of in Insiza district
Crop Crop production Crop production
(bad season) (!kg) (good season) (kg)
Maize grain 15-50 750-1 000
Sorghum grain 0-50 750-1 500
Cowpea seeds 5-12 28-35
Groundnut pods 0 150-200
Bambara nut pods 0 150-200
94
Participants were asked to classify selected farming seasons and give reasons for the
crops that were grown during that time (Table 5.5). Farmers pointed out that they have
observed no relationship between the winter temperatures and rainfall pattern of the
following farming season. The choice of crop grown depended on the crop species that
were available for selection. However, the area under each crop especially maize and
sorghum would vary from season to season depending on the rainfall distribution and
amount received in the previous summer season. The area under sorghum would be
increased in a growing season followed a drought the previous year. The growing seasons
classified as bad by the participants were drought years (Chapter 4). The 1999/00 and
2005/06 seasons, classified as good by participants, both did receive above average
rainfall in the Mzingwane catchment.
Soil fertility amendments used in ward 1 were basal compound D (8: 14:7 - N:P20S:K20)
and top dressing ammonium nitrate (34.5 % nitrogen) fertilizer as well as livestock
manure (Table 5.6). Quantities of soil fertility amendments applied vary from season to
season depending on the prevailing rainfall pattern of the summer season. Farmers
decrease the quantities of basal inorganic fertilizer or livestock manure applied at land
preparation stage if the previous growing .season had poor rainfall. The quantities of top
dressing fertilizer are adjusted depending on the rainfall pattern of the current growing
season, a kind of response farming practice by smallholder farmers in semi-arid southern
Zimbabwe.
95
Table 5.5. Characteristics of different winter and growing seasons and crops grown as
reported by farmers in ward 1 of Insiza district
Season Weather pattern Crops grown Reasons for choice of
crops 2rown
1946/47 Drought (bad season), Pearl millet, sorghum Common crops grown
normal winter (Lundende, Tsweta), and seed was available
temperatures pumpkins
1971/72 Below average rainfall Maize, pearl millet, Common crops grown
but was not very bad sorghum, pumpkins, by the 1960s and
season, normal winter watermelons, 1970s
temperatures groundnuts, bambara
nuts
1979/80 Good season - above Maize, pearl millet, Common crops grown
normal rainfall, sorghum, pumpkins, by the 1970s and
normal winter watermelons, 1980s
temperatures groundnuts, bambara
nuts
1981/82 Severe drought (bad Maize, pearl millet, These were common
season), normal winter sorghum, pumpkins, crops grown by the
temperatures watermelons, 1980s
groundnuts, bambara
nuts
1991/92 Severe drought, Maize, sorghum, These were common
normal winter pumpkins, crops grown and
temperatures watermelons, previous season was
groundnuts, bambara good
nuts, cowpea
1999/2000 Good season, above Maize, sorghum, These were common
(Cyclone normal rainfall, pumpkins, crops grown by the
Eline) normal winter watermelons, 1990s and 2000s
temperatures. groundnuts, bambara
nuts
2005/06 Good season, above Maize, sorghum, These were common
normal rainfall pumpkins, crops grown by the
watermelons, 2000s
groundnuts, bambara
.. nuts, cowpea
Note: Lundende and Tsweta are traditional sorghum vaneties
96
The major risk factors highlighted by participants were drought, dry spells, lack of
appropriate seed varieties and inadequate seed supply, and lack of animal draught power.
Participants ranked the risk factors according to how severe they can impact the
household food security (Table 5.7). Participants unanimously agreed that drought had
the most severe impact on food availability in all households in ward 1. Drought is a
recurrent event in Insiza and Gwanda districts as shown from the rainfall analysis in
Chapter 4.
Table 5.6. Soil fertility amendments applied during bad and good seasons by farmers in
ward 1, Insiza district
Soil fertility amendment Quantity applied in
bad season good season
Manure 600-800 kgha" 1 200-1 600 kgha"
Compound D 0 40 kgha'
Ammonium nitrate 0-25 kgha' 60 kgha"
Table 5.7. Risk factors and their severity on household food security as identified by
farmers in ward 1 of Insiza district
Risk factor Ra IIIk
Drought 1
Dry spells 2
Lack of draught power 3
Lack of appropriate and adequate seed 4
4 = least severe; 1= most severe
5.5.1.4 Coping strategies in ward 1of Insiza district
The survival strategies used by farmers in the area during periods of drought were to
divert attention and look to off farm activities such as gold panning and selling of
livestock. Gold panning, although illegal in Zimbabwe, is the main survival activity
during drought years as ward 1 of Insiza district lies in the gold belt of Zimbabwe.
Farmers-tumed-miners sell their gold to legal and/or illegal gold dealers who frequent
97
their area. The major livestock species sold during periods of drought are cattle and goats.
Goats are normally sold first with cattle being sold when school fees or medical bills
have to be paid. Livestock are sold to private abattoirs or the Cold Storage Company of
Zimbabwe. Remittances, money sent home by household members working in cities or
other countries, cushion households especially those with family members working in
South Africa, Botswana and other countries. Household members working outside
Zimbabwe bring cash, groceries, clothes, household furniture and even building materials
when they return for visits to the family.
5..5.2 Gwanda district (Agroecological region J?
5.5.2.1 Livestock system
More than 90 % of the participants owned at least one of the following livestock species:
cattle (Bos indicus), goats (Capra hircus), sheep (Ovis aries), donkeys (Equus asinus),
chickens (Gal/us domesticus) and turkeys (Mel/eagris gal/opavo). Cattle, goats, sheep
and donkeys graze on communal pastures but cattle are sometimes also taken to rented
grazing paddocks (Mulageni) during the dry season. Donkeys are always grazed on the
communal pastures even in years of severe drought. They are sometimes fed on the crop
residues that were stacked at the side of the kraal from the crop field. World Vision and
Practical Action have started forming farmer groups to process stover collected after
harvesting into stock feed for dry season feeding.
98
5.5.2.2 Cropping system
5.5.2.2.1 Soil types and crops grown
The major soil type is sand (inhlabathi in Ndebele) with small areas having black clay
(isidaka in Ndebele) and red clay (isibomvu in Ndebele). Same areas have sadie soils
(isikwakwa in Ndebele) and these are not suitable for cropping, however, some
participants said they grow pearl millet on sadie soils. The main cereal crops grown are
pearl millet, sorghum and maize (Table 5.8). The most common grain legume is cowpea
with groundnuts and bambara nuts being grown to a lesser extent. Pumpkins and
watermelons are intercropped with cereals and a few farmers also grow sunflower and
sweet potatoes.
Table 5.8. Major soil types, crop areas and possible grain production in a good growing
season in Gwanda district
Crop Soil type Area Crop production
(ha) (ke)
Pearl millet grain Sand, red clay, 0.8-1.2 750-1 000
black clay, sadie
Sorghum grain Sand, red clay, 0.8-1.6 1 250-1 500
black clay
Maize grain Sand, red clay, 0.8-1.6 250-500
black clay
Groundnuts pods Sand, red clay, 0.1 100-400
black clay
Cowpea grain Sand, red clay, 0.1 25-100
black clay
Bambara nuts pods Sand, red clay, 0.1 50-250
black clay
5.5.2.2.2 Land preparation and planting
The total land area owned by households ranges from 3.2 to 4.9 ha. Although winter
ploughing is recommended for Gwanda south, none of the participants practiced winter
ploughing either for soil water conservation or for incorporating crop residue and weeds,
99
at the time of these discussions in March 2005. Conventional ploughing of the land starts
in November/December each year after the first effective rainfall and when draught
animals are strong. Farmers usually waste initial soil water because draught animals are
very weak as a result of inadequate feeding during the dry season.
Planting of cereals starts after the first effective rains usually late November to
December. The growing season in Gwanda district normally starts in December. Planting
of maize can stretch from December to February if the season is promising to be a good
one like the 2005/06 cropping season. Small grains (sorghum and pearl millet) are
planted before the end of December even if the season promises to be good. More than 80
% of the participants pointed out that they also do dry planting of pearl millet from mid-
October onwards before the first rain occurs. Legumes are also planted during November
and December each year at the onset of the rains.
Maize is planted in every 3rd furrow while sorghum and pearl millet are either broadcast
or dribbled behind the plough. In a year with low rainfall farmers prefer dribbling all
cereals behind the plough than planting in furrows opened by a plough. Participants
pointed out that in seasons with poor rainfall distribution during the planting period,
cereal crop establishment is better where the seed was dribbled behind the plough than in
furrows opened by a plough and covered with hand hoes. Maize is spaced at 30 - 40 cm
within the rows. In-row spacing for sorghum and pearl millet varies between 30 and 40
cm. Rows of cowpea and bambara nuts are spaced two furrows apart. For bambara nuts
this facilitates earthing up at pod development stage. Groundnut is planted in every
100
furrow. In-row spacing for groundnut and bambara nuts varies between 15 and 20 cm
while cowpea is spaced at 30 - 40 cm.
5.5.2.2.3 Weeding
Weeding starts in December every year especially in maize fields. Sorghum and pearl
millet are usually weeded twice, between late December and February. Maize is weeded
thrice especially in a season with above average rainfall pattern. The first weeding of
maize is done in December followed by a second weeding in January or February. A third
weeding in maize fields is done between the end of February and March. Legumes are
weeded once, usually in January, except for groundnut which can have a second weeding
between February and March in a season with good rains.
Draught power owners use animal drawn cultivators as well as hand hoes for weeding
cereals. When an animal drawn cultivator is used for weeding, farmers follow the
cultivator with hand hoes cleaning up any remaining weeds especially in the in-row
spaces. Non-draught animal owners can only use the hand hoe for weeding both cereals
and legumes.
5.5.2.2.4 Soil fertility management
The majority of participants appreciated the importance of managing soil fertility in their
farming system. However, livestock manure and inorganic fertilizers are not used despite
evidence of poor soil fertility such as declining crop yields and the high striga infestation
on most farms. Farmers have experienced crop burn in some growing seasons when they
101
have used manure and inorganic fertilizer and this makes them cautious. Livestock
manure and fertilizer are applied in gardens where soil water availability is guaranteed.
Those farmers that once applied manure or fertilizer to crops had done so basing their
decision on the rainfall distribution during the previous farming season. Farmers believe
that the legumes they are growing are improving the fertility of their soils.
5.5.2.2.5 Harvesting
Harvesting of cereal crops stretches from April to June and legumes are also harvested at
the same time. Harvesting of cowpea dry pods starts in March and continues up to May
depending on the variety. Local varieties are usually the last to be harvested. Farmers use
natural indicators like leaf colour of certain trees to make decisions on when to start
harvesting certain crops. In Gwanda farmers use the' Umvimila' tree as a signal of when
to harvest legumes especially bambara nuts. When the leaves of a 'Umvimila' tree start
turning yellow, it is the ideal time to harvest groundnuts and bambara nuts if these
legumes were planted between November and December.
5.5.2.2.6 Crop rotation and intereropping
Participants pointed out that crop rotation is only triggered off by the appearance of
striga. Striga usually appears in a season with a good rainfall pattern and this usually
happens after five or more years of continuous growing of a cereal crop in a given field.
Those that practice rotation pointed out that the rotation usually does not cover the whole
area that was under a cereal crop because of smaller legume areas. Usually legumes are
planted on less than 25 % of the cropped area in one season. Allocation of less land for
102
growing legumes has also been reported in other districts of southern Zimbabwe (Ncube,
2007). Maize and sorghum are usually intercropped with legumes (bambara nuts, cowpea
and groundnut). The intercrops are usually arranged such that a row of cereal alternates
with two rows of a legume. On some farms single cereal and legume rows alternate.
Sunflower, watermelons and pumpkins are intercropped with any cereal or legume.
5.5.2.3 Soil water management
5.5.2.3.1 Ward 17
There were no in-field soil water management practices in ward 17 in March 2005. All
participants had dead level contours with and without infiltration pits. Farmers have
observed that crops on the downslope side and close to the dead level contour with
infiltration pits take longer to show water stress in the event of a dry spell. Participants
unanimously agreed that dead level contours are suitable for both wet and droughty
seasons. In a wet season there would be more rainwater to collect resulting in a good crop
while in a droughty year the little rainwater received is captured by the dead level
contours and infiltration pits.
Participants pointed out that they took part in the dead level contours project because
Practical Action for Southern Africa was providing them with tools and food as the
farmers constructed the contours. The NGO encouraged them to form groups to assist
each other in the construction of the contours. The challenge facing the dead level
contours project is that it is labour intensive. An ordinary family in Gwanda south has
two to four able-bodied people and construction of the contours becomes an impossible
103
task for the family alone. The tools distributed by Practical Action have sparked friction
within some groups in all wards where focus group discussions were held.
5.5.2.3.2 Ward 18
The inter-field water management practices were dead level contours with and without
infiltration pits on most fields in ward 18. A few farmers did not have any inter-field soil
water management practices. None of the participants had in-field soil water management
practices. The dead level contours with infiltration pits that are present are being
promoted by NGOs. Farmers are organized in groups of 20 and they are given the
implements such as picks, shovels and wheelbarrows to use by Practical Action. World
Vision has been promoting the dead level contours through its Food for Work
Programme since 2004. Some farmers report that the crops look less water stressed in the
fields with the dead level contours but the yield benefits are still to be realized. Farmers
suggested improvements like reducing the field width through digging the dead level
contours closer to each other and making the infiltration pits larger.
5.5.2.3.3 Ward 20
There were no in-field soil water management practices being used on farmers' fields in
2005. However, a few farmers were using inter-field structures such as dead level
contours andfanyajuus. These practices were being used by less than 20 % of ihe people
who took part in the focus group discussions. Those farmers implementing the dead level
contours and fanya juus said they had received training from World Vision under the
Food for Work Programme in 2004. Those farmers that had dug the dead level contours
104
and infiltration pits said they had not yet seen any crop yield benefits from the dead level
contours, hence they were reluctant to undertake the strenuous work when crop yield
benefits were not guaranteed. Farmers who had dead level contours on their fields
suggested improvements of the dead level contour structure such as bunding the contour
so that more water could be collected in the contour during heavy rainfall events. None of
the participants in the focus group discussions were using winter ploughing as a soil
water management strategy.
5.5.2.4 Gwanda district farmer risk perceptions
Farmers classified good and bad seasons based on crop production as had been done in
Insiza district (Table 5.9). The decision to grow a certain crop depends on the rainfall
characteristics of the previous season. After a drought, farmers said that they grew more
drought tolerant crop varieties especially local varieties and decreased the area under
maize. From the farmers' assessment of different seasons there was no relationship
between winter temperatures and the rainfall pattern of the following farming season
(Table 5.10).
Table 59.. Crop pro duc IOn ill goo d andb ad farmmg seasons m·G. wan da d·istnet
Crop Crop production
Bad season Good season
Pearl millet grain 0- 50 kg 300-400 kg
Maize grain 0 500-725 kg
Sorghum grain 0- 50 kg 1 000-1 500 kg
Cowpea seeds 0 10 kg
Groundnuts pods 0 100-150 kg
Bambara nuts pods 0 150-200 kg
The majority of participants voted that crop failure in most years was caused by (i) low
105
and erratic rainfall; (ii) delayed start of the cropping season due to late rains; (iii) long dry
spells especially in January when the crop is entering the reproductive phase; and (iv)
inappropriate crop varieties (Table 5.11). Inappropriate crop varieties usually come
through donations by NGOs, church organizations or the Grain Marketing Board (GMB).
Table 5.10. Characteristics of different winter and growing seasons and crops grown as
observed by farmers in Gwanda district
Season Weather pattern Crops grown Reasons for choice of
crops
1946/47 Severe drought, below Pearl millet, sorghum, These were common
average rainfall, pumpkins, crops grown by then
normal winter watermelons
temperatures
1947/48 Good rain season, Pearl millet, sorghum, The season 1946/47
normal rainfall pattern, pumpkins, was very dry. Local
normal winter watermelons varieties such as
temperatures Tshweta and Lundende
sorghum were grown
1971/72 Below average rainfall, Maize, pearl millet, These were common
normal winter sorghum, pumpkins, crops grown by the
temperatures watermelons, 1970s
groundnuts, bambara
nuts
1991/92 Drought, normal Maize, pearl millet, These were common
winter temperatures sorghum, pumpkins, crops grown by the
watermelons, 1990s
groundnuts, bambara
nuts, sunflower
1999/2000 Very good season, Maize, pearl millet, These were common
(Cyclone above normal rainfall, sorghum, pumpkins, crops grown by 2000
Eline) normal winter '. watermelons, .
temperatures. groundnuts, bambara
nuts, sunflower
2005/06 Above normal rainfall Maize, pearl millet, These were common
sorghum, pumpkins, crops grown now
watermelons,
groundnuts, bambara
.. nuts, sunflower
Note: Lundende and Tsweta are traditional sorghum vanettes grown III southern ZImbabwe
106
Table 5.11. Risk factors and their severity on household food security as identified by
farmers in Gwanda district
Risk factor Rank
Low and erratic rainfall 1
Dry spells 2
Inappropriate crop varieties 3
Late start of season 4
1 = most severe; 4 = least severe
5.5.2.5 Copying strategies for farmers in Gwanda district
Households in wards 17, 18 and 20 survive by generating cash from livestock sales,
selling mopane worms from January to April, and remittances during periods of drought
induced hardships. All types of livestock are sold including donkeys when there is severe
drought. Goats and sheep are normally sold first and cattle follow especially when food
has to be bought and school fees paid. Poultry are sold most of the time whether there is
drought or not. Harvesting and selling Mopane worms is a big activity during the time
when the caterpillars are available from January to April. Households sell the worms to
other villagers and to buyers who come from Bulawayo and other cities. Recently an
NGO cailed SAFIRE formed community clubs that buy mopane worms from villagers
and resell in cities. The mopane worms also have lucrative markets further afield in the
Democratic Republic of Congo and Zambia. However, Mopane worms breed better when
the rainy season is average to above average. During times of severe drought these
worms are only available in very small populations. Almost every household in wards 17,
18 and 20 of Gwanda district has a family member working in one of the Zimbabwean
cities or other countries such as South Africa and Botswana. The support comes on a
regular basis in the form cash, clothes or groceries.
107
5.6 Conclusion
Despite the evidence of crop yield benefits derived from using some of the research
developed technologies, smallholder farmers in Insiza and Gwanda districts have not
adopted the available technologies. In ward 1 of Insiza district the only soil water
management technique used are the poorly maintained graded contours which direct
excess water from the field. In Gwanda district dead level contours, infiltration pits and
underground tanks are the technologies smallholder farmers are using for rainwater
harvesting in the field. These inter-field structures are being promoted predominantly by
Practical Action who provide farmers with implements for constructing the rainwater
harvesting structures.
The use of soil fertility amendments such as inorganic fertilizers and manure is still very
low in Insiza and Gwanda districts of the Mzingwane catchment. Only the smallholder
farmers in the relatively wetter Insiza district (NR IV) are using manure and inorganic
fertilizers. The farmers in Gwanda district are not using manure despite having more
manure from cattle and goats compared with Insiza district. The farmers fear crop bum
by inorganic fertilizer and manure, and they only apply the soil fertility amendments in
vegetable gardens where soil water is guaranteed through watering by buckets.
Farmers in both districts consider lack of rainfall to be their biggest challenge. Drought
was ranked as the major challenge to successful cropping in ward 1 of Insiza district. The
smallholder farmers in Gwanda district ranked low seasonal rainfall and its distribution
during the growing season as the major challenges to successful rainfed crop production.
108
The mid-season dry spells were ranked as the next serious constraint to cropping in both
Insiza and Gwanda districts. During the periods of drought, the farming families in Insiza
and Gwanda districts rely on off-farm activities for livelihoods. In Insiza district gold
panning is the main income generating activity while selling mopane worms dominates in
Gwanda. Households also rely on selling livestock in years of severe drought.
Households also rely on remittances as each household as a family member working in
the cities of Zimbabwe or more importantly, in neighbouring countries such as Botswana
and South Africa.
109
CHAPTER6
Farm Characteristics and Evaluation of Soil Water and Fertility Management Options
for Smallholder Semi-Arid Cropping Systems
6.1 Introduction
Smallholder farming systems in sub-Saharan Africa comprise crop and livestock sub-systems
with crops benefiting from livestock through manure and draught power. The livestock derives
feed and bedding from the crop residues. Crop and livestock production on these smallholder
farms is controlled by a wide range of socio-economic conditions and access to resources
(Defoer, 2002). Developing technologies aimed at increasing productivity of smallholder
farming systems is meaningless without input from farmers (Defoer et al., 2000).
A number of technologies have been developed for the smallholder farming system. These have
been developed from studies on soil water and fertility management, weed management and crop
varieties (Vogel, 1992; Nyamudeza, 1993; Nyakatawa et al., 1996; Twomlowand Bruneau,
2000). However, levels of adoption have remained disappointingly low despite the benefits
shown from experimentation. The reasons for lack of adoption vary from household to household
with different farming objectives, wealth status, labour availability and access to inputs reported
as some of the major drivers to technology uptake (Petersen et al., 1999).
Despite the lack of technology adoption, research and development work still needs to continue
as more challenges crop up in the smallholder farming sector. More recently, International
research organisations, NGOs and extension service departments have embarked on the
promotion of conservation agriculture in semi-arid smallholder farming areas of Zimbabwe. The
110
farming practices being promoted involve the use of both ripper and planting basin tillage
systems. At the end of three seasons of testing these technologies on their farms, smallholder
farmers in Gwanda and Insiza districts of southern Zimbabwe had an opportunity to evaluate the
performance of the four tillage systems. This chapter presents results on the characterization of
agricultural resource allocation using resource flow mapping and evaluation of tillage systems
through focus group discussions.
6.2 Objectives
The overall objective was to allow farmers to evaluate soil water and fertility management
options that are being promoted in Gwanda and Insiza districts. The specific objectives of the
study were:
(1) to identify agricultural resource allocation on smallholder farms m semi-arid southern
Zimbabwe; and
(2) to evaluate the performance of single and double conventional ploughing, ripper and planting
basin tillage systems after three growing seasons.
6.3 Materials and Methods
6.3.1 Resource now mapping
Resource flow maps were drawn following guidelines outlined by Defoer et al. (2000). The farm
was considered the unit of analysis, comprising three sub-systems namely household, cropping
and livestock. The flow of agricultural resources from one sub-system into another was indicated
by arrows on the map. During the mapping exercise each farmer drew maps of their homestead
and field on the ground. The maps indicated where various components of the farm were located
111
and types of crops grown in the different fields during 2005/06 and 2007/08. The maps were then
transferred onto manila sheets by researchers. Each farmer related sources and quantities of seed
for each crop grown and soil fertility amendments applied to different fields. The farmer was
asked to give an estimate of production obtained for each crop type grown. The flows of
nutrients, crop harvests and other resources on the farm were indicated on the map by arrows of
different line styles. After the drawing the map, the farm was toured to confirm the outlined set-
up and components of the farm given by the farmer (Plate 3.2). Farmers also participated in the
exercise documenting the overall cropping calendar for each district.
6.3.1 Focus group discussions
A checklist of guiding questions was developed by researchers before focus group discussions
were held (Appendix 1). One researcher from ICRlSAT facilitated the group discussions while
an AGRITEX officer assisted with facilitation and translation. During group discussions notes
were recorded on flipcharts by the facilitator while two other researchers recorded proceedings in
notebooks. Each group discussion lasted two to two and halfhours in each district.
6.4 Results and Discussion
6.4.1 Resource flow mapping
6.4.1.1 Resource endowment
The resource endowment analysis showed that four of 14 households had no cattle and for those
with cattle, the average number of cattle was 11 in 2005/06 and eight in 2007/08 per household
(Table 6.1). The Magaya family of Humbane village in Gwanda owned 40 cattle and all the
cattle were grazed at a rented paddock (mulageni) in the same district. The cattle numbers
112
obtained from this study were within similar ranges as those reported by Ncube (2007) in
Tsholotsho district through a resource flow mapping study and Homann et al. (2007) in a survey
in Beitbridge, Matobo, Binga and Nkayi districts of Zimbabwe. In the semi-arid Chivi district
cattle numbers were lower as reported by Chibudu et al. (2001) ranging from one to eight for
different resoureed households. A resource flow mapping study by Zingore (2006) also revealed
that smallholder farmers in the more humid northern districts of Zimbabwe own 10 to 16 cattle.
Goat numbers averaged between 18 and 25 per household during the two seasons studied. The
highest number of goats, 76, was recorded at the farm of John Ncube while the Mlilo family also
in Gwanda district had 28 donkeys. More goats and donkeys per household were recorded at end
of the 2005/06 season than the 2007/08 season (Table 6.1) and these numbers were higher than
those reported for Tsholotsho district by Ncube (2007). Goats are a main source of income in
Gwanda and Beitbridge districts of southern Zimbabwe (Homann et al, 2007). Donkeys are the
main source of draught power in Gwanda district with only a few households using oxen for
agricultural land preparation. Only the Magaya family in Gwanda district used oxen for draught
power while in Insiza district the main source of draught power was cattle.
The farming equipment analysis showed that each household had at least one plough and four
hand hoes (Table 6.1). Ownership of these farming equipment meant that each household could
prepare land for cropping through conventional tillage or embark on the use of planting basins.
Family size averaged eight people per household at the end of 2005/06 and seven after the
2007/08 season. The size of families ranged from six to 13 people per house at the end of
2005/06 and four to 10 in 2007108 growing seasons with an average of three family members
providing labour for agricultural activities. The household size was generally similar to results
113
from Tsholotsho, Chivi and Murewa districts of Zimbabwe (Chibudu et al., 2001; Zingore, 2006;
Ncube, 2007).
Table 6.1. Farmer resource status as an average of Gwanda and Insiza districts of southern
Zimbabwe
Criteria 2005/06 season 2007/08 season
Cropped area (ha) not measured 2.1 (0.34)
Livestock numbers
Cattle 11 (3.8) 8 (2.5)
Goats 25 (7.4) 18 (6.8)
Donkeys 7 (2.7) 5 (1.6)
Chickens 12 (2.8) 11 (2.9)
Asset numbers
Plough 2 (0.3) 1 (0.2)
Scotch carts 1 (0.2) 1 (0.2)
Hoes 4 (0.5) 4 (0.4)
Wheelbarrows 1 (0.3) 1 (0.4)
Bicycles 2 (0.6) 1 (0.4)
Family size 8 (0.7) 7 (0.6)
Number of members 3 (0.31) 3 (0.27)
providing labor
Grain required (kgyear') 912 (82.4) 829 (69.5)
Grain produced (kgyear') 891 (231) 262 (85.3)
Surplus/deficit (kgyear') -21 (221) -567 (113)
The numbersin bracketsindicatestandarderrorsofmeans.
Familysize includesadultsand children.
The major farming objective at all households was to provide enough food for the family. Cereal
grain requirements per household per year were average 912 kg in 2005/06 and 829 kg in
2007/08. In both seasons households had a grain deficit despite 2005/06 being a good
agricultural season and 2007/08 having fair rainfall distribution in the first half of the season.
Grain surplus/deficit was calculated as the difference between household annual grain
requirement and grain harvested at end of each season from the main cereals namely maize,
sorghum and pearl millet. All farms had a grain deficit in both 2006 and 2008 and this is typical
of semi-arid areas of Zimbabwe. The severity of the situation in 2008 could have been
114
aggravated by a general decline in the Zimbabwean economy and the political unrest, meaning
that little seed was available to plant in 2007. Surveys conducted by ICRISAT have revealed that
the majority of households in semi-arid southern Zimbabwe rarely produce enough grain to meet
their own household requirements (Twomlowand Rohrbach, 2006). Crop yields are always poor
regardless of rainfall and this has also been reported in other districts of Zimbabwe (Ncube,
2007).
6.4.1.2 Flow of agricultural resources
The flow of agricultural resources on two farms, one representing farmers hosting research trials
and another one representing farms without trials, is presented and discussed in this section,
others are given in Appendix 6.2. The resource flow maps show how agricultural resources were
allocated during the 2007/08 growing season and crop produce harvested from each field. Mr.
John Ncube's farm as illustrated in Figure 6.1 is located in Fumukwe village of ward 17 in
Gwanda district. The family of Mr. John Ncube comprises six adults and three primary school
going children. Two of the six adult family members are working in South Africa and
occasionally come to Fumukwe at the end of some months. Mr. John Ncube's total area is about
five hectares. The 72 year old, Mr. John Ncube started interacting with researchers at the
beginning of 2005/06 growing season and was not related to any member of the research team.
Before the arrival of researchers in 2005/06 season, Mr. John Ncube had already been using
livestock manure and inorganic fertilizer particularly as a topdressing fertilizer. Cattle and goat
manure are the major soil fertility amendments used at Mr. John Ncube farm. Farmers who have
had positive experiences with manure and inorganic fertilizer continue to use them despite the
lack of proper recommendations about application rates for semi-arid areas. The widely
115
promoted micro-dosing program in semi-arid areas of Zimbabwe has exposed farmers to the use
of low rates of inorganic fertilizers (Twomlowet al., 2008b). Before the start of the 2007/08
growing season, Mr. John Ncube bought 25 kg of urea from a neighbour who had received the
fertilizer through World Vision's agricultural relief program. Researchers gave Mr. John Ncube
11 kg of ammonium nitrate which was to be applied to the maize on the research field. Urea was
applied to maize while sorghum and pearl millet received neither inorganic fertilizer nor
livestock manure during the 2007/08 growing season. Results from a survey by Chibudu et al.
(2001) in semi-arid Chivi district showed farmers' preference of applying manure to maize
fields. According to Mr. John Ncube, cattle and goats provided 20 carts of manure each and all
the manure was broadcast onto the maize fields before ploughing operations were carried out.
Chicken manure and household waste were applied to the vegetable garden. All maize residues
were collected and stacked at the kraal for dry season cattle feed. Grain legume residues were fed
to the goats with donkeys having no stover allocation. All cereal and legume grain produced was
consumed at the household with pearl millet providing the bulk of cereal grain harvested at the
end of 2007/08 season.
116
25 kg urea
Toilet
o
o o o
0.22 ha Millet 0.18 ha Sorghum:
o
o
10 o
(20 carts (20 carts (180 kg grain) (45 kg grain) :
chicken
manure) manure)
(8.1 kg grain)
17 o
.,00
donkeys 0.27 ha G/nuts 0.27 ha B/nuts 0
(100 kg unshelled) (50 kg unshelled)
Nutrient flow
ooo~ Harvest
=+ Crop residue G/nuts - groundnuts
o Hutlkitchen B/nuts - bambara nuts
11 kg ammonium nitrate
Figure 6.1. Resource flow map for Mr. John Ncube of Gwanda district. The map represents flow
of resources at a farm that interacted with researchers for three cropping seasons
117
The farm of the 76 year old, Mrs. Malotha is located in Humbane village of ward 17 in Gwanda
district (Fig. 6.2). The map summarizes allocation of agricultural resources during 2007/08
growing season. The household of Mrs. Malotha comprises three adults and two primary school
going children. One of the three adult family members is working in Botswana and comes home
occasionally during the course of the year. The farm size for the Malotha family is about three
and half hectares. Despite having 16 goats Mrs. Malotha emphatically stated that she does not
use livestock manure and inorganic fertilizers because they bum her crops. This is a widely held
view by most smallholder farmers in semi-arid areas of Zimbabwe (Chibudu et al., 2001;
Mapfumo and Giller, 2001; Twomlowet al., 2008b). Mrs. Malotha said she once used
topdressing fertilizer that was donated by government in the 1980s and her sorghum crop wilted.
She attributed the poor performance of that sorghum crop to the fertilizer that she had applied.
During the 2007/08 growing season no soil fertility amendment was used at the farm and all
maize residues were stored for feeding goats. The farming system at Malotha farm could lead to
nutrient mining and consequently reduced soil productivity and crop yields. Continuous nutrient
depletion without replacement through organic and inorganic sources has been observed on
smallholder farms in sub-Saharan Africa (Smaling et al., 1997). Legume stover was grazed in
situ by her animals and those from the neighbourhood. The 300 kg of sorghum harvested was
consumed at the farm. Both farms relied on family members for all agricultural activities during .
2007/08 growing season.
118
0.45 ha Sorghum (300 kg grain)
o
•
o l 0.2 ha 0.2 ha 0.2 ha 0.3 haToilet
Granary J Maize GInuts BInuts Maize FII House
(0 kg) (0 kg) (0 kg) (0 kg)
o EJ \
\
\ "
House o \
\
4 chickens II "16 goats li> I I 2 donkeys
o 0 0 0 0[:> Harvest GInuts - groundnuts
,_, ... Crop residue BInuts - bambara nuts
o Hutlkitchen F - fallow field
<==:> Grazing in situ
Figure 6.2. Resource flow map for Mrs. Malotha of Humbane village, Gwanda district. The map
represents flow of resources at a farm that had no interaction with researchers from 2005/06 to
2007/08 seasons
6.4.1.3 Sources of seed for all farms
Pearl millet was only grown by farmers in Gwanda district and they planted retained seed in
2005/06, 2006/07 and 2007/08 growing seasons. The majority (58 %) of the farmers planted
119
retained sorghum seed while the bulk of maize seed (40 %) came from ICRISAT for the research
field (Fig. 6.3 and 6.4). The Grain Marketing Board (GMB) through its 'Operation Maguta/lnala'
program is also emerging as a major source of cereal seed (14-15 %) in both Gwanda and Insiza
districts. The 'Operation Maguta/Inala' was officially launched in November 2005 to assist
smallholder farmers with farming inputs. Farmers who were not hosting research trials planted
seed received from various NGOs namely World Vision, Practical Action and Organization of
Rural Associations for Progress (ORAP). All legume seed planted in the three seasons was either
seed retained on the farm or was bought from neighbours with quantities ranging from 0.25 to
two kg. Seed availability is a great constraint to farmers in Gwanda and Insiza districts, and this
observation is consistent with results obtained from both sub-humid and semi-arid districts by
other researchers (Ncube, 2007). NGOs have assumed a major role in providing both cereal and
legume seed through a variety of Agricultural Relief programs in recent years. Access to good
quality seed of appropriate crop types is crucial to increase production on smallholder farms.
Sorghum
Figure 6.3. Average sources of sorghum seed planted by 14 households in 2005/06,2006/.07 and
2007/08 seasons in ward 1 of Insiza and ward 17 ofGwanda district
120
Maize
21%
ITIGIMB LINGO !SI AREX ~ ICRISAT LIRetained § Bought
Figure 6.4. Average sources of maize seed planted by 14 households in 2005/06, 2006/07 and
2007/08 seasons in ward 1 of Insiza and ward 17 of Gwanda district
6.5 What the farmers said
6.5.1 Land preparation
Crop production in Gwanda and Insiza districts is done at a subsistence level. Tables 6.2 and 6.3
summarize agricultural activities and time when they were carried out on the conventionally
ploughed farmer's fields and research plots during 2005/06 and 2007/08 growing seasons. On the
research plots, single conventional ploughing was carried out during November and December
depending on the onset of the rains. The first ploughing in double ploughing tillage system (for
the research plots) was done between August and October each year (Table 6.4) with farmers in
Insiza district generally carrying out the operation earlier than those in Gwanda district. All
farmers broadcast cattle manure before ploughing in both single and double ploughing tillage
systems. Ripping was done in October and November (Table 6.5) while digging of planting
121
basins was carried out during September and October each year (Table 6.6). Manure was banded
along the ripped furrow in the ripper system and applied soon after ripping. In the basin system
manure was placed in each planting basin soon after the basins were opened (Table 6.6).
Table 6.2. Cropping calendar for the farmer practice fields in Gwanda and Insiza districts during
2 2006/07 and 2007/08 .
Table 6.3. conventional ploughing in Gwanda and Insiza
Table 6.4. Cropping calendar for the double conventional ploughing in Gwanda and Insiza
2005/06 2006/07 and 2007/08 flTr\'lI"n
122
Table 6.5. Cropping calendar for the ripper system III Gwanda and Insiza districts during
200 2006/07 and 2007/08 urn'.."n
Table 6.6. Cropping calendar for the basin system in Gwanda and Insiza districts during 2005/06,
2006/07 and 2007/08 .
6.5.2 Planting
Planting period for all tillage systems stretches from November to December and is also being
extended into January by farmers under their traditional methods. Planting date relies on the
onset of the effective rains which normally vary from season to season. Planting basins and
ripper tillage systems were planted first during the three seasons of experimentation. The earliest
planting in these systems was done during the first week of November whilst the last planting
date recorded between 2005 and 2008 was 29 December. Thus planting basin and riper tillage
systems allow timely land preparation and planting (Nyagumbo, 2007; Twomlowand Hove,
2007). Farmers both with or without draught power can make effective use of the planting rains
in November-December if land is already prepared for planting. Maize planting in farmers'
fields, single and double conventional ploughing plots was done in December across all the
seasons. Dry planting of pearl millet on other fields of some farms was done in October with
123
sorghum and maize being planted between November and December. Cropping calendars
presented by Twomlowet al. (2006b) showed that planting on smallholder farms in semi-arid
areas can begin in October while harvesting stretches up to May. Planting of maize was done in
furrows opened by donkey/ox drawn plough. Although smallholder farmers normally plough-
plant in the third furrow (Chapter 5; Twomlowet al., 2006b) farmers and researchers agreed at
the beginning of the trials that third furrow plough-planting will not be done on the research plots
so maize seed was planted in furrows that were opened by a donkey drawn plough.
6.5.3 Weed management
In the farmers' fields weeding was done by donkey-drawn cultivator first, followed by hand
weeding. In the four experimental tillage systems weeding was done by hand hoeing throughout
the three seasons· of study. In the planting basin and ripper systems first weeding was done soon
after planting between end of November and early December (Tables 6.5 and 6.6). The second
weeding was done in the first two weeks of January on farmer practice, single and double
ploughing treatments (Tables 6.2, 6.3 and 6.4). In 2005/06 growing season farmers reported high
weed pressure and the majority of them weeded for the fourth time in March 2006. Weed
management is one of the major challenges to adoption of conservation agriculture systems
(Nyagumbo, 2007). Frequency of weeding depends .on the rainfall pattern for each season with
single weeding common in a dry cropping season and three or more weeding in wet seasons
(Van der Meer, 2000). In the single and double ploughing systems (Tables 6.3 and 6.4), and
farmer practice field (Table 6.2) weeding started in January and second weeding was done in
February. None of the participants practiced winter weeding in the planting basin and ripper
tillage systems. The reasons highlighted by the farmers were (i) there will be no crop in field; (ii)
124
there are other activities that are more important than weeding during winter period; and (iii) it is
'socially strange' to weed in winter when there are no crops growing 'neighbours will laugh at
you'.
6.5.4 Crop establishment
Farmers observed that the best maize crop establishment was in the basin system regardless of
rainfall distribution during the November-December period. Maize crop establishment in the
ripper system was ranked second with single and double conventional ploughing having similar
plant stands despite different rainfall patterns observed during 2005/06, 2006/07 and 2007/08
growing seasons.
6.5.5 Labour requirements
Based on a plot size of 50 m x 50 m (0.25 ha) farmers estimated the labour requirements for land
preparation and weeding for the different tillage systems (Tables 6.7 and 6.8). Planting basin
system had the highest labour demand across all households during 2005/06, 2006/07 and
2007/08 growing seasons. The farmers unanimously agreed that reopening planting basins was
generally easier than establishing them for the first time. Participants in the group discussions
described the soil as 'softer' when they were reopening planting basins at the beginning of
2007/08 growing season. It has been observed in other districts of Zimbabwe that once
established, reopening planting basins in the same position in successive years becomes easier
(Twomlowand Hove, 2006). Ripping to a depth of 0.15-0.18 m at 0.9 m inter-row spacing
required the least number of people and took the shortest time.
125
Table 6.7. Farmer assessment of labour demands for land preparation on a 0.25 ha plot of
different tillage systems tested on their farms averaged for three growing seasons
Tillage system Number of people Time taken (days or hours)
Plough-planting 3 3 hours
Double conventional ploughing 2 (3) 3 (3) hours
Ripping 2 1 hour
Planting basins (first time) 10 3 days
Planting basins (reopening) 7 3 days
Figures in brackets stand for number of people and time taken for the second conventional ploughing
The highest labour demands for weeding were in the ripper and planting basin tillage systems
regardless of the rainfall pattern (Table 6.8). Some farmers resorted to weeding using ox/donkey
drawn cultivator on their fields especially during 2005/06 season. Hand weeding or ox-drawn
cultivation followed by hand hoeing are common practices for weed control (Chapter 5).
Frequency of weeding depends on the rainfall pattern for each season with single weeding
common in a dry cropping season and three or more weedings in a wet year (Van der Meer,
2000). Data available from research studies conducted in Zimbabwe clearly show that timely and
effective weed control improves soil water conservation and crop yields (Vander Meer, 2000).
Studies conducted by Shumba et al. (1992) at Matopos and Makoholi research stations also
showed that early.weeding resulted in 40% more maize grain than the farmers' practice which
was used as a control.
126
Table 6.8. Farmer assessment of weeding methods, frequency and labour demands during
seasons W.Ith diIff erent ram. iia11pattern m. Gwan da and Inssiiza d·istnets
Tillage Weeding method Weeding frequency Labour requirement
system (mandays)
Below Above Below Above Below Above
normal normal normal normal normal normal
rainfall rainfall rainfall rainfall rainfall rainfall
Plough Animal Animal 1 2-3 4 5
plant drawn drawn
cultivator cultivator
followed by followed by
hand hoeing hand hoeing
Single Animal Animal 1 2-3 4 5
ploughing drawn drawn
cultivator cultivator
followed by followed by
hand hoeing hand hoeing
Double Animal Animal 1 2-3 4 5
ploughing drawn drawn
cultivator cultivator
followed by followed by
hand hoeing hand hoeing
Ripper Hand hoeing Hand hoeing 2 3-4 5 8
and hand
pulling
Planting Hand hoeing Hand hoeing 2 3-4 5 8
basins and hand
pulling
6.5.6 Soil fertility management
The majority of participants in focus group discussions in Gwanda district did not use livestock
manure or inorganic fertilizers in their farming system. The farmers believe that inorganic
fertilizers and manure bum their crops, while others within the same group of farmers were of a
strong opinion that their soils contain enough plant nutrients to sustain crop production. This
confirmed earlier survey reports by Ahmed et al. (1997) that revealed low fertilizer usage by
smallholder farmers in semi-arid southern Zimbabwe. Given the results from resource flow
mapping and focus group discussions nutrient- mining from agricultural fields could be rampant.
127
~
\
m Gwanda district. Smallholder farmers in Insiza district had been usmg both inorganic
fertilizers and livestock manure before researchers arrived in their area. Compound D fertilizer is
applied at planting by dribbling along planting furrows or placed in each planting hole next to the
seed. Livestock manure is applied in September/October as small heaps and broadcast later at
ploughing after effective planting rains. Quantities of manure applied vary from 4 to 6 tha·l.
Although the national extension service recommends 8 tha-I for semi-arid districts (Mapfumo
and Giller, 2001), studies conducted by Ncube et al. (2007) in Tsholotsho district showed that 3
to 6 tha-I supplemented with low rates of nitrogen fertilizer significantly increase cereal yields.
Topdressing fertilizer in Insiza district is applied to maize from the end of January through to
February. Some farmers use teaspoons for measuring quantities of fertilizer (either basal or
topdressing type) to apply per planting station while others use tips of their fingers. Farmers in
the focus group discussions normally applied 25 kg of either compound D or ammonium nitrate
received from NGOs to a 50 m x 50 m plot of maize, sorghum or pearl millet. This is the
equivalent of 200 kgha' of either the compound D or topdressing fertilizer, which is the blanket
recommendation for smallholder farming systems of southern Zimbabwe from AGRITEX.
6.5. 7Soil water conservation and the effect of dry spells on a maize crop
Participants were asked to share their observations on the effect of dry spells on the maize crop
growing under single and double ploughing, planting basins, ripper and farmer practice fields.
The ripper system was overwhelmingly voted as the tillage practice most affected by dry spells.
Only one farmer from Gwanda district, Mr. John Ncube, observed the least water stressed maize
crop in the ripper system during dry spells over the three seasons of experimentation. During
2006/07 which was a drought year, farmers observed maize plants wilting first in the ripper
128
plots. The double ploughing and planting basin tillage systems were nominated by the farmers as
the best systems that managed to extend the soil water availability to crops during dry periods.
The impact of dry spells on maize grown under single conventional ploughing was less than that
on ripper tillage system. The fact that there was lower soil water availability under the ripper
system compared with double ploughing and planting basin systems was confirmed by field
measurements taken during 2006/07 growing season (Mupangwa et al., 2008; Chapter 7).
6.5.8 Crop yield assessment
Farmers used maize yields achieved in 2005/06, 2006/07 and 2007/08 growing seasons, to rank
the four tillage systems and results are summarized in Figure 6.7. The double ploughing tillage
system was ranked the highest yielding system by 46 % of farmers in the focus group
discussions. Planting basin system was ranked second while all participants observed that the
single ploughing tillage system was the lowest yielding across three seasons (compare yields in
Chapter 7).
129
OP
46%
Ripper
18%
Figure 6.5. Farmer ranking of basin, double ploughing (OP) and ripper tillage systems tested on
their fields for three growing seasons in Gwanda and Insiza districts. Ranking was based on
maize grain yields achieved over three growing seasons
6.5.9 Pests and diseases
The pests observed across the three seasons of experimentation are shown in Table 6.9. Mice
were a menace in planting basins particularly during 2006/07 season when there was a drought.
This was also reported in other semi-arid districts where planting basins were being promoted
during the same season. The armored cricket attacked the maize grain at milk dough stage and
the attack was more severe in 2005/06 and 2007/08 seasons. No diseases were observed on the
maize crop in the on-farm research plots during the three growing seasons.
130
Table 6.9. Pest species in different tillage system affected as observed by farmers in Gwanda and
Insiza districts during 2005/06,2006/07 and 2007/08 growing seasons
Pest Tillage system Crop growth stage affected
Green grasshopper All Vegetative and reproductive
Mice Basins Between planting and crop emergence
Cutworms All Soon after crop emergence
Birds All Between planting and crop emergence
Termites All Harvest maturity
Annored cricket All Milk dough
6.5.10 Preferred til/age system and crops to be grown
The participants chose tillage system(s) that they would prefer to continue using In future
without the researchers. The choice of tillage systems revealed some sharp differences between
husbands and wives participating in focus group discussions. The female farmers were equally
divided between selecting to continue with double ploughing and planting basin tillage systems
(Fig. 6.6). Female farmers who preferred planting basins gave the following reasons for their
choice: (a) crop establishment is better in planting basins and if the rainfall season is good there
are better chances of high yields and (b) manure and topdressing fertilizer can cover a larger area
of cropped fields when maize is grown in planting basins as it is placed only in each basin.
However, the female farmers highlighted the challenge of the high labour needed to dig planting
basins. They suggested working in groups, as the best way forward, during land preparation
stage.
131
Women
Basins DP
43% 43%
Ripper
Men
OP
62%
Figure 6.6. Tillage systems chosen for adoption by female and male farmers during focus group
discussions in Gwanda and Insiza districts
The majority (62 %) of male farmers strongly voiced their concern over labour requirements for
the basin system and they selected the double ploughing system. The opinion of male farmers on
using of planting basins can be summarized by this statement which came from a farmer who
132
worked with researchers for three seasons 'Ukhulima ngamagodi hukhuhambisa umuntu ejele
leKhami' which literally means 'farming using planting basins is like taking someone to Khami
prison'. Male farmers gave the following reasons for choosing the double ploughing system: (a)
maize plants in double ploughed plots showed less soil water stress during dry spells and highest
grain yields were obtained from the double ploughing system; (b) weeding was not a problem
even in growing seasons with above-normal rainfall pattern such as 2005/06 and (c) land
preparation is done using ox/donkey drawn plough. However, all the farmers proposed growing
pearl millet, sorghum, traditional spreading cowpea varieties and groundnuts under the ripper or
basin tillage systems.
6.5.11 Technical and material support
The farmers who participated in the focus group discussions identified sources of farming advice
that they had received before ICRISAT researchers came to the area and during the period of
hosting research trials. They also highlighted the support they had received from each
organization and the support expected in future (Table 6.10). Ranking of organisations based on
support rendered to the farmers was done based on a scale of one to five with the former being a
lot of support and the latter representing least support. The AGRITEX was given a low rank in
both Gwanda and Insiza, as they are facing many challenges ranging from lack of financial and
material resources to high staff turnover. The high staff turnover has a negative impact on
agricultural development programmes such as conservation agriculture (CA) (Hove, 2006).
Practical Action was ranked highly in Gwanda district because farmers appreciated the rainwater
harvesting program that brought dead level contours, tools such as wheelbarrows and shovels,
and food rations during the construction of the contours.
133
Table 6.10. Farmer ranking of organisations working in Gwanda and Insiza, and support
rendered and expected by farmers in Gwanda and Insiza districts. (1 = lot of support; 5 = little
support)
Organisation Rank Support provided Future support required
AGRlTEX 4 (Gwanda o Pegging contour ridges I) Training qn fertilizer and
and Insiza) manure use
o Farmer training in crop o Crop pests and disease
production control
o Animal pests and disease
control
o Striga control
o Frequent visits to farmers'
fields
Practical Action 1 (Gwanda) o Setting up dead level o Setting up storage tanks
contours and infiltration along dead level contours
pits
o Food and tools for dead
level contours' program
World Vision 2 (Insiza) o Seed and fertilizer o Seed and fertilizer
4 (Gwanda) o Training on fertilizer and
o CA training manure use
o Crop pests and disease
control
ICRISAT 1 (Insiza) o CA training o Extension materials
2 (Gwanda) o Extension materials e.g. o Modify the ripper tine so that
planting basins and it creates bigger furrows .
ripper calendars
o Frequent visits to
farmers' fields
Grain Marketing 3 (Gwanda o Seed and fertilizer o Seed and fertilizer
Board and Insiza)
Lutheran 5 (Gwanda) o Seed and fertilizer for o Seed and fertilizer for
Development nutrition gardens nutrition gardens
Services o Training on vegetable
production
6.6 Conclusion
All households who participated in the resource flow mapping exercise own a piece of land
where crop production is done at a subsistence level. Each household owns at least one of the
following animal species: free range chickens, goats, donkeys and cattle. Each participating
134
household had access to a plough and hand hoes, and an average of three family members
provide labour for agricultural activities.
The common agricultural resources used by better resoureed households in Gwanda and Insiza
districts are seed, manure and inorganic fertilizer. The low resoureed households particularly in
Gwanda district (NR V) do not use any soil fertility amendments regardless of the in-season
rainfall pattern. This implies that low resoureed households miss out on opportunities of better
crop yields when rainfall conditions are favourable as was the case in the 2005/06 growing
season. The households that do not apply organic and inorganic soil fertility amendments still
believe that manure and fertilizer bum their crops and should not be used in southern Zimbabwe.
These farmers require training on manure and fertilizer application quantities and field days
(green and brown shows) should be organized at the fields of their colleagues who are applying
soil fertility amendments. Research and extension agents have some work to do to demystify
such perceptions as there are dangers of serious nutrient mining and hence soil degradation in
semi-arid smallholder cropping' systems. Technologies such as microdosing that have proved
beneficial in other semi-arid districts should be promoted more widely to reach out areas such as
the southern parts of Gwanda district. The households who are using soil fertility amendments
have observed the benefits of addressing soil fertility and now know the quantities of manure and
fertilizer to apply. The accessibility of agricultural inputs such as inorganic fertilizers needs to be
improved for the benefit of farmers such as Mr. John Ncube.
During the evaluation exercise, farmers did not just consider maize yields, they also took into
consideration auxillary benefits such as good crop establishment, quantities of inputs used, area
135
covered by a given quantity of inputs and labour demands. Farmers appreciated that farming
using planting basins and ripper systems allows timely planting because land is prepared well
before the onset of the rains. This therefore reduces demand on limited family labour during the
planting period. The focus group discussions showed that husbands and wives, coming from the
same household, differed in their choice of technologies to be adopted on their farms. This
demonstrates how individual farmers, viewed by change agents as all vulnerable, differ on what
they see as opportunities and constraints of different soil and fertility management technologies.
A concerted effort by all service providers in Insiza and Gwanda districts should be made to
reach out to farmers with proven technologies. It has been seen that some of the technology
transfer methods have been successful with specific farmers. An investigation into those methods
should be made so as to use them again. An alternative is to use farmer to farmer transfer by
promoting groups and visits to each other's farms so as to enable learning and capacity building
and still transfer the technologies.
136
CHAPTER 7
Integrated Tillage and Nitrogen Management for
Improved Soil and Water Productivity on Smallholder Farms in Semi-Arid
Zimbabwe
7.1 Intreductlen
Smallholder agriculture in semi-arid areas is facing many challenges including low,
erratic and highly variable rainfall between and within seasons (Rockstrëm et al.,
2002; Twomlowet al., 2006b). Droughts and intra-seasonal dry spells are common in
semi-arid southern Zimbabwe. To reduce the impact of the frequently occurring
droughts and dry spells on smallholder cropping systems, soil water management
technologies have been developed (Vogel, 1992; Nyamudeza, 1993; Nyakatawa et al.,
1996; Twomlowand Bruneau, 2000). The findings from these studies have indicated
that the period of soil water availability to crops can be prolonged even on granitic
sandy soils.
The low crop yields achieved from the smallholder cropping systems are often blamed
on the rainfall pattern during the growing season. However, when rainfall is adequate
crop production is limited by the low fertility status of the predominantly granitic
sandy soils in the smallholder sector. There is continued nutrient off take in harvested
crops without replenishment through use of organic and inorganic nutrient sources
(Stocking, 2003). The major chemical and physical constraints with the sandy soils in
smallholder farming system include the inherent low levels of organic matter,
nitrogen and phosphorus, and the poor soil structure (Grant, 1981; Nyamangara and
Mpofu, 1996; Twomlowet al., 2006b). Studies conducted in the northern
137
(Mtambanengwe and Mapfumo, 2005) and southern (Ncube et al., 2007) parts of
Zimbabwe demonstrated that livestock manure is a key resource for improving
productivity of the sandy soils on the smallholder farms. Crop yields in the
smallholder farming systems can be increased substantially by supplementing manure
with nitrogen fertilizer (Mubonderi, 1999; Nhamo, 2003; Ncube et al, 2007). The
wide scale promotion of the use of low rates of nitrogen fertilizer (10 kgNha-l) in
semi-arid smallholder cropping systems demonstrated that cereal yields can be
increased by 30 to 100 % in the conventional tillage system (Twomlow er al., 2008b).
Deep winter ploughing is a technique which has been recommended by the
government's department of agriculture research and extension (Mupangwa et al.,
2006). The technique involves ploughing to a depth of 0.2 - 0.23 m soon after
harvesting. However, smallholder farmers only plough to a depth of 0.1 - 0.15 m
resulting in the formation of a hardpan within the soil profile (Tsimba et al., 1996).
The plough pan restricts root penetration and rainwater penetration to deeper soil
layers. Crops grown under shallow ploughing cannot withstand extended periods of
soil water stress during mid-season dry spells.
The planting basin and ripper tillage techniques are being promoted on smallholder
farms in semi-arid Zimbabwe (Mupangwa et al., 2006). These tillage systems
combine water with nutrient management practices. Planting basins are targeted at
households that have little or no access to animal draught power (Twomlowand
Hove, 2006) while the ripper tillage system was designed for-households with some
draught power. The two tillage systems emphasize the spreading of labour
requirement throughout the year, the precise application of nutrients and timeliness of
138
management operations. Land preparation in the planting basin and ripper systems
starts in October (Chapter 3). Rain water collects into the basins or ripper furrows
during the early season rainfall events (October and November) allowing those parts
of the field to be wetter than the rest of field, even with little rain. Studies conducted
in Botswana and Zimbabwe have demonstrated that double spring ploughing can
increase crop yields in a wide range of soil types under semi-arid conditions
(Twomlowand Bruneau, 2000). This chapter presents results obtained from three
seasons of participatory on-farm experimentation in Insiza NR IV and Gwanda NR V
districts of semi-arid Mzingwane catchment of southern Zimbabwe.
7.2 Objectives
This study was designed to use on-farm trials in farmers' fields
(1) to quantify soil water content under single and double conventional ploughing,
ripping and planting basin tillage systems;
(2) to quantify runoff water losses from single and double conventional ploughing,
ripper and planting basin tillage systems;
(3) to determine effect of the four tillage systems and nitrogen fertilizer on maize crop
performance; and
(4) to determine water use and agronomic nitrogen use efficiencies under the four
tillage systems.
7.3 Materials and Methods
7.3.1 Experimental design and layout
The experimental design was split plot with tillage method as the main plot factor,
nitrogen fertilizer level as the sub-plot and each farm was considered as a replicate
139
with districts being the blocks in the study. The four tillage treatments were planting
basins (Basins), tine ripping (Ripper), single (CP) and double (DP) ploughing. The
three nitrogen rates were 0, 10 and 20 kgblha' applied as ammonium nitrate (34.5%
N). Each tillage main plot measured 20 m x 10m in 2005/06 season and 30 m x 10m
in 2006/07 and 2007/08 seasons. In the first season only 0 and 10 kglvha" were used
while a third rate of 20 kgNha-1 was introduced in the second and third growing
seasons.
In 2005/06 season plots were pegged out in October 2005 before all tillage operations
commenced. In 2006/07 season new plots were pegged out in the same fields because
main plot size had increased from 20 m x 10m to 30 m x lOm. The plots established
in 2006/07 season were maintained for 2007/08 growing season. Planting basins were
dug at 0.9 m x 0.6 m spacing using a hand hoe and each basin measured 0.15 m
(length) x 0.15 m (width) x 0.15 m (depth). Rip lines were opened at 0.9 m inter-row
spacing using a ZimPlow ripper tine attached to the beam of an ox or donkey-drawn
mouldboard plough (Plate 3.4c, Chapter 3). The first conventional ploughing for the
double ploughing treatment was carried out in September/October each year. The
second ploughing was done at the same time as the single conventional ploughing
treatment soon after the first effective rain (20 to 30 mm) in NovemberlDecember
each year using a donkey-drawn mouldboard plough. Shallow planting furrows were
opened by donkey-drawn plough in the single and double ploughing treatments (open-
plough furrow planting).
140
7.3.2 Experimental management
Planting on all farms occurred between late November and December depending on
the start of the rainy season. A short season hybrid maize variety, SC403, was planted
at all farms during each of the three cropping seasons. Three kernels were planted per
basin in the basin tillage system. Two kernels were planted per station at 0.3 m in-row
spacing in the ripping, single and double conventional tillage treatments. Plants were
thinned to two per basin in planting basin and one plant per station in the ripping,
single and double conventional tillage treatments two weeks after planting.
Cattle manure was applied each year at a rate of 3 tha·1 in all plots under planting
basins, ripper, single and double ploughing tillage treatments as basal soil fertility
amendment. Manure was placed in the planting basins and dribbled along the ripline.
In the single and double ploughing tillage systems manure was broadcast just before
the ploughing operation. The annual 3 tha-1 application rate for manure is the current
recommendation under the Protracted Relief Program (PRP) for the smallholder
farming sector where conservation agriculture is being promoted. Ammonium nitrate
was spot applied on top of the soil between January and February and each tillage
main plot was subdivided into two during the first season and into three subplots
during the 2006/07 and 2007/08 seasons. The 10m x 10m subplots were for
accommodating the N treatments. Weeds were controlled by hand hoe at all farms
across all three seasons.
141
7.3.3 Data collection
7.3.3.1 Soil analyses
All soil properties were determined using procedures outlined by Anderson and
Ingram (1993). Soil pH was determined by the water method in a 2:5 soil:water ratio.
Soil organic carbon, and total N and P were determined by the Sommers and Kjedahl
methods (Anderson and Ingram, 1993). Available P was extracted by the bi-carbonate
procedure and determined calorimetrically. Total Nand P in plant tissue were also
determined using the Kjedahl procedure. Soil texture was determined by the
hydrometer method as outlined by Anderson and Ingram (1993).
7.3.3.2 Soil water content and daily rainfall
During the first season (2005/06) soil water was measured monthly by the gravimetric
method at five farms namely Moyo and Mpofu in Insiza district, and J. Ncube,
Sibanda and Siziba in Gwanda district. Soil samples were collected at 0.15 m depth
intervals up to 0.6 m by steel cores measuring 0.03 m internal diameter by 0.95 m
length. Soil samples were oven dried at 105°C for 48 hours before determining
gravimetric water content. Volumetric water content was calculated using bulk
density and gravimetric water content of each soil layer (Anderson and Ingram, 1993).
During the 2006/07 season, soil water was measured at Mpofu and Nyathi farms in
Insiza district and, 1. Ncube, Sibanda and Siziba farms in Gwanda. In the 2007/08
season, soil water was measured at Mpofu, N. Ncube and Nkomo farms in ward 1 of
Insiza district. In Gwanda district soil water was measured at 1. Ncube, Sibanda and
Siziba farms. During the 2006/07 and 2007/08 seasons volumetric soil water content
was measured using a capacitance probe (microgopher sensor type). Depth of access
tubes varied between 0.6 and 0.8 m and three access tubes were installed per tillage
142
main plot. Depth of access tubes was restricted by presence of a lateritic layer in the
profile especially at experimental sites in Insiza district. Soil water was measured
fortnightly at 0.1 m depth increments during 2006/07 and 2007/08 seasons. Soil water
content in millimetres was determined by multiplying volumetric water content by
thickness of each layer from which soil water was measured. Each farmer was given
an ordinary plastic raingauge for recording daily rainfall during the growing season.
7.3.3.3 Surface runoff
Runoff water collected from each tillage treatment plot was measured soon after each
runoff-generating rainfall event. The height of water in the drum was measured by
staff gauge graduated from 0 to 1.5 m. The drums were emptied after each rainfall
event. Runoff water depth from each treatment was calculated by subtracting the
volume of water contributed by the 10m x 1.7 m triangular receiving surface from
total water volume measured in the drum.
7.3.3.4 Maize plant stand and yield
In every season plant counts after thinning in each treatment were done two weeks
after crop emergence. At harvest, grain and stover (above-ground biomass minus
grain) yields were measured from 10m x lOin subplot for each tillage treatment. The
weight of cobs and stover from the subplot of each tillage treatment was determined
in the field before taking sub-samples for moisture correction. Grain and stover
samples were dried at 60°C for 48 hours for moisture adjustment using formulae
outlined in Equation 7.1. The maize shelling percentage was determined for each
treatment for converting cob weight into grain weight. Grain weight was converted to
a per hectare basis at 12.5% moisture content as final grain yield. The harvest index
143
was determined as a proportion of grain mass to total above-ground biomass mass
(grain, cores and stover).
Grain yield (kgha', @12.5%) = [(l-MC)I(l-O.125)} x FGW Equation 7.1
Where MC = moisture content (%) measured in the cobs after drying for 48 hours at
60°C and FGW = grain weight at field moisture content (kgha").
7.3.3.5 Water and agronomic nitrogen use efficiencies
Water use efficiency (WUE) was calculated based on grain yield (kgha"), changes in
profile soil water (mm) and rainfall data (Fatondji et al., 2006). Water use efficiency
was calculated as:
WUE= YI(P-(6.sW+D)) Equation 7.2
where WUE = water use efficiency (kghalmm')
Y = grain yield (kgha")
P = precipitation (mm)
tJ.SW = changes in soil water (mm)
D = drainage (mm) which was assumed to be negligible.
Agronomic nitrogen use efficiency (ANUE) was calculated based on grain yield and
quantity of nitrogen applied (Fatondji et al., 2006):
ANUE = IJ.YIN applied Equation 7.3
where tJ.Y = difference in grain yield between N treated plot and zero N plot.
144
7.3.4 Statistical analysis
Statistical analysis was conducted on baseline soil characteristics, maize yield, soil
water and runoff data. All analyses were performed with ANOVA procedure using
Genstat Discovery Edition 3 (www.vsni.co.uk). Baseline soil characteristics were
analyzed using General Analysis of Variance with 'farmer' as the only factor in the
model. Maize yield data were analyzed using a split plot design with tillage as main
factor and nitrogen as sub-plot factor. The yield data were analysed for each season
individually and later pooled and analysed across the three growing seasons. Soil
water and runoff data were analyzed using the Unbalanced Treatment structure as the
depths of access tubes for soil water measurements varied across the farms.
Unbalanced design was applied in the analysis of runoff data because the four farms
recorded different number of runoff events during 2006/07 and 2007/08 seasons.
Least significant difference (Lsd) values calculated at 5 % significance level were
used to compare the treatment means.
7.4 Results
7.4.1Soil properties of soils at the experimental sites
Soil properties measured at the beginning of the experiment are given in Table 7.1.
Significant (P < 0.001) differences in pH, organic carbon (O.C.) and total nitrogen
(Tot. N) were observed across the farms. Total phosphorus content (Tot. P) of the
soils used in the study also differed significantly (P = 0.001) across the farms. Siziba
farm in Gwanda district had the highest pH and organic carbon content. Soil texture
ranged from sand to loamy sand across the farms used during the three seasons of
experimentation.
145
Table 7.1. Selected initial soil chemical and physical properties (0 - 0.6 m) at farms
used from 2005106 to 2007/08, Insiza and Gwanda districts
District Farmer pH O.e. Tot. N Tot. P B.density Texture
(water) (%) (%) (%) (g!cm~
Insiza Mguni 6.0 0.56 0.05 0.02 1.50 Sand
Moyo 5.9 0.58 0.06 0.006 1.49 L.sand
Mpofu 5.2 0.41 0.03 0.007 1.46 Sand
Nkomo 5.6 0.36 0.02 0.006 1.50 Sand
Mlalazi 5.5 0.36 0.02 0.009 1.52 Sand
NcubeN 6.5. 0.34 0.02 0.002 1.52 L.sand
Nyathi 6.3 0.29 0.03 0.008 1.51 .L. sand
Gwanda Ncube J 5.3 0.33 0.03 0.009 1.49 Sand
Sibanda 5.6 0.46 0.03 0.021 1.51 Sand
Siziba 6.1 0.71 0.06 0.019 1.56 L. sand
Lsdo.o5 0.53 0.18 0.033 0.007 0.020
CV(%) 7.3 29 40 43 1.1
7.4.2 Seasonal rainfall and profile soil water regimes
7.4.2.1 2005/06 growing season
7.4.2.1.1. 1nsiza district
Insiza district which lies in NR IV received more rain than Gwanda district (NR V)
during the three seasons the study was conducted. Rainfall was evenly distributed
between November 2005 and February 2006 in ward 1 of Insiza district where
experimental sites were located. The total seasonal rainfall was 541 and 535 mm for
Mpofu and Mayo farms (Fig. 7.1). These seasonal rainfall totals were similar to the
74 year average for Filabusi which lies in Insiza district (Chapter 4). The growing
season started on day 61 (30 November 2005) and ended on day 168 (17 March 2006)
which is two weeks before the long term average at Filabusi. Thirty-one wet days
were recorded across the two farms during 2005106 growing season which is above
the long term mean of 22 days but is not extreme (Chapter 4). The heaviest daily
rainfall event of 53mm was recorded on 11 February 2006 (day 134).
146
600
,-.. 1-
ê 500
'-'
;:§ 400
.s
~
Q) 300
.;~. Insiza
:; 200
ê
;:l 100
U
0
0 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2005)
Figure 7.1. Cumulative rainfall distribution at Mpofu and Moyo farms of ward 1 in
Insiza district during 2005/06 growing season
In the 2005/06 growing season basin and ripper tillage systems had higher (P < 0.001)
profile water content than DP and CP systems on day 103 (11 January) when the first
measurement was taken (Fig. 7.2). Profile water content ranged from 50 mm in the
CP system to 86 mm recorded in the ripper and basin tillage systems. The DP system
had 70 mm in the profile on the same day. Profile soil water content increased by 4,
11,1 and 28 % in CP, DP, ripper and basin systems following 149 mm of rain that
was accumulated between soil water measurements on days 125 (2 February) and 145
(22 February). The basin system had consistently more soil water in the soil profile
from day 145 (22 February) until the last soil water measurement was taken on day
189 (7 April), suggesting better rainwater harvesting in the basin system.
147
160
140 1-0- CP 0 -eo 0 DP - .. 0 Ripper ---- Basins I
120
100
80
60
40
20
0+----,-----,----,----,----,----,-----,----,----,----,
o 25 50 75 100 125 150 175 200 225 250
Figure 7.2. Profile soil water changes (0 - 0.50 m) averaged across two farms (Moyo
and Mpofu) in ward 1 of Insiza district during 2005/06 growing season. Error bars
represent standard errors of means
7.4.2.1.2 Gwanda district
Rainfall was evenly distributed between 28 November 2005 and 28 March 2006
across the three farms used in the 2005/06 growing season. The total seasonal rainfall
was 286, 271 and 375 mm for Sibanda, Siziba and J. Ncube farms respectively (Fig.
7.3). The seasonal rainfall recorded at Sibanda and Siziba was 24 and 28 % less than
the 50 year average rainfall for Beitbridge which is usually drier than Gwanda district
(Chapter 4). The growing season started on day 70 (9 December) and ended on day
179 (28 March) with 22 wet days being recorded during the season. The length of the
2005/06 growing was 109 days and the highest daily rainfall event was 50 mm which
was recorded on day III (19 January 2006) at Siziba farm. Two 19-day dry spells
were recorded at Sibanda and Siziba farms from days 79 to 98 and 111 to 130.
148
600
j b Sibanda - Sizlba - - - J Ncube I500
'-'
~ 400c a'--CIII--
1~300
Q)
..>;:;
ro
"3 200
E Gwanda
::l
U 100
o -
0 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2005)
Figure 7.3. Cumulative rainfall distribution at Sibanda, Siziba and J. Ncube farms of
ward 17 in Gwanda district during 2005/06 growing season
In the 2005/06 growing season basin tillage system had higher (P < 0.001) profile
water content (78 mm) compared with 60 mm recorded in CP system, 63 mm in the
DP and 54 mm in the ripper system when the first measurement was taken (Fig. 7.4).
Profile soil water content increased by 18, 4 and 9 % in the DP, ripper and basin
systems following 104 mm of rain that was accumulated between days 99 (7 January)
and 136 (13 February). Profile soil water content decreased by 8 % in the
conventional system between days 99 (7 January) and 136 (13 February). The tillage
systems had similar (P > 0.05) profile water content from day 99 (7 January) until day
225 (13 May) when the last measurement was taken.
149
160 1-0- CP - -e- - DP -. - Ripper ----- Basins1
140
Ê 120
5.... 100
~ 80
~
Q)
ta 60
0
c..,.. 40
20
0
0 25 50 75 100 125 150 175 200 225
Time(days after 1October 2005)
Figure 7.4. Profile soil water changes (0 - 0.60 m) averaged across three farms in
ward 17 of Gwanda district during 2005/06 growing season. Error bars represent
standard errors of means
7.4.2.2 2006/07 growing season
7.4.2.2.1 Insiza district
Rainfall was poorly distributed at each of the farms used in the 2006/07 growing
season. The total seasonal rainfall was 456, 402, 297 and 343 mm for Nyathi, Mpofu,
Mguni and Mlalazi farms respectively (Fig. 7.5). The growing season started on day
50 (19 November) and ended on day 192 (10 April), and an average of 19 wet days
were recorded across the farms in ward 1 of Insiza district. Although the growing
season was longer than the 74 year record (Chapter 4), rainfall distribution during the
season was very erratic resulting in crop failure on most of the farms used in the
2006/07 season. The highest daily rainfall event was 75 mm recorded on day 92 (31
December 2006) at Nyathi fann. A 42-day dry spell was recorded between days 92
and 134 in ward 1 ofInsiza district during 2006/07 growing season (Fig. 7.5).
150
600 1- -Mpofu -- Nyathi D - _ Mguni - - Mlalazi I
~ 500
-.._,~ 400 I - - •
.c~ 300 _"ff'--
<l) ),----'
.~ ...
;0::;1 200
E
::l
U 100
0
0 25 50 75 100 125 150 175 200 225 250
Time (daysafter1 October2006)
Figure 7.5. Cumulative rainfall distribution at Mpofu, Nyathi, Mguni and Mlalazi
farms of ward 1 in Insiza district during 2006/07 growing season
Despite the below average rainfall observed across the study sites, soil water regimes
did respond to each rainfall event at Mpofu and Nyathi farms. The basin tillage
system consistently had the highest profile water content during 2066/07 season (Fig.
7.6). Profile water content decreased in all tillage systems from day 90 (29 December)
to day 145 (22 February) The highest profile water content was recorded on day 90
(29 December) with CP having 57 mm, DP recording 70 mm, ripper had 75 mm and
80 mm for the basin system. The high profile water content recorded on day 90 (29
December) is attributed to the 60 - 63 mm of rain that was accumulated across the
farms between the soil water measurements on day 75 (14 December) and 90 (29
December).
151
120 Insiza
1-0- CP - -e- - DP - .. - Ripper • Basin 1
100
~..._,
1-0 80
Q)
tij
~
Q)
tE 60
p8...
40
20
100 120 140 160 180 200
Time (days after 1 October 2006)
Figure 7.6. Profile soil water changes (0 - 0.60 m) averaged across two farms (Mpofu
and Nyathi) in ward 1 of Insiza district during 2006/07 growing season. Error bars
represent standard errors of means
7.4.2.2.2 Gwanda district
Rainfall was poorly distributed at each of the three farms used in the 2006/07 season.
Total seasonal rainfall was 238, 243 and 317 mm for Sibanda, Siziba and J. Ncube
farms respectively (Fig. 7.7). The growing season started on day 54 (23 November)
and ended on day 187 (5 April) with 14 wet days being recorded during the season.
As observed in Insiza district, rainfall distribution was erratic during the season
despite the length of the season being greater than the long term averages reported in
Chapter 4. A 55 day dry spell was recorded in ward 17 of Gwanda district between
days 94 and 149 during the 2006/07 growing season (Fig. 7.7). The highest daily
rainfall event was 50 mm recorded on 31 December 2006 and 28 March 2007 at John
Ncube farm while the other farmers recorded less.
There were little differences in the total profile water content in the four tillage
systems throughout the whole season, with values remaining around 50 mm. The
152
profile soil water responses to rainfall events were quite similar in the four tillage
systems in a season that had below average rainfall (Fig. 7.8).
600
1- = Sibanda ~ Siztba = = = J Ncube I
,-.. 500
1
:= 400
~c::
.§ 300
Cl)
.~
- 200
u~ 100
o -----r---
o 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2006)
Figure 7.7. Cumulative rainfall distribution at Sibanda, Siziba and J. Ncube farms of
ward 17 in Gwanda district during 2006/07 growing season
120 Gwanda
1- 0- CP - -€)- - DP - ~ - Ripper ==6 Basin 1
100
I.... 80
Q)
~
~
Q) 60
lea
~
40
20
100 120 140 160 180 200
Time (days after 1 October 2006)
Figure 7.8. Profile soil water changes (0 - 0.60 m) averaged across three farms
(Sibanda, Siziba and J. Ncube) in ward 17 of Gwanda district during 2006/07 growing
season. Error bars represent standard errors of means
153
7.4.2.32007/08 growing season
7.4.2.3.1 Insiza district
Total seasonal rainfall was 549, 541, 444, 319 and 288 mm for Mpofu, Nkomo, N.
Ncube, Mguni and Mlalazi farms respectively compared to long term mean at Filabusi
of 546 mm (Fig. 7.9). The growing season started on day 54 (23 November) and
ended very early on day 116 (24 January) with an average of 24 wet days being
recorded across the five farms. The first half of the growing season was wetter than
the second half with five farms receiving more than 60 % of their total seasonal
rainfall during the October to December period. The highest daily rainfall event was
50 mm recorded at Mpofu farm on day 100 (8 January 2008).
600
,.....,
ê 500..__,
~ 400
s::: _ _ _ a _ _ _ _ m _ •
1§ 300
Q)
.>~
"3 200
E ~Mpofu --NNcube
;:I
U 100 -===0- Nkomo ~ m ~ Mguni
=Mlalazi
0
0 25 50 75 100 125 ISO 175 200 225 250
Time(days after I October 2007)
Figure 7.9. Cumulative rainfall distribution at Mpofu, N. Ncube, Nkomo, Mguni and
Mlalazi farms of ward 1 in Insiza district during 2007/08 growing season
The rainfall events of November 2007 to January 2008 recharged the soil profiles
under all four tillage systems at the three farms during 2007/08 growing season (Fig.
7.10). The conventional tillage system had lowest profile water content from the time
the rains ended in January 2008 up to the end of the 2007/08 growing season. Profile
water content decreased in each tillage system between days 79 (18 December 2007)
154
and 99 (7 January 2008). The rainfall events that occurred between days 99 (7
January) and 113 (13 January) resulted in waterlogging in the basin system at Mpofu
and Nkomo farms. The soil profile had been recharged previously by 180 mm of rain
that was received between days 70 (9 December) and 79 (18 December).
180
160 1-c:- CP - -0- - DP -~ - Rjpper --l:r- Basin I
.,-., 140
i.... 120
~ 100
~
;B 80
J: 60
40
20 +--------,--------,--------,--------,--------,
o 50 100 150 200 250
Time(days after 1 October 2007)
Figure 7.10. Profile soil water changes (0 - 0.60 m) averaged across three farms
(Mpofu, N. Ncube and Nkomo) in ward 1 of Insiza district during 2007/08 growing
season. Error bars represent standard errors of means
7.4.2.3.2 Gwanda district
The growing season in Gwanda district was also characterised by an early cessation of
the rains in January 2008 as observed in Insiza district. The total seasonal rainfall was
265, 225, 243 and 303 mm for Sibanda, Siziba, John Ncube and Tlou farms
respectively (Fig. 7.11). The growing season started on day 78 (17 December) and
ended on day 118 (26 January) and only 15 wet days were recorded during the season.
The highest daily rainfall event of 75 mm was recorded at John Ncube farm on day
100 (8 January). As observed in Insiza district the first half of the season was wetter
than the January-March period. Each farm in Insiza district received more than 65 %
of its total seasonal rainfall during the first half of the growing season.
155
600
,.-.,
I 500 1- Sibanda - SizIba - - J Ncube - D TIou1
~ 400
$:=:
1~300
(1)
.~
"3 200
:E:s
U 100
o -
0 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2007)
Figure 7.11. Cumulative rainfall distribution at Sibanda, Siziba, J. Ncube and Tlou
farms of ward 17 in Gwanda district during 2007/08 growing season
Profile soil water responses to rainfall events were quite similar in the four tillage
systems across the three farms where soil water was monitored during the 2007/08
season (Fig. 7.12). The highest profile water content was recorded on day 101 (9
January) following 40-75 mm rainfall events recorded across the three farms (J.
Ncube, Sibanda and Siziba). The CP system had the lowest profile water content for
most periods of the 2007/08 season when soil water was measured. In contrast to
observations made in Insiza district, there was no waterlogging in the basin system at
any of the three farms in Gwanda district.
156
180 E-CP - .. - DP - ~ Ripper -e-- BasinI
160
A 140
1120
~
~ 100
~
;Ë 80
8
p.. 60
40
20 +--------,---------,--------,---------,--------,
o 50 100 150 200 250
Time (days after 1 October 2007)
Figure 7.12. Profile soil water changes (0 - 0.60 m) averaged across three farms (J.
Ncube, Sibanda and Siziba) in ward 17 of Gwanda district during 2007/08 growing
season. Error bars represent standard errors of means
7.4.3 Surface runoff
7.4.3.1 2006/07 growing season
The observed seasonal runoff water losses from four tillage treatments are
summarized in Fig. 7.13. Runoff water losses differed significantly (P < 0.001) across
the four farms. The highest runoff water losses were recorded at Mpofu and Nyathi
farms of Insiza district with sandy to loamy sand soil types. The least runoff water
losses were recorded at Sibanda farm which has sandy soil and lies in the southern
most part of ward 17 of Gwanda district. Tillage treatment had a significant (P =
0.003) effect on runoff water losses across the four farms. Planting basins had the
least water losses. Runoff water losses from single and double conventional
ploughing, and ripper tillage treatments were comparable at all farms. At Ncube farm
runoff water losses from all four tillage treatments were not significantly different.
157
25 I_ CP lIDDP ~ Ripper e:!l BasinsI
I20
~o 15c:
2
ëcii: 10o
r:/l
CI:S
~ 5
o
Mpofu Nyathi JNcube Sibanda
Farm
Figure 7.13. Seasonal runoff losses from plots under four tillage systems in 2006/07
growing season, Insiza (Mpofu and Nyathi) and Gwanda (J. Ncube and Sibanda)
districts. Error bars are standard error of differences between means
7.4.3.22007/08 growing season
Total seasonal runoff measured differed significantly (P < 0.001) across the four
farms. Mpofu and N. Ncube farms located in Insiza district had higher surface runoff
losses than J. Ncube and Sibanda of Gwanda district (Fig. 7.14). Planting basin tillage
system had the lowest measurable surface runoff in both districts. In the CP and DP
tillage systems Mpofu and N. Ncube farms experienced double.the amount of surface
runoff than that measured at J. Ncube and Sibanda in the same tillage treatments.
158
30 I. CP m DP IZI~per g!J Basins 1
,-.., 25
120
~o
2 15
o
Mpofu N Ncube JNcube Sibanda
Farm
Figure 7.14. Seasonal runoff losses from plots under four tillage systems in 2007/08
growing season, Insiza (Mpofu and N. Ncube) and Gwanda (J. Ncube and Sibanda)
districts. Error bars are standard error of differences between means
7.4.4 Maize performance
7.4.4.J Maize plant stand
In the 2005/06 growing season, planting basin system had better maize crop
establishment at all the five farms used in 2005/06 season (Table 7.2). In the 2005/06
season, the basin system had 11 % more plants per m2 than CP and DP systems. In the·
relatively dry 2006/07 season, the basin system had 75 % and 31 % more plants than
CP, and both DP and ripper systems. In 2007/08 season DP, ripper and basin systems
had 11, 22 and 39 % more plants per unit area than CP system.
Table 7.2. Average maize plant stands (plants m") under four tillage systems (CP,
DP, ripper and basin) measured two weeks after crop emergence during the three
seasons of experimentation in Insiza and Gwanda districts
Tinlage system 2005/06 2006/07 2007/08
CP 2.7 1.2 1.8
DP 2.7 1.6 2.0
Ripper 2.9 1.6 2.2
Basins 3.0 2.1 2.5
Lsdo.o5 0.20 0.37 0.27
CV (%) 5.1 27 15
159
7.4.4.2 Maize yields
7.4.4.2. J 2005/06 growing season
The main tillage and nitrogen effects on maize crop performance are given in Table
7.3. The tillage system and nitrogen fertilizer had no significant (P > 0.05) influence
on maize grain production and harvest index during the relatively wet 2005/06
growing season. Nitrogen increased maize stover (P < 0.05) production regardless of
tillage system used. The two-way tillage system x nitrogen interaction had no
significant (P > 0.05) influence on grain and stover production.
Table 7.3. Maize responses to four tillage systems (CP, DP, ripper and basin) and
nitrogen fertilizer (0 and 10 kgNha-') averaged across five farms in 2005/06 season,
Insiza and Gwanda districts
Treatment Grain yield Stover yield Harvest index
(kgha-1) (kgha-1)
Tillage CP 1 173 2236 0.34
DP 1403 2316 0.39
Ripper 1086 2093 0.36
Basins 1250 2398 0.36
Lsdo.05 442 328 0.063
Nitrogen 0 1 123 2015 0.36
10 1 315 2499 0.35
Lsdo.o5 313 232 0.045
CV(%) 32 13 18
7.4.4.2.2 2006/07 growing season
The tillage system had no significant (P = 0.25) influence on maize grain production
during the 2006/07 season. In contrast, nitrogen fertilizer increased grain (P = 0.023)
and stover (P < 0.001) production during the 2006/07 growing season (Table 7.4).
Maize harvest index was not influenced by the tillage system nor amount of nitrogen
applied. The two-way tillage system x nitrogen interaction had no significant (P >
0.05) effect on maize grain and stover production, nor harvest index.
160
Table 7.4. Maize responses to four tillage systems (CP, DP, ripper and basin) and
nitrogen fertilizer averaged across three farms in 2006/07 season, Insiza and Gwanda
districts
T'reatment Grain yield Stover yield Harvest index
(kgha-1) (kgha")
Tillage CP 304 704
DP 593 1 013
Ripper 349 1 130
Basins 418 937
213 343
Nitrogen o 282 710 0.28
10 413 943 0.29
20 554 1 184 0.30
Lsdo.o5 185 297 0.080
CV(%) 40 29 25
7.4.4.2.3 2007/08 growing season
In the 2007/08 season the tillage system had no significant (P > 0.05) effect on maize
grain and stover production at the seven farms in Gwanda and Insiza districts (Table
7.5). Nitrogen significantly (P < 0.001) increased maize grain and stover production
during the 2007/08 growing season (Table 7.5). The two-way tillage system x
nitrogen interaction had a significant (P = 0.004) influence on grain production across
seven farms scattered within Insiza and Gwanda districts (Table 7.5). For the 0
kgNha-1, conventional system had the lowest grain yield while DP system had the
highest yield across the seven farms. Under the 10 kglvha' treatment, the basin
system had the highest grain production. Maize grain production decreased with
further increase in nitrogen application rate from 10 to 20 kgNha-1 in the DP, ripper
and basin tillage systems.
161
Table 7.5. Maize responses to four tilla!fe systems (CP, DP, ripper and basin) and
nitrogen fertilizer (0, 10 and 20 kgblha ) averaged across seven farms in 2007/08
season, Insiza and Gwanda districts
Tillage N applied Grain yield Stover yield Harvest
s;ystem {kgha-t} {kgha-t} {kgha-t} index
CP 0 413 889 0.25
10 704 1 030 0.31
20 714 1027 0.33
DP 0 782 1209 0.31
10 700 1434 0.29
20 196 747 0.20
Ripper 0 518 1 211 0.28
10 652 1 417 0.28
20 447 867 0.28
Basins 0 702 1 164 0.30
10 875 1290 0.34
20 254 729 0.17
Lsdo.o5(tillage) 285 524 0.078
Lsdo.o5(N) 144 79 0.032
Lsdo.o5(interaction) 361 536 0.092
CV(%) 45 14 23
7.4.4.2.4 Across seasons comparison
The seasons had a significant (P < 0.001) influence on maize grain and stover
produced, and harvest index observed under each tillage system. Maize yield was
significantly lower in 2006/07 season at all N quantities applied compared with
2005/06 and 2007/08 seasons. The two-way interaction of season and tillage system
significantly influenced grain production (P = 0.001). In all three seasons the DP
system gave higher grain yields than the other tillage treatments (Table 7.6). The
two-way season x nitrogen fertilizer interaction had a significant (P = 0.006) influence
on grain yield (Table 7.7). Nitrogen increased maize grain production in all tillage
systems even in the 2006/07 and 2007/08 seasons with below average rainfall pattern.
162
Table 7.6. Effects of season and tillage system on maize crop performance during
three seasons of experimentation in Insiza and Gwanda districts
Season Tillage system Grain yield Stover yield Harvest index
(kgha-I) (kgha")
2005/06 CP 1 144 2 163 0.34
DP 1 510 2588 0.37
Ripper 1 057 2020 0.36
Basins ~.._- 1 222 2325 ~--0.3-6-....,
2006/07 CP 304 704 0.29
DP 593 1013 0.35
Ripper 349 1 130 0.23
Basins ---- -4-18--------- 937-------- 0.29
2007/08 CP 495 0.29
DP 732 0.36
Ripper 456 0.29
..---..._..,.~--~- Basins 675 -.,--_- --~- 0.34
Lsdo.o5(tillage x season 498 0.082
interaction)
CV(%) 43 35 19
Table 7.7. Effects of season and nitrogen fertilizer on maize crop performance during
three seasons of experimentation in Insiza and Gwanda districts
Season N applied Grain yield Stover yield Harvest index
(kgha-1) (kgha-1) (kgha-I)
2005/06 0 1 123 2015 0.36
10 1315 2499 0.35
2006/07 0 282 710 0.28
10 413 943 0.29
r--------- 20 554 1 184 0.30
2007/08 0 432 843 0.28
10 604 1 118 0.32
20 733 1293 0.35
Lsdo.o5(nitrogen) 223 273 0.063
Lsdo.05(nitrogen x season 326 354 0.054
interaction)
CV(%) 26 35 18
163
7.4.4.3 Water use efficiency (WUE)
There was no significant (P > 0.05) differences in water use efficiency across the four
tillage treatments in 2005/06 growing season (Figs. 7.15 and 7.16). Nitrogen fertilizer
had no significant effect on water use efficiency during the relatively wet 2005/06
season. In the 2006/07 growing season water use efficiency could not be calculated
because soil water was not monitored on farms where maize was harvested. In the
2007/08 season, N fertilizer significantly (P < 0.001) increased WUE in all tillage
systems (Fig. 7.16). At 20 kgNha-1 CP, DP, ripper and basin tillage systems recorded
WUE of 1.2,2.0,2.0 and 2.1 kghalmm".
4.5
'ê 4.0
--;"~3.5
Ol)
C 3.0
:g>. 2.5
Q)
ë!E3 20.
~ 1.5
en
;: 1.0
~~ 0.5
0.0
CP DP Ripper Basins
Tillagesystem
Figure 7.15. Water use efficiency as affected by four tillage systems (CP, DP, ripper
and basin) and nitrogen fertilizer (0 and 10 kgNha-l) across five farms (Moyo, Mpofu,
J. Ncube, Sibanda and Siziba) in 2005/06 season, Insiza and Gwanda districts. Error
bars stand for standard error of differences between means
164
,,-., 4.5
"7 4.0 1111 0 IC 10 III20 I
-~ê 3.5
COD 3.0
g;>- 2.5
Q)
'u 2.0
!:E
Q) l.5
Q)
;'.:".::.l l.0
Q)
:~s: 0.50.0
CP DP Ripper Basins
Tillage system
Figure 7.16. Water use efficiency as affected by four tillage systems (CP, DP, ripper
and basin) and nitrogen fertilizer (0, 10 and 20 kgNha-1) across six farms in 2007/08
season, Insiza and Gwanda districts. Error bars stand for standard error of differences
between means
7.4.4.4 Agronomic nitrogen use efficiency (ANUE)
Agronomic nitrogen use efficiency (ANUE) calculated for the different growing
seasons and tillage systems are summarized in Table 7.8. In 2005/06 growing season
with above average rainfall pattern single ploughing treatment had the highest ANUE
with double ploughing recording the lowest. In the 2006/07 season with below-
average rainfall pattern, double ploughing treatment gave the highest ANUE and
single ploughing recorded the lowest agronomic nitrogen use efficiency. In 2007/08
season, the ripper system had higher (P = 0.054) ANUE compared with the other
three tillage systems.
165
Table 7.8. Agronomic nitrogen use efficiency of maize in four tillage systems (CP,
DP, ripper and basin) averaged across farms for each season during 2005/06,2006/07
and 2007/08 growing seasons Insiza and Gwanda districts
THBagemethod 2005/06 2006/07 2007/08
CP 26.0 9.1 11.0
DP 6.7 24.0 3.1
Ripper 9.8 10.0 28.0
Basins 12.0 11.0 23.0
Lsdo.05 19
CV(%) 42 53 18
7.5 Discussion
7.5.1 Seasonal rainfall patterns
Rainfall distribution at each farm in Insiza and Gwanda districts differed during the
2005/06, 2006/07 and 2007/08 seasons. The 2005/06 season recorded above average
and well distributed rainfall between December 2005 and April 2006 in both Insiza
and Gwanda districts. The total rainfall recorded at each farm in Insiza district was
similar to the 74 year average for Insiza district (Chapter 4). However, for Gwanda
district total rainfall recorded at each farm was less than 376 mm which is the 50 year
average rainfall for Beitbridge which is drier than Gwanda district (Chapter 4).
Rainfall was below average in 2006/07 season when 47 to 150 mm was received
between 28 December 2006 and 1 January 2007. In this drought year, total rainfall at
each farm was less than 546 mm which is the 74 year average for the Filabusi
meteorological station located in Insiza district. Similar dry spells observed during the
2006/07 growing season in southern Zimbabwe have been reported in other years
(Oosterhout, 1996). The dry spells ranging from 42 days in Insiza district and 50 days
in Gwanda led to significant yield reductions and complete crop failure at 75 % of the
farms. The large variation in rainfall observed in the 2006/07 and 2007/08 growing
seasons is typical of the Mzingwane catchment and has been reported in other studies
(Chibulu, 2007; Masvaya et al. 2008; Love et al., 2008).
166
7.5.2 Soil water regimes and surface runoff
The four tillage systems harvested different quantities of rainwater with planting
basins having consistently higher soil water content at the start of season (Figs. 7.2,
7.4 and 7.10). The high initial soil water content in planting basin system probably
explain the better maize crop establishment observed at most of the farms (Table 7.2).
From the on-station experiments, relatively high soil water content in the basin system
was also observed on clay soil where mulch cover was superimposed on top of
conventional, ripper and basin tillage systems (Chapter 9a; Mupangwa et al., 2007).
The tillage treatments affected soil surface microtopography to varying degrees. The
grid of 0.9 m x 0.6 m spaced planting basins created a higher surface roughness
compared to the other tillage treatments resulting in superior rainwater harvesting and
better promotion of infiltration. As observed by Guzha (2004) the higher the surface
roughness the greater the potential for surface depression water storage. As the season
progressed the planting basin structure did not collapse completely as the basins were
partially filled with soil. This allowed continued harvesting of rainwater up to the last
rainfall event of the season in April. In the 2007/08 season, waterlogging was
experienced only at Mpofu and Nkomo farms of Insiza district. The soil profile under
the basin system maintained relatively higher soil water content until the last
measurement was taken on 12 April 2008 (Fig. 7.12). No waterlogging was
experienced in Gwanda district because of the lower rainfall received in the district
compared to Insiza during the 2007/08 season.
Although the double spring ploughed system started off with higher soil water content
at some farms (Figs. 7.8 and 7.12), its rainwater harvesting ability decreased
substantially as the growing season progressed. This initial high soil water content
167
under ploughing can be attributed to increased porosity of the soil following loosening
during tillage operations and the short lived surface depressions (Bruneau and
Twomlow, 1998). Initially ploughing reduces bulk density, increases porosity and
infiltration capacity of the surface soil layer (Unger and Stewart, 1983; Das and
Chopra, 1988; Sasal et al., 2006; Van der Meer, 2000). However, as the season
progresses compaction begins in the surface soil layer resulting in a reduction in soil
porosity. Consequently infiltration capacity of the surface soil layer decreases as a.
result of reduced pore space (Shinde et al., 1982). Furrows created by tine ripping
were easily filled with soil as the season progressed and this resulted in reduced
rainwater harvesting ability of that tillage system.
The loosening of soil during ploughing coupled with surface crusting could have
promoted high runoff losses measured under single and double conventional tillage
systems. The heavy rainfall events observed could have promoted soil surface sealing
and crusting, and reduced infiltration (AI-Qinna and Abu-Awwad, 2001; Bouman,
2007) thereby increasing surface runoff from ploughing and ripping tillage treatments.
As the season progressed soil surface depressions disappear in all four tillage
treatments but the decline was least in planting basin experimental plots. The better
surface storage in planting basin system explains reduced surface runoff measured in
2006/07 and 2007108 seasons in the basin system (Figs. 7.13 and 7.14).
7.5.3 Maize crop performance
Seasonal rainfall pattern had a strong impact on maize crop performance during the
three year study period. Soil water could have been the major growth limiting factor
in 2006/07 and 2007108 given the rainfall distribution patterns observed during the
168
two seasons. The 2006/07 season gave the lowest yields after experiencing dry spells
of up to 42 and 50 days in Insiza and Gwanda districts. The.dry spells occurred during
the flowering stage of the hybrid maize variety (SC403) grown in the trials. Maize is
more sensitive to water stress at flowering and grain filling stages (NeSmith and
Ritchie, 1992; Otegui et al., 1995; Otegui and Bonhomme, 1998; Pandey et al., 2000).
Cakir (2004) conducted a three year study on a silty loam soil to determine the most
water stress sensitive stage of maize. The results from the study showed that soil
water stress during flowering and ear formation stages significantly reduce grain
yield. Otegui et al. (1995) also report that soil water stress during maize flowering and
silking stages reduces yield because these are the stages when kernel number on each
maize cob is defined.
Maize crop responded differently to tillage systems during the 2005/06, 2006/07 and
2007/08 growing seasons. Incessant rains were observed between planting in
December 2005 and February 2006 during the 2005/06 growing season. Soil water
was not a limiting factor to maize crop growth and the effect of tillage systems on
rainwater harvesting was eliminated in that season with above average rainfall. In the
2006/07 and 2007/08 seasons with below average rainfall, planting basin, ripper and
double ploughing systems improved maize performance compared with single
conventional ploughing system (Tables 7.4 and 7.5). The improved performance of
maize crop can be attributed to more soil water availability in the double ploughing,
planting basin and ripper systems compared to single conventional ploughing. In the
ripper and basin tillage systems the improved yields could also be attributed to
improved soil fertility along rip furrows arid in basins. The same rip furrows arid
planting basins received manure before planting in 2006/07 and 2007/08 seasons.
169
Addition of nitrogen fertilizer increased maize yield even in the dry 2006/07 growing
season. Nitrogen is one of the major yield limiting nutrients in Zimbabwean soils
(Mapfumo and Giller, 2001). Maize yield responses to N fertilizer were small in the
wet 2005/06 season and this can attributed to the smaller quantity of 10 kgNha-1
applied in a growing season that received a lot of rainfall between January and
February 2006. Leaching could have been a pathway through which some of the
applied N was lost during a growing season with above average rainfall. Mapfumo
and Giller (2001) report that leaching is a common pathway of N loss especially on
granite derived sandy soils. Positive maize yield responses to N in our study are
consistent with observations made in previous studies in semi-arid southern
Zimbabwe (Ncube et al., 2007; Twomlowet al., 2008b). The wide scale promotion of
N fertilizer use in semi-arid southern Zimbabwe showed maize yield gains of 30-100
% following application of 10 kgNha-1 on smallholder farms (Twomlowet al.,
2008b). Studies by Ncube et al. (2007) in western Zimbabwe revealed a significant
maize yield gain through combining 3-6 tha" of manure with 10 kgNha-l. In the
2006/07 growing season 20 kgl-lha" gave an extra maize yield benefit over the 10
kgNha-1 treatment across the farms. The decrease in grain yield following the
application of 20 kgNha-1 in the double ploughing, ripper and basin tillage systems is
unusual given the low fertility status of soils used in this study (Table 7.1). The low
maize harvest index and suppressed grain yield could have been a result of soil water
stress during the grain filling stage of the crop. Effective grain filling in maize is
highly dependant on the availability of assimilates and these are significantly
influenced by soil water availability (Maddonni et al., 1998).
170
7.5.4 Water and Agronomic Nitrogen Use Efficiency
Nitrogen fertilizer improved water use efficiency in all tillage treatments (Fig. 7.15
and Fig. 7.16). Increase in water use efficiency following application of N fertilizer
could be attributed to faster root and shoot development which promotes more
effective water and nutrient uptake (Gaiser et al., 2004). The soil surface gets covered
quickly thereby reducing non-productive water losses such as soil evaporation. Our
study showed highest water use efficiency in double ploughed, ripper and basin
systems in response to 10 and 20 kgNha-'. This observation could be attributed to
maize roots being able to explore deeper for water and nutrients in the three systems.
Plant nutrients placed in the vicinity of the crop in the ripper and basin systems and
this, coupled with more favourable soil water conditions, could have led to better crop
performance and hence higher water use efficiency. Water use efficiency increased
with increase in applied nitrogen. Increase in water use efficiency following
application of N fertilizer under semi-arid conditions has also been observed by
Ncube et al., (2007) in Tsholotsho district of south western Zimbabwe.
Calculated agronomic nitrogen use efficiencies averaged 14, 14 and 16 kg grain per
kg of N applied for 2005/06, 2006/07 and 2007/08 seasons. The applied N was
probably used more efficiently in 2005/06 season when soil water was not limiting
compared to 2006/07 and 2007/08 seasons which had soil water deficits during
critical stages of maize growth. The observed ANUE in our study were lower than 31-
53 kg grain per kg N achieved in a study by Ncube (2007) in Tsholotsho district of
western Zimbabwe. Differences in seasonal rainfall amounts and their distribution
during the growing season could explain the differences in ANUE observed in
Matebeleland South (Insiza-Gwanda districts) and Matebeleland North (Tsholotsho
171
district). The efficient use of applied N could have been significantly influenced by
management especially effective weed control on the farms.
7.6 Condusion
This study has demonstrated that under semi-arid smallholder conditions, the
performance of single and double conventional ploughing, ripper and basin tillage
systems depends on the characteristics of the seasonal rainfall. The recommended
double spring ploughing and the widely promoted planting basins and ripper tillage
systems have similar soil water patterns during seasons with above and below average
rainfall pattern. This implies there are equal chances of experiencing reduced crop
yields or total crop failure in the event of a drought or prolonged dry spell if a farmer
is anyone of the tested four tillage systems. Planting basin tillage system harvests
more rainwater at the beginning of the season and this promoted better maize crop
establishment compared with the ripper and conventionally ploughed systems. The
basin tillage system also reduced surface runoff significantly across all farms and soil
types. The surface depressions created by planting basins reduced surface runoff and
promoted infiltration resulting in better soil profile recharge compared with ripper,
single and double ploughing. However, the higher soil water content early in the
season and reduced surface runoff in the basin system was not translated into
substantial differences in water use efficiency, agronomic nitrogen use efficiency and
grain yield gains compared with the other tillage systems.
The rainwater harvesting ability of ripper and basin systems is reduced during the
season as rip furrows and basins fill up with soil. Double ploughing does have the
potential of harvesting and storing early rains as effective as planting basin and ripper
172
tillage systems but its soil water collection ability is lost as the growing season
progresses. Timing of the first ploughing in the double ploughed system could be
critical so that the early spring rains are captured and stored in the soil profile. Single
and double ploughing systems give similar crop yields to ripper and planting basin
systems in a growing season with above-average rainfall. However, in a season with
below-average rainfall the double ploughing tillage system gave similar or slightly
more maize yield benefits compared with ripper and planting basin systems.
Nitrogen fertilization with 10 kglvha' on granitic sandy soils in a season with above-
average rainfall in semi-arid areas gives insignificant yield gains. In a drought year
nitrogen fertilizer improves maize yields especially in double ploughing, ripper and
planting basin systems. Application of nitrogen fertilizer increases water and
agronomic nitrogen use efficiency, and yields regardless of the tillage system used.
Results from our study further suggest that soil water management techniques and
nitrogen fertilizer are required concurrently in semi-arid smallholder cropping
systems.
173
CHAPTERS
Effect of Dead lLevelContours and Infiltration Pits on Soil Water Content and
Crop Yields on Smallholder Farms in Gwanda District, Southern Zimbabwe
8.1 Introduction
Water harvesting is a method of collecting surface runoff from a catchment area and
storing the water in the root zone of a cultivated area for crop growth (Li et al., 2004). In
rainfed agriculture of the semi-arid areas rainwater harvesting could be a source of water
for smallholder cropping systems. Traditionally smallholder farmers in Zimbabwe have
been using the graded contour ridge for safely diverting excess water from the field
(AGRITEX, undated). A standard graded contour ridge is pegged at a gradient of 1:250
and contour ridges are usually spaced at 20 to 30 m apart on gentle slopes (AGRITEX,
undated; Hughes and Venema, 2005).
In semi-arid areas dead level contours, pegged at zero gradient, are an appropriate field
technique for harvesting runoff water for crop production. The purpose of a dead level
contour is to retain runoff water in the contour and the water will move into cropped field
by lateral flow (Hughes and Venerna, 2005). To improve the ability of the dead level
contour in retaining runoff, water infiltration pits are dug along the contour often at lOm
interval (AGRITEX, undated). A standard infiltration pit measures 2 m long, 1 m wide
and 0.5 - 1 m deep (AGRITEX undated; Hughes and Venema, 2005).
174
Efforts by international research organisations, NGOs and Government Agricultural
Departments continue in fmding a combination of appropriate technologies that try to
reduce the impact of harsh climatic factors on people's livelihoods. In Gwanda district of
southern Zimbabwe dead level contours and infiltration pits are being promoted by
Practical Action Southern Africa. The design of the contours includes having an
infiltration pit or storage tank at certain intervals along the contour (Plate 8.1). Dead level
contours and infiltration pits have been explored in other semi-arid districts of Zimbabwe
(Mot si et al., 2004; Mugabe, 2004). Results from these earlier studies have shown that
dead level contours and infiltration pits can contribute towards soil water status in the
cropped field. This chapter reports on a study to investigate the in-field soil water content
supplied by dead level contours with and without infiltration pits in semi-arid Gwanda
district of southern Zimbabwe.
Plate 8.1. Dead level contours with (a) storage tank at Magaya farm and (b) with
infiltration pit at Ncube farm in Humbane village of ward 17, Gwanda district
175
8.2 Objectives
The current study in Gwanda district was designed to quantify soil water content supplied
by dead level contours with and without infiltration pits at four farms in ward 17. At the
end of the 2007/08 summer growing season pearl millet and maize yields were measured
along the transects between specified distances from the contours. The specific objectives
of study were:
(1) to measure the profile soil water content at 7 and 2 m upslope, and at 3, 8, 13 and 18
m downslope of the dead level contours with and without infiltration pits through the
growmg season;
(2) to determine the extent to which soil water moves laterally from contours into the
cropped field;
(3) to determine the soil water balance at different distances from the dead level contour
with and without infiltration pit; and
(4) to measure crop yields and water use efficiency at different distances from the dead
level contour with and without infiltration pit.
8.3 Materials and Methods
8.3.1 Experimental set up
The average depth of the dead level contour at Ncube farm was 0.30 m with a width of
1.0 m. At Moyo farm, the dead level contour had an average depth of 0.3 m and width of
0.9 m. The infiltration pit at Dube farm measured 1.5 m long, 1.0 m wide and 0.4 m deep
compared with the pit at Siziba which measured 1 m long, 0.5 m wide and 0.3 m deep.
176
Moyo and Ncube farms are located 4 - 5 km apart while Dube and Siziba farms are about
3 km apart.
8.3.2 Soil water and crop yield data collection
During both the 2006/07 and 2007/08 seasons volumetric soil water content was
measured using a capacitance probe imicrogopher sensor type). Depth of access tubes
varied between 0.6 and 0.8 m across the four farms. Depth of access tubes was restricted
by the presence of a lateritic layer in the profile at Ncube and Moyo farms in Gwanda
district. Soil water was measured fortnightly at 0.1 m depth increments during both
2006/07 and 2007/08 seasons. Soil water content in millimetres was determined by
multiplying volumetric water content by thickness of each layer where the soil water was
measured. The soil water content in the 0 - 0.6 m profile at each farm was calculated
during the 2006/07 and 2007/08 growing seasons.
Soil texture was determined by the hydrometer method as outlined by Anderson and
Ingram (1993). The lower limit of the soil at each farm was derived from the soil water
content on a day when the lowest profile soil water content was measured through the
season although this date varied from farm to farm. The drained upper limit was derived
from the day that had the highest profile soil water content at least 18 hours after a
rainfall event of 30 mm or more. The plant available water capacity (PAWC) at each
farm was calculated as the difference between the drained upper limit (DUL) and lower
limit (LL). The DUL was derived from soil water content measured on 19 December
2007 (day 70) at Dube, Siziba and Ncube farms. The DUL for Moyo farm was derived
177
from soil water content measured on 9 January 2008 (day 101). The LL for Dube and
Siziba farms was derived from soil water content measured on 14 November 2007 (day
45). The LL for Ncube and Moyo farms was derived soil water content measured on 5
December 2007 (day 66). The components of the water balance measured were rainfall
(P) and change in profile soil water content (LlS) from the beginning to the end of season
and these were used to derive evapotranspiration (ET) at each distance from the contour
(Equation 8.1).
ET = P +/- LlSW Equation 8.1
where calculated ET (mm) is seasonal evapotranspiration, P (mm) is total seasonal
rainfall and LlSW (mm) is the sum of differences in soil water content measured on
different days of the growing season when soil water measurements were taken.
Crop yields were measured from a net plot area of 10m2 between adjacent access tubes.
Water use efficiency was calculated as follows:
Equation 8.2
where WUE is water use efficiency (kgha'lmm"), Y is either above ground biomass or
grain yield (kgha') and ET is the evapotranspiration (mm). The evapotranspiration value
used to calculate WUE between two adjacent access tubes was an average of the two ET
values for the adjacent access tubes along the ploughed transect.
178
8.3.3 Soil water and crop yield data analysis
Soil water data were analyzed by analysis of variance (ANOV A) using unbalanced
design because depth to which access tubes were sunk differed across the four farms
(Dube, Siziba, Moyo and Ncube). Crop data was analyzed by ANOVA using the general
analysis of variance design (Genstat Discovery Edition 3). All statistical analyses were
conducted using Genstat Discovery Edition 3 (www.vsni.co.uk). The least significant
difference (Lsd) at 5 % significance level was used to compare means.
8.4 Results and Discussion
8.4.1 Characteristics of the soils for the four farms
Soil water characteristics derived from the soil water data collected during 2007/08
season from Moyo, Ncube, Dube and Siziba farms are given in Figure 8.1. Dube and
Siziba farms had dead level contours with infiltration pits while Ncube and Moyo farms
had only dead level contours. Plant available water capacity (PAWC) for Moyo, Ncube,
Dube and Siziba farms were 36, 40, 48 and 40 mm based on a soil profile of 0.65 m
depth. The higher plant available water recorded at Dube farm could be attributed to the
high clay content in the soil (Table 8.l) and the bigger infiltration pit on the farm in
2007/08 season. The PAWC for Ncube farm was similar to that measured at Siziba farm
although the former had much lower clay content in the profile. The soil profile was
recharged by 54 mm of rainfall that was accumulated in two days before soil water
measurements were taken at Ncube farm for DUL. At Siziba farm the soil profile was
recharged by 40 mm of rain that was accumulated over two days before soil water
measurements were taken for DUL.
179
VoJwnetric water content (%)
o 5 10 15 20o +- ~ _L L_ ~
10
E20
~30
%40
Q)
Cl 50 Moyo
60
70 I -x-LL -o-DUL I
Volumetric water content (%)
o o 5 10 15 20+- ~L_ _L L_ ~
10
- x~E-- 20 x
u
'-' 30 /.s
fr 40 /x Ncube
Cl 50 xJ
60 \x
70 -x-LL -DUL
Volumetricwater content (%)
o 5 10 15 20
o +- ~L_ _L ~ __ ~------~
10
E20
~30
%40
Q)
Cl 50
60 Dube
70 -x- LL - DUL
Figure 8.1. Drained upper limit (DUL) and lower limit (LL) derived from 2007/08 soil
water data for Moyo, Ncube and Dube farms in Gwanda district
180
Volwnetric water content (%)
0 5 10 15 20
0
,..).0..
..§._, 20
,%30
Cl)
°40
50
60 Sizba
70 -x- LL -0- DUL I
Figure 8.2. Drained upper limit (DUL) and lower limit (LL) derived from 2007/08 soil
water data for Siziba farm in Gwanda district
Table 8.1. Soil textural variation with soil depth at the four farms used for quantifying
soil water supply from dead level contours and infiltration pits in Gwanda district
Farm Structure De~th {cm} % c1a~ % silt % sand
Moyo Only dead 0-15 5 3 92
level 15-30 6 4 90
contour 30-45 8 4 88
45-60 7 8 85
Ncube Only dead 0-15 4 6 90
level 15-30 3 5 92
contour 30-45 2 8 90
45-60 4 7 89
Dube Dead level 0-15 10 4 86
contour and 15-30 11 3 86
infiltration 30-45 16 6 78
pit 45-60 18 6 76
Siziba Dead level 0-15 6 13 81
contour and 15-30 12 7 81
infiltration 30-45 18 6 76
Eit 45-60 20 7 73
8.4.2 Seasonal rainfall at selected farms
8.4.2.1 2006/07 growing season
The four farms received different amounts of rainfall during the 2006/07 growing season
(Fig. 8.3). The number of rainfall events recorded at the four farms ranged from eight
recorded at Ncube farm to 17 received at Siziba farm. The total season rainfall recorded
181
during the 2006/07 growing season ranged from 145 mm recorded at Ncube farm to 242
mm received at Siziba farm. Moyo and Dube farms received 208 and 203 mm of rain
during the 2006/07 growing season. The soil water data obtained from Ncube farm
suggests that the farmer did not record some rainfall event(s) that occurred between day
143 (20 February 2007) and day 160 (9 March 2007) as there are changes in soil water
content. The rainfall records from Sibanda farm that is located about 3 km north-east of
Ncube farm, indicate that there was a rainfall event on 26 February (day 149) and
Sibanda farm recorded 20 mm on that day. The other farms (Tlou, Nare and Magaya)
which are located within a 5 km radius from Ncube farm also received rainfall on 25
February 2007 (day 143) during the 2006/07 season. So it is reasonable to assume that
Ncube farm received some rain that day. The highest daily rainfall event during the
2006/07 season was 38 mm received at Moyo farm on 26 February 2007 (day 149).
A 54-day dry spell was experienced from day 94 (2 January 2007) to day 149 (26
February 2007) when 22 mm of rain was received. Dry spells occurring during the
cropping period are common in semi-arid areas (Oosterhout, 1996; Rockstrom et al.,
2003). The total seasonal rainfall recorded at each farm during the 2006/07 season was
below the 50 year average of 376 mm for Beitbridge district (NR V) (Chapter 4). The
2006/07 season was classified as a drought in semi-arid southern Zimbabwe.
182
400 I D D D Dube - Moyo - - Ncube - SizlbaI
300
,-..
~ê
:::l 200
.~ro5
~
100
o -
0 25 50 75 100 125 150 175 200 225 250
Time(days after 1 October 2006)
Figure 8.3. Cumulative rainfall distribution at Mayo, Ncube, Dube and Siziba farms of
ward 17 in Gwanda district during 2006/07 growing season
8.4.2.2 2007/08 growing season
The number of rainfall events recorded during the 2007/08 season ranged from nine
recorded at Mayo farm to 20 recorded at Siziba farm. Dube and Ncube farms received 12
and Il rainfall events each during the 2007/08 growing season. The total seasonal rainfall
ranged from 204 mm recorded at Mayo farm to 298 mm received at Dube farm during
the 2007/08 season (Fig. 8.4). The total seasonal rainfall recorded at Ncube and Siziba
farms was 289 and 225 mm during the 2007/08 season. Most of the rainfall events were
concentrated in the October to December period with just five events occurring during
January to April period. The highest daily rainfall event was 70 mm recorded on day 100
(8 January) at Dube, Moyo and Ncube farms. The last rainfall event was recorded on 24
January 2008 (day 116) with 18 to 20 mm being received across Dube, Moyo and Ncube
farms. Seasonal rainfall at both farms was below the 50 year 376 mm calculated for
Beitbridge district (NR V) (Chapter 4).
183
400
I ~ a ~ Dube -===0 Moyo - - Ncube ---SiZIba I
300
,.-..
-..ê_, 200 s:~~
.5
<Il
~
100
o 25 50 75 100 125 150 175 200 225 250
Time(daysafter1 October2007)
Figure 8.4. Cumulative rainfall distribution at Moyo, Ncube, Dube and Siziba farms of
ward 17 in Gwanda district during 2007/08 growing season
8.4.3 Soil water content in fields with dead level contours only
8.4.3.1 2006/07 season
There was no significant (P > 0.05) difference between treatments in profile soil water
content measured at the six distances from the dead level contour on each day that soil
water measurements were taken along the ploughed transect. Profile water content
decreased significantly (P < 0.001) from day 117 (25 January) to day 142 (19 February)
at all distances from the dead level contour (Fig. 8.5). The decrease in profile water
content between these two dates is attributed to soil water uptake by crops growing
around the access tubes as well soil evaporation. Profile water content decreased by 41 -
48 mm upslope of the dead level contour and 42 - 48 mm downslope of the contour. A 7
mm rainfall event was recorded at Moyo farm between days 117 (25 January) and 142
(19 February) and this amount was negligible to cause any notable changes in profile soil
184
water content. The soil profiles at Moyo and Ncube farms were recharged by 49 - 57 mm
of rain that was received from day 179 (28 March) to day 194 (13 April).
120 --0--7 ~-2 - -6- -3 -8 -::.:-13 -0- 18
100
A
..ê......., 80
Q)
1a
~ 60
Q)
ta 40
~e
20
0
0 20 40 60 80 100 120 140 160 180 200
Time(daysafter1October2006)
Figure 8.5. Profile soil water changes along ploughed transects averaged across two
farms with dead level contours only (Moyo and Ncube) during the 2006/07 growing
season
The soil water content measured at each depth was the same (P = 0.116) at the six
distances from the dead level contour. On 19 February (day 142), the 0 - 0.15 m soil
layer was wetter at 3 m from the dead level contour than at the other distances from the
contour (Fig. 8.6). The 0.15 - 0.25, 0.35 - 0.45 and 0.55 - 0.65 m soil layers were wettest
at 18 m from the dead level contour. On 12 January 2007 (day 104), soil water content at
each soil depth was the similar (P = 0.821) at five of the six distances from the dead level
contour along the ploughed transect. The 0 - 0.15 m layer had more (P < 0.001) soil
water than the deeper layers at each distance from the dead level contour on 12 January
2007 (day 104).
185
Soilwater content(nun)
0 5 10 15 20 25 30
0
10
Ê 20
(.)
.'-s" 30
p,
Q)
Cl 40
50
60
19/2/2007 770 12/1/2007 -0-- ~-2 :":o-A- - 3 11 -0- 8 -)(- 13 18.
Figure 8.6. Soil water distribution with respect to depth at different distances from the
.dead level contour only averaged across two farms (Moyo and Ncube) on the driest day
along unploughed (a) and ploughed (b) transects during 2006/07 season
8.4.3.2 2007/08 growing season
Along the unploughed transect, the highest profile soil water content was recorded at 3 m
between 14 November 2007 (day 45) and 9 January 2008 (day 101) (Fig. 8.7a). The
ploughed transect had more soil water in the profile between 14November 2007 (day 45)
and 24 January 2008 (day 116) (Fig. 8.7b). After receiving 70 mm of rain the night of 8
to 9 January 2008, the measured profile water content ranged from 80 to 83 mm upslope
of the dead level contour and 81.mm to 97 mm on the downslope of the dead level
contour along the unploughed transect. The highest profile water content of 97 mm was
recorded at 3 m downslope of the dead level contour along the unploughed transect.
Along the ploughed transect, profile water content ranged from 91 mm to 95 mm on the
upslope of the dead level contour and 81 mm to 98 mm on the downslope of the contour.
The highest profile water content was recorded at 3 m downslope, suggesting that water
186
had moved laterally from the dead level contour. There was notable lateral soil water
movement following the 70 mm rainfall event recorded at both Moyo and Ncube farms.
Conventional ploughing further increased profile water content along the ploughed
transect. Ploughing loosens the soil and creates surface depressions that collect rainwater
and give it more time to infiltrate (Van der Meer, 2000; Sasal et al., 2006)
160
140 1-0--7 --<>--2 - -I:!.- - 3 --e--8 -)(-13 -0- 181 G
A
..ê 120.,__, 100 •• I:!.
~~
~ 80
.~·~a' ,
Q)
PB 60 ,.
c2, 40 Unploughed
20
0
0 20 40 60 80 100 120 140 160 180 200
Time(days after 1 October 2007)
160
140 1-0--7 ~-2 - -I:!.- -3 ----8 -)(-13 -~ G
A
ê 120..,__, 100
.Q..)~.
:::: 80
Q)
rE 60
c2, 40 Ploughed :::3
20
0
0 20 40 60 80 100 120 140 160 180 200
Time(days after 1October 2007)
Figure 8.7. Profile soil water changes averaged across two farms (Moyo and Ncube)
along unploughed (a) and ploughed (b) transects across farms with only dead level
contours during 2007/08 growing season
187
On one of the driest days during 2007/08 growing season (5 December 2007), the soil
water content measured at each soil depth and distance from the dead level contour was
the same (P = 0.205) along the unploughed and ploughed transects (Fig. 8.8 and 8.9). In
the 0.25-0.5 m soil layer, soil water content at 3 m was significantly (P < 0.001) higher
than at the other distances from the contour along the unploughed transect showing that
the difference could be due to lateral flow from the dead level contour. In the 0.25-0.45 m
soil layer, soil water content was also higher at 3 m from the contour compared with the
other distances along the ploughed transect (Fig. 8.9). The dead level contour supplied
soil water to soil layers below 0.25 m.
Soilwater content (mm)
0 5 10 15 20 25 30 35
0
10
-. 20·
S
..o_,
-s 30
0.. 5/12/2007 Unploughed
Q)
Cl
40
50
60 1~-7 ~ -2 - -t::.- - 3 ~8 )I( 13 -0- 181
Figure 8.8. Soil water distribution with respect to depth at different distances from the
dead level contour only averaged across two farms (Moyo and Ncube) on the driest day
along unploughed transect during 2007/08 season
188
Soilwater content (rrnn)
0 5 10 15 20 25 30 35
0
10
8'20
~ 5/12/2007 Ploughed
t30
d)
Cl
40
50
60 -0--7 --<>---2 - -6,- - 3 --e-8 ~ 13 -0- 18
Figure 8.9. Soil water distribution with respect to depth at different distances from the
dead level contour only averaged across two farms (Moyo and Ncube) on the driest day
along ploughed transect during 2007/08 season
The 0 - 0.15 m soil layer had significantly (P < 0.001) higher soil water content than the
deeper layers at each distance from the contour following the 70 mm rainfall event
recorded at both farms. The distance from the dead level contour had a significant (P <
0.001) influence on soil water content measured at each soil depth along the unploughed
and ploughed transects. The highest soil water content at each soil depth was recorded at
3 m along the unploughed transect (Fig. 8.10a). Along the ploughed transect, soil layers
starting from 0.2 m depth were wettest at 3 m (Fig. 8.10b). The top soil layer had the
highest water content at 2 m upslope. This further confirms that there was lateral soil
water movement following the 70 mm rainfall event along both the unploughed and
ploughed transects.
189
Soil water content (rrnn)
0 5 10 15 20 25 30 35
0
10
,.-2.0.
E
.~s 30
0..
Q)
Cl40 \6, 9/1/2008 Unploughed
50
60 1-0--7 ~-2 - -6,- - 3 ~8 -)lEo -13 -0- 181
Soil water content (mm)
o 5 10 15 20 25 30 35
o +- ~ ~ L_ ~ ~ L_ ~
10 G
,.-2.0.
E
~
.s 30 ~
0.. •
Q) 9/1/2008 Ploughed
Cl 40 ,A
50 j
60 -0- -7 --<>- -2 - -6,- - 3 --G- 8 )I( 13 -0- 18
Figure 8.1O. Soil water distribution with respect to depth at different distances from the
dead level contour only averaged across two farms (Moyo and Ncube) on wettest day
along unploughed (a) and ploughed (b) transects during 2007/08season
190
8.4.4 Farms with dead level contours and infiltration pits
8.4.4.1 2006/07 growing season
As observed at farms with dead level contours only, there was no significant (P > 0.05)
difference in profile soil water content measured at the six distances from the dead level
contour and infiltration pit on each day soil water measurements were taken during the
2006/07 season (Fig. 8.11). Profile water content decreased significantly (P < 0.001) from
day 117 (25 January) to day 142 (19 February) at all the six distances from the dead level
contour and infiltration pit. Profile water content decreased by 47 mm upslope of the
contour and 36 - 43 mm downslope of the dead level and infiltration pit. The soil profiles
were recharged following 22 mm of rainfall recorded at Dube farm and 26 mm at Siziba
farm on day 149 (24 February). The rainfall events that occurred between days 179 (28
March) and 194 (13 April) also recharged the soil profiles at both farms with dead level
contours and infiltration pits. Dube farm received 60 mm of rain over a nine day period
while Siziba farm accumulated 69 mm during the same period. The farms with dead level
contours only also received rainfall during the same period. More profile water was
gained downslope at farms with dead level contours and infiltration pits than at those
farms with dead level contours only. This indicates that the dead level contours and
infiltration pits at Dube and Siziba farms captured more rainwater that moved laterally
into the field than the dead level contours at Moyo and Ncube farms. Our observation at
Dube and Siziba farms are consistently with the findings of Mugabe (2004) that also
indicated lateral soil water upslope and downslope of a 7 m long and 1 m deep infiltration
pit.
191
120 1~-7 ~-2 - -6- - 3 --8 ~13 -0- 181
.-... 100
ê
'-...'. 80
$~I
~ 60
.£
t.e::: 40
A-
20
0
0 20 40 60 80 100 120 140 160 180 200
Time(days from I October 2006)
Figure 8.11. Profile soil water changes at different distances from the dead level contour
with infiltration pit averaged across two farms (Dube and Siziba) during 2006/07 growing
season
Soil water distributions with respect to depth measured on the driest (19 February) and
wettest (12 January) days during 2006/07 season at farms with dead level contours and
infiltration pits are given in Figure 8.12. The 0 - 0.25 m soil layer had more (P = 0.042)
soil water at 18 m from the contour than at the other five distances from the contour on
one of the driest days (19 February). Soil water content in the 0.25 - 0.6 m layer was
similar at the six distances from the dead level contour. The distance from the dead level
contour had no significant (P = 0.985) influence on soil water content measured at each
soil depth along the ploughed transect on 12 January 2007 (day 104). The top soil layer
had more (P < 0.001) soil water than the deeper layers on 12 January, nine days after a 35
mm rainfall event. The 0 - 0.15 m soil layer had more soil water at 2 m upslope on 12
January (day 104) than at the other distances nine days after the last rainfall event. This
suggests that soil water could have moved upslope from the dead level contour through
capillary flow.
192
Soilwater content (mm)
0 5 10 15 20 25 30
0
10
,..-.2, 0
E
~ 30
.f£r 40 1-::--87 ~-2 ~310 . -):-13 -0- 1~
50
60
70 19/2/2007 12/1/2007
Figure 8.l2. Soil water distribution with respect to depth at different distances from the
dead level contour and infiltration pits averaged across two farms (Dube and Siziba) on
driest and wettest days during 2006/07 season
8.4.4.22007/08 growing season
As observed at Moyo and Ncube farms, the soil profile along the ploughed transect had
more soil water than the unploughed transect. The unploughed transect had the highest
profile water content at 18 m from the contour after receiving 60 mm of rainfall during
the night of 8 to 9 January 2008 (Fig. 8.13a). There could have been some surface
depressions near the access tube at 18 m that were created by ploughing in previous
seasons. These depressions collected rainwater during the rainfall event of 8 January (day
100) resulting in the higher soil water measured than at the other five distances from the
contour. Soil water content in the profile was the same at 2 and 7 m upslope of the
contour but ranged from 111 mm to 118 mm on the downslope of the dead level contour.
The lowest profile water content was recorded at 2 m upslope between 14 November
2007 (day 45) and 12April 2008 (day 194). Along the ploughed transect the highest soil
water content in the profile was recorded at 18 m between 9 January 2008 (day 101) and
12 April2008 (day 194) (Fig. 8.13b). Profile water content ranged from 124 mm to 128
193
mm on the upslope of the dead level contour while the downslope side recorded 126-138
mm. The highest profile water content was recorded at 18 m from the dead level contour.
160 [-0--7 ~-2 - -/).-- 3 ~8 -)1(-13 -0- 18[
140
A
ê 120
'-..'. 100
*~ 80Q)
~ 60.0..
t:l-. 40 Unploughed
20
0
0
~~~~~2_0~_4_0~~60~~8 Time (da_y0_s~a_f1t_er0_10O~c_t1o2b_e0r~2_010_74)0~~16_0~_1 _8_0~01 j
L..
160 ~-7 ---<>--2 - -/).-- 3 ----B-8 -)1(-13 -0- 1~
140
A 120
~... 100
~
<:':l
~ 80
~
tee 60c, 40 Ploughed
20
0
0 20 40 60 80 100 120 140 160 180 200
Time (days after 1 October 2007)
Figure 8.13. Profile soil water changes along unploughed (a) and ploughed (b) transects
at farms with dead level contours and infiltration pits averaged across two farms (Dube
and Siziba) during 2007/08 growing season
194
On 5 December (day 66), the ploughed transect had more (P < 0.001) soil water in the
profile than the unploughed transect. The distance from the dead level contour and
infiltration pit had no significant (P = 0.084) influence on soil water measured at each
depth on 5 December (day 66) along either the unploughed or ploughed transects (Figs.
8.14 and 8.15). The wettest soil layers between 0.2 and 0.6 m depths were observed at l3
and 18 m from the contour along the unploughed transect on 5 December 2007. The
driest soil layers were recorded at 7 m upslope. Lateral movement upslope was probably
negligible. Along the ploughed transect the 0 - 0.15 m soil layer was the wettest at 3 m
and driest at 8 m from the contour (Fig. 8.15).
Soil water content (mm)
o 5 10 15 20 25 30 35
o +- ~ J_ ~ _L L_ _L ~
10
20
j
t- 30 5112/2007 Unploughedo 40
50
60
70 1-0--7 ~ -2--6.--3~8 )1( 13-0- 181
Figure 8.14. Soil water distribution with respect to depth at different distances from the
dead level contour and infiltration pits averaged across two farms (Dube and Siziba) on
driest day along unploughed transect during 2007/08 season
195
Soil water content (mm)
0 5 10 15 20 25 30 35
0
10
20
'?
3 30 5/12/2007 Ploughedt
Q)
Cl 40
50
60
70 1-0-- -7 --<>--- -2 - -6.- - 3 ~ 8 )1( 13-0- 181
Figure 8.15. Soil water distribution with respect to depth at different distances from the
dead level contour and infiltration pits averaged across two farms (Dube and Siziba) on
driest day along ploughed transect during 2007/08 season
On 9 January (day 101) the soil profile along the ploughed transect had more (P < 0.001)
soil water than the unploughed transect. This indicates that ploughing increased soil
water content through collecting rainwater on in the soil surface depressions created
during the tillage operation. The distance from the dead level contour and infiltration had
no significant (P = 0.485) influence on soil water content measured at each soil depth
along the unploughed transect. The 0 - 0.15 m soil layer had higher soil water content
than the deeper layers along the unploughed and ploughed transects (Figs. 8.16a and
8.13b). Along the ploughed transect the wettest 0 - 0.15 m layer was at 18 m while the
same layer was driest at 2 m upslope (Fig. 8.16b). Soil layers below 0.35 - 0.45 m were
driest at 13 m from the dead level contour and infiltration pit.
196
Soil water content (mm)
0 5 10 15 20 25 30 35
0
10
20
Ê
'-o' 30 9/1/2008 Unploughed
~
Q 40
50
60
70 1-0---7 ~-2 - -t:;- - 3 --I!!>--8 )1( 13 -0- 18/
Soil water content (mm)
0 5 10 15 20 25 30 35
0
10
§ 20
'-' 30 9/1/2008 Ploughedt
Q)
Q 40
50
60
70 1-0---7 --<>--2 - -t:;- -3 ~8 )1( 13 -0- 181
Figure 8.16. Soil water distribution with respect to depth at different distances from the
dead level contour and infiltration pits averaged across two farms (Dube and Siziba) on
wettest day along unploughed (a) and ploughed (b) transects during 2007/08 season
8.4.5 Crop yields
The changes in soil water content (i1S) from the first day of measurement (14 November
2007) to the last day (12 April 2008) are given in Tables 8.2 and 8.3. At farms with dead
level contour only, soil water increased at 2, 3 and 8 m from the contour along the
197
ploughed transect (Table 8.2). In contrast soil water decreased at 3 and 8 m from the
contour along the ploughed transect at farms with dead level contour and infiltration pits
(Table 8.3). Crops planted along the ploughed transect extracted more soil water at farms
with dead level contours and infiltration pits than at farms with dead level contours only.
Evapotranspiration values were similar at all distances from the contour along the
unploughed and ploughed transects during the 2007/08 season.
Table 8.2. Components of water balance averaged across two farms (Moyo and Ncube) at
each distance along unploughed and ploughed transects the during 2007/08 season
Rainfall Distance ~SW unploughed ~SWploughed lEl'unploughed lEl'ploughed
{mm~ {m~ {mm~ {mm} {mm~ {mm~
247 -7 -1.1 -1.2 246 246
-2 -7.9 5.2 239 252
3 -15.0 2.0 232 249
8 -3.9 3.9 243 251
l3 -2.4 -4.5 245 243
18 -2.l -1.3 245 246
Mean -3.9 0.7 242 248
s.e. 2.7 1.5 2.2 1.4
Table 8.3. Components of the soil water balance averaged across two farms (Dube and
Siziba) at each distance along unploughed and ploughed transects across dead level
contours and infiltration pits during 2007/08 season
Rainfall Distance ~SWunploughed ~SW ploughed lEl'unploughed lEl'ploughed
(mm) (m) (mm) (mm) (mm) (mm)
262 -7 0.2 -14.l 262 248
-2 -6.0 -0.4 256 262
3 -2.7 -20.0 259 242
8 -9.l -15.4 253 247
l3 -17.4 -14.9 245 247
18 -14.6 -10.5 247 252
Mean -8.2 -12.5 254 250
s.e. 2.8 2.7 2.7 2.8
198
Distance from dead level contour had no significant (P > 0.05) influence on maize and
pearl millet yields observed at the end of 2007/08 season along the ploughed transect
(Tables 8.4 and 8.5). There was no maize grain harvested during 2007/08 growing at the
two farms because of drought (Table 8.5). The farms with dead level contours (Dube and
Siziba) had higher pearl millet and maize yields than farms with dead level contour only
(Moyo and Ncube) (Tables 8.4 and 8.5). Comparing the WUE of the four farms shows
that farms with no infiltration pit had the lower WUE than farms with dead level contour
and infiltration pit. The higher WUE at Dube and Siziba farms compared with Moyo and
Ncube farms can be attributed to the higher yields achieved at the farms with dead level
contours and infiltration pits as the ET values were quite similar (Tables 8.2 and 8.3).
Generally yields and WUE observed at these farms were lower than reported elsewhere
(Chapter 7; Ncube, 2007).
Table 8.4. Pearl millet and maize yields and water use efficiency between different
positions from dead level contour averaged across two farms (Moyo and Ncube) along
the ploughed transect at the end of the 2007/08 growing season
Position Crop Grain yield Biomass WUEgrain WUEbiomass
from (kgha") yield (kgha" (kgha-t
contour (kgha-I) mm") mm")
(-7)to (-2) P.millet 160 865 0.6 3.5
3 to 8 95 845 0.4 3.4
8 to 13 150 1 005 0.6 4.1
.1~3 to 18 IIJ~240 _1 2I65__1.0 5.2Mean 161 995 0.7 4.0s.e. 66 97 0.1 0.4ra
(-7) to (-2) Maize 0 820 0 2.9
3 to 8 0 700 0 2.5
8 to 13 0 1 300 0 4.7
13to18 0 1400 0 5.2
Mean 0 1 055 0 3.8
s.e. 0 173 0 0.7
199
Table 8.5. Pearl millet and maize yields and water use efficiency between different
'positions from dead level contour averaged across two farms (Dube and Siziba) with
dead level contours and infiltration pits along the ploughed transect at the end of 2007/08
season
Position Crop Grain yield Biomass WUEgrain WUEbiomass
from (kgha') yield (kgha-t (kgha"
contour (m) (kgha-t) mm") mm")
(-7)to (-2) P. millet 300 1 850 l.2 7.3
3 to 8 325 2 175 l.3 8.9
8 to 13 400 2500 l.6 10.1
13 to 18 390 2340 l.6 9.4
Mean 354 2216 1.4 8.9
s.e. 73 138 0.1 0.6
(-7) to (-2) Maize 0 1400 6.7
3 to 8 0 1 100 5.8
8 to 13 0 1200 6.2
13 to 18 0 900 4.5
Mean 0 1 150 5.8
s.e. 0 104.0 0.47
8.5 Conclusion
As expected the upward side of the dead level contour with or without infiltration pit was
drier than downslope on most occasions when soil water was measured. The dead level
contours with infiltration pits captured more rainwater following heavy rainfall events
(more than 40 mm). After the contours and infiltration pits had filled with rainwater
following rainfall events of 60 and 70 mm on the night of 8 to 9 January 2008, the wettest
parts of the field were at -2 and 3 m from the contour: The lateral movement of soil water
from the contour could also be detected at 8 m on the down slope at some farms.
The soil layers that benefited most from the rainwater captured by the dead level contours
with/without infiltration pit were in the middle of the soil profile. Crops such as maize,
sorghum, pearl millet and legumes commonly grown in semi-arid areas can access soil
200
water from these soil layers. The bulk of the active maize roots have been observed in the
0-0.8 m soil layer (Amato and Ritchie, 2002) although roots can reach 1.5 m deep
depending on the soil type. For sorghum and pearl millet most of the roots are
concentrated in 0 - 0.50 m soil layer (Zaongo et al., 1997). Roots for grain legumes such
as cowpea are also concentrated in the 0 - 0.50 m soil layer (Moroke et al., 2005).
Given the crop-livestock systems of Gwanda district, it is probably worth exploring strip
cropping of food and fodder crops on the downslope of the dead level contours and
infiltration pits. Deep rooted fodder crops would probably benefit more from the soil
water supply derived from dead level contours and infiltration pits. Combining structures
such as dead level contours and infiltration pits with in-field rainwater harvesting
techniques such as planting basins and tine ripping could be explored to assess if it will
bring better returns to smallholder farmers where dead level contours and infiltration pits
are being promoted and used.
201
CHAPTER9a
Soil Water and Maize Yield Responses toMinimum Tillage and Mulching
on Clayey and! Sandy Soils in Semi-Arid Southern Zimbabwe
9a.1 Introduction
Water is one of the major factors limiting agricultural production in semi-arid areas of
sub-Saharan Africa (Tsubo et al., 2005). In the smallholder farming system of the
semi-arid environments, rainfall is the only source of water for crop production. High
in-season spatial and temporal rainfall variability makes crop production risky under
the semi-arid conditions. Recurrent droughts and frequent mid-season dry spells are
some of the major causes of reduced crop yields and total crop failures observed in
smallholder cropping systems (Rockstrërn et al., 2003). Reducing total crop failure or
substantial yield reductions due to droughts or mid-season dry spells remains a
challenge for semi-arid crop production (Tabor, 1995).
The impact of low and poorly distributed rainfall on smallholder crop production can
be reduced by praeticing soil water management techniques. A wide range of
techniques for soil water management have been developed for improved semi-arid
smallholder crop production. Minimum tillage is one such technique that has been
explored for improved and prolonged soil water supply during the cropping period.
The minimum tillage techniques that have been developed include clean and mulch
ripping, no-till tied ridging and zero tillage. Currently a planting basin tillage system
is being widely promoted in the smallholder sector of the semi-arid districts of
Zimbabwe (Twomlowet al., 2008a; Chapter 7). These minimum tillage practices
202
simultaneously conserve soil and water resources, and increase or stabilise crop
production.
The soil water and crop yield benefits derived from using minimum tillage under
semi-arid conditions can be enhanced by using mulch cover in the cropping system.
For the semi-arid areas mulching maybe a suitable agronomic practice for conserving
soil water and controlling soil temperature regimes (Charkraborty et al., 2008). The
presence of crop residue mulch at the soil-atmosphere interface has a direct influence
on infiltration of rainwater into the soil and evaporation from the soil (Erenstein,
2002). Mulch cover reduces surface runoff and holds rainwater at the soil surface
thereby giving it more time to infiltrate into the soil. Soil biota increase in a mulched
soil environment thereby improving nutrient cycling and organic matter build up over
a period of several years (Holland, 2004).
The effect of the basin and ripper systems, and mulching on soil water dynamics and
crop yields have not been quantified under the semi-arid conditions of southern
Zimbabwe. This trial was designed to assess the soil water and maize yield benefits
derived from praeticing ripper and basin tillage systems combined with mulching
under the semi-arid conditions of southern Zimbabwe. The chapter reports the
findings from two seasons of experimentation on a sandy soil at Lucydale
experimental site and four seasons using a clay soil at Matopos Research Station.
203
9a.2 Objectives
This study uses on-station field experiments:
(1) to determine soil water responses to conventional ploughing, ripper and basin
tillage systems, and mulching;
(2) to determine maize (Zea mays L.) yield responses to conventional, ripper and
basin tillage systems, and mulching; and
(3) to identify the optimum rates of mulch application for low potential areas of
southern Zimbabwe.
9a.3 Materials and Methods
9a.3.1 Experimental set up
The experiment was set up with a factorial treatment structure consisting of three
tillage methods (conventional ploughing, ripping and planting basins) and seven rates
of mulch cover (0, 0.5, 1, 2, 4, 8 and 10 tha"). Plots at Lucydale and Matopos were
initially pegged out in October 2004 and then maintained in subsequent seasons. The
treatments were arranged in a split-plot design with three replications at each field
location. The main plot factor was tillage (63 m x 6 m) and seven mulch levels were
randomly allocated in sub-plots (8 m x 6 m) on each tillage treatment. Each plot was
separated by a 1 m pathway to avoid movement of residue from one plot to the next
when tillage was undertaken. At Matopos research site a new field was established for
the experiment in 2005/06 as cowpea was planted on the 2004/05 field as part of a
rotation (Table 9a.l). Unfortunately, due to logistical problems it was not possible to
establish a new field at Lucydale in 2005/06. Residues at the location were not
enough to make fresh mulch applications. As a consequence in 2005/06 the residual
effect of the previous season's mulch levels was assessed.
204
Digging of planting basins and ripping were carried out after applying mulch cover in
August/September of each year, with the exception of the first year when all
operations were carried out in October. Planting basins were dug at 0.9 m x 0.6 m
spacing using a hand hoe and each basin measured 0.15 m (length) x 0.15 m (width) x
0.15 m (depth). Rip lines were opened at 0.9 m inter-row spacing using a
commercially available ripper tine (ZirnPlow) attached to the beam of a donkey-
drawn mouldboard plough. The ripping depth achieved on both soils, with a single
pass of the implement, varied between 0.15 and 0.18 m. Cattle manure (40 % organic
carbon, 0.43 % N, 0.21 % P) was applied in October each year at a rate of3 tha" in all
plots as basal soil fertility amendment. Manure was placed in the planting basins in
the basin system, dribbled along the rip line in the ripper system and broadcast in the
conventional system. Conventional ploughing was done soon after the first effective
rain (30 to 50 mm) in December each year using a donkey-drawn VS 100 mouldboard
plough. Planting furrows were then opened by donkey-drawn conventional plough at
inter-row spacing of 0.9 m in the conventional system. During the ploughing process
most of the crop residue applied as mulch was partially incorporated into the soil.
Table 9a.l. Experimental fields used and crops grown in each field from 2004/05 to
2007/08 seasons at Matopos Research Station
SeasOlIn lExperimenntal fneid Crops growlI1 Crop phase
2004/05 1 Maize 1
2005/06 1 Cowpea 1
2005/06 2 Maize 2
2006/07 1 Sorghum 1
2006/07 2 Cowpea 2
2006/07 3 Maize 3
2007/08 1 Maize 4
2007/08 2 Maize 4
2007/08 3 Maize 4
2007/08 4 Maize 4
205
9a.3.2 Experimental management
Planting on both soils was done between November and December in each season
depending on the onset of the rains. At Matopos a hybrid maize variety, SC403, was
planted in both seasons whereas at Lucydale an open pollinated variety, ZM421, was
planted in 2004/05 season and SC403 in 2005/06. An open pollinated variety had to
be planted at Lucydale in 2004/05 season in order to avoid contamination of breeding
trials that were established near our experimental field. Plant spacing was 0.9 m x 0.6
m with three kernels per station for planting basins. In-row spacing was 0.3 m with
two kernels per station for the ripping and conventional tillage treatments. Plants were
thinned two weeks after crop emergence to two per basin in planting basin and one
plant per station in the ripping and conventional tillage treatments. Ammonium nitrate
(34.5% N) was applied to all plots at 20 kgNha-1 as topdressing when the maize had
reached the 6 leaf stage. Weeds were controlled by hand hoe as required in all seasons
at each site.
9a.3.3 Data collection and analysis
In every season plant counts in each treatment was done two weeks after crop
emergence. At harvest grain and stover (above-ground biomass minus grain) yields
were estimated from a net plot consisting of five middle rows with a running length of
6 m. The weight of cobs and stover from the net plot of each treatment were
determined in the field before taking sub-samples for moisture correction. Grain and
stover samples were dried at 60°C for 48 hours for moisture adjustment. The maize
shelling percentage was determined for each treatment so as to be able to convert cob
weight into grain and core weights. Grain weight was converted to a per hectare basis
at 12.5% moisture content as final grain yield
206
In the 2004/05 and 2005/06 seasons gravimetric soil water was determined by
collecting soil samples fortnightly between planting basins, along riplines and rows in
the basin, ripper and conventional tillage systems, respectively. Soil samples were
collected at 0.15 m depth interval up to a maximum depth of 0.6 m. The soils were
weighed before oven drying them at 105°C for 48 hours for determining gravimetric
water content. Gravimetric water content for each soil layer was calculated using the
procedure outlined by Anderson and Ingram (1993). Gravimetric water content was
converted to volumetric water content for each respective depth using the measured
bulk density for each soil layer. Soil water content in millimetres was determined by
multiplying volumetric water content by thickness of each layer from which soil water
was measured. In 2006/07 and 2007/08 growing seasons soil water was measured
weekly at 0.1 m depth interval using a microgopher type capacitance probe.
Statistical analysis of maize crop data was done using split plot design with tillage as
main plot factor and mulch application rate as subplot factor. Regression analysis was
conducted to assess the relationship between mulch treatments and measured
parameters of the maize crop. Soil water data were analyzed using unbalanced
treatment design of ANOVA because of soil depth differences of access tubes
installed for soil water monitoring. All statistical analyses were conducted using
Genstat Discovery Edition 3 (www.vsni.co.uk). The least significant difference (Lsd) at
5 % significance level was used to compare treatment means.
207
9a.4 Results and Discussion
9a.4.1 Site characteristics
Table 9a.2 summarises the major soil physical and chemical characteristics at thetwo
experimental field sites, based on detailed soil analyses undertaken by Moyo in 2001.
The effective soil depth of the clay loam at Matopos was 1.3 m, compared to 0.9 m
for the sandy soil at Lucydale. Irrespective of soil texture observed at each site, clay
content increases with depth (Table 9a.2). The slopes were 0.5 - 1 % at Matopos and 3
% for Lucydale (Moyo, 2001).
Table 9a.2. Physical and chemical characteristics of Matopos and Lucydale soils, after
Moyo (2001)
Soil Matepos Lucydale
property
Classification Chromic-Leptic Cambisol Eutric Arenosol
Depth (cm) 0-6 6-16 16-40 40-60 0--12 12-24 24-35 35-57
Clay (%) 41 38 47 52 4 5 6 10
Silt (%) 20 23 17 17 4 5 4 3
Sand (%) 38 39 36 31 91 91 99 87
Gravel (%) - - - - 5 7 8 17
pH (CaCh) 7.5 7.6 7.7 7.8 5.0 4.9 4.8 5.5
O.C. (%) 0.46 0.80 0.37 0.48 0.00 0.00 0.04 0.00
Ca 40.2 40.9 32.3 33.4 1.2 0.80 0.70 3.1
(Cmol.kg")
Mg 14.8 15.4 16.6 19.7 0.40 1.00 0.70 2.2
(Cmol.kg")
K 1.98 1.77 1.64 1.67 0.02 0.03 0.03 0.04
(Cmol.kg")
90.4.2 Lucydale site
9a.4. 2.1 Seasonal rainfall
Total rainfall recorded during 2004/05 growing season at Lucydale was 320 mm,
which was 56 % of the long term average rainfall for the Matopos area (Chapter 4).
The October-December and January-March periods recorded 160 mm each. The
rainfall received during October-December and January-March periods was less than
208
the 69 year average for the first and second halves of the growing season (Chapter 4).
The 69 year October-December and January-March rainfall averages are 248 and 291
mm respectively. There was a 31-day dry spell that stretched from 26 January (day
118) to 25 February 2005 (day 148) (Fig. 9a.l). This 31-day dry spell started 43 days
after planting of the maize crop (Table 9a.3). In the 2005/06 season rainfall was well
distributed during the maize growing period and the total rainfall amount recorded
was 787 mm. The October-December and January-March periods recorded 393 and
394 mm of rain during 2005/06 season and these were higher than the 69 year average
for the first and second halves of the season for Matopos (Chapter 4).
In both 2004/05 and 2005/06 seasons, the soil profile at Lucydale had been recharged
following rainfall events that occurred before the planting dates. In the 2004/05
season 104 mm of rain was accumulated between 1 December (day 62) and 14
December 2004 (day 75). The soil profile was even wetter in the 2005/06 season than
2004/05 after 288 mm of rain was received between 1 December and 14 December
2005. In the 2004/05 season planting was followed by an 8 day dry spell while in the
2005/06 a 30 mm rainfall event was recorded two days after planting. This allowed
better maize crop establishment in the 2005/06 season than 2004/05 season (see later
in Tables 9a.4 and 9a.5).
209
1000 1- 2004/05 - - 2005/061
900
.---
..ê 800 - --_., 700 .J,J
;$ 600 _,.e5 ,,-500
<1) I
.;~;. 400 _.J
"3
E 300 I"
u::3 200 I
100 ir:
0
0 25 50 75 100 125 150 175 200 225 250
Time(days after I October)
Figure 9a.1. Daily rainfall distribution at Lucydale experimental site during 2004/05
and 2005/06 growing seasons
Table 9a.3. Characteristics of the 2004/05 and 2005/06 growing seasons at Lucydale
expenm. en taISIite
Season 2004/05 2005/06
* Start 9/12/2004 1/12/2005
* End 26/2/2005 4/3/2006
Number of rainy days 18 40
Maximum 24 hour rainfall (mm) 55 75
Total seasonal rainfall (mm) 320 787
Planting date 14/12/2004 14/12/2005
* The cntena for start and end of the growmg season and ramy day have been defined m
Chapter 4.
9a.4.2.2 Soil water regimes
The tillage system had no significant (P = 0.231) influence on soil water content
measured in the 0 - 0.30 m soil profile during the 2004/05 season. However, soil
water content increased (P < 0.005) with an increase in mulch cover in the 0.30 m
layer of this sandy soil in 2004/05 (Fig. 9a.2). In the conventional and ripper systems,
soil water content was significantly higher at 2 and 4 tha-I mulch cover than the basin
system. Further increase in mulch cover over and above 4 tha" did not give additional
benefits in soil water content under any of the three tillage treatments.
210
60
1- 0- Plough - -(>- - Ripper --r- Basins 1
50
,-... _.;t _
I - .-cr-. T........,;; ....,_ / 1.. ..-- - _-_---........ ...-......_.... --_... -Q)~ 40 ,
~
'0
C/J
30
20 +---------r--------,---------.--------~------__.
o 2 4 6 8 10
Mulch rate (tha")
Figure 9a.2. Average seasonal soil water content in the 0-0.30 m profile at Lucydale
during 2004/05 cropping season. Bars indicate standard error
In the 2005/06 season neither tillage system nor the residual effects of the previous
season's mulch treatments had any significant (P > 0.05) influence on soil water
content measured in the Ó - 0.3 m profile (Fig. 9a.3). Soil water content measured on
7 March (day 158) was significantly (P < 0.001) higher than soil water content
observed on other sampling dates. This was attributed to a 45 mm rainfall event that
was received on 4 March 2006 (day 155). However, the water content was similar for
all tillage treatments with 69, 70 and 72 mm being recorded in the 0 - 0.60 m profile
under conventional, ripper and planting basin tillage systems respectively. This
represented a 38, 24 and 42 % increase in profile soil water content from the previous
measurement in the CP, ripper and basin systems. This water was just as quickly lost
as the values measured on day 173 (23 March) was about 28 mm.
211
100 1-0- CP - -0- - Ripper -- Basins I
80
20
o +----,----,----,-----,----,----,----,----,----,----,
o 25 50 75 100 125 150 175 200 225 250
Time (days after I October)
Figure 9a.3. Soil water content in the 0-0.60 m profile at Lucydale during 2005/06
cropping season. Bars indicate standard error
9a.4.2.3 Maize crop performance
The basin tillage system had significantly (P = 0.002) higher number of plants than
conventional and ripper tillage systems during 2004/05 growing season (Table 9a.4).
Planting basin system had an average of 2.8 plants per m2 compared with 1.3 and 1.2
plants per m2 for the conventional and ripper tillage systems. Planting basins collect
rainwater and gives it more time to infiltrate into the soil. This leaves the soil in the
basin moist for much longer than rip furrows and conventional ploughing. The seed
planted in a basin has access to soil water for a slightly longer period compared to
ripper and conventional systems. The better maize crop establishment observed in the
basin system at Lucydale has also been reported on farmers' fields in semi-arid
districts of Zimbabwe where conservation farming is being promoted (Chapter 7).
Maize grain production was not significantly influenced (P > 0.05) by the tillage
system during the 2004/05 growing season at Lucydale. The maize yields were lowest
at 466 kgha" for the conventional, with 619 kgha" in the ripper and 1 069 kgha" for
212
basin tillage systems. Maize harvest index was not significantly (P > 0.05) influenced
by the tillage system used. Mulching had no significant (P > 0.05) effect on maize
grain and stover production, .harvest index and plant stand during the 2004/05
growing season. One would have expected the mulching benefits on maize yields in
2004/05 given the dry spells recorded between January and February 2005 when the
maize crop was in its reproductive stages. Although soil water increased with increase
in mulch cover (Fig. 9a.2), this soil water advantage was not translated into grain
yield on a sandy soil. Probably soil fertility constraints became more limiting to crop
growth than soil water availability in this particular season on a sandy soil.
Table 9a.4 summarises maize yield responses to the two-way tillage and mulch
interaction during the 2004/05 season. Under the conventional ploughing system
grain yield decreased with increase in mulch treatment from 1 to 8 tha'. Under the
..
ripper system, la tha-1 mulch treatment gave 514 kgha" more grain than 1 tha-1 which
was the lowest yielding mulch treatment. In the planting basin tillage system 2 tha"
mulch treatment gave highest grain yield of 1 470 kgha" with 10 tha-1 achieving 711
kgha". The lowest (P = 0.010) harvest index were recorded at 8 tha' mulch treatment
in the conventional ploughing, at zero mulch in the ripper and at 10 tha" mulch in the
planting basin tillage systems. Maize stover production and plant stand were not
significantly affected by the tillage and mulch rate interaction on the Lucydale sandy
soil.
213
Table 9a.4. First year effects of tillage and mulch treatments on maize crop
performance at Luc~dale eXEerimental site in 2004/05 growing season
Tillage Mulch rate Grain Stover Harvest Plant stand
(tha-1) yield yield index (m")
{kgha-1} {kgha-1}
Plough 0 693 568 0.47 1.6
0.5 269 654 0.31 1.3
1 503 568 0.42 1.2
2 423 617 0.37 1.4
4 406 642 0.35 1.1
8 213 592 0.26 1.1
10 754 889 0.40 1.5
Mean 466 647 0.37 1.3
Ripper 0 415 803 0.31 1.4
0.5 781 913 0.43 1.5
1 323 482 0.37 0.9
2 812 716 0.45 1.4
4 542 815 0.32 1.1
8 624 568 0.46 1.0
10 837 568 0.45 1.3
Mean 619 695 0.40 1.2
Basins 0 1 259 1037 0.51 2.8
0.5 949 988 0.42 2.8
1 972 1 062 0.45 2.7
2 1470 1 173 0.50 2.9
4 1276 1 197 0.47 3.0
8 1026 753 0.52 2.7
10 711 1 161 0.36 2.7
Mean 1094 1053 0.46 2.8
Lsdo.o5(tillage) 512 183 0.21 0.49
Lsdo.o5(mulch) 284 338 0.07 0.41
Lsdo.05(tillage x mulch 607 557 0.21 0.74
interaction)
CV(%) 40 44 19 24
In the 2005/06 season, maize crop stand was significantly (P < 0.001) influenced by
the tillage system used. The ripper system had more plants per m2 than conventional
and basin tillage systems (Table 9a.5). In the 2005/06 season soil water was not a
limiting factor for maize crop establishment and growth in each tillage system
compared with the 2004/05 season. Maize grain and stover production were not
significantly (P > 0.05) influenced by the tillage system and residual mulch cover
during 2005/06 season. Most the maize residue applied in 2004/05 season as mulch
214
had rotten away by the beginning of the 2005/06 season. Only the plots that had
received 8 and 10 tha-I in 2004/05 season had some maize residue left by the
beginning of 2005/06 season. Maize residues decompose rapidly especially during the
summer period when temperature and soil water conditions are conducive (Nhamo,
2007). Nhamo (2007), using the litter bag technique, observed a 45 - 50 % weight
loss of maize residues applied as mulch during summer.
Table 9a.5. Second year effects of tillage and residual mulch treatments on maize crop
performance at Lucydale experimental site in 2005/06 growing season
Tillage Mulclll rate Grain yield Stover yield Harvest Plant stand
(tllna-I) (kgha -I) (kgB:na-l) index (m-2)
Plough o 1 796 2533 0.35 2.0
0.5 2714 2610 0.45 1.9
1 1 880 2933 0.35 1.9
2 1 834 3082 0.32 2.3
4 2961 3022 0.46 1.7
8 1 785 4033 0.30 1.9
10 2231 2832 0.41 1.6
Ripper o 2851 3245 0.43 3..3 -
0.5 3 843 4 652 0.44 3.3
1 2606 3 096 0.45 3.1
2 3654 3911 0.45 3.1
4 3 122 3 259 0.45 2.3
8 2626 3 837 0.39 3.2
10 2263 2356 0.47 3.3
.j11_1i1.IIlIli~JJIIIII~:~4IIII~
Basins o 1 505 2321 0.42 2.6
0.5 2596 3210 0.43 2.1
1 3226 3 975 0.40 2.1
2 2604 4420 0.36 2.1
4 2 425 2 827 0.43 2.0
8 1 410 2049 0.39 2.1
10 1 665 2 185 0.40 2.0
.mmt.t.lI:i~iltliIIl~~t~IIf~_
Lsdo.os(tillage) 2233 3 129 0.13 0.31
Lsdo.os(mulch) 1 134 1 394 0.07 0.43
CV (%) 48 46 17 19
215
9a.4.3 Matopos site
9a.4.3.1 Seasonal rainfall
Total seasonal rainfall for 2004/05 was 359 mm at Matopos (Table 9a.6) and this was
61 % of the long term average rainfall reported by Ncube (2007) for the site. The
October-December and January-March periods received 199 and 160 mm of rain
during the 2004/05 season. These rainfall amounts for the both first and second halves
of the season were below the long term averages for the October-December and
January-March periods (Chapter 4). During the 2004/05 cropping period rainfall
distribution was poor between mid-January and 1 March (day 153) (Figs. 9a.4). This
coincided with end of the vegetative and start of reproductive stages of the hybrid
maize variety grown at Matopos. As observed at Lucydale, rainfall was also well
distributed at Matopos experimental site during the 2005/06 growing season with a
total of 832 mm being recorded for that season (Table 9a.6). The first and second
halves of the season received 412 and 420 mm of rain and these were 66 and 44 %
more than the long term average for these periods of the growing season (Chapter 4).
The 2005/06 growing season had the highest number of wet days and largest daily
rainfall event compared to the other three seasons (Table 9a.6). There was more
variation with the end of growing season than its start during the four seasons of
experimentation.
216
Table 9a.6. Characteristics of the 2004/05, 2005/06, 2006/07 and 2007/08 growing
seasons at Mato pos experimental site
Season 2004/05 2005/06 2006/07 2007/08 Long term
mean
Start 8/12/2004 30/11/2005 17/11/2006 23/11/2007 . 2/12
End 27/2/2005 19/3/2006 4/4/2007 26/1/2008 29/3
Number of 18 41 25 23 22
rainy days
Maximum24 55 80 52 46
hour rainfall 18/11/2006 17/1212007
(mm) 25/1/2008
Total seasonal 359 832 465 364 573
rainfall (mm)
Planting date 13/12/2004 13/12/2005 21/11/2006 12/12/2007
for ripper
and basin;
8/12/2007
for plough
1000 1~2004/05 ~ m ~ 2005/06 2006/07 ~ - 2007/081
900
,......_ 800 dlaDCDaClDê 0700 11'-'
.~5 600 ,."
~ 500 Jl>
Cl)
.;p> 400 = ~~ /_<':S ~ ="'5
E 300 1/ ~ ",; "',. ~::I
U 200 Jl q JL r.JJ.....,q
100 r_/'Df
o -
0 25 50 75 100 125 150 175 200 225 250
Time (days after I October)
Figure 9a.4. Cumulative rainfall distribution at Matopos experimental site during
2004/05, 2005/06, 2006/07 and 2007/08 growing seasons
Rainfall was poorly distributed during 2006/07 growing season and 465 mm was
received between October 2006 and April 2007 (Table 6). The first and second halves
of the growing season received 252 and 169 mm during 2006/07. The October-
December period received slightly more than the 69 year average rainfall for Matopos
217
while the second half received 42 % less than the 69 year average rainfall (Chapter 4).
In January 2007 only 12 mm of rainfall was recorded on 17 January (day 109).
Following a 16 day dry spell was experienced between 1 January 2007 and 16 January
2007. This dry spell was immediately followed by a 20 day dry spell which started on
18 January 2007 and ended on 6 February 2007. There is a5-11 % chance of getting
21 day dry spells in January at Matopos (Chapter 4). The first dry spell started 41
days after planting in the ripper and basin tillage systems and 24 days after planting in
the conventional system (Table 9a.6) so the plants were still small. The next rainfall
event after 17 January 2007 was 3.8 mm received on 7 February 2007 (day 130) and
this was followed by a 5 mm rainfall event that occurred on 10 February 2007 (day
133). These would have little effect on crop growth as they are less than 10 mm. The
highest daily rainfall event was 52 mm (Table 9a.6) and was received on 18
November 2006.
Total seasonal rainfall for 2007/08 growing season was 364 mm which was 38 % less
than the long term average rainfall for Matopos. The October-December and January-
March periods received 250 and 114 mm of rain respectively. The first half received
slightly more than the 69 year average rainfall while the January-March period
recorded 62 % less rainfall than the 69 year average (Chapter 4). The last rainfall
event of season was 46 mm recorded on 25 January 2008 (day 117) but does not set a
new record as previous extreme was on 13 December (Chapter 4). The maize crop
had to rely on stored soil water between 25 January 2008 (day 117) and the
physiological maturity stage of the crop. The highest daily rainfall was 46 mm
received on 17 December 2007 (day 78) and also on 25 January 2008 (day 117)
(Table 9a.6).
218
As observed at the Lucydale site in the 2004/05 and 2005/06 seasons, the soil profile
at Matopos was adequately recharged prior to planting. In the 2004/05 season 75 mm
of rain were accumulated in five days before the planting date. In the 2005/06 season
81 mm of rain were accumulated in five days before the planting date. In the 2006/07
season the ripper and basin systems were planted earlier than conventional system. It
was too wet to plough the red clay soil after receiving 68 mm of rain five days prior to
21 November 2006 as the soil profile was recharged by a 52 mm rainfall event
received on 18 November 2006 (day 49). In the 2007/08 season, planting was done
after accumulating 39 mm of rain four days prior to planting.
Generally the January-March half of the season was drier than the October-December
period during 2004/05, 2006/07 and 2007/08 growing seasons. This observation is at
odds with earlier results obtained in Chapter 4 which show that in the long term the
January-March period was wetter than the October-December period. The 2004/05,
2006/07 and 2007/08 seasons were probably some of the seasons with below average
rainfall that occur in semi-arid areas. The abrupt end of the 2007/08 growing season
on 25 January 2008 had a large detrimental impact on maize yields achieved under all
the different tillage systems and mulching treatments showing that in some seasons it
is not possible to prevent crop failure. The unpredictability of start and end of the
growing seasons is one of the major challenges in rainfed cropping in semi-arid areas
(Dennett, 1987). The dry spells of more than two weeks observed at Matopos during
the seasons with below average rainfall have also been reported in other semi-arid
environments of Africa (Sivakumar, 1992; Rockstrom et al., 2003).
219
9a.4.3.2 Soil water regimes
Figure 9a.5 gives the soil water content in the 0 - 0.30 m soil profile at Matopos
during 2004/05 season. Despite the below-average rainfall, the three tillage methods
had no significant (P = 0.866) influence on the average soil water content observed in
the top 0.30 m. There was a general increase in average seasonal soil water content
with increasing mulch rate from 0 to 4 tha", A mulch rate of 4 t ha-I appears to be a
threshold point on this soil for the season, with a significant (P < 0.001) increase in
soil water content, compared to the lower mulch rates. The low soil water content
observed at 1 tha" mulch cover in the ripper system can probably be attributed to
human error during the processing of soil samples for gravimetric water
determination.
In the 2005/06 season neither tillage nor mulching treatments had any significant (P >
0.05) influence on soil water content measured in the top 0.60 m depth during this
wetter-than-average season. The basin system started with marginally higher soil
water content in the top 0.60 m soil depth (Fig. 9a.6), suggesting some initial water
harvesting. However, soil water content in the top 0.60 m of the profile gradually
declined as the season progressed (Fig. 9a.6).
220
70
60 1-0- Plough - -.- - Ripper -- Basin 1
50 1I:_-'-- - - - -f .. - - ., __-a- .---.._.,... -- ....-A ..
#.~ ---
40 ,\;Z •
'f'
30
20 +----------r---------.---------.----------~------__.
o 2 4 6 8 10
Figure 9a.5. Average seasonal soil water content in the 0 - 0.30 m profile under three
tillage systems and mulching treatments at Matopos during 2004/05 season. Bars
indicate standard error of means
180
160 1-0- Plough - -0- - Ripper -- Basins 1
,.......
140
~
.._."
.... 120
Q)
rt!
~ lOO
Q)
rea
80
A.. 60
40
20
0 25 50 75 lOO 125 150 175 200
Time (days after 1 October 2005)
Figure 9a.6. Soil water content in the 0 - 0.30 m profile at Matopos during 2005/06
cropping season. Bars indicate standard error
In the 2006/07 season, profile soil water dynamics were quite similar in conventional,
ripper and basin tillage systems (Figs. 9a.7). Day 142 (20 February 2007) had the
lowest profile water content during the growing season as the small showers (18 mm
in total) had no effect after the long dry spells. On that day conventional tillage
system had 114 mm at 0 tha·1 mulch, 100 mm at 2 tha", 109 mm at 4 tha-I and 99 mm
221
at 10 tha-I mulch treatment The ripper system had 97 mm for 0, 2 and 4 tha-1while 10
tha' mulch treatment had 104 mm. Planting basins had 100, 108, 87 and 113 mm of
soil water at 0, 2, 4 and 10 tha-Imulch treatments. The higher profile water content at
10 tha' mulch cover in the basin and ripper systems is attributed to mulching being
able to reduce soil evaporation and hence conserve soil water.
Changes in soil water content in the top four soil layers followed a similar pattern in
the three tillage systems (Fig. 9a.8). The 0 - 0.15 m soil layer responded to wetting
and drying cycles more dramatically than the other layers. The 0.15 - 0.25 m layer
consistently held the lowest amount of soil water throughout the growing season in all
tillage systems. The low soil water content in the 0.15 - 0.25 m soil layer is attributed
to root extraction by the maize crop. Studies conducted by Vogel (1992) and Chikowo
et al. (2004) in the relatively wetter northern Zimbabwe indicated that a large
proportion of maize roots occur within the 0 - 0.2 m plough layer.
In the 2007/08 season soil water content in the 0 - 0.85 m profile responded to rainfall
events in all three tillage systems throughout the growing season (Fig. 9a.9). In the
conventional system, 10 tha' mulch treatment had slightly more soil water than other
mulch treatments particularly from day 85 (24 December) up to the end of season. In
the ripper system there was more soil water under 10 tha" treatment than the other
mulch treatments following rainfall events of December 2007 and January 2008. In
the basin system 10 tha" mulch treatment had more (P = 0.017) soil water throughout
the growing season particularly during extended dry period after day 150 (27
February) to the end of the season. As observed in other seasons at the Matopos site,
the 0.15 - 0.25 m layer had the lowest water content in all tillage systems throughout
222
2007/08 growing season (Fig. 9a.10). The top soil layer (0 - 0.15 m) also showed
significant responses to wetting and drying cycles. The deepest layer (0.35 - 0.45 m)
had the highest soil water content in the second half of the growing season.
180 -0 -+-2 - -0- -4 -- 10
I160
~ 140
Q)
«i 120
~
~ 100
t;::: Plough
c8, 80
60+---~----~--~--~----~--~----~--~----~~
o 25 50 75 100 125 150 175 200 225 250
Time(days after 1October 2006)
180 1-0 -+-2 - -0- -4 -0- 101
K--- 160,_ 140
Q)
«i 120
~
ïB 100
~8 80
60+---~--~----~--~--~----~--~--~----~--~
o 25 50 75 100 125 150 175 200 225 250
Time(days after 1October 2006)
180 1-0 -+-2 - -0- -4 -0- lol
I160
,_ 140
~
~ 120
ïB 100
c8, 80
60-·~--~--~----~--~--~--~~--~--~----~~
o 25 50 75 100 125 150 175 200 225 250
Time(days after 1 October 2006)
Figure 9a.7. Profile soil water changes in the conventional ploughing, ripper and basin
system and four mulch treatments (0, 2, 4 and 10 tha") at Matopos experimental site
during 2006/07 growing season
223
30 ---+-0-15 - -6,- -15-25 ~25-35 --t- 35-45
·0
en 10 Plough
5 +----~---,----.----.----.----.---,----,----.----,
o 25 50 75 100 125 150 175 200 225 250
Time(days after 1 October 2006)
30 ~0-15 - -6,- - 15-25 ~25-35 --t- 35-45
ï 25
1-0 20
~
-~ 15·0
If.J 10 Ripper
5 +---~----~---.---,~--,----.----.----,----.---~
o 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2006)
30 G 0-15 - -6,- - 15-25 ~25-35 --t- 35-45
A 25
ê
':" 20
~
.-.~o... 15
If.J 10 Basins
5 +---~---,~--,----.----.----.---,----.----.---,
o 25 50 75 100 125 150 175 200 225 250
Time(days after 1October 2006)
Figure 9a.8. Soil water changes in four different layers of the soil profile under the
conventional ploughing, ripper and basin tillage systems during 2006/07 growing
season at Matopos site
224
400 /-0 -+- 2--0- -4 -- 10/
-- 350 .
~ 300 Plough
~t 250
~ 200
~J:! 150
£ 100
50+----.----,---~---,~--~---.----~--~----,---~
o 25 50 75 100 125 150 175 200 225 250
Time(days after 1 October 2007)
400 1--0 -+- 2 - -0- -4 -0- 101
-- 350
~ 300....
~ 250
~ 200
~.£e 150c, 100
50
0 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2007)
-- 400 /-0 -+- 2 - -0- -4 -Ii>- 10/ê 350
'-...". 300
~ 250
~ 200
Q)
~ 150
ce, 100
50
0 25 50 75 100 125 150 175 200 225 250
Time(days after I October)
Figure 9a.9. Profile soil water changes in the conventional, ripper and basin tillage
systems and four mulch treatments (0, 2, 4 and 10 tha") at Matopos experimental site
during 2007/08 growing season
225
50 .. 0-15 - -t::.- - 15-25 -<>- 25-35 --I- 35-45
t 40
':;' 30
(IJ
(ij
~ 20
:-;::::I
0
r:/.J 10
0
0 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2007)
50 -0--0-15 - -t::.- -15-25 -0-25-35 --!- 35-45
I 40 Ripper
[) 30
(ij
~ 20
:-;::::I
o0: 10
0
0 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2007)
50 ---e- 0-15 - -t:r - 15-25 ---<r- 25-35 -+- 35-45
t 40
'-~" 30
~
~ 20
.......
'0
r:/.J 10
0
0 25 50 75 100 125 150 175 200 225 250
Time (days after 1 October 2007)
Figure 9a.10. Soil water changes in four different layers of the soil profile during
2007/08 growing season at Matopos experimental site
9a.4.3.3 Maize crop performance
The tillage system had no significant (P > 0.05) effect on maize grain production in
the 2004/05 season. Tillage system and mulch rate interaction had no significant
226
influence on maize stover produced and harvest index. The two-way tillage and mulch
interaction significantly (P = 0.035) influenced maize grain production at Matopos in
2004/05 growing season (Table 9a.7). The highest grain yield was recorded at 4 tha-1
mulch treatment in the conventional tillage system (Table 9a.7). The 1 tha-1 mulch
treatment gave the highest yield in the ripper system while 10 tha" mulch cover out
yielded the other mulch treatments under the basin system.
Table 9a.7. First year effects of tillage and mulch treatments on maize grain yield
(kgha') at Matopos experimental site in 2004/05 growing season
Tillage Mulch cover
tha"
o 0.5 1 2 4 8 10 Mean
Plough 2 106 2269 2527 2429 3243 2547 2514 2519
Ripper 1 633 2074 3699 3413 2086 2777 2968 2664
Basins 1 800 2 549 2 322 2320 3203 2822 3470 2641
Mean 1 846 2 297 2 849 2721 2844 2715 2984 2608
Lsdo.o5= 1 165 kgha ,CV = 26 %
The two-way tillage x mulch interaction had a significant influence on maize crop
establishment (Table 9a.8). At 4 - 10 tha" mulch levels, the basin tillage system had a
higher (P < 0.05) plant stand than the conventional and ripper systems. Under the 0
tha" mulch treatment, the conventional system had the lowest (P < 0.05) plant stand.
Table 9a.8. First year effects of tillage and mulch treatments on maize crop
establishment (plants m-2) at Matopos experimental site in 2004/05 growing season
Tillage Mulch cover
(tIIla-I)
o 0.5 1 2 4 8 la Mean
Plough 1.7 3.0 3.3 3.0 2.0 2.0 2.3 2.5
Ripper 2.0 2.7 3.0 2.7 2.3 2.3 2.7 2.7
Basins 2.0 3.6 3.7 2.3 3.7 3.7 3.0 3.1
Mean 1.9 3.1 3.4 3.0 2.7 2.7 3.0 2.8
Lsdo.o5= 0.92 plants per m ,CV = 21 %
227
Contrary to our results presented previously from the Lucydale site in 2004/05 season
(Table 9a.4), grain production was significantly (P = 0.006) influenced by mulch
cover during 2004/05 season at Matopos. In all tillage systems grain yield recorded at
o tha·1 mulch treatment was lower than grain achieved at all other higher mulch cover.
The clay soil at Matopos has a higher water holding capacity than the granitic sandy
soil at Lucydale and the soil profile at Matopos was recharged by 128 mm of rain
received in December and 119 mm recorded from 1 January to 25 January. There was
more soil water held in the soil profile at Matopos than in the soil profile at Lucydale
which the maize roots could access. Despite the 31 day dry spell recorded at the two
sites the mulch cover could have conserved more soil water at Matopos than
.Lucydale, prolonging the period of soil water availability to the maize plants. Other
researchers elsewhere have reported that mulching reduces soil evaporation thereby
extending the period that soil water is available to crops (Zhai et al., 1990; Sauer et
al., 1996; Erenstein, 2002).
At the end of a good rainfall 2005/06 season, neither tillage nor mulch treatments nor
any combination of the two had a significant (P > 0.05) effect on all the maize
production parameters measured at the Matopos site (Table 9a.9). Although there
were no statistically significant effects of tillage on either stover or grain production,
the basins out performed conventional ploughing at all except 2 and 10 tha-I mulch
levels, and the ripper at mulch levels of 1 to 8 tha-I. The two-way tillage and mulch
cover interaction did not have a significant influence on maize crop performance in a
growing season with above average rainfall pattern. As observed at the Lucydale site
soil water availability was not a constraint to maize crop growth during the 2005/06
growing season.
228
Table 9a.9. First year effects of tillage and mulch treatments on maize crop
perfonnance at Matopos experimental site in 2005/06 growing season
Tillage Mulch rate Grain Stover Harvest Plant
(tha-1) yield yield index stand
(kgha-1) (kgha") (m-2)
Plough 0 3 795 6 124 0.35 2.8
0.5 3 076 5 728 0.34 2.7
1 4 388 5 765 0.40 2.6
2 6342 8 148 0.41 2.9
4 5 108 7 161 0.39 2.7
8 5 359 6296 0.43 2.5
10 4947 6543 0.42 2.7
Ripper 0 4947 5 642 0.43 2.8
0.5 5 177 4938 0.49 2.7
1 5 536 6333 0.43 2.7
2 4986 6494 0.41 2.7
.t__ 4 4950 5 580 0.44 3.08 3 170 4 827 0.36 2.610 5369 6 124 0.44 2.7I!@.JIIiIlII!41[17~B.fI~~mr~iIII
Basins 0 4739 6 662 0.38 3.0
0.5 4975 6 617 0.40 3.2
1 5 754 6469 0.43 3.0
2 5687 7408 0.41 3.2
4 5498 7049 0.40 3.3
8 5379 7284 0.39 2.9
10 4048 5852 0.38 2.9
___ ~~~ll~~i$_J!.f~~~iIII
Lsdo.o5(tillage) 2 382 3 332 0.049 0.95
Lsdo.o5(mulch) 1 212 1 071 0.047 0.24
Lsdo.05(tillage x mulch 2710 3338 0.083 0.94
interaction)
CV (%) 26 18 12 9.0
During the 2006/07 season, the tillage system had no significant (P > 0.05) influence
on maize grain and stover production, and plant stand (Table 9a.10). However, in this
season the basin system had significantly (P = 0.011) lower harvest index than
conventional and ripper systems (Table 9a.10). The mean harvest indices observed
were 0.36, 0.29 and 0.25 for conventional, ripper and basin systems. Mulching had a
significant (P < 0.001) influence on maize grain production during 2006/07 growing
season at Matopos. Maize grain production significantly increased with increase in
229
--------------------------------------------------------------------------
mulch cover particularly in the ripper and basin systems (Fig. 9a.13). The 2006/07
season received below average rainfall and there were dry spells lasting more than
two weeks during January and February 2007. The increase in grain yields with
increase in mulch cover (Table 9a.10) could be ascribed to the extension of the period
when soil water was available to the maize crop under high mulch cover. Prolonging
the period of soil water availability under high mulch treatments facilitated better
grain filling as this is when dry spells occurred late in the season. Mulching had no
significant influence on maize stover production and plant stand in the 2006/07
season.
In the 2007/08 growing season, tillage system and mulch treatment had no significant
(P > 0.05) influence on maize grain and stover production, harvest index and plant
stand (Table 9a.II). There were wide variations in plant stands across the treatments
due to rodent damage particularly in the basins early in the season. Waterlogging in
the ripper and basin plots during late December and January also resulted in the
suppression of maize yields observed at the end of 2007/08 season. Lack of yield
response to mulching is rather puzzling given the fact that the last rainfall event was
recorded on 25 January 2008 but shows that the vegetative growth of the crop has a
strong influence on the final crop yield. The maize crop endured more than two
months without rain. The possible explanation is that the clay soil profile was
saturated by the rains received in December 2007 and early January 2008, and there
was enough soil water to take the maize crop to maturity regardless of the mulch
cover.
230
Table 9a.10. First year effects of tillage and mulch treatments on maize crop
performance at Matopos experimental site in 2006/07 growing season
Tillage Mulch rate Grain yield Stover Harvest Plant stand
(tha-I) (kgha-I) yield index (m-2)
(kgha")
Plough o 676 1 494 0.31 0.8
0.5 1045 1 543 0.38 1.1
1 1 712 3 136 0.34 1.4
2 1 536 2420 0.36 1.4
4 902 2086 0.28 0.8
8 2018 2025 0.46 1.1
10 1 554 2099 0.40 1.4
~
Ripper o 1 060 3 086 0.25 2.6
0.5 1 148 1 679 0.38 1.4
1 919 3654 0.23 1.8
2 1456 3790 0.26 1.3
4 2030 4086 0.31 2.0
8 1 857 3444 0.34 1.5
_w~Yi~10_m11 7~51~~.414441 0.27 1.3
Basins 0 735 4 111 0.14 2.5
0.5 973 4371 0.18 2.6
1 1 210 3 876 0.24 1.9
2 928 3 642 0.25 1.6
4 1 570 3000 0.33 1.5
8 2 123 5 025 0.27 2.2
10 2 129 3 852 0.33 1.9
~ ..
Lsdo.o5(tillage) 592 1 807 0.053 1.2
Lsdo.o5(mulch) 409 937 0.056 0.36
Lsdo.os(tillage x mulch 794 2074 0.098 1.2
interaction)
CV(%) 31 31 20 23
231
2500 Plough
,-.. 2000 D
-(II
'c"Éb 1500 !ill
'0
~
;>.
.~ 1000 c
c
500 Y 0.094x 3= - 9.244x-? + 149.49x + 1027.8
R2 = 0.32
0
0 2 4 6 8 10
Mulch cover (tha")
2500 Ripper
e
,-.. 2000
-(II D
'c"Éb 1500
::9
QJ
'>,
.~ 1000
o 3 ?
500 Y = -0.4074x - 19.381 x" + 311.84x + 929.3
R2 = 0.85
0
0 2 4 6 8 10
2500 Basins
2000
-('"II"'
~
C 1500
::9
QJ
'>,
1000
.~
o
500 3y= -2.1441x + ?23.338x- + 109.04x + 851.82
R2 = 0.94
0
0 2 4 6 8 10
Figure 9a.ll. Maize grain yield responses to mulching on a red clay soil under
conventional, ripper and basin tillage system during the 2006/07 growing season
232
Table 9a.ll. First year effects of tillage and mulch treatments on maize crop
performance at Matopos experimental site in 2007/08 growing season
Tillage Mulch rate Grain Stover Harvest Plant
(tba-I) yield yield index stand
(kgha-I) (kgha') (m-2)
CP 0 1 504 1 741 0.39 1.9
0.5 1 507 1 321 0.46 1.7
1 1 129 1 210 0.40 1.6
2 664 815 0.40 1.5
4 1 218 1 272 0.41 1.5
8 1 163 1 284 0.42 1.4
10 1 878 1 543 0.47 1.9
_~~§ïDfI.tll~<1m.~_
Ripper 0 1 257 1 296 0.42 1.8
0.5 2211 1 790 0.48 2.2
1 1 194 1 210 0.43 1.8
2 1 202 1 198 0.42 1.8
4 1 162 1 333 0.41 1.8
8 1 686 1 679 0.44 2.0
1If_~ 10 •1 1•00 1 210 0.43 1.8_£8!g ll~l.
Basins 0 1 123 1 296 0.40 1.6
0.5 964 1 173 0.41 1.5
1 971 1 185 0.41 1.8
2 1 009 1 136 0.41 2.2
4 1 076 1 271 0.40 1.7
8 781 1 037 0.39 1.3
10 519 704 0.30 1.1
~~~~E~Jm~~,~g~~~~'\!f~lmiI~:"%"~wl~,~~~~~ ;9~~
Lsdo,o5(tillage) 613 738 0.060 1.45
Lsdo,o5(mulch) 428 419 0.066 0.32
Lsdo,o5(tillage x mulch 828 883 0.114 1.42
interaction)
CV(%) 37 34 17 13
9a.4.3.4 Maize crop performance across four growing seasons
The maize crop performance differed significantly (P < 0.001) across the four seasons
of experimentation. All measured production parameters were significantly (P <
0.001) influenced by season with 2005/06 having the highest grain and stover yields,
and plant stand (Table 9a.12). Measured mean grain yield was 2 656, 4 937, 1 397 and
1 206 kgha" for 2004/05,2005/06, 2006/07 and 2007/08 seasons. Mean stover yields
were 2948,6376,3 136 and 1 272 kgha" for 2004/05,2005/06,2006/07 and 2007/08
233
seasons. Maize harvest indices ranged from 0.30 to 0.43 while the highest plant stand
recorded across the four seasons was 3.7 plants m -2.
Table 9a.12. Effect of the growing season on maize crop performance at Matopos
experimental site
Season Grain Stover Harvest index Plant stand
(kgha-1) (kgba-1) (plants m-2)
2004/05 2656 2948 0.43 2.7
2005/06 4937 6376 0.41 3.7
2006/07 1 397 3 136 0.30 1.8
2007/08 1206 1 272 0.40 1.7
Lsdo.o5 390 438 0.023 0.20
CV(%) 43 36 17 23
The two seasons at Lucydale and four seasons at Matopos were characterised by
different rainfall distributions (Figs. 9a.1 and 9a.4). The relatively wet 2005/06 season
had the highest yields on both soil types regardless of the tillage system used. Studies
conducted in northern Zimbabwe by Smith (1988) over a 10 year period indicated that
rainfall characteristics during the season has more influence on maize crop
performance than tillage systems such as conventional ploughing, ripping and zero
tillage. The dry spells observed in the 2004/05, 2006/07 and 2007/08 seasons
substantially reduced maize yields. The dry spells experienced during this project all
coincided with flowering and grain filling stages of the maize crop during the second
half of the season. In West Africa Rockstrom and de Rouw (1997) observed
significant pearl millet yield reduction when dry spells occurred during the flowering
stage of the cereal crop.
The tillage system had no significant (P > 0.05) influence on any of the measured
parameters across the four seasons. Generally maize grain production increased (P =
0.010) with increase in mulch application rate up to 8 tha-I (Table 9a.13). The lowest
234
(P = 0.045) harvest index was recorded at 0 tha-I mulch treatment while the least (P =
0.005) plant stand was observed at 0 and 8 tha-I mulch treatments. The two-way
tillage system and mulch treatment interaction had no significant (P = 0.325)
influence on maize crop performance. The three-way tillage, mulch and season
interaction also failed (P = 0.580) to influence any of the measured parameters of the
maize crop.
Table 9a.13. Effects of tillage and mulching on maize performance averaged across
four growing seasons at Matopos experimental site
Treatment Grain yield Stover yield! Harvest Plant stand!
(kgha-l) (lkgllna-l) index (plants m-2)
_Tillage~CP 2470 3251 0.40 2.6Ri~pper 12 602~_3 386 I~0.39 2.3Basins 2575 3 661 0.38 _2.5
Lsdo.os 818 1 107 0.022 0.52
Mulch o 2 153 3237 0.36 2.3
0.5 2 331 3 093 0.40 2.6
1 2 613 3 602 0.38 2.6
2 2 664 3 654 0.38 2.5
4 2 670 3 466 0.39 2.4
8 2 723 3 540 0.40 2.3
10 2 687 3 439 0.40 2.4
R!~!~_....
Lsdo.os 344 460 0.029 0.18
CV(%) 43 36 17 23
9a.5 Conclusion
The four seasons of experimentation demonstrated that rainfall distribution during the
growing period is the key to successful cropping in semi-arid southern Zimbabwe.
Rainfall in two of the four seasons of experimentation started well in November but
that early start was not a guarantee for good distribution during the growing season.
Mid-season dry spells and abrupt cessation of rains occurred in seasons with below
average rainfall pattern and that had a negative impact on maize yields even on well
managed on-station experiments. This means that the level of farmer management and
235
care of the crop should be increased because any mishap at the beginning of the
season can be carried with the crop through to the yield.
Soil water dynamics were quite similar under the conventional ploughing, ripper and
basin tillage systems and mulch cover treatments even in seasons with below average
rainfall. This implies that there are equal chances of having reduced crop yields or
total crop failure if a smallholder farmer in southern Zimbabwe is using single
conventional ploughing, ripper or basin tillage systems combined with mulch cover in
the event of a drought or prolonged dry spell. Researchers have a task of modifying
the CA systems so that they can deal with the variable rainfall pattern of semi-arid
areas such as southern Zimbabwe. The 10 tha-I mulch treatment had more soil water
in the profile than the lower mulch treatments in basin and ripper systems but this soil
water advantage was not translated into maize yield. This study revealed that crop
stands and ultimately yields in the basin system can be reduced by rodents in some
drought years and by waterlogging in wet seasons. Smallholder farmers using the
basin system can therefore fail to benefit from the seasons with good rainfall
distribution if crop stand is poor.
The conventional, ripper and basin systems gave similar maize yields regardless of
the rainfall pattern during the growing season. Although the tillage system had no
significant influence on maize yields in seasons of varying rainfall patterns, the use of
ripper and planting basin tillage systems offers smallholder farmers an opportunity to
plant with the first effective rains. Therefore smallholder farmers in semi-arid
southern Zimbabwe have a range of tillage options to choose from depending on their
resource status. As demonstrated in 2004/05 and 2006/07 seasons mulching improves
236
maize yields in drought years. Although the highest maize yields were achieved at 8
thaot mulch rate, there were no significant yield benefits derived from increasing
mulching rate from 1 to 10 tha". Given the low cereal and legume biomass production
of the semi-arid smallholder systems and crop-livestock competition for crop
residues, mulch levels of 4 - 10 tha" might be impossible despite the little soil water
and crop yield benefits observed in this study. Smallholder farmers can therefore
target using 2 thaot mulch cover when embarking on conservation farming practices.
237
CHAPl'ER9b
Minimum Tillage and Mulching Effects on Soil Water Regimes, and!
Cowpea and Sorghum Yields on a R.edClay Soil in Semi-Arid Southern Zimbabwe
9b.l Introduction
Cowpea (Vigna unguiculata (L.) Walp) and sorghum (Sorghum bicolor (L.) Moeneh) are
traditionally grown as subsistence crops in the smallholder sector of sub-Saharan Africa.
Both crops are considered to be drought tolerant and commonly grown as sole or
intercrops in the smallholder farming systems of Zimbabwe (Nhamo et al., 2003), but
rarely in a planned rotation (Ncube et al., 2007). However, where rotation of cowpea with
cereals such as sorghum do occur, utilization of soil water and nutrients from different
soil depths may be promoted both within and between seasons (Moroke et al., 2005;
Ncube et al., 2007). For example, some improved cowpea varieties mature early, and
have the potential to allow some carryover of stored soil water from one season to the
next, for the following cereal crop.
Agricultural land management practices such as mulching and minimum tillage can
improve soil water supply to crops through reduced runoff and soil evaporation,
increased infiltration and water storage (Zaongo et al., 1997; Hatfield et al., 2001;
Erenstein, 2002; Lipic et al., 2005). Studies conducted in the higher potential areas of
Zimbabwe between 1988 and 1995 indicated that mulching significantly reduced surface
runoff and hence soil loss (Erenstein, 2002). The mulching material at the
soil-atmosphere interface holds rainwater at the soil surface thereby giving it more time
238
to infiltrate into the soil. Mulch cover also shields the soil from solar radiation thereby
reducing the energy available for evaporation from the soil.
Conservation agriculture is a system that encourages minimum soil disturbance and the
use of crop residue as mulch cover for improved soil water and weed management among
other benefits. It emphasizes rotation of cereals and legumes in a permanent grid of
planting positions. A fixed grid of planting positions is established during the dry season
and nutrients are applied to those fixed planting stations (Twomlowet al., 2008a).
However, the challenge lies in accommodating legumes and cereals in similar planting
positions each growing season without compromising on the principle of minimum soil
disturbance or yield of either the cereal or legume crop. Basically does a permanent
planting grid designed for cereals compromise yields for legume crops? This chapter
focuses on the cowpea and sorghum yield responses in consecutive seasons following the
planting of an initial maize crop.
9b.2 Objectives
This on-station experiment was designed to determine
(1) effect of planting basin, ripper and conventional tillage systems and mulching on soil
water regimes under cowpea and sorghum crops;
(2) effect of planting basin, ripper and conventional tillage systems and mulching on
cowpea and sorghum yields; and
(3) optimum mulching rate for semi-arid Zimbabwe.
239
9b.3 Materials and Methods
9b.3.1 Experimental set up
The experiment was established at the International Crops Research Institute for Semi-
Arid Tropics (ICRISAT), Matopos Research Station, for two growing seasons on a clay
loam soil. In 2004/2005 season the maize phase 1 was established and planted to a sole
maize crop (see Table 9a.1 in Chapter 9a). In 2005-2006, as part of planned rotation for
the experiment, a cowpea crop (cowpea phase 1) was planted in the field that was
previously under maize phase 1. In the same 2005/06 season, maize phase 2 was
established on an adjacent plot of land and planted to a sole maize crop. As part of the
planned rotation a cowpea crop (cowpea phase 2) was planted in the 2006/2007 season on
the field which was previously under maize phase 2. In 2006/07 season a sorghum phase
was established in the field that was previously under cowpea phase 1. Maize phase 3
was established on a new field that was adjacent to maize phase 1 and planted to a sole
maize crop. In the 2007/08 season maize phase 4 was established on a new field and also
on the three fields that were previously under maize phase 3, cowpea phase 2 and
sorghum phase the previous year. The experiment was set up using a split plot design
with three replications. The main plot factor was tillage method (planting basins, ripper
and conventional plough) with mulch cover (0, 0.5, 1, 2, 4, 8 and 10 tha-I) as a subplot
factor. Plots were pegged out in October of each maize phase, and then maintained in
subsequent seasons.
Tillage was established on 63 m x 6 m strips and seven mulch levels were randomly
allocated to sub-plots measuring 8 m x 6 m on each tillage treatment. Each plot was
240
separated by a 1 m pathway to avoid movement of residue from one plot to the next when
tillage was undertaken. All tillage operations were carried out after applying mulching
material to each sub-plot. Planting basins were dug at 0.6 m x 0.9 m spacing using a hand
hoe and each basin measured 0.15 m (length) x 0.15 m (width) x 0.15 m (depth),
disturbing less than the equivalent of 1000 m2 ha-I. Rip lines were opened at 0.9 m inter-
row spacing using a ripper tine attached to the beam of a donkey-drawn mouldboard
plough. The ripping depth achieved varied between 0.15 and 0.18 m, with a surface width
disturbance of 0.12 m, disturbing less than the equivalent of 1 440 m2 ha-I. Conventional
ploughing was done soon after the first effective rain in December each year using a
donkey-drawn VS 100 mouldboard plough. Planting furrows were then opened at the
appropriate inter-row spacing for the crop of 0.6 m. An in-row spacing of 0.2 m between
seeds was used in conventional and ripper systems.
The cowpea variety used was a semi-determinate 86D 719 which matures in 120 days
(Ncube, 2007). The sorghum Macia variety was used as it is recommended for the semi-
arid environments of southern Zimbabwe. In 2005/06 the cowpea phase 1 of the rotation
was planted on 16 December 2005. In 2006/07 season, planting in the basin and ripper
systems was done on 22 November 2006 for, the cowpea phase 2. Planting of cowpea in .
the conventional system was done on 8 December 2006 because the soil was too wet to
plough. Planting in the sorghum phase, was done on 14 December 2006. In the cowpea
and sorghum phases, five seeds were planted per basin while two seeds were planted per
station in the conventional and ripper systems. For the cowpea phase of the rotation,
plants were thinned to four per basin in planting basin and one plant per station in the
241
ripping and conventional tillage treatments two weeks after emergence. For the sorghum
phase of the rotation, plants were thinned to two per basin in the planting basin system
and one per station in the ripper and conventional tillage treatments two weeks after
emergence. Weeds were controlled by hand hoe as required in all seasons.
In the 2005/06 season gravimetric soil water was determined by collecting soil samples
fortnightly between planting basins, along riplines and in rows for conventional tillage
treatments. Soil samples were collected at 0.15 m depth interval up to a maximum depth
of 0.60 m. The soils were weighed before oven drying them at 105°C for 48 hours for
determining gravimetric water content. Gravimetric water content for each soil layer was
calculated using the procedure outlined by Anderson and Ingram (1993). Gravimetric
water content was converted to volumetric water content for each respective depth using
the measured bulk density for each soil layer. Soil water content in millimetres was
determined by multiplying volumetric water content by thickness of each layer from
which soil water was measured. In the 2006/07 growing season soil water was measured
weekly at 0.1 m depth interval using a capacitance probe (microgopher type).
9b.3.2 Data collection and analysis
Daily rainfall was measured by a standard raingauge throughout each season. At harvest
grain yield was estimated from a net plot consisting of the five middle rows of 6 m
length. Grain samples were dried at 60°C for 48 hours for moisture adjustment. Grain
weight was converted to a per hectare basis at 12.5% moisture content. All data were
analyzed by ANOVA using Genstat Discovery Edition 3 (www.vsni.co.uk)with split plot
242
design for cowpea and sorghum yield data, and unbalanced design for soil water data.
Then the least significant difference (Lsd) at 5 % significance level was used to compare
tillage treatment means.
9b.4 Results and Discussion
9b.4.1 2005/06 season
9b.4.1.1 Seasonal rainfall and soil water changes
The total seasonal rainfall for 2005/06 growing season was 832 mm, with 298 mm of this
having fallen before the cowpea phase 1 was planted on 16 December 2005 (day 77). The
2005/06 season's rainfall was 45 % more than the 69 year average rainfall for Matopos
(Chapter 4). Rainfall events, totalling 294 mm observed between 1 December (day 62)
and 12 December (day 73) filled up the 0 - 0.30 m soil profile of all tillage systems (Fig.
9b.1).
100
2005/06
80
,-...
~
'-' 60
~
.S 40
t':I
~
20
0 .nL '" L
0 25 50 75 100 125 150 175 200 225
Time(days after 1 October 2005)
Figure 9b.l. Daily rainfall distribution at Matopos site during the 2005/06 growing
season
243
Drying of the 0 - 0.30 m soil profile started after 28 December 2005 (day 89) until 8
February 2006 (day 131) despite a further 256 mm of rainfall falling during this period.
Soil water was being extracted by cowpea roots as the crop was in its vegetative and
flowering stages between these dates. Soil water contents increased steadily in response
to rainfall received between 8 February (day 131) and 6 March (day 157). On 6 March
2006, the 4 tha·1 mulch treatment had on average 17,28, and 30 mm more soil water than
10, 2 and 0 t ha-I mulch treatments under the conventional ploughing tillage system (Fig.
9b.2) following a 50 mm rainfall event recorded on 3 March (day 154). On the same day
the ripper system had 16, 24 and 16 mm more soil water in the 4 tha-I mulch treatment
than 0, 2 and 10 t ha-I treatments respectively. Under the planting basin system the 0, 2
and 4 tha-I mulch treatments had similar soil water content but the lOt ha-I mulch
treatment had 22 % less water than the other three mulch treatments on 6 March 2006
(day 151) (Fig. 9b.2).
244
160 ---0 -x-2 - -0- -4 - ...... 10
120 Plough
100
80
~
t..::: 60
2
A.. 40
20+----.----.----,----.----,,---~----,----,--~
o 25 50 75 100 125 150 175 200 225
Time(days after 1 October 2005)
160 1---0 -:.:-2 - -0- - 4 -Ii!)- 10 I
120
100
80
60
40
20 +-----,----,----,-----,----,----,-----,----,-----
o 25 50 75 100 125 150 175 200 225
Time(days after 1 October 2005)
160 1--0 -x-2 - -0- - 4 -Iiil- 10I
M~ Basins
100
80
Q) ~ Ó~····~
tea 60
I~V..m __~
~ 40 ± I
20 +----,----.-----,----,----~---.----.-----,---~
o 25 50 75 100 125 150 175 200 225
Time(days after 1 October 2005)
Figure 9b.2. Soil water changes in 0 - 0.30 m profile in the cowpea field under
conventional ploughing tillage system and four mulch treatments (0, 2,4 and 10 t ha-I) on
a clay soil during 2005/06 growing season. Error bars represent standard error of means
245
9b.4.2 2006/07 season
9b.4.2.1 Seasonal rainfall and soil water regimes
The total seasonal rainfall for 2006/07 growing season was 465 mm of which 104 mm
was accumulated between l3 November (day 44) and 22 November (day 53) (Fig. 9b.3).
The 2006/07 total season's rainfall was 23 % less than the 69 year average rainfall for
Matopos (Chapter 4). The month of January 2007 received very low rainfall of only 12
mm recorded on 17 January (day 110). February and March 2007 received 80 and 77 mm
of rain with the 2006/07 season ending on 4 April 2007 (see Table 9a.6 in Chapter 9a).
The cowpea crop experienced soil water stress during the 14-16 day dry spells
experienced in January 2007 (Fig. 9b.3) and went on to give similar yields to those
observed by Ncube et al. (2007) in 2004/05 season on a sandy soil. This can be attributed
to the ability of cowpea to extract soil water from soils with low water content. Cowpea
has a uniform rooting system that increases its surface area and enables the plant to
extract soil water from a mieroseale even in drying soils (Hall, 2004).
100
80 2006/07
60
40
Time (days after 1 October 2006)
Figure 9b.3. Daily rainfall distribution at Matopos site during the 2006/07 growing
season
246
Profile soil water content to a depth of 0.55 m responded to rainfall events (Fig. 9b.4) and
the pattern of soil water changes was quite similar in the three tillage systems and four
mulch treatments. On 15 November (day 46) profile water content was higher (P < 0.001)
under 10 tha-I mulch treatment than 0, 2 and 4 tha-I in the conventional tillage system
(Fig. 9b.4). The 10 tha-I mulch cover probably captured more rainwater from the 18 mm
rainfall event observed on 13 November 2006 (day 44). A total of 50.5 mm rainfall was
received between 15 November (day 46) and 8 December (day 69). Profile soil water
measured on 8 December 2006 .showed that under the conventional tillage system there
was an average of 105 mm of water under the 0 t ha-I mulch treatment, 90 mm under the
2 t ha-\ 127 mm under the 4 t ha-I and 116 mm under the 10 t ha-I. In the basin system,
profile water content on 8 December was 149, 169, 185 and 153 mm under 0, 2, 4 and 10
tha-I mulch treatments which are all much higher than the conventional plough. The
ripper system had similar profile water content under 0, 2, 4 and 10 tha-I mulch
treatments on 8 December 2006. Profile water content recorded on 12 February 2007
(day 134) showed that under the conventional system there was 44 mm at 0 tha" mulch
cover, 36 mm at 2 tha', 48 mm at 4 tha" and 51 mm being recorded at 10 tha-1• In the
ripper tillage system profile water content on 12 February was 46, 49, 52 and 52 mm
under 0, 2, 4 and 10 tha-I mulch treatments which were marginally higher than the
conventional system. Profile water content under the planting basin system was 44 mm at
o tha" mulch cover, 46 mm at 2 tha', 54 mm at 4 tha-I and 56 mm at 10 tha-I mulch
treatments.
247
200 -0 -x-2 - -0- -4 -- 10
Plough
o +-----,-----,-----,-----,-----,-----,-----,-----.
o 25 50 75 100 125 150 175 200
Time (days after 1 October 2006)
200 -0 -x-2 - -0- -4 -El- 10
lj Ripper
I=J' "0 .8»,g,o:: .£:8
... :s:
o +-----~----~----~----~----~------~----~--~
o 25 50 75 100 125 150 175 200
Time (days after 1 October 2006)
200 -0 -x-2 - -0- -4 -~ 10l
• x Basins
"i5~ -: Q: Ji. r"-.~. -I
~x ~lK~~ ~x . ... ~,'"
~ ,
Cl)
~ 50 ~......
0...
o +-----.-----,---~.-----,-----,------~----,-----~
o 25 50 75 100 125 150 175 200
Time (days after 1 October 2006)
Figure 9b.4. Profile soil water changes in 0 - 0.55 m profile in the cowpea field under
conventional, ripper and basin tillage systems and four mulch treatments (0, 2, 4 and lOt
ha") on a clay soil during 2006/07 growing season. Error bars represent standard error of
means
248
9b.4.2.2 Water changes in different soil layers under three tillage systems (cowpeafield)
Soil water content in the different layers varied significantly (P < 0.001) especially
between November 2006 and February 2007 (Figs. 9b.5 and 9b.6). In each tillage system,
the top layer (0-0.15 m) responded to rainfall events more strongly than deeper layers as
expected. The top layer also lost soil water faster by soil surface evaporation during
drying cycles. The 0.15-0.25 m layer was the driest throughout the season indicating that
there was good cowpea root distribution within this layer to extract the water. Most
cowpea roots are reportedly concentrated in the top 0.5 m soil layer and root density
decreases progressively with soil depth (Zaongo et al., 1997; Moroke et al., 2005). Soil
water could be saved in the top soil layers for the cowpea crop by applying surface mulch
cover.
60
1--0.-0-15 __ 15-25 - -<>- -25-35 --t- 35-451
50
Plough
o +------,------,------,------,------,-----,,-----,------,
o 25 50 75 100 125 150 175 200
Time (days after 1 October 2006)
Figure 9b.5. Average changes in equivalent soil water depth in different layers of a clay
soil under conventional ploughing tillage system in the cowpea field at Matopos Research
Station during 2006/07 growing season. Error bars indicate standard error of the means
249
60 --- 0-15 _ 15-25 - -.- - 25-35 --I- 35-451
50
I, 40 Ripper30
~ 20
10
0
0 25 50 75 100 125 150 175 200
Time (days after 1 October 2006)
60 ~ 0-15 15-25 - -0- - 25-35 --I- 35-45
50
ê 40
'-'
i 30
:::l
~ 20
10
0
0 25 50 75 100 125 150 175 200
Time (days after 1October 2006)
Figure 9b.6. Average changes in equivalent soil water depth in different layers of a clay
soil under ripper and basin tillage systems in the cowpea field at Matopos Research
Station during 2006/07 growing season. Error bars indicate standard error of the means
9b.4.2.3 Soil water regimes in the sorghumfield (sorghumphase)
The soil water measurements in the sorghum field on 15 November 2006 were the
highest throughout the growing season across all treatments which is a month before
planting (Figs. 9b.7). There were little differences in the total profile water content
between mulch treatments under the same tillage system and across tillage systems
throughout the whole season, with values remaining around 100 mm. The soil water
250
patterns in a clay soil under cowpea and sorghum crops were quite similar in a season
with below average rainfall. This is consistent with results from Ncube et al. (2007) who
observed similar soil water dynamics (0 - 0.25 m soil layer) in sorghum and grain legume
fields on a sandy soil. On 15 November (day 45) ripper and basin tillage systems had a
greater (P < 0.001) depth of equivalent soil water in the 0 - 0.65 m profile than
conventional tillage system. Conventional ploughing had not yet been done by 15
November 2006 while ripped furrows and planting basins had been prepared by 15
October 2006. The basin and ripper tillage systems collected rainwater from the rainfall
events recorded between 1 October and 14November 2006 amounting to 34 mm. The 10
tha-I mulch treatment had 39, 40 and 38 mm of soil water more than 0, 2 and 4 t ha-I
treatments in the conventional system. On the same day in the ripper system, the 10 tha-I
I
mulch treatment had 23, 30 and 25 mm more soil water than 0, 2 and 4 tha-I mulch
treatments. Under basin system, the 10 tha-I mulch treatment had 17,9 and 21 mm more
soil water than 0, 2 and 4 tha" treatments. This can be due to the effect of 10 tha-I mulch
cover being able to protect the soil from solar radiation and reduce soil surface
evaporation. Hatfield et al. (2001) reported that mulching can reduce soil evaporation by
34 to 40 %.
251
250
1-0 -)1(-2 - -0- -4 -- lol
I 200
J 150
11)
~~ 100
50 +------.-----,------,-----~----_,------._----,_----_,
o 25 50 75 100 125 150 175 200
Time (daysafter1 October 2006)
250
I, 1-0 -*-2 - -0- -4 -19- lol200150
~ 100
50 +------r-----,------,-----~----~------r_----._----~
o 25 50 75 100 125 150 175 200
Time (daysafter1 October 2006)
250
I 1-0 -)1(-2 - -o~ -4 -Il)- lol200
J 150
11)
~ 100
50 +------r-----,------,-----~----~------._----._----~
o 25 50 75 100 125 150 175 200
Time (daysafter1 October 2006)
Figure 9b.7. Soil water changes (0 - 0.65 m profile) in response to conventional, ripper
and basin tillage systems and four mulching treatments (0, 2, 4 and 10 tha") in the
sorghum field during 2006/07 growing season at Matopos Research Station. Error bars
represent standard error of means
252
9b.4.2.4 Water changes in different soil layers under three til/age systems (sorghum
phase)
The 0 - 0.15 m soil layer responded to rainfall events more than the deeper layers of the
profile (Fig. 9b.8). The same layer also lost soil water faster than the other layers which
can be seen by the steeper slope in Figure 9b.8. The 0.15 - 0.25 m layer had the lowest
soil water content throughout the growing season in each tillage system in the sorghum
field suggesting more extraction of soil water by the sorghum roots from this layer. The
results from a study by Sow et al. (1997) indicated that most of the roots of sorghum
plants are concentrated in the 0.1 - 0.25 m soil layer. At the beginning of soil water
measurements, the top soil layer had 36, 41 and 43 mm in the conventional, ripper and
basin tillage systems. The 0.15 - 0.25, 0.25 - 0.35 and 0.35 - 0.45 m soil layers had
similar water content at the first day of measurements (day 45) long before planting. The
0- 0.15 m soil layer was recharged in all tillage systems between days 136 (13 February)
and 162 (Il March) following 59 mm of rain received. The same soil layer was further
recharged in all tillage systems between days 168 (17 March) and 184 (2 April) following
86 mm accumulated between these two dates. Under the three tillage systems the 0.15 -
0.25 m soil layer was also recharged between days 168 (17 March) and 184 (2 April).
253
50
EO-IS --0--15-25 - -.- - 25-35 --I- 35-451
r 40 Plough'-" 30
i 20
~
10
o +-----_,-------.-----~----~------._----_.----_,r_--__,
o 25 50 75 100 125 150 175 200
Time (days after 1 October 2006)
50 1-0--0-15 --0-- 15-25 - -0- - 25-35 --I- 35-451
40
§ Ripper
._, 30
, 20
~
10
o +------,------.-----~----~------._----_,------,_--__,
o 25 50 75 100 125 150 175 200
Time (days after 1 October 2006)
50
1-- 0-15 -- 15-25 - -0- - 25-35 --i- 35-451
40
Basins
10
o +------,------.------.-----,------~-----,------,-----~
o 25 50 75 100 125 150 175 200
Time (days after 1 October 2006)
Figure 9b.8. Average changes in equivalent soil water depth in different layers of a clay
soil under three tillage systems (conventional, ripper and basin) in sorghum field at
Matopos Research Station during 2006/07 growing season. Error bars indicate standard
error of the means
(
254
9b.4.3 Crop performance
9b. 4.3.1 Cowpea and sorghum plant stands
In the 2005/06 season planting basins had a higher (P = 0.035) plant stand than the ripper
and conventional ploughing tillage treatments (Table 9b.l). In the drier 2006/07 season,
there was better cowpea establishment in the ripper tillage system compared to the basin
and conventional systems. There was rodent attack before cowpea crop emergence
particularly in the basin system. The rodent damage on cereal crops grown in planting
basins has also been observed by some smallholder farmers praeticing conservation
farming in semi-arid districts of Zimbabwe. Some smallholder farmers have devised
rodent traps in an effort to curb the challenge (Plate 9b.l). Although the level of
mulching had no significant influence on cowpea crop establishment, over the two
seasons there was a trend of decreasing plant stand as the mulch cover rate was increased
from 4 to 8 and 10 tha" (Table 9b.l). Soil water content was adequate for good cowpea
crop establishment in both 2005/06 and 2006/07 seasons. Sorghum crop stand was lower
(P = 0.078) under the basin system than the conventional ploughing and ripper systems
(Table 9b.2). There was serious attack by rodents in the basin system at crop
establishment stage compared to the conventional ploughing and ripper systems.
255
Table 9b.l. Effect of tillage method and mulch cover on cowpea plant stand two weeks
after planting (plants m") on a red clay soil at Matopos Research Station
Treatment 2005/06 2006/07 Across seasons
Tillage method
Plough 4.5 2.1 3.3
Ripper 4.5 4.0 4.2
Planting basins 6.3 2.9 4.6
Lsdo.05 1.3 1.5 0.57
Mulch (tha-1)
0 5.l 3.4 4.3
0.5 5.4 3.3 4.4
1 5.4 3.2 4.3
2 4.8 3.2 4.0
4 5.3 3.2 4.3
8 4.7 2.5 3.6
10 5.0 2.4 3.7
Lsdo.05 0.81 1.1 0.70
CV(%) 17 37 28
Plate 9b.l. AIO litre bucket half filled with water being used as a rodent trap in a
groundnut field in Gokwe district of Zimbabwe
256
Table 9b.2. Effect of tillage system and mulch treatment on sorghum stand two weeks
after planting (plants m") on a red clay soil at Matopos Research Station
Treatment 2006/07
Tillage Plough 2.1
i_~"J(ILRipper 2.0Basin 1.0Lsdo.o5~_. 0.43
Mulch (t ha-I) 0 1.6
0.5 1.6
1 1.7
2 1.7
4 1.7
8 1.9
10 1.6
Lsdo.o5 0.20
CV (%) 12
9b.4.3.2 Cowpea and sorghum yields
During the relatively wet 2005/06 growing season (832 mm) the tillage and mulch
treatments had no significant (P > 0.05) influence on grain production from the cowpea
phase 1. However, in the 2006/07 season which had below-average rainfall (465 mm),
there was significantly (P = 0.003) lower grain yield for the cowpea phase 2 in the
conventional system, compared to yields obtained from the ripper and planting basin
tillage treatments (Table 9b.3). This lower yield is attributed to the fact that the
conventional tillage treatment for the cowpea phase 2 was planted 16 days after planting
in the ripper and basin systems because the clay soil at Matopos was too wet to plough.
In contrast to the 2005/2006 cowpea phase 1 yields, in the dry season 2006/07 the
cowpea phase 2 grain yields were significantly (P = 0.021) affected by mulch cover
(Table 9b.3). The highest yields were recorded for the 4 tha" mulch cover (Table 9b.3),
which gave an average 228 kg ha-I more grain than the 8 tha-I mulch treatment, and 221
kg ha-I than for the 10 t ha-I. The lower cowpea yields obtained at 8 and 10 tha-I mulch
257
cover can be attributed to inferior plant stand recorded under these two mulch treatments
(Table 9b.I). The two-way interaction of tillage system and mulch cover had no
significant (P > 0.05) influence on cowpea grain production during either 2005/06 or
2006/07 growing seasons.
Table 9b.3. Cowpea grain yield responses (kgha') averaged across three tillage systems
and seven mulch levels at Matopos Research Station during 2005/06 and 2006/07
growing seasons
Treatment 2005/06 2006/07 Across seasons
'fillage method
Plough
Ripper
Planting basins
Mulch (tha )
o 1 732 403 1067
0.5 1 564 423 993
1 1 875 440 1 157
2 1474 564 1 019
4 1 800 616 1 208
8 1 686 388 1037
10 1479 395 937
Lsdo.05 639 154 319
CV(%) 40 35 48
Across the season comparison revealed that tillage and mulch cover had no significant
influence on cowpea yield (Table 9b.3). However, the season had a significant (P <
0.001) effect on cowpea yields achieved on the clay soil. Cowpea yield differences
between 2005/06 and 2006/07 season can be attributed to differences in rainfall pattern
observed at Matopos during the two seasons (Figs. 9b.l and 9b.3). The two long dry
spells recorded in January 2007 negatively impacted on cowpea yields achieved from
each tillage system. Delayed planting in the conventional system in 2006/07 season also
258
resulted in lower cowpea yields in the system compared to the ripper and basin systems
as the leaf area had not developed full canopy.
The tillage system had a significant influence on sorghum grain (P = 0.014) and stover (P
= 0.018) production (Table 9b.4). The lowest yields were observed under the planting
basin system, with the ripped plots yielding 600 kg ha", and conventionally tilled plots
yielding 416 kg ha-1 more grain. The rodent attack in the basin system of the sorghum
phase was more severe than in the cowpea phase 2. Ripper and conventional tillage
systems also produced 1 266 and 1 186 kgha' more stover than planting basins. The
measured plant stand in planting basin was 43 and 45 % of plant stands achieved in the
conventional and ripper tillage systems and attributed to the lower yields produced.
Therefore, as the results were influenced by an external factor of rodents, they cannot be
used to assess the treatment differences. This experiment would need to be repeated
before such results can be used in practice. Mulching had no significant effect on
sorghum grain production, harvest index or plant stand. However, sorghum stover yield
increased (P = 0.032) with increase in mulch cover. The highest sorghum stover produced
was 4 481 kgha' and was achieved at 8 tha-1 mulch treatment. The two-way tillage
system and mulch cover interaction had no significant (P = 0.626) influence on any of the
measured parameters of the sorghum phase crop during 2006/07 growing season.
259
Table 9b.4. Sorghum yield responses (kgha') averaged across three tillage systems and
seven mulch treatments at Matopos Research Station during 2006/07 growing season
Treatment Grain yield Stover yield Harvest index
(kgha-1) (kgha-I)
Tillage
Plough 2 005 4 207 0.32
Ripper 2 189 4 287 0.34
Basins 1 589 3 021 0.32
Lsdo.05 311 774 0.055
~~~~~~$~,~~~~~~~~~J~"~~~"~W:%~.,~ïil~,"{,,~~~~\~!i!~c~l.~fjï)i;~~~ï;ff'i!ll1l!1!i!~!Fj;.~~§~:\~~~~.B~--~
Mu/ch
o 1 756 3477 0.34
0.5 1 764 3416 0.34
1 2046 3684 0.32
2 1 844 3877 0.32
4 1976 3794 0.34
8 2050 4481 0.31
10 2059 4140 0.33
Lsdo.05 376 663 0.044
CV(%) 20 18 14
9b.5 Conclusion
The in-season rainfall distribution had a strong influence on cowpea and sorghum yields
achieved during 2005/06 and 2006/07 seasons. Dry spells lasting 14 to 16 days negatively
impacted on cowpea and sorghum yields as these dry spells coincided with the vegetative
and reproductive stages of the crops. Delayed planting of cowpea due to unfavourable
soil conditions resulted in lower cowpea yields in the conventional system compared to
the basin and ripper systems due to lower leaf area at peak radiation. Planting basin and
ripper systems offered an opportunity for timely planting in both seasons resulting in
higher cowpea yields particularly in the 2006/07 season which had below average
rainfall. Soil water dynamics in the conventional, ripper and basin tillage systems under
cowpea and sorghum crops were similar in a clay soil regardless of the rainfall pattern.
260
Cowpea and sorghum crops both extract soil water predominantly from the 0.15 - 0.25 m
depth on a clay soil.
The fmdings from our study indicate that the wider spacing of the permanent planting
positions in the basin and ripper systems does not compromise the yields of cowpea and
sorghum crops. This result suggests that grain legumes and small grains can be grown
successfully under semi-arid conditions of Zimbabwe using the planting basin and ripper
tillage techniques. The basin and ripper systems can give even higher cowpea and
sorghum yields than a well managed conventional ploughing system in years with below
and above average rainfall pattern as observed in 2005/06 and 2006/07 seasons.
However, rodent attack in the basin system poses a big challenge for successful cropping
especially in seasons with below average rainfall. In drought years mulching improves
cowpea yields with 2 and 4 tha-1 giving similar yields. Higher mulch levels of 8 and 10
tha', which are unachievable under the current smallholder conditions, suppress cowpea
crop establishment, and subsequently lower yields, so should not be recommended
anyway. Smallholder farmers in semi-arid southern can target 2 tha' mulch cover if crop
residues are available.
261
CHAPTER 9c
Cumulative Effects of Minimum Tillage, Mulching and Rotation on
Selected SoillProperties and Maize Yield on a Clayey Soil
in Semi-Arid Southern Zimbabwe
9c.l Introduction
Minimum tillage, mulching and crop rotation change soil conditions through
modification of physical, chemical and biological properties (Arshad et al., 1999; Dexter,
2004). Positive changes normally occur in soil organic carbon content (HaIvorson et al.,
2002), soil physical properties such as bulk density and porosity (Arshad et al., 1999),
and microbial biomass and activities (Salinas-Garcia et al. 1997). Mulching combined
with minimum tillage is effective in reducing surface runoff, maintaining soil structure,
conserving soil water and adding organic matter to the soil (Liebig et al., 2004; Glab and
Kulig, 2008).
The widespread promotion of conservation agriculture (CA) practices has shown
significant cereal yield increases on smallholder farms in semi-arid districts of Zimbabwe
(Nyagumbo, 2007; Twomlowel al., 2008a). While information on crop yield benefits of
conservation agriculture practices continues to be generated, there is a need to assess the
other benefits a farmer could derive from adopting CA practices. This is particularly
crucial for smallholder farms where farming takes place on fragile and highly degraded
soils. Conservation agriculture practices, it is hypothesized, will improve soil fertility,
infiltration of rainwater, soil water storage and prolong the period of soil water
262
availability to crops (Nyagumbo, 2007). This will in turn improve crop productivity from
both the soils and limited rainfall received during cropping period.
9c.2 Objectives
The main objective of this study was to assess improvements in soil properties following
different periods of praeticing conservation agriculture on a clay soil. The specific
objectives were to determine the cumulative effect of minimum tillage, mulching, crop
rotation and period of exposure to CA practices on the following:
(1) soil organic carbon content;
(2) soil bulk density;
(3) infiltration; and
(4) maize yield.
9c.3 Materials and Methods
9c.3.1 Experimental design and layout
The experiment was set up on four fields from the 2004/05 till the 2007/08 growing
season as outlined in Chapter 3. The same fields and some of the plots used for chapters
9a and 9b were used for assessing changes in soil properties. Crops grown in each field
and their sequence are outlined in Table 9c.l.
263
Table 9c.1. Experimental fields used and crops grown in each field from 2004/05 to
2007/08 seasons at Matopos Research Station
Field Period! Season(s) Crop Tillage Mulch rate
number under CA field was sequence system used (tha-t)
(#seasons) used all years
1 2007/08 Maize Plough, 0,4and 10
(M) ripper,
basins
•__ II_~.dE&_
2 2 2006/07 and Maize - Plough, 0, 4 and 10
2007/08 maize ripper,
(MM) basins
~~~._IIIIt4!.iI
3 3 2005/06, Maize - Plough, 0, 4 and 10
2006/07 and cowpea - ripper,
2007/08 maize basins
(MCM)
~1i:~:'m~\~fi,~W~~~~~~~~~~~~~1~1~I4~~t4_11~q~~~,,__,~*,i~ill~~:·t;·t:;~:~,~1%t,~;'-··'~a~~~
4 4 2004/05, Maize - Plough, 0, 4 and 10
2005/06, cowpea - npper,
2006/07 and sorghum - basins
2007/08 maize
(MCSM)
9c.3.2 Soil sampling and measurements taken
In planting basin system measurements were made within the basin and in the <'"space
between two basins in the same row. In the ripper system measurements were made
within the rip line and at the inter-row space. In the conventional ploughing system
measurements were made along the planting row. In this chapter the following
abbreviations are used to denote the tillage treatments; CP (plough) - conventional
ploughing, BRL - between riplines, RL - within riplines, BB - between planting basins
and WB - within planting basin. Soil samples for organic carbon and bulk density
analysis were collected from 24 to 28 March 2008. Infiltration measurements in all four
fields were done from 31 March to 4 April 2008.
264
9c.3.2.1 Organic carbon
In each plot, soil samples were collected from 0 - 0.10 m and 0.10 - 0.20 m depths using
steel cores measuring 0.03 m internal diameter by 0.95 m length. Three sampling
positions were used in each plot for each soil depth. A composite soil sample for each
soil depth was made by mixing soil samples collected from the three sampling positions.
A sub-sample from each treatment was packed into khaki pockets and taken to the
laboratory for air drying. After air drying the soil was sieved through a 2 mm sieve before
doing chemical analysis for organic carbon using the Sommer's method (Anderson and
Ingram, 1993).
9c.3.2.2 Bulk density
Steel cores measuring 0.05 m diameter and 0.05 m height were used to collect soil
samples from each tillage x mulch treatment combination in all four fields. Undisturbed
soil samples were collected in duplicates from each plot at 0.05 m depth interval up to
0.30 m depth. The samples were oven dried at I05°e for 48 hours before measuring
weight of the dried samples and calculating bulk density at each soil depth according to
the procedure outlined by Anderson and Ingram (1993).
9c.3.2.3 Infiltration, hydraulic conductivity and sorptivity
Infiltration measurements were conducted using a minidisk infiltrometer produced by
Decagon Devices (2007) (Plate 9c.l). The minidisk infiltrometer consists of three major
components namely sintered steel disc of 0.022 m radius at the base of infiltrorneter, a 95
mm water reservoir and a bubble chamber. Infiltration runs were conducted in duplicate
265
for each tillage x mulch treatment combination for all four experimental fields. Particular
care was taken during the preparation of the soil surface before each infiltration run. A
relatively flat surface was identified in each tillage x mulch treatment combination and all
residues of weeds and maize were removed from the identified soil surface. The micro-
topography of the surface was leveled with a spatula in order to ensure adequate
horizontal contact between soil surface and the sintered steel base of the infiltrometer.
Infiltration measurements were made at a single supply head of -20 mm as recommended
by the manufacturer for clay soils (Decagon Devices, 2007) and it took 10 to 15 minutes
to reach a steady state flow from the infiltrometer. Changes in water volume in the water
reservoir of the infiltrometer were recorded at 30 s interval.
Plate 9c.l. Minidisk infiltrometer used for infiltration measurements
266
Infiltration results were fitted to the following equations in order to calculate hydraulic
conductivity and capillary sorptivity. The procedure followed in our calculations of
hydraulic conductivity and sorptivity was proposed by Zhang (1997) in Decagon Devices
(2007).
1= Cl + C2. ...Jt Equation 9c. J
where I is cumulative infiltration (mm), Cl (ms') is related to hydraulic conductivity, C2
(ms-O·s)is related to soil sorptivity, and t is time taken for each infiltration run. The
hydraulic conductivity (K) of the soil is then computed from
Equation 9c.2
where Cl is slope of the curve of cumulative infiltration vs. square root of time, and A is a
value relating the van Genuchten parameters for a given soil type to the suction rate and
radius of the infiltrometer disk. The A is computed from
A = [(1J.65(n"o.1 - J))exp.[2.92(n-J)áhol}/(áro)"O·91 Equation 9c.3
where n and á are the van Genuchten parameters for the soil, ro is disk radius, ho is the
suction at the disk surface. Values of A (van Genuchten parameter for a 2.2 cm radius
minidisk infiltrometer) computed for the minidisk infiltrometer are given in Table 9c.2.
C2 which gives the soil's sorptivity is computed from
Equation 9c.4
9c.3.2.4 Soil water content before and after infiltration
Soil water content at 0 - 0.10 and 0.10 - 0.20 m depths was measured in each plot using a
capacitance probe before infiltration runs were conducted. Soon after each infiltration run
a soil sample was taken from 0 - 0.10 m depth on the position where the base of the .
267
infiltrometer was sitting. The soil samples were weighed and then oven dried at 105°C for
48 hours and gravimetric water content was determined using a procedure outlined by
Anderson and Ingram (1993). Gravimetric water content was converted to volumetric
water content using measured bulk density for each soil layer.
Table 9c.2. van Genuchten parameters for 12 textural classes and A values for 2.2 cm
disk radius and suction values from 0.5 to 6.0 cm (Adapted from Carsel and Parrish,
1988)
ho
-0.5 -1.0 -2.0 -3.0 -4.0 -5.0 -6.0
Texture á n A
Sand 0.145 2.68 2.9 2.5 1.8 1.3 0.9 0.7 0.5
Loamy sand 0.124 2.28 3.0 2.8 2.5 2.2 1.9 1.6 1.4
Sandy loam 0.075 1.89 4.0 4.0 4.0 4.0 4.0 4.1 4.1
Loam 0.036 1.56 5.6 5.8 6.4 7.0 7.7 8.4 9.2
Silt 0.016 1.37 8.1 8.3 8.9 9.5 10.1 10.8 11.5
Silt loam 0.020 1.41 7.2 7.5 8.1 8.7 9.4 10.1 10.9
Sandy clay loam 0.059 1.48 3.3 3.6 4.3 5.2 6.3 7.6 9.1
Clay loam 0.019 1.31 6.0 6.2 6.8 7.4 8.0 8.7 9.5
Silt clay loam 0.010 1.23 8.1 8.3 8.7 9.1 9.6 10.1 10.6
Sandy clay 0.027 1.23 3.4 3.6 4.2 4.8 5.5 6.3 7.2
Silty clay 0.005 1.09 6.2 6.3 6.5 6.7 6.9 7.1 7.3
Clay 0.008 1.09 4.1 4.2 4.4 4.6 4.8 5.1 5.3
Where n and á are the van Genuchten parameters for a given sOiltexture, ho IS the suction at the
disk surface.
9c.3.2.5 Maize crop performance
Plant counts were taken two weeks after planting. At harvest grain and stover (above-
ground biomass minus grain) yields were estimated from a net plot consisting of five
middle rows with a running length of 6 m. The weight of cobs and stover from the netplot
of each treatment were determined in the field before taking sub-samples for moisture
correction. Grain and stover samples were dried at 60°C for 48 hours for moisture
adjustment. The maize shelling percentage was 'determined for each treatment to convert
268
cob weight into grain and core weights. Grain weight was converted to a per hectare basis
at 12.5 % moisture content as final grain yield following procedure outlined in Chapter 3.
9c.3.3 Statistical analysis
All statistical analysis was performed using Residual Minimum Likelihood Method
(REML). The REML method has the capability of analyzing data with an unbalanced
design. The tillage treatment in our study was unbalanced because ripper and basin
systems had two sampling positions each while plough had only one level. Tillage, mulch
and field and their interactions formed the fixed model while each replicate for the
measurements was applied as the random model. Soil water data collected before and
after infiltration were analyzed using the General Analysis model of ANOV A. Genstat
Discovery Edition 3 (www.vsni.co.uk) was used for statistical analysis. Then the least
significant difference (Lsd) at 5 % significance level was used to compare treatment
means.
9c.4llResults and Discussion
9c.4.1 Soil organic carbon
Soil organic carbon content in the 0 - 0.2 m depth .measured in each field at the end of
2007/08 growing season is given in Table 9c.3. There was a significant (P = 0.042)
increase in soil organic carbon content in CP and WB with increase in period of exposure
to minimum tillage, mulching and rotation (Table 9c.3). The maize-cowpea-maize (field
3) and maize-cowpea-sorghum-maize (field 4) fields had higher (P = 0.042) organic
carbon content than the maize (fields 1) and maize-maize (field 2) fields in CP, as well as
269
between and within the planting basin. The lowest organic carbon content was 0.67 %
observed in CP tillage system at 0 tha-1 mulch level following only one year of CA
treatments maize field (field 1). The maize-cowpea-maize (field 3) and maize-cowpea-
sorghum-maize (field 4) fields benefited from cowpea stover that was left on the surface
after harvesting at the end of 2005/06 and 2006/07 growing seasons (Chapter 9b). In
fields 1 and 2 only manure, mulching material and weeds contributed towards organic
carbon during 2006/07 and 2007/08 seasons.
Table 9c.3. Effect of tillage and reriod of exposure to CA practices averaged across three
mulch levels (0, 4 and 10 tha" ) on soil organic carbon content (%) of a clay soil at
Matopos Research Station
Tillage & Field number and crop sequence
location 1(M) 2 (MM) 3 (MCM) 4 (MCSM)
CP 0.67 0.71 0.90 1.00
BRL 0.97 0.92 0.91 0.90
RL 0.98 0.90 0.92 0.95
BB 0.76 0.80 0.90 0.83
WB 0.93 0.90 1.13 1.06
Lsdo.os= 0.19; CV = 21 %
There was a higher (P < 0.05) organic carbon content in the 0- 0.10 m layer than 0.10 -
0.20 m under the ripper and planting basin systems regardless of sampling position (Fig.
9c.I). In the top 0; IOm layer, there was more organic carbon within planting basin than
all other tillage treatments. More organic carbon within planting basin can be attributed to
manure that was applied in the basins every year. Along the rip line manure was dribbled
each year with the area between rip lines and basins receiving no manure in any of the
seasons. Studies by Belder et al. (2007) also showed an increase in organic carbon in
basins following three seasons of praeticing conservation farming on smallholder farms.
270
Carter (1996) observed organic matter increase in 0 - 0.10 m soil layer with use of
reduced tillage after nine years of experimentation. Similar organic carbon content was
observed in the 0 - 0.10 and 0.10 - 0.20 m layers under the conventionally ploughed
system. The ploughing operation incorporated the surface applied manure and some of
the maize residue applied as mulch. As ploughing was conducted in subsequent seasons
manure and decomposing maize residues were probably distributed evenly within the 0 -
0.20 m plough layer.
1.4 I El 0-10010-20 I
1.2
1.0
~ 0.8
'-'
U
0 0.6
0.4
0.2
0.0
CP BRL RL BB WB
Tillage treatment
Figure 9c.1. Effect of tillage and soil depth on organic carbon content (%) of a red clay
soil at Matopos Research Station. Error bars represent standard error of means
9c.4.2 Soil bulk density
In the one-season maize field (M), the lowest soil bulk density was recorded in the ripline
while the area between planting basins had the most compacted soil (Table 9c.4). In this
same field, soil bulk density at 0 and 4 tha-I mulch treatment was similar but lower than
271
soil bulk density observed at 10 tha-I mulch cover in the CP system. In the ripper tillage
system, 10 tha-I mulch treatment had lower soil bulk density compared to 0 and 4 tha-I
mulch cover in the ripline. Soil bulk density was similar under the three mulch treatments
between the riplines. In the planting basin (WB), soil bulk density was lower under 4 tha'
I than 0 and 10 tha-I mulch treatments. The 0 and 10 tha-I mulch treatments had similar
soil bulk density in the planting basin. In the area between planting basins, the soil was
more compacted at 0 tha" mulch treatment than at 4 and 10 tha-Imulch cover.
Table 9c.4. Effect of tillage, mulching and year interaction on soil bulk density of a red
clay soil at Matopos Research Station
. Mulch Tillage Field number and crop sequence
(tha -1)
]. 2 3 4
{M) ~MM} {MCM} (MCSM)
0 CP 1.51 1.55 1.57 1.50
RL 1.46 1.57 1.47 1.44
BRL 1.45 1.51 1.52 1.52
WB 1.57 1.59 1.48 1.48
BB 1.63 1.55 1.48 1.50
4 CP 1.51 1.62 1.54 1.48
RL 1.48 1.53 1.45 1.43
BRL 1.48 1.54 1.54 1.45
WB 1.44 1.55 1.50 1.46
BB 1.54 1.59 1.51 1.49
10 CP 1.63 1.66 1.51 1.50
RL 1.39 1.46 1.46 1.36
BRL 1.46 1.54 1.52 1.46
WB 1.58 1.47 1.50 1.41
BB 1.55 1.62 1.49 1.47
Lsdo.os=0.096; CV = 7.2 %
In the field with two seasons of conservation agriculture practices and monoerop maize
(MM), the least compacted soil was in the ripline regardless of the mulch treatment
(Table 9c.4). The conventional tillage system had the most compacted soil followed by
272
the area between planting basins. In the conventional system, 4 and 10 tha' mulch
treatments had similar soil bulk densities (1.62 and 1.66 gcm") with the 0 tha-1 mulch
treatment having lower soil bulk density (1.55 gcm") compared to the mulched plots. In
the ripline and planting basin, soil bulk density decreased with increase in the quantity of
mulch applied. However, in the area between riplines, soil bulk density was similar under
0, 4 and 10 tha-1 mulch treatments. Soil bulk density was higher at 10 tha-1 mulch
treatment than at 0 and 4 tha" mulch cover on the area between planting basins.
In the field with three seasons under conservation agriculture with cowpea rotation
(MCM), the least compacted soil was also observed in the ripline while the conventional
system had the most compacted soil. In the conventional system, soil bulk density
decreased with increase in the amount of mulch cover applied. In the ripper system, the
three mulch treatments had similar effect on the measured soil bulk density in the ripline.
In the area between riplines, 0 and 10 tha-1 mulch had similar effect on soil bulk density
and the 4 tha-1 mulch cover had higher soil bulk density than the other two mulch
treatments. In the planting basin, 4 and 10 tha' mulch treatments had a similar effect on
soil bulk density. In the area between planting basins, the soil under 4 tha-1 mulch cover
was more compacted compared with soil under 0 and 10 tha-1 mulch treatments.
In the field exposed to conservation agriculture practices for four seasons with cowpea
and sorghum in rotation with maize (MCSM), the soil in the ripline was the least
compacted followed by soil in the planting basin. The conventional system and the area
between planting basins had the most compacted soil. In the conventional system, 0 and
273
10 tha-Imulch treatments had similar soil bulk density (1.5 gcm") and the 4 tha-Imulch
cover had 1.48 gcm-3. In the ripper system, soil bulk density decreased with increase in
mulch cover applied in the ripline. In the area between riplines, the soil under 4 and 10
tha-I mulch cover was less compacted compared with the soil in unmulched plots. In the
basin system, soil bulk density decreased with increase in the quantity of mulch cover
applied. Soil in the planting basin was less compacted than the soil from the area between
basins. The comparison across the four fields used in the 2007/08 season indicates that
soil bulk density decreased when conservation agriculture practices had been
implemented a longer period of time. The most noticeable decrease in soil bulk density
was measured in the ripline and within the planting basin.
The reduced bulk density between basins and rip lines at 4 and 10 tha-I mulch in fields 3
and 4 could be attributed to the combined effect of minimum tillage and mulching on
modifying soil physical properties. Minimum tillage and the decomposing maize residues
probably promoted soil aggregation as reported elsewhere by Hadas et al., (1994) and
Mulumba and Lal (2008). As soil aggregation takes place more pores develop and
findings from a study by Glab and Kulig (2008) revealed a significant increase in
porosity within the top 0.10 m soil layer with reduced tillage and mulching. Soil bulk
density was generally lower in field 1 (M) than field 2 (MM) across all mulch levels and
tillage methods. This could be attributed to the fact that field 1 (M) was fallow during
2006/07 cropping season. At the start of 2005/06 season field 1 was tractor ploughed and
maize was grown during that season. During 2006/07 season the field lay fallow with
274
maize stover from 2005/06 season left in the field and it decomposed during the course of
2006/07 season. Then maize for this experiment was planted in 2007/08 season.
Figures 9c.2 and 9c.3 presents soil bulk density variation with respect to soil depth in the
four fields exposed to conservation agriculture practices for different number of seasons.
After the first season of CA treatments (illustrated by field 1), in the surface soil layer,
the ripline the lowest soil bulk density as the soil was disturbed during ripping. Soil bulk
density between planting basins in the surface layer is as high as the conventional tillage
system. This would be expected as the area between basins has remained unploughed and
compaction can also be caused by the movement of people during field operations. In the
deeper soil layers (20-30 cm), the highest soil bulk density was measured in the basin
tillage system. This shows that the effect of the planting basin on soil disturbance is only
confined to the 0-15 cm soil layer.
In the field with two seasons of CA treatments (illustrated by field 2), soil bulk density
was similar in the top 15 cm layer within the ripline and planting basin. The area between
planting basins had the most compacted soil in the top 5 cm layer and this is attributed to
the fact that this area had not been ploughed for two seasons. The conventional system
had more compact soil from 10 cm layer and below compared with the other treatments.
The high soil bulk density at 25 to 30 cm depth can be attributed to a plough pan that
could have developed over the years during ploughing operations before our trial was
established.
275
1.25 1.5 1.75 2
0
().
5 ,,
<>x
10 ..',,..-., ,
E, x. \) -..._........._.._...
Field 1
15 \ \
...s::: )X" <>
0..20 "·1
Q)
"'0 o····xI .
25 I
e
30 1---0-- CP ... X· •• BRL - -<>- - RL ---6- BB ---6- Wij
35
Figure 9c.2. Soil bulk density distribution with respect to soil depth in different tillage
treatments in field 1 (first year of CA treatments) at Matopos Research Station
In the field exposed to conservation agriculture for three seasons (Field 3), soil bulk
density within ripline and planting basin was lowest and similar in the top 5 cm soil layer.
However, from the 10 cm to 25 cm depth, soil in the ripline was the least compacted. As
observed in fields 1 (M) and 2 (MM), the lowest soil bulk density in the ripline is
attributed to the ripper being able to disturb the soil much deeper than the other tillage
treatments. In the planting basin, soil bulk density increased more rapidly from the 10 cm
layer to the 15 cm than the other tillage treatments. This is rather unusual as the 0-15 cm
soil layer was disturbed three times during the opening of planting basins. As observed in
the other two fields, the conventional system had more compacted soil than the other
treatments.
276
Bulk density (gcm')
1.25 1.5 1.75 2
0
5
,..-., 10 Field 2
E
..C_J., 15
;5
c.. 20
Q)
"'0 25
30
35 I-o-Cp ---x'-- BRL--.--RL -tz-BB ~ WBI
Bulk density (gcm")
1.25 1.5 1.75 2
0
5
10
I 15 Fiekl 3
£
c..
Q) 20
"0
25
30
35 1--0- CP -- -:I(. - - BRL - -.- - RL -er-- BB --I!r- WB 1
Bulk density (gcm")
1.25 1.5 1.75 2
0
5
10
,..-.,
E
~ 15
t Field 4
Q) 20
"0
25
30 -
35 1--0- CP -- -:1(' - - BRL - -- - RL -er-- BB _..__ WB I
Figure 9c3. Soil bulk density distribution with respect to soil depth in different tillage
treatments and fields at Matopos Research Station
277
As observed in the other three fields (M, MM and MCM), the least compacted soil in the
top 15 cm soil layer was within the ripline and planting basin in a field exposed to
conservation agriculture for four seasons (Field 4). Soil bulk density was consistently
lower in the ripline in the entire 0-30 cm profile than the other tillage treatments (Fig.
9c.2). In contrast to the ripline, soil bulk density within the planting basin increased in the
15 to 20 cm layers, further highlighting the fact that the basin effect is confined in the top
0-15 cm layer. As observed in fields 1 (M) and 2 (MM), the area between planting basins
had the most compacted soil compared with the other treatments.
9c.4.3 Cumulative infiltration
Cumulative infiltration measured under the 0 tha-1 mulch treatment in fields 1 (M), 2
(MM), 3 (MCM) and 4 (MCSM) under conventional (CP), within ripline (RL) and within
planting basin (WB) tillage treatments is presented in Figures 9c.4 and 9c.5. The highest
amount of water infiltrating into the soil was 60 mm recorded in the planting basin from
the maize-cowpea-maize field (Field 3). The low soil bulk density observed in the top 5
cm in the planting basin (WB) from field 3 (MCM) (Fig. 9c.3) suggests better porosity
and this could have enabled more water to infiltrate into the soil. Under the 0 tha-1 mulch
treatment, field 1 (M) had the lowest cumulative infiltration measured from the CP
system during the IS-minute infiltration run compared with the other three fields. In the
ripline, the maize-cowpea-maize field (Field 3) recorded more water infiltrating into the
soil than the other fields (Fig. 9c.4 and 9c.5). The low cumulative infiltration in field 1
(M) could be explained by the relatively high initial soil water content. Field 1 (M) had
the highest soil water content in the top 0.10 m before infiltration measurements were
278
conducted. Das and Chopra (1988) state that in some soils, infiltration rate is high if the
initial soil water content is low. Field 1 was waterlogged during part of December 2007
and January 2008.
70 ~ Field 1 ~ Field 2 -~ Field 3 - 04- - Field 4
,-..
ê 60 Plough 55
'-='0 50 53
.~
40
t;::::
.5 33
(!) 30
.>~ 24
"'5 20
E~ 10
U
0
0 5 10 15 20 25 30 35
Time (--./s)
70 --0- Field 1 --+- Field 2 -~ Field 3 - -8- - Field 4
,-...
E
E 60 RL
'-'
..=9.. 50
~
tB 40
.5
(!) 30
..2..:..~:s 20
~E 10
U __ -
0
0 5 10 15 20 25 30 35
Time (--./s)
Figure 9c.4. Effect of conventional ploughing and ripper systems, 0 tha-1 mulch cover and
period of exposure to CA practices on cumulative infiltration of a clay soil
279
70 ~ Field 1 ~ Field2 -Ir- Field 3 - .. - - Field 4
,-ê._ 60
..._,
r:: 50
.'0::
.ê 40
~
.5 30
<I)
:>
.~ 20
"3
:E:s 10
U
0
0 5 10 15 20 25 30 35
Time(.ys)
Figure 9c.5. Effect of basin system, 0 tha-1 mulch cover and period of exposure to CA
practices on cumulative infiltration of a clay soil
Under the 4 tha-1 mulch treatment, the maize-cowpea-maize field (field 3) had 62 mm of
water infiltrating into the soil from the CP tillage treatment (Fig. 9c.6). Within the ripline
the maize-cowpea-maize field (field 3) had more water infiltrating into the soil during the
I5-minute infiltration run. In contrast, the highest volume of water (47 mm) infiltrating
into the soil in the planting basin was recorded from the maize-cowpea-sorghum-maize
field (field 4). The high cumulative infiltration in the CP treatment of field 3 (MCM) at 4
tha-1 mulch could be linked to high capillary sorptivity observed under this treatment
(Table 9c.5). This observation could be explained by the size and orientation of soil pores
in field 3 (MCM) in the CP system. Moreno et al. (1997) state that if initial soil water
content is not significantly different across treatments under consideration, infiltration
can be influenced considerably by pore size distribution in the soil matrix. The field with
first season of CA treatments (Field 1) had the lowest cumulative infiltration across the
tillage treatments and this could be due to low sorptivity of the soil (Table 9c.5):
280
70 1~Field 1 -_-Field 2 -- Field 3 - ....- - Field 41
I 60 62
"-'
i5040 Plough30
i2010
0
0 5 10 15 20 25 30 35
Time (Vs)
70
I60 1 ~ Field 1 ---- Field 2 -t- Field 3 - -0- - Field 41
e
0 50
~ 40
~
(1) 30
.:~>
'"5 20
E
;::l 10
U
0
0 5 10 15 20 25 30 35
Time (Vs)
70
I 1~Field 1 ----Field 2 -t- Field 3 - -Q- - Field 4160
'-"'
c:
..0c 50
~ 40
~ 30
..~c
to
'"5 20
§ 10
U
0 . -
0 5 10 15 20 25 30 35
Time (vs)
Figure 9c.6. Effect of tillage, 4 tha-I mulch cover and period of exposure to CA practices
on cumulative infiltration of a clay soil
281
Under the 10 tha" mulch treatment, the maize-cowpea-maize field (Field 3) had the
highest cumulative infiltration (62 mm) measured in the planting basin during the 15
minute infiltration run (Fig. 9c.7). The field exposed to CA practices for four seasons
(Field 4) had the lowest cumulative infiltration (30 mm) in the CP and within rip line at
10 tha-1 mulch treatment. This is at odds with hydraulic conductivity observed in CP and
RL in field 4 (Table 9c.5). In field 2, representing maize monocropping, planting basins
recorded a lower volume of water infiltrating into the soil than the other tillage
treatments. This observation on cumulative infiltration is inconsistent with soil bulk
density, hydraulic conductivity and capillary sorptivity results (Tables 9c.4 and 9c.5).
Hydraulic conductivity was relatively high in CP and RL tillage treatments (Table 9c.5).
Low water absorption in the CP and RL at 10 tha-1 could be a result of trapped air within
the soil matrix. The advance of water through the soil during infiltration may result in
sealing of pathways through which air could escape (Peck, 1969; Azooz and Arshad,
1996). This results in a significant reduction in infiltration rate as water is forced to
follow other paths that could be more tortuous.
The mulch treatment had no significant (P = 0.634) influence on the amount of water
infiltrating into the soil during the 15 minute infiltration run across the four fields. Under
the 0 and 10 tha-1 mulch treatments, 41 mm of water infiltrated into the soil while 37 mm
were recorded at 4 tha' mulch cover. This observation suggests that water movement
through the soil matrix is influenced by soil properties regardless of the level of mulch
cover applied.
282
70 I-=--Field 1 - .... -Field 2 -~ Field 3 - ... - - Field 41
'[ 60
'-' Plough
é 50
.~ 40
~ 30
.~
~ 20
~ 10
o - --
o 5 10 15 20 25 30 35
Time (../s)
I-=-- Field 1 - Field 2 - ~ Field 3 - -(1)- - Field 41
57
RL
38
37
34
o 5 10 15 20 25 30 35
Time (../s)
I~ Field 1 --0-- Field 2 -~ Field 3 - -0- - Field 41
62
o 5 10 15 20 25 30 35
Time (../s)
Figure 9c.7. Effect of tillage, 10 tha-( mulch cover and period of exposure to CA practices
on cumulative infiltration of a clay soil
283
9c.4.4 Hydraulic conductivity and capillary sorptivity
Hydraulic conductivity and capillary sorptivity were significantly (P < 0.001) influenced
by the three way interaction of mulching, tillage system and period of exposure to CA
practices. In the field under CA practices for the first time, conventional system had
higher hydraulic conductivity and sorptivity at 0 and 4 tha-1 mulch treatments than the
other tillage treatments (Table 9c.5). However, at 10 tha' mulch cover the highest
hydraulic conductivity and sorptivity were observed in the ripline from field I (M). In the
continuous maize monoeropping field (Field 2), the measured hydraulic conductivity was
higher in the conventional system. In contrast, highest sorptivity was 1.1 mmJ--Js in the
planting basin at 0 tha-1 mulch, 1.26 mmJ--Js in the ripline at 4 tha" mulch and 0.93
mmJ--Js between planting basins at 10 tha-1 mulch cover.
In the maize-cowpea-maize field (field 3), hydraulic conductivity was higher in the
ripline under 0 and 10 tha-1 mulch treatments. In the same field, hydraulic conductivity
measured in the planting basin was slightly lower than in the ripline under 0 tha-1 mulch
treatment. Under the 4 tha" mulch treatment, the CP and BB (between basins) treatment
had similar hydraulic conductivity in the maize-cowpea-maize field. The sorptivity was
high in the BRL (between riplines) treatment under 0 tha" mulch cover, in the CP
treatment under 4 tha-1 and RL (in riplines) under 10 tha-1 mulch treatment. In the maize-
cowpea-sorghum-maize field (field 4), the soil between planting basins had higher
hydraulic conductivity under 4 and 10 tha-1 mulch cover. In the unmulched tillage
treatments, soil in the ripline had higher hydraulic conductivity compared with other
284
treatments. Sorptivity was high in the planting basin at each mulch treatment in field 4
(MCSM).
Table 9c.5. Effect of minimum tillage, mulching and period of exposure to CA practices
on soil hydraulic conductivity (K) (10-3mms-l) and capillary sorptivity (S) (mm/Vs) of a
cl soil at Research Station
Mulch Tillage
(tha-t)
0 CP 9.80 3.43 5.00 2.57 1.31 0.73 1.36 0.56
BRL 3.00 1.98 4.05 3.27 0.78 0.60 1.39 0.84
RL 4.02 2.00 7.59 3.36 0.43 0.93 1.31 0.95
BB 3.77 2.32 3.93 1.59 0.65 0.97 0.91 0.84
WB 6.23 2.02 6.98 2.07 1.13 1.10 0.64 1.06
4 CP 3.46 3.66 4.00 4.84 0.52 0.84 1.39 0.92
BRL 2.95 2.78 2.93 2.93 0.23 0.51 0.43 0.67
RL 1.59 2.68 3.09 3.09 0.40 1.26 1.13 0.73
BB 1.80 1.68 4.00 6.18 0.40 1.07 1.07 0.78
WB 3.02 3.00 3.48 4.02 0.37 0.71 0.79 0.96
10 CP 3.89 4.98 2.42 4.50 0.59 0.49 0.74 1.09
BRL 1.27 0.98 0.71 3.57 0.57 0.43 0.51 0.84
RL 4.32 2.00 2.68 3.96 0.68 0.85 1.45 0.62
BB 1.61 4.64 0.80 4.52 0.50 0.93 0.82 1.01
WB 2.14 2.91 1.72 4.11 0.48 0.65 0.42 1.10
Field 1 (M) is first season of CA treatments, field 2 is maize monoeropping for two seasons under
CA, field 3 is maize-cowpea-maize rotation and three seasons of CA, field 4 is maize-cowpea-
sorghum-maize rotation and four years under CA.
The observed changes in hydraulic conductivity with the introduction of CA practices
could be attributed to soil structural improvements that were indicated by changes in bulk
density (Table 9c.4 and Fig. 9c.2). Das and Chopra (1988) noted that changes in soil bulk
density bring about variations in pore geometry which can significantly influence water
movement within the soil matrix. Hydraulic conductivity is a soil property that is
dependant on soil water content (Hanks and Ashcroft, 1980). Very high hydraulic
285
conductivity in field 1 (first season CA treatments) could be attributed to more soil water
that was observed in the top 0.10 m under CP system (Table 9c.6). In fact hydraulic
conductivity of soil decreases considerably as soil water content changes from saturation
to permanent wilting point (Hanks and Ashcroft, 1980). Field 3 (MCM) had lowest
capillary sorptivity in planting basins at 0 tha" mulch. The highest sorptivity in CP and
RL was recorded in fields 4 (MCSM) and 3 (MCM). Capillary sorptivity is usually high
when the soil is dry at the start of infiltration measurements (Bissett and O'Leary, 1996)
as was the case in this study.
9c.4.5 Soil water content before and after infiltration runs
Under the 0 tha-1 mulch treatment, field 2 (MM) had highest soil water content observed
in RL treatment while field 1 (M) had lowest water content at the same mulch level
(Table 9c.6). Under the 4 tha" mulch cover, the highest soil water content was observed
in RL of field 4 (MCSM). The maize-cowpea-maize field (Field 3) under BB tillage
treatment had marginally higher initial soil water content under 10 tha-1 mulch than the
other tillage treatments. One of the major roles played by mulch cover could be reducing
soil evaporation. Hatfield et al. (2001) observed a 34-50 % reduction in soil water
evaporation as a result of crop residue mulching.
286
Table 9c.6. Volumetric water content (%) measured by capacitance probe in 0 - 0.10 m
soil layer before infiltration runs at Matopos Research Station
Mulch Tillage Field number
(tha-I)
o CP 10 9.8 7.7 8.2
RL 9.4 18 13 Il
BB 8.5 6.6 9.3 12
4 CP 13 13 11 6.5
RL 7.3 15 14 16
BB 9.3 10 15 6.4
10 CP Il 10 6.1 8.2
RL 11 7.6 Il 12
BB Il 13 14 13
Lsdo.os = 5.8; CV = 26 %
After the infiltration runs field 1 had more (P = 0.871) soil water content under 0 tha-l
mulch in CP compared to other fields (Table 9c.7). The same field had slightly more soil
water in the CP system than others before the infiltrations runs were made (Table 9c.6)
under the 0 tha-l mulch treatment. The high soil water content after infiltration in CP at 0
tha-l mulch cover (Table 9c.7) could also be a result of better aligned and linked soil
pores created during the ploughing operation resulting in high hydraulic conductivity
(Table 9c.5). Vauchin and Chopart (1997) observed increases in hydraulic conductivity
and sorptivity in conventionally ploughed treatments and attributed this to better pore
connectivity due to turning of the soil. Under the 4 tha-l mulch soil water content ranged
from 10 to 15 % across the five tillage treatments. Under the 10 tha" mulch treatment, 22
% soil water content was recorded in basins in field 3 (MC M) while lowest soil water
content was 11 % observed in BRL of fields 1 (M) and 2 (MM).
287
Table 9c.7. Volumetric soil water content (%) in 0 - 0.l0 m layer measured after
infiltration runs in the four fields at Matopos Research Station
Mulcb Tillage Field number
(tba-I) 1 2 3 4
o CP 22 12 12 14
BRL 13 14 14 13
RL 14 16 12 12
BB 18 15 Il 15
WB 16 14 15 14
~~ilïfu~:li~ilt~~:~,'iilffil'lI!!!?ji!l~1:!g:1m:~'W~:;j11\lN'i.:ti~ii$ffil~~~~~~~~~m~~~m,1i\\m,~~
~~~~~~~~~~~llfi~~~i?f.W.~~~
4 CP 15 15 14 15
BRL 10 13 13 11
RL 13 14 14 14
BB 10 14 12 13
WB 15 15 13 13
10 CP 17 17 15 16
BRL Il Il 13 15
RL 12 19 16 12
BB 15 14 14 14
WB 16 14 22 16
Lsdo.os= 6.1; CV = 22 %
9c.4.6 Maize crop performance
Table 9c.8 summarizes the main effects of period of exposure to CA practices on maize
crop performance at the end of the 2007/08 growing season. The performance of maize
crop differed significantly (P < 0.001) across the four fields during 2007/08 season. Field
4 (MCSM) produced the highest grain and stover yields and field 1 (M) gave the lowest
yields, suggesting improved soil productivity with more years of praeticing CA. Field 4
(MCSM) produced 1 970, 1 880 and 679 kgha' more grain than fields 1 (M), 2 (MM)
and 3 (MCM). Field 4 (MCSM) also produced 1 905, 1 681 and 946 kgha' more stover
than fields 1 (M), 2 (MM) and 3 (MCM). Harvest index was highest in field 3 (MC M)
while fields 3 (MCM) and 4 (MCSM) had an average plant stand of 2.6 plants per m2.
Higher maize yields in field 4 (MCSM) could be attributed to improvements in soil
288
physical and chemical conditions. Annual application of manure and cowpea stover that
was left at the surface after harvest contributed towards soil organic carbon content as
well as adding nutrients such as Nand P. Cowpea grown as part of a rotation could have
fixed N that benefited maize crop in fields 3 (MCM) and 4 (MCSM). Giller et al. (1997)
reported that in general cowpea grown as sole crop fixes 11 - 210 kgNha" in African
cropping systems. Changes in soil bulk density improve the soil environment for root
growth and exploration for nutrients and water.
Table 9c.8. Maize crop performance measured at the end of the 2007/08 season as
influenced by different periods of exposure to CA practices on a clay soil at Matopos
Research Station
Field! Grain yield! Stever yield Harvest index Plant stand
(kgha-1) (kgha-1) (m-2)
1 (M) 1 028 1277 0.43 1.7
2(MM) 1 118 1 501 0.42 1.9
3 (MCM) 2319 2236 0.50 2.6
4 (MCSM) 2998 3 182 0.48 2.6
Lsdo.o5 248 210 0.024 0.17
CV(%) 38 29 15 22
The two-way field x tillage system interaction had a significant (P < 0.001) influence on
grain and stover yields, and plant stand measured across the four fields (Table 9c.9). The
lowest grain and stover yields were recorded in the planting basin system in field 1 (M) ..
Field 4 (MCSM) had the highest grain production in all tillage systems. Planting basins in
field 4 gave 668 and 665 kgha" more grain than CP and ripper tillage systems. The two-
way interaction had no significant (P = 0.113) effect on maize harvest index. Planting
basins had the highest plant stand in field 3 (2.8 plants per m2) and 4 (3.1 plants per m2)
while field 1 (M) had similar plant stand in CP and basin tillage systems (1.6 plants per·
289
m2). During December 2007 and January 2008 fields 1 (M) and 2 (M) were affected by
waterlogging in planting basins. Rodent attack at crop establishment and waterlogging
during the vegetative stages affected the observed yields in fields 1 (M) and 2 (MM).
Plant stand was higher in basins than other two tillage systems in fields 3 and 4 during
2007/08 seasons. The better maize crop establishment under the basin system was
reported at on-farm and on-station experiments during the other seasons of
experimentation (Chapter 7; Chapter 9a).
Table 9c.9. Maize yield (kgha"), harvest index and plant stand (plants m-2) responses to
three tillage methods at Matopos Research Station in 2007/08 season
Field Tillage Grain Yield Stever Harvest Plant stand
method kgha' kgha' index #per m2
Plough 1 092 1 312 0.45 1.6
Ripper 1 222 1 406 0.46 1.9
Basins 769 1 115 0.40 1.6
2 Plough 1245 1 610 0.44 1.8
Ripper 1 097 1 515 0.42 2.0
Basins 1 013 1 379 0.42 1.9
3 Plough 2554 2310 0.51 2.4
Ripper 1 993 1 977 0.50 2.6
Basins 2410 2422 0.48 2.8
4 Plough 2774 3046 0.47 2.2
Ripper 2777 2924 0.49 2.4
Basins 3442 3577 0.49 3.1
Lsdo.o5 414 327 0.047 0.70
CV(%) 38 29 15 22
9c.5 Conclusion
The study showed that significant changes in both chemical and physical soil properties
occurred during the one to four years the clay soil was exposed to minimum tillage,
mulching and crop rotation. Soil organic carbon increased significantly particularly in
290
planting basins during the four year study period. The maize residue applied as mulch
and cowpea stover left in the field after harvesting contributed towards the soil organic
carbon pool even between rip lines and planting basins. Minimum tillage resulted in
significant positive changes in soil bulk density and hence improved soil physical
environment for plant roots. The improved soil conditions increased water infiltration
into the soil as indicated by positive changes in soil hydraulic conductivity and capillary
sorptivity. Improvements in soil structure due to CA practices give better chances of
minimizing rainwater losses from the fields. This potentially increases soil water supply
and extends the period of soil water availability to crops in smallholder cropping systems.
Improvements in soil physical and chemical environment were also reflected in maize
yield differences observed in fields exposed for different periods to CA practices. The
field with four years of exposure to CA practices outperformed the other three fields in
all maize parameters measured. Improvements in soil fertility due to annual application
of manure and crop residues left on the surface resulted in higher maize yields with a
longer period of praeticing CA. The improved soil physical environment, as shown by
positive changes in bulk density, enables plant roots to explore more soil volume for
water and nutrients resulting in improved productivity in semi-arid smallholder cropping
systems.
291
CHAPTER 10
Productivity of Planting Basin Tillage System and Nitrogen
under lHlighlyVariable Rainfall Regimes of Semi-Arid Southern Zimbabwe:
A Modelling Assessment
in.r Introduction
In semi-arid regions high spatial and temporal rainfall variability presents a big challenge
to rainfed smallholder agriculture. Reduced crop yields and total crop failures are a
common feature in semi-arid farming systems. In growing seasons with above average
rainfall crop yields still remain low due to soil fertility constraints. In wet seasons
smallholder farmers fail to make effective use of available soil water and cereal yields
from unfertilized granitic sands remain below 500 kgha' (Shamudzarira and Robertson,
2002; Chapter 7).
The advent of conservation farming using planting basins brought a ray of hope to
smallholder farmers in semi-arid districts of Zimbabwe. The planting basin tillage system
enables farmers to prepare land early spread the limited farm labour and plant on time
with respect to the effective planting rain. The planting basins created in a grid of 0.9 m x
0.6 m spacing harvest rainwater and reduce surface runoff from cropping fields
(Mupangwa et al., 2008). Substantial crop yields obtained from the planting basin system
have been reported in several semi-arid districts of Zimbabwe (Nyagumbo, 2007;
Twomlowet al., 2008a) and on agricultural research stations (Chapter 9a):
292
Although planting basins on smallholder farms have been in use for only four growing
seasons, the use of simulation modeling can help understand the long-term impact of the
tillage system under semi-arid conditions. The Agricultural Production Simulator Model
(APSIM) is a deterministic, process based model that has been used for simulating crop
production in smallholder cropping systems. The APSIM model has performed well in
predicting crop production and its interaction with climate, soil and management factors
(Keating et al., 2003). In smallholder farming systems, APSIM has been used with
success to simulate nitrogen (N) dynamics of manure inputs (Delve and Probert, 2004),
maize response to N (Shamudzarira and Robertson, 2002), water use efficiency (Dimes
and Malherbe, 2006), and N and water stress dynamics in cereal-legume rotations (Ncube
et al., 2008). However, no description of the effect of mulch on crop yields and soil water
dynamics under conventional, ripper and planting basin tillage systems is included in any
of the previous studies.
llO.2 Objectives
The main objectives of the current study were to evaluate the capability of APSIM
(version 6.0) cropping systems model to simulate maize yield and soil water responses to
different seasons, mulch levels and three tillage systems on two soil types. Data obtained
from on-station experiments (Chapter 9a) will be used to verify model performance. Then
to use the validated APSIM model to assess the long term interaction effects of N and
two tillage systems,' conventional and planting basins, on maize yields and selected
components of the soil water balance for a granitic sand soil in semi-arid areas of
southern Zimbabwe. The specific objectives were:
293
(1) to evaluate APSIM capability in predicting the seasonal and mulching effects on
maize grain and total biomass yields;
(2) to evaluate APSIM capability in predicting effects of conventional, ripper and
planting basin tillage techniques on soil water regimes; and
(3) to use the validated APSIM model to assess the long term interaction effects of N
fertilizer, and conventional and planting basin tillage systems on maize yields, surface
runoff and deep drainage.
11.0.3Materials and Methods
10.3.1 Set up of the model
10.3.1.1 Climate parameters
Daily rainfall, temperature and radiation data were collected from Matopos Research
Station weather station which is located 3 km from Matopos experimental site and up to
10 km from the Lucydale site. The climate record used for APSIM calibration stretched
from 1 October 2004 to 30 June 2008. The simulation was run from 1 October 2004 to 30
June 2008 and the model was reset every 1 July to initial soil organic carbon, nitrogen
and water content.
10.3.1.2 Soil water and soil characteristics
Soil parameters used for calibrating the APSIM model are given in Tables 10.1 and 10.2.
As the experiment for Chapter 9a had a new field established in each season, the PAWC
for 2004/05 field 4 was 116 mm, 2005/06 was 84 mm for field 3, for 2006/07 was 61 mm
for field 2 and for 2007/08 was 84 mm for field 1 in the 0 - 0.85 m soil profile. The
294
drained upper limit (DUL), saturation (SAT) and lower limit (LL) were derived from soil
water measurements made in the basins with zero mulch. For Lucydale PAWC in the 0 -
0.70 m soil profile was 84 mm for 2004/05 and 2005/06 seasons. The Lucydale soil water
parameters were adopted from Masikati (2006). The field used by Masikati (2006) was
adjacent to our experimental field for both 2004/05 and 2005/06 seasons.
Soil nitrate-N for Matopos clay soil was assumed total N was 25 kgha' (20 kg N03- and
5 kg NH4l. Soil water and organic matter were reset to zero on the first of July each year
while N was reset to 25 kgha" on the same date. It was assumed that the 3 tha-I of
manure used in our experiment supplied 9.6 kgNha-l. Runoff curve number for bare soil
was set at 85. For the plough and ripper tillage systems the curve number was adjusted
downwards by 10 units which were lost after 50 mm of rainfall, so it then reverted to 85.
For the basins the curve number was adjusted downwards by 20 units which were lost
after 250 mm of rainfall was received. The C:N ratio of mulching material was set at 60
(0.67 % N) and alO % incorporation of the mulching material in the CP system was
assumed. For basins and ripper tillage a 0 % incorporation of the mulching material was
assumed.
295
Table 10.1. Soil chemical and physical properties of the clay soil used for Matopos
Researc h Station expenm. enta l site (iirom ICRISAT unpu bliIShed data)
Depth pH N03-N Organic Bulk nUL LL
(cm) (ppm) carbon density (mm/mm) (mm/mm)
(%) (gcm")
0-15 6.0 6.50 1.2 1.4 0.20 0.10
15-25 6.0 2.10 1.0 1.4 0.24 0.10
25-35 6.0 2.10 0.86 1.4 0.26 0.13
35-45 6.0 1.70 0.83 1.4 0.27 0.16
45-55 6.0 1.70 0.58 1.4 0.29 0.20
55-65 6.0 1.70 0.54 1.4 0.29 0.21
65-75 6.0 1.70 0.54 1.4 0.30 0.23
75-85 6.0 1.70 0.5 1.4 0.31 0.25
Table 10.2. Soil chemical and physical properties of the sand soil used for Lucydale
experimental site (from Masikati, 2006)
Depth pE[ N03-N Organic Bulk DUlL lLlL
(cm) (ppm) carbon density (mm/mm) (mm/mm)
(%) (g/cm'')
0-20 6.3 4.3 0.8 1.66 0.15 0.05
20-30 6.3 3.2 0.7 1.65 0.22 0.13
30-40 6.9 1.5 0.7 1.60 0.28 0.20
40-50 6.9 1.5 0.7 1.55 0.34 0.27
50-60 6.9 1.5 0.7 1.51 0.37 0.32
60-70 6.3 1.1 0.6 1.34 0.41 0.36
10.3.1.3 Crop parameters
APSIM crop module contains a description of the short season hybrid variety SC401 used
in Zimbabwe. In our experiment a short seasoned hybrid variety SC403 was planted at
Matopos in all seasons and at Lucydale in 2005/06. An open pollinated variety ZM421
was planted at Lucydale in 2004/05 season because there was a maize breeding
experiment close to our research field. The three varieties are drought tolerant and
recommended for semi-arid areas of Zimbabwe. Hence APSIM crop parameters for
SC401 were selected to describe both SC403 and ZM421 used in the study.
296
10.3.1.4 Experimental management
Experimental management in the model was according to the experimental procedures
(Chapter 9a). Sowing, manure application and topdressing dates for Matopos and
Lucydale experimental sites are given in Tables 10.3 and 10.4. For plough treatments at
Lucydale manure was applied a day before sowing in each season, sowing being 14 and
13 December for 2004/05 and 2005/06 seasons according to the rain received. For
planting basins treatment at Lucydale manure was applied on 26 October 2004 and 14
September 2005 for 2004/05 and 2005/06 growing seasons. A sowing depth of 50 mm
was used in the simulation for each tillage system. Average plant stands of 1.8 and 3.1
plants per m2 were used for Lucydale in 2004/05 and 2005/06 seasons. For Matopos 3.0
plants per m2 was used for all seasons. All plots were kept weed free during period of
experimentation. The APSIM model simulated maize yield and soil water balance until
the crop was mature.
Table 10.3. Dates for field activities carried out at Matopos Research Station during the
four seasons of experimentation (Chapter 9a)
Season TinKage Mulch Manure Sowing Topdressing
method application am!_lication date date
2004/05 Plough 10/11/2004 12/12/2004 13/12/2004 21/1/2005
Ripper 10/11/2004 26/10/2004 13/12/2004 21/1/2005
Basins 10/11/2004 20/10/2004 13/12/2005 21/1/2005
2005/06 Plough 15/9/2005 13/12/2005 13/12/2005 24/1/2006
Ripper 15/9/2005 18/9/2005 13/12/2005 24/1/2006
Basins 15/9/2005 17/9/2005 13/12/2005 24/1/2006
2006/07 Plough 28/7/2006 7/12/2006 8/12/2006 2/1/2007
Ripper 28/7/2006 30/8/2006 21/11/2006 2/1/2007
Basins 28/7/2006 28/8/2006 21/11/2006 2/1/2007
2007/08 Plough 25/7/2007 12/12/2007 12/12/2007 101112008
Ripper 25/7/2007 5/8/2007 12/12/2007 101112008
Basins 25/7/2007 27/9/2007 12/12/2008 10/1/2008
297
Table 10.4. Dates for field activities carried out at Lucydale experimental site during the
two seasons of experimentation (Chapter 9a)
Seasolll Tillage Mulch Manure Sowing Topdressing
method application application
2004/05 Plough 17/10/2004 13/12/2004 14/12/2004 21/1/2005
Ripper 17/10/2004 25110/2004 14/12/2004 21/1/2005
Basins 17/10/2004 26/1012004 14/12/2004 21/1/2005
2005/06 Plough * 12/12/2005 13/12/2005 24/1/2006
Ripper * 8/9/2005 13/12/2005 24/1/2006
Basins * 14/9/2005 13/12/2005 24/1/2006
* No fresh mulch was applied
10.3.2 Long term simulation
The long term simulation was run using soil properties of the Lucydale granitic sandy soil
(Table 10.2). The 69 year climate record (1939 - 2008) derived from Matopos Research
Station weather station was used in the long term simulation. The following scenarios
were used in long term simulation:
o Conventional ploughing plus fourN rates (0,10,20 and 52 kgha')
G Planting basins plus four N rates (0, 10, 20 and 52 kgha")
The N rates of 0, 10 and 20 kgha' were similar to N levels used in the on-farm
experiments conducted in Gwanda and Insiza districts of southern Zimbabwe between
2005/06 and 2007/08 seasons. The 52 kgNha-1 is the national recommendation for
smallholder cropping systems of Zimbabwe (Twomlowet al., 2008a) so it was included
to provide a comparison with what may be considered as providing optimal yields.
298
10.3.3 Reporting frequency
For the on-station experiments the model was set to report selected variables on a daily
time step. The reported variables for the on-station experiments were total biomass and
grain yield, and profile water content in the 0 - 0.25 m layer. In the long term simulation
the model was set up to report variables at harvest stage of the maize crop. Total biomass
and grain yields were reported at 0 % moisture content and are compared to observed
yields at this moisture content. In the long term simulation the reported variables were
grain yield, pre-sowing and in-crop surface runoff, and in-crop deep drainage.
Genstat Discovery Edition 3 (www.vsni.co.uk) was used to analyze maize yield and soil
water data. The root mean square deviation (RMSD) values were calculated for
comparison of observed and predicted data. The RMSD was calculated as follows:
Equation 10.1
where Xi is the observed yield or soil water content, Yl is the predicted yield or soil water
content and III is the number of observations.
Modeling Efficiency (ME) was calculated as follows:
MEJt.(o, -ay - t.(l~- 0,)']
Ï(O; -0)
;=1 Equation 10.2
where Pi and 0 are predicted and observed values respectively, 0 is observed mean value
(Rinaidi et al., 2003).
299
10.4 Results and Discussion
10.4.1 Maize yields
The total growing seasonal rainfall for 2004/05,2005/06,2006/07 and 2007/08 was 320,
915,467 and 364 mm respectively. The predicted seasonal and mulching effects on grain
and biomass yields at Matopos are shown in Figures 10.1 and 10.2. APSIM was able to
predict closely the seasonal effects on grain (ME = 0.86) and biomass (ME = 0.84) yields
for the four seasons with different rainfall pattern at Matopos. The model gave a good
prediction of grain yield in 2004/05 which was an average season in terms of rainfall
received. For the below average rainfall seasons 2006/07 and 2007/08, APSIM predicted
low maize grain and biomass production on the Matopos clay soil which did not really
match the observed values. The model poorly predicted grain (ME = 0.27) and biomass
(ME = 0.49) production at Lucydale for 2004/05 and 2005/06 seasons under the set
model input conditions (Fig. 10.3).
300
4500 [] Observed liPredictedI RMSD= 786
4000 ME= 0.86
3500
,-...
-~ 3000
00
...::.0:
:;- 2500
4)
.;;' 2000 /),.
.s
~ 1500
l~~~ 0ldl-H"H~0ldl-INI+I~i,j,r,,,,~O,~ldH+I~I~
2004/05 2005/06 2006/07 2007/08
Seaso.nandmulchcover -I(tha )
4000 o 0.5 t:. 1 0 21
--. 3500
° 8 x 10 - 1:1 o
-<:Ij
'~ 3000
'-"
'""0.>0. 2500
.5 2000 o
~
'"0 1500 o
:~.a 1000 0<>
8
~
500
o ~---,,----.----0~A----0.---~-.--0---.----~----,
o 500 1000 1500 2000 2500 3000 3500 4000
Observed grain yield (kgha")
Figure 10.1. Observed and predicted grain yield from different mulch levels over four
growing seasons at Matopos Research Station. Error bars stand for standard error of
means for the different mulch levels in each season. R2 = 0.67 (averaged across seven
mulches)
301
12000 o Observed Il Predicted I RMSD= 2313
ME= 0.84
10000
--~ 8000
00
ë
-0 At;.
.0;);:' 6000 ~
Cl)
~
.9 4000
o::l
2004/05 2005/06 2006/07 2007/08
Season and mulch cover (tha")
12000 o 0 e 0.5 IJ::,. 1 E> 2
III 4 o 8 )I( 10 --1:1
10000
o
8000
o <>
6000
4000 o )I(
2000
o ~------.-------,-------,------,-------.-------,
o 2000 4000 6000 8000 10000 12000
Observed biomass yield(kgha")
Figure 10.2. Observed and predicted total biomass yields from different mulch levels
over four growing seasons at Matopos Research Station. Error bars stand for standard
error of means for the different mulch levels in each season. R2 = 0.50 (averaged across
seven mulch levels)
302
The mulching effect on grain and biomass yields in the four different seasons at Matopos
is shown in Figure 10.1. For 2004/05, 2005/06 and 2006/07 seasons the model over
predicted grain yield at 0 and 0.5 tha-I mulch cover. The model under predicted grain
production at 8 and 10 tha-I mulch cover in the same seasons. In 2007/08 the model under
predicted grain production at low mulch levels « 4 tha') while over predicting it at 8 and
10 tha-I (Fig. 10.1). For the wetter 2005/06 season the model under predicted biomass
production and indicated a decrease in yield with increase in mulch cover on the clay soil
(Fig. 10.1). This is in contrast to what happened in the field experiment at Matopos
(Chapter 9a). The mulch cover could have promoted a proliferation of soil micro-
organisms given the better supply of soil water during 2005/06 season. To be able to
decompose the maize residue mulch soil micro-organisms need energy and therefore out-
competed the maize plants in extracting soil N. The 29.6 kgblha" was probably not
enough to meet microbial and crop requirements thereby resulting in inadequate N supply
to the maize crop and thus a lower predicted yield at higher mulch levels. Yellowing of
maize foliage indicating N deficiency was quite evident at 8 and 10 tha" mulch level
particularly at the Matopos experimental site resulting in lower yield being achieved at 8
tha" mulch (Fig. 10.1). Degradation of the mulching material could have been driven by
termites which, unlike soil micro-organisms, do not need to take up N .from the soil in
order to degrade the plant residues (Konig and Varma, 2006).
303
4000 lo I RMSD= 16003500 Observed ~ Predicted ME=0.27
3000
2500
:.E~ 2000
:>..
. 1500
oe
s
1000
500
o +~~~~-~~~~~-~~~~~~~~I,,~-.L~r~~~~~
10
2004/05 2005/06
Season and mulchcover (tha")
9000 lo Observed IJ:. Predicted I RMSD= 3212
8000 ME= 0.49
""-;"~ 7000
OD
C 6000
"0
Q.~) 5000
~'"
4000
E 3000
c.2o 2000 -
1000
0
2004/05 2005/06
Season and mulch cover (tha")
Figure 10.3. Observed and predicted grain and total biomass yields for different mulch
levels on a sand soil over two growing seasons at Lucydale experimental site. Error bars
stand for standard error of means across three replications
The model under predicted maize biomass production with more mulch cover during the
2005/06 growing season at Matopos (Fig. 10.2). In the 2007/08 season the model
predicted an increase in biomass yield with higher mulch cover at Matopos (Fig. 10.2).
The model is probably indicating soil water benefits derived from mulching in 2007/08
304
season that was characterized by an abrupt end of rain in January 2008. The higher mulch
cover conserved and supplied soil water to take the maize crop to maturity. However,
observed maize yield data did not show any significant influence of mulch cover on total
biomass or grain yields. This is rather puzzling given the early cessation of rainfall at a
time when maize crop was at its reproductive stage. The over prediction of grain yield in
2004/05, 2005/06 and 2006/07 seasons at low mulch levels could be an indication of
better N supply because there was no immobilization of applied nitrogen.
At Lucydale sandy soil site, the model under predicted no grain yield responses to freshly
applied mulch cover in 2004/05 and residual mulch cover in 2005/06 seasons (Fig. 10.3).
Lack of grain yield responses to freshly applied and residual mulch in 2004/05 and
2005/06 is consistent with observed results. Field observations made in all seasons
showed that maize residue applied as mulch decomposed quite fast particularly in a
season with a lot of rain like 2005/06. At the end of growing season (April/May) there
was hardly any maize residue left in the experimental plots especially where 0.5, 1 and 2
tha" had been applied. An estimated 30 - 40 % mulch cover would be remai~ing in the 8
and 10 tha-1 treatments. Maize residue, with a C:N ratio averaging 52 (Nhamo, 2007),
decomposes fast when conditions of drivers. of decomposition such as rainfall,
temperature and soil micro-organisms are ideal (Heal et al., 1997)_
10.4.2 Soil water regimes
Soil water predictions from APSIM were compared with observed data at Matopos site
for the 2006/07 and 2007108 growing seasons, In the both seasons, the model over
305
predicted soil water content of the top (0 - 0.25 m) soil layer under the conventional,
ripper and basin tillage systems (Figs. 10.4 and 10.5). The soil water patterns were quite
similar in plough, ripper and basin tillage systems. The relationship between observed
and predicted soil water data (Figs. 10.4 and 10.5) indicates that APSIM was not always
able to closely predict soil water content in the plough, ripper and basin systems under
the conditions that we set up the model the prediction for 2006/07 season was better than
2007108.
50 OQ
00 0 o 0 E)
40 o
.--., 00
ê
'-...'.
~ 30
:C:d:
ë3
Cl)
"'0
~ 20
:oa
~
,:l.;
10
O~------~------~I~0==CP~=0=R=ip=per==0 ~B=asi=n-=-1=:1=1=~
o 10 20 30 40 50
Observedsoilwater(nnn)
Figure 10.4. Observed and predicted soil water content in a clay soil (0 - 0.25 m layer)
under plough, ripper and basin tillage systems at Matopos Research Station during
2006/07 growing season
306
60
®o
50
,-..
gE 40
...s.. "ti' 0 cc
<:1:l 00 El
-~'0 30
en
".0s
(.)
:T_; 20
~
p..
10
I 0 CP 0 Ripper 0 Basin -- 1:1 I
0
0 10 20 30 40 50 60
Observed soil water (rrnn)
Figure 10.5. Observed and predicted soil water in a clay soil (0 - 0.25 m layer) under
plough, ripper and basin tillage systems at Matopos Research Station during 2007/08
growing season
In all tillage systems, the model predicted closely the long term drying after 29 February
2008 (day 152) until the last soil water measurement was taken (Fig. 10.6). The
similarities in predicted soil water regimes under plough, ripper and basin tillage systems
are consistent with results from on-station and on-farm experimentation (Chapters 7, 9a
and 9b; Mupangwa et al., 2008). Earlier studies at Lucydale experimental site showed
that the model can also predict soil water regimes in legume-cereal rotations (Ncube et
al.,2008).
307
70 RMSD= 12
.60 ME=0.47
,.-..
..ê_, 50
..4.0.
~
~ 30
£Cl) 20 e ••
10
o +------------r----------~------------~----------~
o 50 100 150 200
Time (days after 1 October 2007)
70 1-- Predicted <> Observed 1 RMSD= 12
60 ME=0.76
~
..~_, 50
14030
~ 20
10
o +-----------,,-----------.-----------.----------~
o 50 100 150 200
Tune (days after 1 October 2007)
70 1-- Predicted 4> Observed I RMSD= 10
'-:'_60 ME= 0.76150
B~40
~ 30
Cl)
ïa 20
8
c, 10
o +-----------~-----------.-----------.----------~
o 50 100 150 200
Time (days after 1 October 2007)
Figure 10.6. Observed and predicted soil water in a clay soil (0 - 0.25 m layer) under
plough (a), ripper (b) and basin (c) tillage systems at Matopos Research Station during
2007/08 growing season
308
10.4.3 Long term impact of conventional and planting basin tillage systems
J 0.4.3. J Grain yield
The relationship between maize grain yields from conventional ploughing and planting
basin tillage systems is shown in Figure 10.7. The model predicted marginally higher (P
= 0.11) grain yield in the basin system than the conventional system at each nitrogen
level used (Fig.lO. 7). The model predicted mean grain yield of 945 kgha' without N
application, 1 112 kgha' at 10 kgblha' and 1 216 kgha' at 20 kgNha-1 under the
conventional ploughing system. At the recommended 52 kgblha" application, the model
predicted a mean yield of 1 296 kgha" over the 69 year period under the conventional
ploughing system. Under the basin tillage system, the model predicted mean grain yield
of 1 025 kgha' without nitrogen, 1 175 kgha" with 10 kgNha-1 and 1 280 kgha" at 20
kgNha-l. At the standard recommended 52 kgNha-l, the model predicted a mean grain
yield of 1 364 kgha' over the 69 year period. The additional maize yield benefits from
increasing N application rate could not have been realized because soil water was
limiting under semi-arid conditions in both the conventional ploughing and basin
systems. Other studies have shown that in drought-prone environments of Zimbabwe
crop response to fertilizer is highly dependent on seasonal rainfall distribution (Shumba
et al., 1992;Nyakatawa et al., 1996).
309
4500
4000
3500
-:J~e:n: 3000
C
"0
:-§ 2500
>-.
t::
.~ 2000
ril
.5
rroil 1500co
1000
500
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Conventionalplough grain yield(kgha")
Figure 10.7. Comparison of grain yield achieved in the conventional and basin tillage
systems at four N application rates (0, 10, 20 and 52 kgNha-l) on a granitic sandy soil
under semi-arid conditions
The model predicted near-normal grain yield of 563 to 1 435 kgha' without N, 508 to 1
761 kgha' at 10 kgNha-1 and 448 to 1 930 kgha' with 20 kgNha-1 under the conventional
ploughing system. At the recommended 52 kgNha-l, the model predicted near-normal
grain yield of 337 to 1 796 kgha' under the conventional ploughing system. Under the
basin system, the model predicted near-normal grain yield of 665 to I 536 kgha" at 0
kgNha-\ 604 to 1 801 kgha" with 10 kgNha-1 and 630 to 1 895 kgha' at 20 kgNha-l• The
predicted near-normal grain yield was 592 to 1 862 kgha' at recommended 52 kgNha-1
under the basin system. It is notable that there is little difference in modeled yield across
the N levels at yield values less than 1 500 kgha" which occurs 60 % of the time. The 33-
66 % maize yield ranges are narrower in the basin system at 10, 20 and 52 kgNha-1
310
compared to the conventional system. This implies that it is less risky to use these
nitrogen fertilizer levels in the basin tillage system than the conventional system.
1.0
.e 0.8
:g
..eD 0.6
0-
Q)
.;~> Plough
's 0.4
E
;::I
U 0.2- 1-- 0 - - - - 10 - - 20 -- 521
0.0
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Grainyield(kgha")
Figure 10.8.Cumulative distribution function for maize grain yield response to four N
application rates (0, 10, 20 and 52 kgblha") on a granitic sandy soil in the conventional
tillage system under semi-arid conditions
In 15 out of the 69 years the model predicted no grain yield at any of the N rates in the
conventional tillage system (Fig. 10.8). In the planting basin system 14 years would not
produce any grain regardless of the N treatment. This is consistent with observations
from on-farm experimentation in Gwanda and Insiza districts during 2006/07 and
2007/08 growing seasons when no maize yields were obtained on most smallholder farms
(Chapter 6; Chapter 7). Considering the extreme maize yields predicted by the model, the
highest predicted grain yield are 1 981 kgha" without N fertilizer, 2 562 kgha' with 10
kgNha-1 and 2 998 kgha' at 20 kgblha' for the conventional tillage system (Fig. 10.8).
The highest predicted grain yield at 52 kgNha-1 is 4 228 kgha' in the conventional
311
ploughing system. Under the basin system, the highest predicted grain yield was 2 090
kgha' with no nitrogen fertilizer, 2 661 kgha' at 10 kgNha-1 and 3 098 kgha" with 20
kghaNha-l. At the recommended 52 kgNha-l, the highest predicted grain yield is 4 223
kgha' in the basin system (Fig. 10.9).
1.0
g 0.8
~
..e0 0.6
0..
Q.) Basins
:>
'P
:(I;j 0.4
E
;:l
U
0.2 - 1-- 0 - - - - 10 - - 20 -- 521
0.0
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Grain yield (kgha-I )
Figure 10.9. Cumulative distribution function for maize grain yield response to four N
application rates (0. 10, 20 and 52 kgNha-I) on a granitic sandy soil in the basin tillage
system under semi-arid conditions
10.4.3.2 Surface runoff and deep drainage
The predicted surface runoff water losses before the start of and during the crop .growing
season are given in Figure 10.10. During the pre-cropping period the model predicted
higher (P < 0.05) runoff water losses from the conventional system compared to the basin
system. The 0.9 m x 0.6 m grid of basins creates surface depressions which help
capturing rainwater. However, during the cropping period runoff water losses are similar
in most years suggesting that the planting basins would have lost the ability to retain
312
rainwater. The effect of planting basin on surface runoff dissipates as the growing season
progresses. The planting basin structure in sandy soils was observed to collapse fast
depending on the frequency and intensities of rainstorms received (Chapter 7; Mupangwa
et al., 2008). Only four out of the 69 years had more predicted in-crop runoff water losses
from planting basins than conventional system. The predicted pre-erop runoff ranged
from 0.3 to 103 mm per year for the conventional tillage system with a mean value of 21
mm. Pre-erop predicted runoff ranged from 0.1 to 88 mm for basin tillage system with a
mean value of 15 mm. APSIM predicted more deep drainage water losses from the basin
tillage system compared to the conventional system during the cropping period (Fig.
10.11). The simulated deep drainage water losses ranged from 0 to 109 mm with a mean
of 19 mm in the conventional system. The basin system had 0 to 160 mm with a mean of
22mm.
9000 - - - Preplanting-Basin --Post-planting-basin
8000 - - Preplanting-plough - - - - Post plantingplough
.ê-.. 7000
'-'
ti::: 6000
oc
2 5000
Q)
.~ 4000
-a
ê 3000
;::l 2000
U
1000 __ -:: :- _-_ :- _---------- - .
o .r'" _.. __ ~ -:td ..
..-r::- """:'"---- ... - --
+---~~~~,-----,-----,-----,-----,-----,------
1930 1940 1950 1960 1970 1980 1990 2000 2010
Time(years)
Figure 10.10. Surface runoff water losses from the conventional ploughing and basin,
tillage systems on a granitic sandy soil under semi-arid conditions
313
1800
__ 1600 - - Basins --CP
..ê_, 1400
~ 1200
r
.c
o
.t..;. 1000
"'0
.-:<:1:)
800
ro
:; 600
E
;::s 400
U 200
0
1930 1940 1950 1960 1970 1980 1990 2000 2010
Time(years)
Figure 10.11. Deep drainage soil water losses from conventional ploughing and basin
tillage systems during the crop growing period under semi-arid conditions
10.5 Conclusion
The APSIM model was used in this study to predict the observed crop yield and profile
soil water and give a long term impact of the planting basin system on maize yield and
selected components of the water balance. For most of the seasons there was reasonable
agreement between observed and predicted maize yield data sets for the Matopos
experimental site (clay soil). However, for the sandy soil Lucydale site, APSIM under
predicted grain and biomass production under the model input conditions set suggesting
that the model still requires more calibration for the CA systems or rainwater harvesting
techniques such as planting basins and ripper that were tested together with mulching.
The model over predicted soil water in the conventional ploughing, ripper and basin
tillage systems regardless of the mulch level applied.
Long term simulations showed that maize productivity under semi-arid conditions can be
improved significantly through application of N fertilization in both conventional and
314
planting basins tillage systems. The predicted maize yield indicated that 0, 10, 20 and 52
kgNha-1 give no yield difference below 1 500 kgha'. The response to N fertilizer varies
from season to season depending on the rainfall characteristics of seasons. In years with
below average rainfall there are no responses to N fertilizer while significant yield
improvements are realized from N when rainfall is not a constraint. This modeling
exercise has shown that the planting basin technology has a potential of reducing water
and nutrient losses from smallholder crop production systems through reduced surface
runoff. The short and long term benefits of this could be improved soil water and nutrient
supply to crops during the growing season and the eventual build up of soil fertility on
smallholder farmers' fields. Complete crop failure was observed in some years meaning
that the smallholder farmer is still at the mercy of rainfall despite the use of planting
basin tillage system. The planting basin technology does not offer a complete protection
of the farmer against low and poorly distributed in-season rainfall.
315
CHAPTER 11
Summary and Recommendations
As this study started in the community with low maize production and erratic rainfall
one should plot the road travelled and lessons learnt so as to be able to select the
direction to head. This study can be used as a springboard for making changes in the
household food security in the semi-arid parts of Zimbabwe.
11.1 Smallholder Cropping Systems: The status quo
11.1.1 Rainfall pattern
The results from our study showed that total annual rainfall along the Bulawayo to
Beitbridge transect has not changed significantly over the past 50 - 74 years (Chapter
4). Our study has shown that total annual rainfall decreases from 584 to 376 mm as
one moves from Bulawayo (NR IV) to Beitbridge (NR V). The number of wet days
per growing season also decreases from 24 to 12 along the Bulawayo to Beitbridge
transect. The growing season is longer in NR IV than V and there is a higher chance
of 14 and 21 day dry spells from February onwards during the growing season.
Despite the insignificant changes in total annual rainfall, smallholder farmers in
. southern Zimbabwe continue to experience reduced crop yields or total crop failure as
a result of inadequate soil water supply to crops during the growing season. One of
the reasons can be that in-season rainfall distribution is having a bigger negative
effect on crop production than the total annual rainfall. The poor rainfall distribution
and the heavy rainstorms that are normally followed by long dry spells during the
growing season mean that smallholder fariners in the Mzingwane catchment need to
make quick decisions on when to plant or apply topdressing fertilizer soon after
316
rainfall event(s) before soil water is lost through evaporation. A false start of the
growing season is sometimes experienced in Gwanda and Insiza districts and this
further highlights the need for smallholder farmers to improve on the timing of
planting and have a 14-day weather forecast readily available.
In semi-arid districts of southern Zimbabwe some of the strategies that smallholder
farmers can use to reduce the impact of in-season rainfall variability include choosing
the right crop types and varieties that are recommended for NR IV and V. During the
three seasons of experimentation in Gwanda district farmers spoke highly of
traditional sorghum variety Lundende which they said is more drought tolerant than
the open pollinated varieties Macia, SV 2 and SV 4 which were introduced by
researchers. Staggering planting dates of maize, sorghum and pearl millet is one
strategy smallholder farmers in southern Zimbabwe can use to overcome the false
start of the growing season and poor rainfall distribution during the growing season.
This means that they stand a better chance of harvesting at least one of the planting
dates as they will be different stages of development during any dry spell and so crops
will be affected differently. Agronomic practices such as timely and effective
weeding, and use of soil fertility amendments in the smallholder cropping systems
improves the productivity of the rainfall received during the growing season.
11.1.2 Land size and management
Land sizes vary considerably in Gwanda and Insiza districts where the available
arable land per household ranges from less than a hectare in Insiza district to more
than 2 ha in Gwanda district (Chapter 6). Management of agricultural land by
smallholder farmers in Gwanda and Insiza districts involves conventional ploughing
317
at the onset of the rains during the October to December period. Ploughing is carried
out by donkey or ox drawn conventional plough and ploughing depth achieved ranges
between 0.1 and 0.15 m. Ploughing depth is shallower than that recommended by
AGRITEX of 0.23-0.25 m because draught animals are often in poor condition during
the October to December period. Poor condition of ploughs and the inappropriate
setting of the plough results in the shallow ploughing depths achieved by most
smallholder farmers. Although winter ploughing is a recommended practice for semi-
arid southern Zimbabwe, it is not being practiced by most farmers in the Mzingwane
catchment (Chapter 5).
The start of growing season stretches from November to January (Chapter 4),
therefore planting of cereals and legumes is done from November to January (Table
11.1). A typical cropping calendar for cereal and legume crops (bambara nut,
groundnut and cowpea) on smallholder farms in semi-arid southern Zimbabwe is
given in Table 11.1. In Insiza district, smallholder farmers plant maize soon after the
first effective rains while in Gwanda district maize, sorghum and pearl millet are
planted after the effective rains. Between three to six weeks after planting, depending
on household farm power resources, weeding is carried out using hand hoes or with
animal drawn cultivators (Chapters 5 and 6; Mbanje et al., 2001). Most often two to
three household members provide permanent labour for farming activities such as
planting and weeding (Chapter 6). The better resoureed households use ox/donkey
drawn cultivators for weeding and sometimes hire extra labour during peak periods.
Topdressing of cereal crops during January-February period is often delayed because
of inadequate soil water for fertilizer application as dry spells occur more frequently
during the January-February period (Chapter 4; Chapter 7; Chapter 9a).
318
Table 11.1. Typical cropping calendar for smallholder cropping systems of semi-arid
southern Zimbabwe
11.1.3 Livestock sub-system
A smallholder farm in Gwanda and Insiza districts is typified by close interaction and
exchanges between crop and livestock sub-systems. The predominant livestock
species are cattle, goats, donkeys and chickens (Chapters 5 and 6). Livestock numbers
vary considerably from household to household. A well resoureed household can own
about 40 cattle, 76 goats and 28 donkeys (Chapter 6). A poorly resoureed household
in both Gwanda and Insiza districts owns only chickens. Our study showed a general
decrease in the number of cattle, goats and donkeys per household from 2005 to 2008
(Chapter 6). These changes in numbers of cattle and donkeys imply households in
Gwanda and Insiza districts are facing more severe shortages of draught power for
ploughing and cultivation. The low resoureed households will plough and plant much
later given the increased shortage of draught power. There are differences in livestock
numbers owned by households in Gwanda and Insiza districts. The households in
Gwanda district (NR V) own more livestock than households in Insiza district (NR
IV). Cattle and donkeys are sources of draught power and access to draught animal
power enables cattle/donkey owners to prepare land and plant during November to
January period. Donkeys also play a crucial role in the transport system within both
the villages of Gwanda and Insiza districts. Livestock use crop residues from the
cropping sub-system as feed - these being either grazed in situ or taken to the
homestead soon after harvesting for systematic dry season feeding.
319
11.1.4 Crop sub-system
The predominant cereals grown during summer in southern Zimbabwe are maize,
sorghum and pearl millet. Smallholder farmers grow both traditional and hybrid cereal
varieties. The open pollinated sorghum varieties grown in Gwanda and Insiza districts
include Macia, SV 2 and SV 4. The open pollinated traditional sorghum varieties,
grown on a smaller scale than Macia, SV 2 and SV 4, include Lundende, Tsweta and
Nkotakota (Chapter 5). The majority of smallholder farmers in Gwanda and Insiza
districts of southern Zimbabwe grow traditional pearl millet varieties. Maize varieties
are derived from commercial seed houses such as Pannar and Pioneer. In most
seasons smallholder farmers plant maize seed retained from the previous season or
gain access to cereal and legume seed at seed fairs organised in conjunction with
NGOs and AGRITEX.
Legumes are grown to a lesser extent, with groundnuts and bambara nuts being the
predominant legume species. Legume production is constrained by lack of seed on
both the local (village level) and commercial markets. Therefore legumes are grown
on less than 10 % of cropped area during most seasons (Chapters 5 and 6). Lack of
legume seed on the local and commercial markets has promoted continued
monoeropping in both the conventional and CA systems that are being widely
promoted in southern Zimbabwe. This could be changed by increasing production of
legume seed through contract farming and/or seed multiplication programs organized
by NGOs and other agriculture research institutions.
320
11.1.5 Current soil water and fertility management
The only notable soil water management teclmiques being used on smallholder farms
are graded contours in Insiza district, and dead level contours and infiltration pits in
Gwanda district (Chapter 5; Chapter 8). The dead level contours and infiltration pits
are being promoted by NOO driven community development projects as the
construction of graded contours was mandatory during the pre-independenee era.
However, these graded contours have hardly been maintained in either Insiza or
Gwanda districts. The absence of in-field soil water management techniques on
smallholder farms in Gwanda and Insiza districts means there is scope for the wide
promotion of planting basins and ripper tillage systems. Smallholder farmers in Insiza
district (NR IV) already use both livestock manure and inorganic fertilizer (Chapters 5
and 6). In contrast, only a small proportion of smallholder farmers in Gwanda district
(NR V) apply manure and inorganic fertilizer. The strong perception that livestock
manure and inorganic fertilizer bum crops needs to be urgently addressed in order to
promote use of fertilizer and address the continuation of nutrient mining and low
productivity in the smallholder cropping systems of southern Zimbabwe.
11.2 Soil Water and Fertility Management: What Research has Developed!
11.2.1 In-situ water and soil fertility management
On-station and on-farm studies were conducted to explore the effect of teclmiques
such as minimum tillage and mulching on soil water dynamics and crop yields.
Planting basins reduced surface runoff water losses from cropped fields and were able
to collect more rain water at the beginning of cropping season (Chapter 7; Chapter
9a). The higher soil water status at the beginning of the growing season allowed
better crop establishment in the basin system compared with single and double
321
conventional ploughing, and ripper tillage systems. However, our study demonstrated
that planting basins on both clay (Matopos) and sandy (On-farm) soils can lead to
waterlogging during periods of the growing seasons that experience incessant rains.
This suppresses crop growth and could lead to significant yield reductions during
seasons when rainfall conditions are favourable to get a crop harvest.
Hand dug planting basins and double spring conventional ploughing, when combined
,
with nitrogen fertilizer (10-20 kglvha') and manure (3 tha") increased maize yields,
and water use efficiency under semi-arid conditions (Chapter 7). Our study
.demonstrated that double spring ploughing gives higher crop yield with similar soil
fertility inputs than single conventional ploughing, ripper and basin tillage systems
(Chapter 7). The on-farm study clearly highlighted the fact that both soil water and
fertility management are required at the same time in the smallholder cropping
systems of southern Zimbabwe (Chapter 7). The evaluation of the tillage systems
promoted over three growing seasons in Gwanda and Insiza districts revealed that
smallholder farmers had obtained higher crop yields from the double ploughing
system and those with draught animals are prepared to continue using the DP tillage
system (Chapter 6). However, the modelling exercise conducted in our study
indicated that in the long term total crop failures in seasons with below average
rainfall can be experienced under both conventional and basin tillage systems
(Chapter 10). The results from the modelling exercise (Chapter 10) and field
experimentation (Chapter 7; Chapter 9a) suggest that food deficits will continue as the
performance of the new technologies such as planting basins is dependant on the
seasonal rainfall pattern. It is critical to carry out a detailed study on in-season rainfall
variability patterns and this will assists farmers and researchers in modifying rain and
322
1- -
soil water management technologies that have been developed for semi-arid
conditions.
11.2.2 Inter-field soil water management
The current study assessed lateral soil water movement from the dead level contour
and infiltration pit into the cropped field (Chapter 8). The findings revealed that dead
level contours with infiltration pits capture more rainwater than dead level contours
only. Lateral soil water movement from the dead level contour with or without
infiltration pit only occur after receiving rainfall event of more than 40 mm. Lateral
soil water movement was detected 3 m downslope of the dead level contour with or
without infiltration pit (Chapter 8). The dead level contours fed soil water in the
middle layers of the 0 - 0.6 m profile studied. Strip cropping along the downslope
side of the contour may be a more appropriate way of using the rainwater collect,~d in
the contour as soil water only moves 3 m from the dead level contour. There were no
significant maize and pearl millet yield differences as distance from the contour
increased, suggesting no significant soil water differences particularly in seasons with
below average rainfall like 2006/07 and 2007/08.
H.3 What ne the Opportunities for Adopting Developed Technologies?
The soil water and fertility management options explored in our study aim at
promoting sustainable farming.
11.3.1 Crop yields
Crop yields can increase when smallholder farmers use double ploughing, ripper,
planting basins, mulching, crop rotation, inorganic and organic fertilizers (Chapters 7,
323
9a and 9b). Crop yields also increase if planting and weed control are done on time.
Increased crop yields are critical in sub-Saharan Africa where food deficits are
experienced even after growing seasons with above average rainfall (Chapter 6). The
use of planting basin system offers households without draught power an opportunity
to plant on time and consequently achieve crop yields similar to better resoureed
households (Chapters 7, 9a and 9b). The low resoureed households can get similar or
even better crop yields by using planting basins than households with draught power
in seasons with below average rainfall pattern, giving a lot of hope to vulnerable
households in semi-arid environments.
11.3.2 Soil health, water and fertility
In CA systems disturbing the soil only where the seed is to be planted allows physical
recovery of the soil on the undisturbed areas. Improvement in soil structure allows
better soil conditions for plant roots to be able to explore the whole soil profile for
water and nutrients. Minimum tillage and mulching promote increase in organic
carbon (Chapter 9c) and a proliferation of soil fauna and flora which play a crucial
role in nutrient cycling (Sprent, 2007). Instead of acting as emitters of carbon through
release of carbon dioxide, soils under conservation agriculture systems can be sinks of
carbon, absorbing carbon dioxide from the atmosphere (Mrabet, 2007; Smith, 2006).
Combining minimum tillage techniques with mulching reduces surface runoff and soil
evaporation, and improves infiltration of rainwater into the soil (Chapter 9c; Smith,
2006). The improved soil organic matter status and structure improve the water
holding capacity of the soil, potentially making more soil water available to crops.
Minimum tillage techniques such as hand dug planting basins and ripping allow more
324
efficient utilization of farming inputs such as seed, inorganic fertilizer and manure.
For example in planting basin system only 10 % of the total land area is dug to insert
the seed and fertilizer/manure placement compared to the whole area under the
conventional tillage system. Smallholder farmers can therefore apply manure to a
large land area through the use of the planting basin system. In addition soil fertility
builds up at the permanent planting positions as soil fertility amendments are not
spread over the whole field. Soil productivity will eventually improve if the farmers
continue to use the planting basin system leading to increased household food
security.
11.3.3 Farm labour and energy
Conservation agriculture techniques spread labour demand over a longer time period
compared to the conventional farming system. The spread of farm labour enables
farmers to carry out the operations such as land preparation, planting.and weeding by
themselves and on time (Chapter 6). By praeticing minimum tillage farmers can
reduce time and labour requirements by 30 to 40 % in subsequent years compared to
conventional systems (Smith, 2006). In the face of high fossil fuel prices commercial
farmers who use mechanized agriculture can substantially reduce their fuel
requirements by adopting conservation agriculture practices.
U.4 What are the Constraints to Adoption of the Technolegles Developed?
Technologies such as planting basins have high labour demand initially for land
preparation and weeding. Farmers normally complain that digging basins is
backbreaking in all seasons and weed pressure unbearable particularly in relatively
wet seasons. Individual households normally have two to three labour persons
325
(Chapter 6) and they cannot cope with the labour demand during peak periods. In
Gwanda and Insiza districts, our study revealed a slight decrease in the number of
people per household between the years 2005 and 2008 (Chapter 6).
New equipment is obviously required when a farmer adopts some of the new
technologies. Farmers might find it difficult to plant and weed such fields using
traditional hand hoes as observed at Matopos research site during the four seasons of
experimentation (Chapters 9a and 9b). In the hyper-inflationary environments
prevailing in sub-Saharan countries such as Zimbabwe, availability of agricultural
inputs such as fertilizer and seed could be a major stumbling block to adoption of
sustainable farming methods. The poor marketing system and low prices offered for
agricultural produce by GMB could further curtail adoption of promising technologies
by smallholder farmers.
The introduction of new technologies also requires a change of the mindset of a
fanner who might consider new ways of farming more risky. The farmer's mindset
could further be aggravated if new technologies are not well explained to them.
However, the farmers' confidence could grow if they are accepted as equal partners
and experts in their own environment (Steiner; 2007). Lack of technical support from
poorly resoureed agricultural extension departments, like is the case in Zimbabwe
.now, could retard the adoption of promising technologies by smallholder farmers.
H.5 The way forward!
A wide range of promising technologies are already available for uptake by the
smallholder farmers. These technologies need to be promoted on a wider scale by all
326
stakeholders involved in research and development. During the promotion process
NGOs and researchers, in conjunction with experienced extension staff, should
continue training new farmer groups and extension officers. During the training
sessions producers of agriculture inputs such as seed and fertilizers should be invited
to participate so that they become aware of what is expected of them in the
development of smallholder agriculture. A study by ICRISAT and the Zimbabwe
Fertilizer Company (ZFC) showed that smallholder farmers now prefer the 10-20 kg
fertilizer pack sizes to the traditional 50 kg which has become difficult to transport
because of the poor transport system in rural areas. The provision of extension
materials such as manuals, flyers and bulletins translated into vernacular is critical for
adoption of promising technologies. These extension materials should continue to be
developed by researchers, extension and NGOs, and distributed well in advance
before the onset of the rainy season.
Capacity building of farmers is critical for the uptake of the promising technologies
that still lie 'idle' in the farmers' basket of options. At some point during the wide
scale promotion of the promising technologies, farmer to farmer extension could play
a crucial role with the extension and research agents playing a facilitator's role in the
process. For example in all districts in Zimbabwe where CA systems are being
promoted, there are farmers who have been hosting CA demonstrations for two to
three seasons. These experienced farmers could lead farmer groups in experimenting
with the technologies such as planting basins, ripper, the use of manure and inorganic
fertilizer. During the experimentation process the farmers could make adjustments to
certain components of the technologies to suit their circumstances. They might even
start with just a few components of a technology and gradually include all the other
327
aspects as they gain experience and other necessary resources. For example the CA
systems being promoted in semi-arid districts of Zimbabwe encourage the use of
mulch (Chapters 9a and 9b) in an environment where livestock plays a bigger role in
the livelihoods of farming families. Mulching can be adopted at a later stage when
more crop biomass is being produced. Smallholder farmers can even test the
performance of their traditional crop varieties in CA systems especially now when
seed availability is one of the major challenge in the smallholder farming sector. Field
days and exchange visits provide an opportunity for agriculture information
dissemination as extension, researchers and producers of agriculture inputs interact
with farmers and other stakeholders involved in agriculture development.
U.6 Future research
The analysis of 50 to 74 year rainfall data revealed that there is no significant decline
in total annual rainfall nor any changes in the start and end time of growing seasons in
the semi-arid southern part of Zimbabwe (Chapter 4). However, the mid-season dry
spells continue to cause reduced yields and total crop failure in the smallholder
farming sector. There is need for a detailed study on within season rainfall variability
in order to come up with more robust ways of reducing the risk of crop failure in
smallholder agriculture of the semi-arid environments. The droughts and dry spells
that continue to be experienced in southern Zimbabwe point to the need for making
more use of seasonal forecasts that are issued every September before the onset of the
rains. The seasonal forecast information should get to the farmers in a simplified
format and well before the onset of the rains so that they can use it to make
appropriate farming decisions together with the type of information found in this
thesis.
328
The study on the tillage and mulching effects on maize, cowpea and sorghum crop
production conducted at Matopos and Lucydale sites used a blanket 20 kgha"
nitrogen rate. There is need to consider a range of nitrogen application rates as the 20
kgha' might not be adequate in some seasons particularly where immobilization of
the applied nitrogen takes place because of the surface applied mulch. The maize
stover applied as mulch adds a lot of carbon that can result in N immobilization such
that farmers fail to benefit from the use of the mulch in their cropping system.
There is need to carry out a more detailed study on soil water dynamics from dead
level contours with current design of the contours and also with other possible
designs. The results from this study on dead level contours and infiltration pits
indicated lateral water movement from the contours after rainstorms of more than 40
mm. It might be useful to determine the threshold rainstorm sizes for meaningful
lateral movement of soil water into the field from dead level contour with or without
infiltration pit. There is need to determine the effect of different sizes of the
infiltration pit and positions along the dead level contour. Information generated from
such studies can be used in future modification of the current design of the dead level
contours and infiltration pits. The results from simulation of the long term impact of
planting basins on crop productivity in smallholder systems suggest that the APSIM
model still requires more rigorous calibration for the planting basin tillage system.
329
References
Abiven, S. and Recous, S. 2007. Mineralization of crop residues on the soil surface or
incorporated in the soil under controlled conditions. Biology and Fertility of Soils,
43: 849-852.
AGRITEX (undated). A guide for farmers on good land husbandry. 64pp.
www.uz.ac.zw/agriculture/cropscience/CD/gff/handplanting.pdf.
AGRITEX (Department of Agricultural Technical and Extension Services). 1990. Agro-
climatology Applied to Crop Production in Zimbabwe. Department of
Agricultural Technical and Extension Services, Ministry of Lands, Agriculture
and Rural Resettlement. Harare. 26pp.
Ahmed, M.M., Rohrbach, D.D., Gono, L., Mazhangara, E., Mugwira, L., Masendeke,
D.D. and Alibaba, S. 1997. Soil fertility management in communal areas of
Zimbabwe: Current practices, constraints and opportunities for change. Results
of a diagnostic survey. Southern and Eastern African Region. Working Paper
N°6.30pp.
Alvord, E.D. 1936. Annual report to the agriculturist. Native Department, Harare. 35pp.
Al-Qinna, M. 1. and Abu-Awwad, A. M. 2001. Wetting patterns under trickle source in
arid and semi-arid soils with surface crust. Journal of Agricultural Engineering
Research, 80 (3): 301-305.
Amato, M. and Ritchie, J.T. 2002. Spatial distribution of roots and water uptake of maize
(Zea mays L.) as affected by soil structure. Crop Science, 42: 773-780.
Anderson, LP., Brinn, P.J., Moyo, M. and Nyamwanza, B. 1993. Physical Resource
Inventory of Communal Lands of Zimbabwe. Natural Resources Institute
Bulletin 60, London. 186pp.
Anderson, J.M. and Ingram, J.S.L 1993. Tropical Soil Biology and Fertility. A Handbook
of Methods. 2nd Edition. C.A.B. International, Wallingford, UK, 221pp.
Anjichi, V.E., Mauyo, L.W. and Kipsat, MJ. 2007. The effect of socio-economic factors
on a farmer's decision to adopt farm soil conservation measures. An application
of the multivariate logistic analysis in Butere/Mumias district, Kenya. pp915-
919. In: Bationo, A., Waswa, B., Kihara, J. and Kimetu, J. (eds.). Advances in
330
Intergrated Soil Fertility Management in Sub-Saharan Africa. Springer
Publishers, Dordrecht, The Netherlands.
Archer, J. 1988. Crop Nutrition and Fertiliser Use. Farming Press Ltd. Suffolk
Arshad, M.A., Franzluebbers, A.J. and Azooz, R.H. 1999. Components of surface soil
structure under conventional and no-tillage in northwestern Canada. Soil and
Tillage Research, 53: 41-47.
Aviad, Y., Kutiel, H. and Lavee, H. 2004. Analysis of beginning, end and length of the
rainy season along a Mediterranean arid climate transect for geomorphic
purposes. Journal of Arid Environments, 59: 189-204.
Azooz, R.H. and Arshad, M.A. 1996. Soil infiltration and hydraulic conductivity under
long term no-tillage and conventional tillage systems. Canadian Journal of Soil
Science, 76: 143-152.
Barron, J. and Rockstrom, J. 2003. Water Harvesting to Upgrade Smallholder Farming:
Experiences From On-farm Research in Kenya and Burkina Faso.
RELMA/SIDA, Stockholm University, Sweden.
Barron, J. 2004. Dry spell mitigation to upgrade semi-arid rainfed agriculture: Water
harvesting and soil nutrient management for smallholder maize cultivation in
Machakos, Kenya. Ph.D Thesis in Natural Resources Management. Department
of Systems Ecology. StockholmUniversity. Sweden.
Baudeon, F., Mwanza, H.M., Triomphe, B. and Bwalya, M. 2007. Conservation
agriculture in Zambia: a case study of Southern Province. Nairobi. African
Conservation Tillage Network, Centre de Coopération Internationale de
Recherche Agronomique pour le Développement, Food and Agriculture
Organization of the United Nations. 57pp. www.fao.org.
Belder, P., Twomlow, S. and Hove, L. 2007. Early evidence of improved soil quality with
conservation farming under smallholder farming conditions in Zimbabwe. Paper
presented at the ICID conference, November 2007, Johannesburg, South Africa.
15pp.
Bescansa, P., Imaz, M. J., Virto, I., Enrique, A. and Hoogmoed, W. B. 2006. Soil water
retention as affected by tillage and residue management in semi-arid Spain. Soil
and Tillage Research, 87: 19-27.
331
Bhattacharyya, R., Prakash, V., Kundu, S. and Gupta, H.S. 2006. Effect of tillage and
crop rotations on pore size distribution and soil hydraulic conductivity in sandy
clay loam soil of the Indian Himalayas. Soil and Tillage Research, 86: 129-140.
Bindraban, P.S., Stoorvogel, J.I., Jansen, D.M., Vlaming, J. and Groot, JJ.R. 2000. Land
quality indicators for sustainable land management: Proposed method for yield
gap and soil nutrient balance. Agriculture, Ecosystems and Environment, 81 (2):
103-112.
Bissett, M.I. and O'Leary, GJ. 1996. Effects of conservation tillage and rotation on water
infiltration in two soils in south-eastern Australia. Australian Journal of Soil
Research, 34: 299-308.
Bosch, H, van den., Gitari, J.N., Ogaro, V.N; Maobe, S. and Vlaming, J. 1998.
Monitoring nutrient flows and economic performance in African farming
systems (NUTMON) 111. Monitoring nutrient flows and balance in three
districts in Kenya. Agriculture, Ecosystems and Environment, 71(1/3): 63-80.
Botha, J.J., van Rensberg, L.D., Anderson, JJ., Hensley, M., Macheli, M.S., van Staden,
P.P., Kundhlande, G., Groenewald, D.G. and Baiphethi, M.N. 2003. Water
conservation technologies on small plots in semi-arid areas to enhance rainfall
use efficiency, food security and sustainable crop production. WRC Report N°
1176/1/03. 338pp.
Bouman, A. B. M., 2007. A conceptual framework for the improvement of crop water
productivity at different spatial scales. Agriculture Systems, 93: 43-60.
Bruneau, P.M.C. and Twomlow, SJ. 1999. Hydrological and physical responses of a
semi-arid sandy soil to tillage. Journal of Agricultural Engineering Research,
72: 385-391.
Cakir, R., 2004. Effect of water stress at different development stages on vegetative and
reproductive growth of corn. Field Crops Research, 89: 1-16.
Carsel, R.F. and Parrish, R.S. 1988. Developing joint probability distributions of soil
water retention characteristics. Water Resources Research, 24: 755-769.
Carter, M.R. 1996. Characterization of soil physical properties and organic matter under
long-term primary tillage in a humid climate. Soil and Tillage Research, 38:
251-263.
332
Chakraborty, D., Nagarajan, S., Aggarwal, P., Gupta, V.K., Tomar, R.K., Garg, R.N.,
Sahoo, R.N., Sarkar, A., Chopra, U:K., Sarma, K.S.S. and Kalra, N. 2008.
Effect of mulching on soil and plant water, and the growth and yield of wheat
(Triticum aestivum L.) in a semi-arid environment. Agricultural Water
Management, doi: I0.1 016/j.agwat.2008.06.001.
Chibudu, C., Chiota, G., Kandiros, E., Mavedzenge, B., Mombeshora, B., Mudhara, M.,
Murimbarimba, F., Nasasara, A. and Scoones, I. 2001. Soils, livelihoods and
agricultural change: The management of soil fertility in the communal lands of
Zimbabwe. pp116-163. In: Scoones, I. (ed). Dynamics and Diversity: Soil
Fertility and Farming Livelihoods in Africa. Earthscan Publication Ltd.,
London.
Chibulu, B. 2007. Effect of rainfall variability on crop yield under semi-arid conditions at
sub-catchment level. M.Sc. Thesis. Department of Civil Engineering, University
of Zimbabwe, Harare. 77pp.
Chiduza, C., 1995. Analysis of rainfall data and their implication on crop production: A
case of Northern Sebungwe. Zimbabwe Journal of Agricultural Research, 33
(2): 175-189.
Chiduza, C. 1997. Effect of between-row and within-row spacing on yield of sorghum
variety SV 2 in the Zambezi valley of Zimbabwe. Zimbabwe Journal of
Agricultural Research, 35(2): 129-140.
Chikowo, R., Mapfumo, P., Nyamugafata, P. and Giller, G.E. 2004. Maize productivity
and mineral N dynamics following different soil fertility management practices·
on a depleted sandy soil in Zimbabwe. Agriculture, Ecosystems and
Environment, 102: 119-131.
Chuma, E. 1993. Effects of tillage on erosion-related soil properties of a sandy soil in
semi-arid Zimbabwe. In Kronen, M. (Ed.). Proceedings of the Fourth Annual
Scientific Conference, SADC Land and Water Management Research
Programme, SACCAR, Gaborone, Botswana.
Chuma, E. and Haggmann, J. 1998. Development of conservation tillage techniques
through combined on-station and participatory on-farm research. In: Blume, H-
P.; Eger, H.; Flieischhauer, E; Hebel, A.; Reij, C and Steiner, K.G. (Eds.).
333
Towards Sustainable Land Use. Furthering Cooperation Between People and
Institutions Vol. II. Advances in Geoecology 31. Catena Verlag GMBH
Reiskirchen, Germany.
Christens en, J.H., Hewitson, B., Busuioc, A., Chen, A., Gao, X., Held, I., Jones, R.,
Kolli, R.K., Kwon, W.-T., Laprise, R., Magafia Rueda, V., Mearns, L.,
Menéndez, C.G., Raisánen, J., Rinke, A., Sarr, A. and Whetton, P. 2007.
Regional climate projections. In: Solomon, S., D. Qin, M. Manning, Z. Chen,
M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller (Eds.). Climate Change
2007: The Physical Science Basis. Contribution of Working Group I to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change
Cambridge University Press, Cambridge, United Kingdom and New York, NY,
USA.
CIMMYT, 1993. Annual Report November 1992 - October 1993. CIMMYT-Zimbabwe.
57pp.
Clay, E., Bohn, L., de Armas, E.B., Kabambe, S. and Tchale, H. 2003. Malawi and
southern Africa: Climatic variability and economic performance. Disaster
Management Facility. Working Paper NQ 7. World Bank, Washington DC.
March 2003. 112pp.
CONTILL, 1998. Conservation tillage for sustainable crop production systems.
Documented Outputs of the AGRITEX/GTZ Project. Annotated bibliography
1988-1996. Institute of Agricultural Engineering, Harare.
Cooper, P.l.M., Dimes, J., Rao, K.P.C., Shapiro, B. Shiferaw, B. and Twomlow, S. 2008.
Coping better with current climatic variability in the rain-fed farming systems of
sub-Saharan Africa: An essential first step in adapting to future climate change?
Agriculture, Ecosystems and Environment, 126: 24-35.
Das, D. K. and Chopra U. K. 1988. Soil water movement, balance and its use by crops as
influenced by soil physical conditions. pp231-254. In: Somani L. L. (ed.). Soil
physical conditions and crop growth, Monograph. Geo-Environ Academia and
Divyajyoti Prakashan, Shatri Nagar, India. 254pp.
Decagon Devices 2007. Minidisk infiltrometer. User's Manual, version 4. 18pp.
334
Defoer, T., Budelman, A., Toulmin, C., and Carter, S.E. 2000. Building Common
Knowledge. A Resources Guide for Participatory Learning and Action
Research. In: Defoer, T. and Budelman, A. (Eds.), Managing Soil Fertility in the
Tropics: A Resource Guide for Participatory Learning and Action Research.
Royal Tropical Institute (KIT), Amsterdam, The Netherlands. 208pp.
Defoer, T. 2002. Learning about methodology development for integrated soil fertility
management. Agricultural Systems, 73: 57-81.
Delve, R.l. and Probert, M.E. 2004. Modelling nutrient management in tropical cropping
systems. Australian Centre for International Agricultural Research (AClAR),
Canberra. ACIAR Proceedings, NQ 114. 138pp. www.aciar.gov.au.
Dennett, M.D. 1987. Variation of rainfall: the background to soil and water management
in dryland regions. Soil, Use and Management, 3(2): 47-51.
Dexter, A.R. 2004. Soil physical quality. Part 1. Theory, effects of soil texture, density
and organic matter, and effects on root growth. Geoderma, 120: 201-214.
Dhliwayo, H.H., Mabasa, S., Twomlow, SJ. and Riches, c.R. 1995. The effects of
weeding methods and water conservation on weed populations in dryland
maize. Brighton Crop Protection Conference .: Weeds, November 1995.
Farnham, Surrey, UK: BCPC., 1: 207-212.
Diagana, B. 2003. Land degradation in sub-Saharan Africa: What explains the
widespread adoption of unsustainable farming practices? Department of
Agricultural Economics, Montana State University, Bozeman, USA.
http://www .tradeoffs.montana.edu/pdf/SD- BD.pdf.
Dimes, J. and Malherbe, J. 2006. Climate variability and simulation modeling -
challenges and opportunities. In: Mgonja, M.A., Waddington, S., Rollin, D. and
Masenya, M. (eds.). Livelihoods in the Limpopo: CGIAR Challenge Program
on Water and Food Project N° 1. Increased food security and income in the
Limpopo basin through intergrated crops, soil fertility and water management
options and links to markets.Proceedings of the CPWFPN 1 Inception
workshop, 25-27 January 2005. Polokwane, South Africa. 132pp.
Dixon, W.J. and Massey, F.J. 1983. Introduction to Statistical Analysis. McGraw-Hill
International Book Co. Ltd., Japan. 678pp.
335
Dryland Fanning Research Scheme (DLFRS) Phase III. 1985. Final report, 6. Digest of
research findings. Overseas Development Administration, London, UK.
ECAF, 1999. Conservation agriculture in Europe: Environmental, Economic and Policy
Perspectives. Report Nol. European Conservation Agriculture Federation. 24pp.
http://www.ecaf.org.
Ehlers, W., 1975. Observations on earthworm channels and infiltration on tilled and
untilled loess soil. Soil Science, 119: 242-249.
Erenstein, O. 2002. Crop residue mulching in tropical and semi-tropical countries: An
evaluation of residue availability and other technological implications. Soil and
Tillage Research, 67: 115-133.
FAO. 1998. World Reference Base for Soil Resources. World Soil Resources Report 84.
Rome.
FAO, 2006. Fertilizer use by crop in Zimbabwe. Land and Plant Nutrition Services. Land
and Water Development Division. Food and Agriculture Organisation of the
United Nations, Rome.
Fatondji, D., Martius, C., Bielders, C.L., Vlek, P.L.G., Bationo, A. and Gerard, B. 2006.
Effect of planting technique and amendment type on pearl millet yield, nutrient
uptake and water use on degraded land in Niger. Nutrient Cycling in
Agroecosystems, 76: 203-217.
Freund, J.E. 1979. Modern Elementary Statistics. 5th Edition. Prentice-Hall, INC., New
Jersey.5l0pp.
Friedrich, T. and KienzIe, J. 2007. Conservation .agriculture: Impact on fanners'
livelihoods, labour, mechanization and equipment. pp25-36. In: Stewart, B.A.,
Asfary, A.F., Belloum, A., Steiner, K. and Friedrich, T. (eds.). Proceedings of
International Workshop on Conservation Agriculture for Sustainable Land
Management to Improve the Livelihood of People in Dry Areas. 7-9 May 2007,
Damascus, Syria.
Gaiser, T., de Barros, I., Lange, F.M. and Williams, J.R. 2004. Water use efficiency of a
maize/cowpea intererop on a highly acidic tropical soil as affected by liming
and fertilizer application. Plant and Soil, 263(1): 165-171.
336
Gicheru, P.T., Gachene, C.K.K., Mbuvi, lP. and Wanjogu, S.N. 2003. Effects of soil
management practices and tillage systems on soil water conservation and maize
yield on a sandy loam in semi-arid Kenya. pp18-24. In: Beukes, D., de Villiers,
M., Mkhize, S., Sally, H. and van Rensberg, L. (eds.). Proceedings of the
Symposium and Workshop on Water Conservation Technologies for Sustainable
Dryland Agriculture in sub-Saharan Africa (WCT). Bloemfontein, South Africa.
8-11 Apri12003.
Giller, K.E., Cadisch, G., Ehaliotis, C., Adams, E., Sakala, W. and Mafongoya, P.L.
1997. Building soil nitrogen capital in Africa. ppI51-192. In: Buresh, J.R.,
Sanchez, P.A. and Calhoun, F. (eds.). Replenishing Soil Fertility in Africa. Soil
Science Society of America Special Publication N° 51.
Giller, K.E., Witter, E., Corbeels, M. and Tittonell, P. 2008. Conservation agriculture and
smallholder farming in Africa: The heretic's view. Paper submitted to the Field
Crops Research journal, 10/7/2008.
Glab, T. and Kulig, B. 2008. Effect of mulch and tillage system on soil porosity
under wheat (Triticum aestivum L.). Soil and Tillage Research, 99: 169-178.
Grant, P., Meikle, G.J. and Mills, W.R. 1979. A comparison of plough types and depths
of annual ploughing for maize monculture with varied manuring. Rhodesian
Journal Agricultural Research, 17: 99-123.
Grant, P.M. 1981. The fertilization of sandy soils in peasant agriculture. Zimbabwe
Agricultural Journal, 78: 169-175.
Gregory, P.J. 1984. Water availability and crop growth in arid regions. Outlook on
Agriculture, 13(4): 208-215.
Guzha, A. C. 2004. Effect of tillage on soil microrelief surface depression storage and
soil water storage. Soil and Tillage Research 76: 105-114.
Hadas, A., Rawitz, E., Etkin, H. and Margolin, M. 1994. Short-term variations of soil
physical properties as a function of the amount of CIN ratio of decomposing
cotton residues. I. Soil aggregation and aggregate tensile strength. Soil and
Tillage Research, 32: 183-198.
337
Haggblade, S. and Tembo, G. 2003. Conservation farming in Zambia. Conference paper
NQl1. p18. Paper presented at the IFPRI, NEPAD, CTA conference, Successes
in African Agriculture. Pretoria, South Africa. 1-3 December 2003.
Haggmann, J. 1994. Lysimeter measurements of nutrient losses from a sandy soil under
conventional-till and ridge-till. pp305-310. In: Jensen, B.E., Schjonning, P.,
Mikkelsen, K.B. (Eds.). Soil Tillage for Crop Production and Protection of the
Environment. Proceedings of the 13th International Conference, International
Soil Tillage Research Organisation (lSTRO), Aalbourg, Denmark.
Hall, A.E. 2004. Breeding for adaptation to drought and heat in cowpea. European
Journal of Agronomy, 21: 447-454.
Halvorson, A.D., Wienhold, B.J. and Black, A.L. 2002. Tillage, nitrogen, and cropping
system effects on soil carbon sequestration. Soil Science Society of America
Journal, 66: 906-912.
Hanks, R.J. and Ashcroft, G.L. 1980. Applied soil physics: soil water and temperature
applications. Advanced Series in Agricultural Sciences 8. Springer-Verlag.
159pp.
Hatfield, J. L., Sauer, T. J. and Prueger, J. H. 2001. Managing soils to achieve greater'
water use efficiency: A review. Agronomy Journal, 93: 271-280.
Hayes, M.J., Svoboda, M.D., Wilhite, D.A. and Vanyarkho, O.V. 1999. Monitoring the
1996 drought using the Standardized Precipitation Index. BAMS, 80: 429-438.
Heal, O.W., Anderson, J.M. and Swift, MJ. 1997. Plant litter quality and decomposition:
A historical overview. pp3-30. In: Cadisch, G. and Giller, K. (eds.). Driven by
nature: Plant litter quality and decomposition. CAB Publishing, Wallingford.
Heinrich, G. 1989. Report on Farmer Testing Groups (ATlP), Francistown, 1988-1989.
Presented at Annual Crops Seminar, September 1989, Sebele, Botswana.
Henao, J. and Baanante, C. 2006. Agricultural production and soil mining in Africa:
Implications for resource conservation and policy development. International
Fertilizer Centre for Soil Fertility and Agricultural Development. Muscle
Shoals, Alabama, USA. http://www.africafertilizersummit.org.
338
Hikwa, D., Nyathi, P., Mugwira, L.M., Mudhara, M. and Mushambi, C.F. (Eds.). 2001.
Integrated Soil Fertility Development for Resource-poor Farmers in Zimbabwe:
The Research and Development Strategy Beyond 2001. Department of Research
and Specialist Services (DRSS). Harare. 48pp.
Holland, I.M. 2004. The environmental consequences of adopting conservation tillage in
Europe: Reviewing evidence. Agriculture, Ecosystems and Environment, 103:
1-25.
Homann, S., van Rooyen, A., Moyo, T. and Nengomasha, Z. 2007. Goat production and
marketing: Baseline Information for Semi-Arid Zimbabwe. International Crops
Research Institute for the Semi-Arid Tropics (ICRISAT), Bulawayo, Zimbabwe.
77pp.
Hove, L. 2006. Agricultural technology transfer under Relief and Recovery Programs in
Zimbabwe: Are NGOs Meeting the Challenge? Briefing Note NQ 6. ICRISAT-
Bulawayo. September 2006.
Hudson, D.A. and Jones, R.G. 2002. Simulation of present day and future climate over
southern Africa using HadAM3H. Hadley Centre Technical Note 38. 37pp.
Hughes, O. and Venema, I.H. (eds.). 2005. Intergrated soil, water and fertility
management in semi-arid Zimbabwe. Farmer Field Schools Manual, vol. 1.
Harare, Zimbabwe:FAO. ftp://ftp.fao.org/agl/agll/docs/ffsfm zim.pdf.
Hulme, M. and Sheard, N. 1999. Climate change scenarios for Zimbabwe. Climatic
Research Unit, Norwich, UK. 6pp. http://www.cru.uea.ac.uk.
Hussein, J. 1987. Agro-climatological analysis of growing season in natural regions Ill,
IV and V of Zimbabwe. In: AGRITEXlGTZ. Proceedings of a Workshop on
Cropping in Semi-arid Areas of Zimbabwe. 24-28 August 1987, Harare.
AGRITEX/GTZ.
Jackson, IJ. 1977. Climate, Water and Agriculture in the Tropics. Longman Inc., New
York. 248pp.
Kanemasu, E.T., Stewart, JJ., van Donk, S.J. and Virmani, S.M. 1990. Agroclimatic
approaches for improving agricultural productivity in semi-arid tropics.
Advances in Soil Science, 13: 273-309.
339
Keating, B.A., Carberry, P., Hammer, G.L., Probert, M.E., Robertson, MJ., Holzworth,
D., Huth, N.l., Hargreaves, J.N.G., Meinke, H., Hochman, Z:, McLean, G.,
Verbur, K., Snow, V., Dimes, J., Silburn, M., Wang, E., Brown, S., Bristow,
K.L., Asseng, S., Chapman, S., McCown, R.L, Freebairn, D.M. and Smith, C.l.
2003. An overview of APSIM, a model designed for farming systems
simulation. European Journal of Agronomy, 18: 267-288.
Klaij, M.C. and Vauchad, G. 1992. Seasonal water balance of a sandy soil in Niger
cropped with pearl millet, based on profile moisture measurements. Agricultural
Water Management, 21: 313-330.
Klute, A. 1986. Water retention: Laboratory methods. In Klute, A. (Ed.). Methods of Soil
Analysis, Part 1. American Society of Agronomy, Inc., Madison.
Konig, H. and Varma, A. 2006. Intestinal Micro-organisms of Termites and Other
Invertebrates. Springer Publishers. 483pp.
Le Barbel, L., Lebel, T. and Tapsoba, D. 2002. Rainfall variability in West Africa during
the years 1950-90. Journal of Climatology, 15: 187-202.
Lee, B. 1993. Escaping From Hunger: Research to Help Farmers in Semi-arid Kenya to
Grow Enough Food. ACIAR Monograph N°. 23. ACIAR. 52pp.
Liebig, M.A., Tanaka, D.L. and Wienhoid, BJ. 2004. Tillage and cropping effects on soil
quality indicators in the Northern Great Plains. Soil and Tillage Research,
78:131-141.
Lipic, J., Kus, l., Slowinska-Jurkiewicz, A. and Nosalewicz, A. 2005. Soil porosity and
water infiltration as influenced by tillage methods. Soil and Tillage Research,
89: 210-220.
Love, D., Twomlow, S., Mupangwa, W., Van der Zaag, P. and Gumbo, T. 2006.
Implementing the millennium development food security goals - Challenges of
the southern African context. Physics and Chemistry of the Earth, 31: 731-737.
Love, D., Uhlenbrook, S., Twomlow, S. and van der Zaag, P. 2008. Changing rainfall and
discharge patterns in the northern Limpopo Basin, Zimbabwe. European
Geophysical Union General Assembly, Viena, Austria, April 2008.
http://www .worldwaterweek/stockholmwatersymposium/ Abstract volume 06/
workshop 9.htm.
340
Maddonni, O.A., Otegui, M.E. and Bonhomme, R. 1998. Grain yield components in
maize. II. Post silking growth and kernel weight. Field Crops Research, 56: 257-
264.
Mandiringana, O.T., Mabi, M. and Simalenga, T.E. 2003. The potential of three water
conservation technologies for adoption and use by communal farmers in Eastern
Cape. pp56-59. In: Beukes, D., de Villiers, M., Mkhize, S., Sally, H. and van
Rensberg, L. (eds.). Proceedings of the Symposium and Workshop on Water
Conservation Technologies for Sustainable Dryland Agriculture in sub-Saharan
Africa. Bloemfontein, South Africa. 8-11 Apri12003.
Mapfumo, P. and Mtambanengwe, F. 1998. Nutrient mining in maize-based systems of
rural Zimbabwe. In: EARO/CIMMYT, 1998. Maize Production Technologies
for the Future: Challenges and Opportunities. The Sixth Eastern and Southern
Africa Regional Maize Conference held in Addis Ababa, Ethiopia. 21-25
September 1998. Program and Abstracts. EARO/CIMMYT.
Mapfumo, P. and Giller, K.E. 2001. Soil fertility management strategies and practices by
smallholder farmers in semi-arid areas of Zimbabwe. International Crops
Research Institute for the Semi Arid Tropics (ICRISAT) and Food and
Agriculture Organization of the United Nations (FAO), Bulawayo, Zimbabwe.
60pp.
Masikati, P. 2006. Tillage and manure interactions under dryland cropping in semi-arid
Zimbawe. Department of Soil Science and Agricultural Engineering, University
of Zimbabwe, Harare, Zimbawe. M.Phil. thesis. 101pp.
Masvaya, E., Mupangwa, W. and Twomlow, S. 2008. Rainfall variability impacts on
farmers' management strategies and crop yields. Water and sustainable
development for improved livelihoods. 9th WaterNet/WARFSAlGWP-SA
symposium, 29-31 October 2008, Johannesburg, South Africa.
Mbanje, E., Twomlow, S. L, and O'Neill, D. H. 2001. Evaluation of animal-drawn
weeders for smallholder maize production in Zimbabwe. In "Weeds 2001"
Proceedings of the BCPC Conference, 12-15 November 2001, Brighton, UK, 2:
913-918.
341
Mellis, D.A, Bruneau, P.M.C., Twomlow, S.J. and Morgan, R.P.C. 1996. Field
assessment of crusting on tilled sandy clay loam. Soil, Use and Management,
12: 72-75.
Moreno F., Pelegrh F., Fernindez J.E. and Murillo, J.M. 1997. Soil physical properties,
water depletion and crop development under traditional and conservation tillage
in southern Spain. Soil and Tillage Research, 41: 25-42.
Moroke, T.S., Schwartz, R.C., Brown, K.W. and Juo, AS.R. 2005. Soil water depletion
and root distribution of three dryland crops. Soil Science Society of America
Journal,69: 197-205.
Morse, K. 1996. A Review of Soil and Water management Research in Semi-Arid Areas
of Southern and Eastern Africa. Chatham, UK. Natural Resources Institute. UK.
Motsi, K.E., Chuma, E. and Mukamuri, B.B. 2004. Rainwater harvesting for sustainable
agriculture in communal lands of Zimbabwe. Physics and Chemistry of the
Earth, 29: 1069-1073.
Moyo, M. 2001. Representative soil profiles of ICRISAT research sites. Chemistry and
Soil Research Institute. Soils Report N°A666. AREX, Harare, Zimbabwe, 97pp.
Moyo, R., Love, D., Mul, M., Twomlow, S. and Mupangwa, W. 2006. Impact and
sustainability of drip irrigation kits in 54 the semi-arid Lower Mzingwane
Subcatchment, Limpopo Basin, Zimbabwe. Physics and Chemistry of the Earth,
31: 885-892.
Mrabet, R. 2007. Lasting benefits from no-tillage systems: Erosion control and soil
carbon sequestration. pp77-92. In: Stewart, B.A, Asfary, AF., Belloum, A,
Steiner, K. and Friedrich, T. (eds.). Proceedings of International Workshop on
Conservation Agriculture for Sustainable Land Management to Improve the
Livelihood of People in Dry Areas. 7-9 May 2007, Damascus, Syria.
Mtambanengwe, F. and Mapfumo, P. 2005. Organic matter management as an underlying
cause for soil fertility gradients on smallholder farms in Zimbabwe. Nutrient
Cycling in Agroecosystems, 73: 227-243.
Mubonderi, T. 1999. Cattle manure and inorganic fertilizer management for dryland
maize production in the smallholder sector of Zimbabwe. M.Phil. Thesis.
342
Department of Crop Science, University of Zimbabwe, Harare, Zimbabwe.
126pp.
Mugabe, F. 2004. Evaluation of the benefits of infiltration pits on soil moisture in semi-
arid Zimbabwe. Journal of Agronomy, 3: 188-190.
Muller-Samann, K.M. and Kotschi, J. 1994. Sustaining growth: Soil fertility management
in tropical smallholdings. GTZ/CT A, Margraf Verlag. 486pp.
Mulumba, L.N. and Lal, R. 2008. Mulching effects on selected soil physical properties
Soil and Tillage Research, 98: 106-111.
Mupangwa, W., Love, D. and Twomlow, S.J. 2006. Soil-water conservation and
rainwater harvesting strategies in the semi-arid Mzingwane Catchment,
Limpopo Basin, Zimbabwe. Physics and Chemistry of the Earth, 31: 893-900.
Mupangwa, W., Twomlow, S., Walker, S. and Hove, L. 2007. Effect of minimum tillage
and mulching on maize (Zea mays L.) yield and water content of clayey and
sandy soils. Physics and Chemistry of the Earth, 32: 1127-1134.
Mupangwa W., Twomlow S. and Walker S. 2008. The influence of conservation tillage
methods on soil water regimes in semi-arid southern Zimbabwe. Physics and
Chemistry of the Earth, 33: 762-767.
Mwenge Kahinda, J.M. 2004. Water productivity and yield gap analysis of water
harvesting systems in the semi-arid Mzingwane catchment, Zimbabwe. M.Sc.
thesis, Water Resources Engineering and Management Programme, University
of Zimbabwe. 124pp.
Ncube, B. 2007. Understanding cropping systems m semi-arid environments of
Zimbabwe: Options for soil fertility management. Ph.D. Thesis, Wageningen
University, Wageningen, The Netherlands. 155pp.
Ncube, B., Dimes, lP., Twomlow, S., Mupangwa, W. and Giller, K.E. 2007. Raising the
productivity of smallholder farms under semi-arid conditions by use of small
doses of manure and nitrogen: A case of participatory research Nutrient Cycling
in Agroecosystems, 77: 53-67.
Ncube, B., Twomlow, S., van Wijk, M.T., Dimes, J.P. and Giller, K.E. 2007. Productivity
and residual benefits of grain legumes to sorghum under semi-arid conditions in
south-western Zimbabwe. Plant and Soil, 299: 1-15.
343
Ncube, B., Dimes, J.P., van Wijk, M., Twomlow, S. and Giller, K. 2008. Productivity and
residual benefits of grain legumes to sorghum under semi-arid conditions in
south-western Zimbabwe: Unravelling the effects of water and nitrogen using a
simulation model. Field Crops Research, doi:10.1016/j.fcr.2008.08.001.
NeSmith, D.S. and Ritchie, J.T. 1992. Effects of soil water deficits during tassel
emergence on development and yield component of maize (Zea mays L.). Field
Crops Research, 28: 251-256.
Ngwenya, T.P. 2006. Effect of land degradation from grazing pressure on rangeland
hydrology. MSc. Thesis. Department of Civil Engineering, University of
Zimbabwe. 57pp.
Nhamo, N. 2003. The effects of combining cattle manure and inorganic nitrogen fertilizer
on sandy soils in Zimbabwe. M.Phil. Thesis. Department of Soil Science and
Agricultural Engineering, University of Zimbabwe, Harare, Zimbabwe. 149pp.
Nhamo, N., Mupangwa, W., Siziba, S., Gatsi, T. and Chikazhunga, D. 2003. The role of
cowpea (Vigna unguiculata L.) and other grain legumes in the management of
soil fertility in the smallholder farming sector of Zimbabwe. In: Waddington,
S.R. (Ed.). Grain Legume and Green Manures for Soil Fertility in Southern
Africa: Taking Stock of progress. Proceedings of a conference held 8-11
October 2002 at Leopard Rock Hotel, Vumba, Zimbabwe. Soil FertNet and
CIMMYT-Zimbabwe, Harare, Zimbabwe. 246pp.
Nhamo, N. 2007. The contribution of different fauna communities to improved soil
health: A case of Zimbabwean soils under conservation agriculture. Ph.D.
Thesis. University of Bonn. 131pp.
Nonner, J. 1997. Principles of Hydrogeology. IRE Lecture Note. IHE. Delft, The
Netherlands.
Northwood, P. J. and McCartney, J. C. 1971. The effect of different amounts of
cultivation on the growth of maize on some soil types in Tanzania. Tropical
Agriculture (Trinidad), 48: 25-33.
Nyagumbo, I. 1997. Effects of tillage systems on soil physical properties with special
reference to infiltration, bulk density and organic carbon. African Crop Science
Conference Proceedings, 3: 359-368. )
344
Nyagumbo, 1. 2002. The effect of three tillage systems on seasonal water budgets and
drainage of two Zimbabwean soils under maize. Ph.D. Thesis, Department of
Soil Science and Agricultural Engineering, University of Zimbabwe. 251 pp.
Nyagumbo, 1. 2007. A review of experiences and developments towards conservation
agriculture and related systems in Zimbabwe. pp345-372. In: Goddard, T.,
Zoebisch, M., Gan, Y., Ellis, W., Watson, A. and Somatpanit, S. (eds.). No-Till
Farming Systems. Special Publication Number 3. World Association of Soil and
.. Water Conservation.
Nyakatawa, E.Z., Brown, M., and Maringa, D. 1996. Maize and sorghum yields under
tied ridges of fertilised sandy soils in semi-arid south-east lowveld of
Zimbabwe. African Crop Science Journal, 4: 197-206.
Nyamangara, J. and Mpofu, S. E. 1996. Soil pH and lime requirement for high potential
communal areas of Zimbabwe. Journal of Applied Science in Southern Africa 2,
77-81.
Nyamangara, J., Mugwira, L.M. and Mpofu, S.E. 2000. Soil fertility status in the
communal areas of Zimbabwe in relation to sustainable crop production. Journal
of Sustainable Agriculture, 16(2): 15-29.
Nyamapfene, K. 1991. Soils of Zimbabwe, NeHanda Publishers, Harare,
Zimbabwe. 179pp.
Nyamudeza, P., Mazhangara, E. and Kamba, E. 1992. Adoption of tied furrow technique
and effects of the technique and previous crop on residual moisture and yields
of sorghum and maize. Paper Presented at the Annual Review Meeting of the
IBSRAM Vertisol Management Network in Africa. June, 1992. Accra, Ghana.
pp69-81.
Nyamudeza, P. 1993. The effects of growing sorghum (Sorghum bicolorï in furrows and
on the flat at three populations and three row widths in a semi-arid region of
Zimbabwe. I. Grain yield and yield components. Zimbabwe Journal of
Agricultural Research, 31: 1-10.
Nyamudeza, P. 1998. Water and fertility management for crop production in semi-arid
Zimbabwe. Ph.D. Thesis. The University of Nottingham, UK.
345
Oldrieve, B. 1993. Conservation farming for communal, small-scale, resettlement and
cooperative farmers of Zimbabwe: a farm management handbook. Harare: Rio
Tinto Foundation. 73pp.
Oosterhout van, S.A.M. 1996. Copying strategies of smallholder farmers with adverse
weather conditions regarding seed deployment of small grain crops during
1994/1995 cropping season in Zimbabwe. Volumes 1 to 3. SADC/GTZ, Harare.
Osunbitan, lA., Oyedele, D. J. and Adekalu, K.O. 2004. Tillage effects on bulky density,
hydraulic conductivity and strength of a loamy sand soil in south western
Nigeria. Soil and Tillage Research, 82: 57-64.
Otegui, M.E., Andrade, F.H. and Suero, E.E. 1995. Growth, water use and kemel
abortion of maize subjected to drought at silking. Field Crops Research, 40: 87-
94.
Otegui, M.E. and Bonhomme, R. 1998. Grain yields components in maize. I. Ear growth
and kernel set. Field Crops Research, 56: 247-256.
Pala, M, Buekes, DJ., Dimes, lP. and Myers, R.lK. 2005. Agronomic Management for
Improved Water Use Efficiency in the Dry Areas of West Asia and North
Africa. Proceedings of a Workshop organized by the Optimizing Soil Water Use
Consortium, Ankara, Turkey. 22-26 April 2002. Aleppo, Syria: ICARDA; and
Patencheru, India: ICRISAT. 288pp.
Palm, C.A., Myers, R.J.K. and Nandwa, S.M. 1997. Combined use of organic and
inorganic nutrient sources for soil fertility maintenance and replenishment.
ppI93-217. In: Buresh, l.R., Sanchez, P.A. and Calhoun, F. (eds.).
Replenishing soil fertility in Africa. Soil Science Society of America Special'
Publication Number 51.
Pandey, R.K., Maranville, J.W. and Chetima, M.M. 2000. Deficit irrigation and nitrogen
effects on maize in a Sahelian environment. II. Shoot growth, nitrogen uptake
and water extraction. Agricuture Water Management, 46: 15-27.
Peck, A.l. 1969. Entrapment, stability and persistence of air bubbles III soil water.
Australian Journal of Soil Science, 7: 79-90.
Petersen, P., Tardin, J.M. and Marochi, F. 1999. Participatory development of non-till
systems without herbicide for family farming: The experience of the centre-
346
south region of Parana. Environment, Development and Sustain ability, 1: 235-
252.
Phillips, 1.G., Cane, M.A. and Rosenzweig, C. 1998. ENSO, seasonal rainfall patterns
and simulated maize yield variability in Zimbabwe. Agricultural and Forest
Meteorology, 90: 39-50.
Pilbeam, CJ., Tripathi, B.P., Sherchan, D.P., Gregory, P.J., and Gaunt, 1. 2000. Nitrogen
balances for households in the mid-hills of Nepal. Agriculture, Ecosystems and
Environment, 79(1): 61-72.
Plucknett, D.L. 1994. Sources of the next century's new technology. pp343-373. In:
Anderson, J,R. (Ed.). Agricultural Technology: Policy Issues for the
International Community. CAB International.
Powell, 1.M., Fernandez-Rivera, S., Hiernaux, P.A. and Turner, M.D. 1996. Nutrient
cycling in integrated rangeland/crop land systems of the Sahel. Agricultural
Systems, 52(2/3): 143-170.
Reij, C., Scoones, I. and Toulmin, C. 1996. Sustaining the Soil: Indigenous Soil and
Water Conservation in Africa. Earthscan Publications Ltd., London. 260pp.
Rinaidi M., Losavio N. and Flagella Z. 2003. Evaluation and application of CROPGRO-
soybean model for improving soybean management under rainfed conditions.
Agricultural Systems, 78: 17-30.
Rockstrëm, J. and Rouw de, A. 1996. Water, nutrients and slope position in on-farm pearl
millet cultivation in the Sahel. Plant and Soil, 195: 311-327.
Rockstrëm, J. Gordon. L., Folke, C., Falkenmark, M and Engwall, W. 1999. Linkages
among water vapor flows, food production, and terrestrial ecosystem services.
Conservation Ecology 3(2):5. http://www.consecol.org/vo13/iss2/art5
Rockstrëm.T, and Falkenmark, M. 2000. Critical reviews in plant sciences, 19(4): 319-
346.
Rockstrom, 1. 2002. Potential of rainwater harvesting to reduce pressure on fresh water
resources. Dialogue on Water for Food and Environment, International Water
Conference, 14-16 October, Hanoi, Vietnam.
347
Rockstrëm, J., Barron, J. and Fox, P. 2002. Rainwater management for increased
productivity among smallholder farmers in drought prone environments.
Physics and Chemistry of the Earth, 27: 949-959.
Rockstrom. J., Barron, J. and Fox, P. 2003. Water productivity in rainfed agriculture:
Challenges and opportunities for smallholder farmers in drought prone
agroecosystems. ppI45-162. In: Kilne, J.W., Barker, R. and Molden, D. (eds.).
Water Productivity in Agriculture: Limits and Opportunities for Improvement.
CAB International.
Rosenzweig, C. 2001. Impacts of the El Nino-Southern Oscillation on agriculture:
Guidelines for Regional Analysis. pp21-30. In: Hatfield, J.L., Volenec, ].J.,
Dick, W.A. and Kral, D.M. (eds.). Impacts of El Niiio and Climate Variability
on Agriculture. ASA Special Publication Number 63. 126pp.
Rusike, J. and Heinrich, G.M. 2002. Integrated soil water and nutrient management
farmer field schools in Zimbabwe. Proceedings of a Review, Evaluation and
Planning Workshop, 11-12 December 2001, International Crops Research
Institute for the Semi Arid Tropics (ICRISAT), Bulawayo, Zimbabwe. 40pp.
Qin R., Stamp P. and Richner W. 2006. Impact of tillage on maize rooting in a cambisol
and luvisol in Switzerland. Soil and Tillage Research, 85: 50-61.
Salinas-Garcia, J.R., Hons, F.M., Matocha, J.E., and Zuberer, D.A. 1997. Soil carbon and
nitrogen dynamics as affected by long-term tillage and nitrogen fertilization.
Biology and Fertility of Soils, 25: 182-188.
Sasal, M. C., Andriulo, A. E. and Taboada, M. A. 2006. Soil porosity characteristics and
water movement under zero tillage in silty soils in Argentinean Pampas. Soil
and Tillage Research, 87: 9-18.
Sauer, T.J., Hatfield, J.L. and Prueger, J.H. 1996. Com residue age and placement effects
on evaporation and soil thermal regime. Soil Science Society of America
Journal, 60: 1558-1564.
Scoones, I and Cousins B. 1989. A participatory model for agricultural research and
extension: The case of vleis, trees and grazing schemes in the dry south of
Zimbabwe. Zambezia, 16: 45-65.
348
SEDAP (South Eastern Dry Areas Project). 2001. Proceedings of a Workshop on
Adaptive Research and Participatory Adaptive Trials held on 16-18 July 2001 at
Montclair Hotel, Nyanga. Department of Research and Specialist Services.
Zimbabwe.
Shamudzarira, Z. and Robertson, MJ. 2002. Simulating the response of maize to nitrogen
fertilizer in semi-arid Zimbabwe. Experimental Agriculture, 38:79-96.
Shinde, S. S., Magar, S. S. and Kale, S. P. 1982. Effect of soil physical conditions and
initial soil moisture content on infiltration into black soil. Journal of Indian
Society of Soil Science, 30: 441-446.
Shumba, E.M., Chisenga, M.M. and Ndebele, C. 1992. Contribution of fertilizer and
management practices to the grain yield of maize in semi-arid areas of
Zimbabwe. Zimbabwe Journal of Agricultural Research, 30(2): 137-143.
Shumba, E.M., Waddington, S.R. and Rukuni, M. 1992. Use of tine tillage with atrazine
weed control, to permit earlier planting of maize by smallholder farmers in
Zimbabwe. Experimental Agriculture, 28:443-452.
Sivakumar, M.V.K. 1992. Emperical analysis of dry spells for agricultural applications in
West Africa. Journal of Climate, 5: 532-539.
Sivakumar, M.V.K. and Salaam, S.A. 1999. Effect of year and fertilizer on water use
efficiency of pearl millet (Pennisetum sativum L.) in Niger. Journal of
Agricultural Science, 132:139-148.
Smaling, E.M., Nandwa, S.M. and Janssen, B.H. 1997. Soil fertility in Africa is at stake.
pp47-61. In: Buresh, RJ., Sanchez, P.A. and Calhoun, F. (eds.). Replenishing
Soil Fertility in Africa. Soil Science Society of America Special Publication N°
51.
Smith, R.D. 1988. Tillage trials in Zimbabwe: 1957 to 1988. Consultancy report,
November 1988. Institute of Agricultural Engineering (AGRITEX and GTZ),
Harare.
Smith, H.l 2006. Development of a systems model facilitating action research with
resource poor farmers for sustainable management of natural resources. Ph.D.
Thesis. University of the Free State, Bloemfontein, South Africa. 297pp.
349
Soil Science Society of America. 1986. Glossary of Soil Science Terms. Soil Science
Society of America, Madison, Wisconsin.
Sow, A.A., Hossner, L.R., Unger, P.W. and Stewart, B.A. 1997. Tillage and residue
effects on root growth and yields of grain sorghum following wheat. Soil and
Tillage Research, 44: 121-129.
Sprent, JJ. 2007. Soil biology is an essential component of conservation agriculture, with
particular reference to mycorrhiza and legume nodulation. ppl03-110. In:
Stewart, B.A., Asfary, A.F., Belloum, A., Steiner, K. and Friedrich, T. (eds.).
Proceedings of International Workshop on Conservation Agriculture for
Sustainable Land Management to Improve the Livelihood of People in Dry
Areas. 7-9 May 2007, Damascus, Syria.
Steiner, K. 2007. Farmer participation in the development of conservation agriculture
technologies. pp227-242. In: Stewart, B.A., Asfary, A.F., Belloum, A., Stein er,
K. and Friedrich, T. (eds.). Proceedings of International Workshop on
Conservation Agriculture for Sustainable Land Management to Improve the
Livelihood of People in Dry Areas. 7-9 May 2007, Damascus, Syria.
Stem, R., Dennett, M.D. and Garbutt, D.J. 1981. The start of the rains in West Africa.
Journal of Climatology, 1: 59-68.
Stem, R., Dennett, M.D. and Dale, I.C. 1982. Analyzing daily rainfall measurements to
give agronomically useful results. II. A modeling approach. Experimental
Agriculture, 18: 237-253.
Stem, R., Knock, J., Rijks, D. and Dale, I. 2003. INSTAT Climatic Guide. 398pp.
http://www.reading.ac. uk/ssc/softw are/ins tat/climatic .pdf.
Stocking, M. 1989. How Lesotho is tackling soil degradation. Appropriate Technology,
15: 14-16.
Stocking, M.A. 2003. Tropical soils and food security: The next 50 years. Science, 302:
56-58.
Tabor, J.H. 1995. Improving crop yields in the Sahel by means of water harvesting.
Journal of Arid Environments, 30:83-106.
Tadross, M., Suarez, P., Lotsch, A., Hachigonta, S., Mdoka, M., Unganai, L., Lucio, F.,
Kamdonyo, F. and Muchinda, M. 2007. Changes in growing-season rainfall
350
characteristics and downscaled scenarios of change over southern Africa:
implications for growing maize. Regional expert meeting report. pp 193-204.
www.csag.uct.ac.zal-mtadross.
Trenberth, K.E., Jones, P.D., Ambenje, P. Bojariu, R. Easterling, D., Klein Tank, A.,
Parker, D., Rahirnzadeh, F., Renwick, J.A., Rusticucci, M., Soden, B. and Zhai,
P. 2007. Observations: Surface and Atmospheric Climate Change. In: Solomon,
S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and
H.L. Miller (eds.). Climate Change 2007: The Physical Science Basis.
Contribution of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 235pp.
Tsimba, R., Hussein, J. and Ndlovu, L. R. 1999. Relationships between depth of tillage
and soil physical characteristics of sites farmed by smallholders in Mutoko and
Chinyika in Zimbabwe. http://www.fao.org/ag/ags/agse/3ero/namibiallcll.htm.
Tsubo, M., Walker, S. and Hensley, M. 2005. Quantifying risk for water harvesting under
semi-arid conditions. Part 1. Rainfall intensity generation. Agricultural Water
Management, 76: 77-93.
Twomlow, S.J. 1994. Field moisture characteristics of two fersiallitic soils in Zimbabwe.
Soil Use and Management, 10: 168-173.
Twomlow, S.J., Riches, C. and Mabasa, S. 1997. Weeding - Its contribution to soil water
conservation in semi-arid maize production. Brighton Crop Protection
Conference.
Twomlow, S. J. and Bruneau, P.M.C. 1998. Soil-water regimes in semi-arid Zimbabwe.
In: Wheater, Hand Kirby, C (Eds.). Hydrology in Changing Environment.
Proceedings of the British Hydrological Society International Conference,
Exeter, July, 1998. Vol. II. British Hydrological Society. John Wiley & Sons.
Chichester.
Twomlow, S.J. and Dhliwayo, H. 1999. Semi-arid maize yield responses to conservation
tillage and weeding. Brighton Crop Protection Conference - Weeds, November
1999. Farnham, Surrey, UK :BCPC.
351
Twomlow, S.J., Riches, C., O'Neill, D., Brookes, P. and Ellis-Jones, l 1999. Sustainable
dryland smallholder farming in sub-Saharan Africa. Annals of Arid Zone, 38
(2): 93-135.
Twomlow, S.J. and Bruneau, P.M.C. 2000. The influence of tillage on semi-arid soil-
water regimes in Zimbabwe. Geoderma, 95: 33-51.
Twomlow, S. and Hove, L. 2006. Is conservation agriculture an option for vulnerable
households? Briefing Note N° 4. International Crops Research Institute for the
Semi-Arid Tropics (ICRISAT). September 2006.
Twomlow, S. and Rohrbach, D. 2006. Achieving sustained gains in the food security of
vulnerable households. Briefmg Note NQ 1. ICRISAT-Bulawayo. September
2006.
Twomlow, S., Rohrbach, D., Rusike, l, Mupangwa, W., Dimes and Ncube, B. 2006a.
Spreading the word on fertilizer in Zimbabwe. GlobalTheme on
Agroecosystems Report N° 24. International Crops Reseach Institute for the
Semi-Arid Tropics (ICRISAT). Bulawayo. l lpp.
Twomlow, S.J., Steyn, lT., du Preez, C.C. 2006b. Dryland farming in southern Africa.
Chapter 19. pp769-836. In: Petersen, G.A., Unger, W.P. and Payne, W.A.
(eds.). Dryland Agriculture 2nd Ed. Agronomy Monograph N°.23. American
Society of Agronomy, Madison, Wisconsin.
Twomlow, S., Hove, L., Mupangwa, W., Masikati, P. and Mashingaidze, N. 2008a.
Precision conservation agriculture for vulnerable farmers in low potential
zones. Paper presented at the Challenge Program on Water and Food Theme 1
Rainfed Topic Workshop, 22-25 September, Tamale, Ghana.
Twomlow, S.,Rohrbach, D., Dimes, l, Rusike, l, Mupangwa, W., Ncube, B., Hove, L.,
Moyo, M., Mashingaidze, N. and Maphosa, P. 2008b. Micro-dosing as a
pathway to Africa's Green Revolution: evidence from broad-scale on-farm
trials. Nutrient Cycling in Agroecosystems, submitted.
Twomlow, S.J., Urolov, lC., Oldrieve, B. and Jenrich, M. 2008c. Lessons from the field
- Zimbabwe's conservation agriculture taskforce. Journal of SAT Agricultural
Research (in press).
352
Unganai, L.S. 1990. Crop adaptability inventory for Zimbabwe: The edaphic and rainfall
characteristics. Mimeo. Department of Meteorological Services, Harare,
Zimbabwe.
Unganai, S.L. 1996. Historic and future climatic changes In Zimbabwe. Climate
Research., 6: l37-145.
Unger, P.W. and Stewart, B.A. 1983. Soil management for water use efficiency: An
Overview. pp419-454. In: Taylor, H.M., Jordan, W.R. and Sinclair, T.R. (eds.).
Limitations to Water Use in Crop Production. Soil Science Society of America.
UNDP. 2002. Preventing land degradation for sustaining livelihoods. Experience from
GEF-UNDP projects, August 2002. http://www.undp.org/gef/undp-gef-
publication/land degradation bronchure.pdf
Van der Meer, F.W., 2000. Modelling tropical soil water regimes In semi-arid
environments. Ph.D. Thesis, Loughborough University. 270pp.
Vauchin, M. and Chopart, J.L. 1997. Multi-disc infiltrometry for in-situ determination of
the surface hydrodynamic features of a gravel soil in Cote-d'Ivoire.
L' Agronomie Tropicale, 46(4): 259-27l.
Verplancke, H. 1994. Water balance studies in the root zone of annual and perennial
crops in rainfed and irrigated agriculture. Refresher Course for Alumni of the
International Training Centre for Post-Graduate Soil Scientists of the Ghent
University. New Waves in Soil Science. ITC-Ghent Publications Series N°S.
BADC/University of Zimbabwe. Harare.
Vincent, V. and Thomas, R.G. 1960. An Agricultural Survey of Southern Rhodesia (now
Zimbabwe), Part 1. Agro-ecological Survey. Government Printers, Salisbury
(now Harare).
Vogel, H. 1992. Tillage effects on maize yield, rooting depth and soil water content on
sandy soils in Zimbabwe. Field Crops Research, 33: 376-384.
Waddington, S.R., Edmeades, G.O., Chapman, S.C. and Barreto, H.J. 1994. Where to
with agricultural research for drought-prone maize environments? A Paper
Presented at the Fourth Eastern and Southern Africa Regional Maize
Conference, 28 March - 1April 1994. Harare, Zimbabwe. 32pp.
353
Willat, S.T. 1967. Moisture status of in Rhodesian soils prior to rams. Rhodesia
Agricultural Journal, 64(1): 4-6.
Woltering, L. 2005. Estimating the influence of on-farm conservation practices on the
water balance: Case of the Mzinyathini catchment in Zimbabwe. M.Sc. Thesis.
Delft University of Technology, The Netherlands, 103pp.
Xiano-Bin, W., Dian-Xiong, C., Hoogmoed, W.B., Oenema, O. and Perdok, U.D. 2006.
Potential effect of conservation tillage on sustainable land use: A review of
global long-term studies. Pedosphere, 16(5): 587-595.
Zaongo, C.G.L., Wendt, C.W., Lascano, R.J. and Juo, A.S.R. 1997. Interactions of water,
mulch and nitrogen on sorghum in Niger. Plant and Soil, 197: 119-126.
Zhai, R., Kachanoski, R.G. and Voroney, R.P. 1990. Tillage effects on spatial and
temporal variations of soil water. Soil Science Society of America Journal, 54:
186-192.
Zhang, R. 1997. Determination of soil sorptivity and hydraulic conductivity from the
minidisk infiltrometer. Soil Science Society of America Journal, 61: 1024-1030.
Zingore, S. 2006. Exploring diversity within smallholder farming systems in Zimbabwe:
Nutrient use efficiencies and resource management strategies for crop
production. Ph.D. Thesis, Wageningen University, Wageningen, The
Netherlands. 258pp.
354
APPENDICES
Appendix li. Guidelines for resource mappmg and evaluation of tillage systems m
Gwanda and Insiza districts
Resource Flow Mapping
o Draw homestead, the different fields and the kraals on-scale. First on the ground
then on paper
o Indicate the soil type in each of the field
o Draw the different plots on each field and indicate what is grown on each plot or
what was grown the previous year (depending on the time of year)
o How much seed was used for each plot?
o Was the seed retained or purchased?
o What is the amount of labour needed for the plots for ploughing, planting,
weeding, etc. (in days and number of people working that day)?
o Who was involved in the different labour activities in each plot?
o Other inputs in the field: manure, fertiliser and pesticides?
e Yields of each plot: how many bags / kg / scotch carts (determine size) were
yielded from each plot?
e Where did the grain go after harvest (draw arrow)
Cj) Where did the stover go after harvest (draw arrow)
Focus Group Discussions
o Land preparation - when is it done for each tillage system? How long it takes?
o Application of soil fertility amendments - what amendments are used? How much
is applied (scotchcarts/wheelbarrows? How are they applied? When applied? How
often are amendments applied? How many people are involved in application of
soil fertility amendments?
o Planting - When is it done for each tillage system? How is it done? How many
people are required to do the operation?
o Crop establishment - which tillage system has best/poorest establishment --- in a
355
bad season and good season?
o Weeding - How is it done? How often is it carried out in below-normal, normal
above-normal rainfall seasons? How long it takes per-unit area?
o Crop response to topdressing -leaf colour, plant vigour, yield?
o Pests/diseases attacks - Which pests/diseases are common in each tillage system?
How severe is the attack? How often is the attack? What are the possible
remedies?
o Rainwater harvesting - ranking of the tillage systems according to ability to
harvest rainwater
o Waterlogging - in a season with above-normal rainfall
o Dry spell effect - which tillage system reduces the impact of dry spells on maize?
o Crop yield - in seasons with below-normal, normal and above-normal rainfall
patterns?
o Technical support - which tillage system requires more support?
o Which organizations are providing technical support?
356
Appendix 2. Farmers who hosted on-farm experiments and participated in resource flow
mapping and focus group discussions in Insiza and Gwanda districts
Name of farmer District Village
S. Mguni Insiza Masi yepambili
M. Mlalazi Insiza Thuthuka
M. Moyo Insiza Mpumelelo
M. MQ_ofu Insiza Mpumelelo
N. Ncube Insiza Mpumelelo
S. Nkomo Insiza Mpumelelo
M. Nyathi Insiza Mpumelelo
S. Nsingo Insiza Thuthuka
J. Dube Gwanda Mnyabetsi S
H. Magaya Gwanda Humbane
s.Malotha Gwanda Humbane
G. Mlilo Gwanda Magaya
K. Mlilo Gwanda Mnyabezi D
A. Moyo Gwanda Humbane
M. Nare Gwanda Mnyabetsi S
A. Ncube Gwanda Gohole
E. Ncube Gwanda Magaya
J. Ncube Gwanda Fumukwe
N. Sibanda Gwanda Mnyabezi D
T. Sibanda Gwanda Mnyabezi D
J. Siziba Gwanda Mnyabezi D
M. Siziba Gwanda Mnyabezi D
T. Tlou Gwanda Mnyabetsi S
357