0/51 Id.{ 8y_
II~Un~iv~erOsitOy ~FrOeeOS~tatIe~~OO~~~~
34300003650748
Universiteit Vrystaat
THE IMPACT OF IRRIGATION WITH COAL MINE
WATER ON GROUNDWATER RESOURCES
by
P.D.VERMEULEN
THESIS
Submitted in the fulfilment of the requirements for the degree of Doctor of Philosophy
in the Faculty of Natural Science and Agriculture, Department of Geohydrology,
University of the Free State, Bloemfontein
November 2006
Promoter: Dr. B.H.Usher
Acknowledgements
I hereby wish to express my sincere thanks to a large number of people who have helped me
to complete this thesis:
o To my promoter, Dr. Brent Usher, thank you for your guidance, advice and friendship,
the hours of intense academic debating and your huge personal effort in this research
project.
o To my great friend and mentor, Prof. Gerrit van Tonder, for his advice,
encouragement and academic input.
o I gratefully express appreciation to the Director, and all my colleagues at the Institute
for Groundwater Studies for assistance, technical expertise, administration, laboratory
analyses, fieldwork and encouragement. Thanks are due to:
o Eelco Lukas for assistance with WISH and with all IT matters.
o Rainier Dennis for assistance with GIMI
o Lore-Mari Cruywagen and Eina de Necker for the laboratory analyses.
o Jane van den Heever for all administrative arrangements. Without her financial and
administrative advice, we would not have remained on track.
o Catherine Bitzer for proofreading
o The Water Research Commission and Coaltech for three years of financial support.
Without their funding, this project would never have been. A special word of thanks to
Mr. Meiring du Plessis of the WRC and Mr. Johan Beukus of Coaltech for their input
and expert advice during steering committee meetings.
o Ingwe, Anglo Coal and Sasol Coal are gratefully acknowledged for access to their
properties, data and personnel, and for funding and assisting in drilling at the sites. In
this regard, Mrs Erika Prinsloo and Messrs. Bertie Botha, Ernie van Biljon, Roux
Meyer and Dr Mark Aken are thanked. Also to the farmers, Smit Broers and Kanhym
for allowing me to dig pits at their irrigation sites.
o Prof. Simon Lorentz and Mr. Cobus Pretorius of the University of KwaZulu-Natal for
designing the deep tensiometers and interpretation of the data.
o Dr.Piet Ie Roux and Dr. Malcolm Hensley for all their assistance and advice in soil
science issues, and the patience of teaching a geohydrologist the ways of a soil
scientist.
o There are many experts who contributed to the ideas and understanding in this thesis,
including Prof. John Annandale, Chris de Jager, Annemarie van der Westhuizen and
countless others, who provided insight into different aspects of soil science,
groundwater hydraulics and hydrogeochemistry.
o My great friends of many years, Gideon and Sussie van Zyl, Willem and Tinkie
Pretorius and Sandra Labuschagne, for all their support, encouragement and for
believing in me.
o To my mother-in-law, Natie Hoffman, I am grateful for your support and
encouragement throughout the project.
o To my father and deceased mother, who would have been very proud.
o To my wife, Wilma, and children, Elana and Tarina, thank you for all your love,
encouragement and endless patience. I am very grateful for the hours you have
sacrificed to enable me to complete the thesis.
Last but not least to my Saviour, who gave me strength and hope. I have moved at such a
hectic pace during the last three years; it is now time to slow down for a while to allow my
soul to catch up with my body!!
Be not afrai~ of 90iHB slowl~i be afraj~ onl~ of stan~in9 still.
chinese Proverb
Table of Contents
1 INTRODUCTION 1
1.1 Overview '" 1
1.2 The problem of mine closure 2
1.3 Irrigation with gypsiferous water 3
1.4 Objectives 4
1.5 Method of Investigation 4
1.6 Thesis Structure 7
2 IRRIGATION WITH COAL MINE WATER 9
2.1 Introduction 9
2.2 Water volumes available for irrigation 10
2.3 Composition of coal mine waters 12
2.3.1 Introduction 12
2.3.2 Acid Mine Drainage 12
2.3.3 Gypsiferous Waters 17
2.4 Perception of the final fate of salts 21
3 GENERAL DESCRIPTION OF IRRIGATION SITES 23
3.1 Introduction : 23
3.2 Kleinkopje Colliery : 23
3.2.1 Surface Hydrology. 24
3.2.2 Geology 24
3.2. 3 Positioning of the monitoring boreholes 26
3.3 New Vaal Colliery 30
3.3. 1 Surface Hydrology 31
3.3.2 Geology 32
3.3.3 Installation of monitoring boreholes 33
3.3.4 Water levels 34
3.3.5 Irrigation water ana/ysis 35
3.3.6 Multi-parameter profiling (see Appendix D for monitoring data). 36
3.4 Syferfontein Colliery 37
3.4. 1 Surface Hydrology 38
3.4.2 Geology 39
3.4.3 Geohydrology at Syferfontein Pivot 41
3.4.4 Water levels 42
3.4.5 Irrigation Water Quality. 42
3.4.6 Groundwater Quality 43
3.5 Optimum Colliery 44
3.5 1 Introduction 44
3.52 Geology 45
3.53 Topography 45
3.54 Geohydrology 46
3.6 Summary 47
4 A REVIEW OF THE HYDRAULIC CHARACTERISTICS AND HYDROCHEMICAL
IMPACTS OF IRRIGATION ON SOil AND SPOilS .48
4.1 Virgin soil irrigation 48
4. 1.1 The Unsaturated Zone (Vadose zone). 48
4.1.2 Unsaturated flow (matrix flow or diffusive flow) 49
4.1.3 Preferential flow (macropore flux). 57
4.1.4 The hydrochemical effects of irrigation with saline water on the soil. 65
4.1.5 Prior-conceptual model for irrigation at virgin soit sites 65
4.2 Spoils irrigation 67
4.2. 1 Introduction 67
4.2.2 Hydraulic Properties at Opencast Collieries 69
5 METHODOLOGY 75
5.1 Introduction 75
5.2 Groundwater monitoring 75
52.1 Geochemical profiling 75
52.2 Chemical analysis 77
5.3 Soil Sampling 78
5.4 Soil Analysis 79
5.4.1 Soil Properties 79
5.4.2 Soil chemical analysis 80
5.4.3 XRD and XRF 80
5.5 Soil water sampling (Porous cup Iysimeters) 81
5.6 Tensiometry 83
5.7 Leaching columns 85
5.8 Humidity cells 88
5.9 Point Dilution Tests 90
5.10 Slug tests 93
5.11 Pumping Tests 94
6 DETAILED CHARACTERISATION OF VIRGIN SOIL IRRIGATION SITES 97
6.1 Introduction 97
6.2 Groundwater monitoring 97
6.2.1 Kleinkopje Colliery 98
6.2.2 New Vaal Colliery 101
6.2.3 Syferfontein Colliery 103
6.2.4 Overall conclusions from the groundwater monitoring at the pivots 104
6.3 Soil analysis 105
6.3. 1 Kleinkopje 105
6.3.2 New Vaal Colliery 107
6.3.3 Syferfontein Colliery 111
6.3.4 Conclusions 114
6.4 Unsaturated soil sampling 114
6.4.1 Introduction 114
6.4.2 Kleinkopje 115
6.4.3 Syferfontein 121
6.4.4 New Vaal 126
6.4.5 Conclusions 126
6.5 Gypsum precipitation from the CaS04-rich water 127
6.51 Introduction 127
ii
6.52 Determination of the presence of gypsum in irrigated soils 130
6.5.3 Sulphate adsorption: 131
6.6 Assessment of the hydraulic behaviour and chemistry in the unsaturated zone at
the virgin soil irrigation sites 136
6.6.1 Measurement of soi' moisture potential 137
6.6.2 Salt Balance of the profile 147
6.6.3 Leaching of the salts in the protite 149
6.6.4 Identification of the aquifer parameters and flow zones through interpretation
of point diution, slug and pumping test data (See Sections 5.9 to 5 11). 157
6.6.5 Determination of flow at the Irrigation site 162
6.6.6 The influence of clay in the unsaturated profile on the soil water that reaches
the water level 169
6.7 Conclusion 173
7 IRRIGATION AT OPENCAST COLLIERIES (SPOILS) 174
7.1 Scenarios considered for spoils irrigation 174
7.2 Laboratory testing of spoils irrigation 174
72.1 Humidity cells 174
72.2 Leaching columns 191
72.3 Field leaching columns 198
72.4 Discussion 202
7.3 Hydraulic response of the spoils 202
7.4 Geochemical Modelling 203
74.2 Reaction modelling 207
7.5 Discussion of geochemical response 226
8 CONSOLIDATED MODELS OF MINE WATER IRRIGATION FOR VIRGIN SOIL
IRRIGATION 228
8.1 Introduction 228
8.2 Virgin soil model 228
8.3 Detailed site evaluation for virgin soil sites 230
8.3.1 Kleinkopje 1 231
8.3.2 Kleinkopje 4 234
8.3.3 Syferfontein 237
8.3.4 New Vaal 240
8.4 Summary of Results 243
9 OPERATION AND MONITORING GUIDELINES 245
9.1 Basic Data Input 245
9.2 Selection criteria 246
9.2. 1 Identifying land suitable for irrigation (Ie Raux, 2006) 247
9.2.2 Procedures for selecting land for irrigation 251
9.2. 3 Site specific and more general data 254
9.3 Monitoring Guidelines 257
9.4 GIMI- A user-friendly decision support application for mine water irrigation. 261
9.4. 1 OverviewofGIMI operation 262
9.4.2 Additional features in GIMI 263
9.4.3 Guidelines for operating GIM/. 266
10 CONCLUSIONS AND RECOMMENDATIONS 271
iii
10.1 Availability of irrigation water: 271
10.2 Water levels and water quality 271
10.3 Soil and Soil Water Analysis 272
10.4 Hydraulic behaviour in the unsaturated zone 273
10.4.1 Kleinkopje 1 274
10.4.2 Kleinkopje 4 274
10.4.3 Syferfontein 274
10.4.4 New Vaal 274
10.5 Detailed site evaluation for opencast (spoils) sites 275
10.6 Overall Recommendation 276
11 REFERENCES 278
iv
List of Figures
Figure 1-1: The South African Coalfields, ajier Erasmus et al. (1981) 1
Figure 1-2: South African Map with the towns near to the irrigation sites 5
Figure 1-3: Colliery map of Mpumalanga, illustrating the collieries with irrigation sites investigated
in this research. (Map courtesy ofGrobbelaar et al., 2002) 6
Figure 2-1. Stages in pH evolution as a result of different buffering minerals (after Morin, 1983 and
Usher et al., 2001) 16
Figure 2-2: Gypsum saturation (SJ) in neutralised mine water used to irrigate at Kleinkopje Pivotl .
................................................................................................................................................. 18
Figure 2-3: Saturation state of gypsum plotted against sulphate concentration (after Usher, 2003) .
................................................................................................................................................. 19
Figure 2-4: Sulphate concentration and gypsum saturation for decant water from Optimum Colliery .
................................................................................................................................................. 20
Figure 3-1: Topographic map of the Kleinkopje area, indicating the position of Pivot 1 and Pivot 4.
................................................................................................................................................. 23
Figure 3-2: Fence diagram of geology at Pivot 1, drawn through the boreholes 24
Figure 3-3: Geology and aquifer parameters at Pivotl 25
Figure 3-4: Geology and aquifer parameters at Pivot 4 25
Figure 3-5: Position of the monitoring boreholes at Pivotl 26
Figure 3-6: Position of the monitoring boreholes at Pivot 4 27
Figure 3-7: Water levels at Pivot I 28
Figure 3-8: Water levels at Pivot 4 28
Figure 3-9: Multiparameter profiling at PVl BH2 30
Figure 3-10: Topographic map of the New Vaal area 31
Figure 3-11: Surface contoursfor the New Vaal area 32
Figure 3-12: Log illustration of geology at New Vaal pivot 33
Figure 3-13: Fence diagram of geology at New Vaal pivot 33
Figure 3-14: Position of the monitoring boreholes at the New Vaal Pivot 34
Figure 3-15: Water levels at New Vaal Pivot 35
Figure 3-16: Durov diagram of the Irrigation Dam 595 35
Figure 3-17: Multi-parameter profiling of PVl 36
Figure 3-18: Regional setting of Syferfontein irrigation site 37
Figure 3-19: Pivot site showing the positions of the monitoring boreholes (with the positions of the
dykes, provided by the mine's geologist) 39
Figure 3-20: A model diagram of the geology at Syferfontein pivot, viewedfrom the south.. 40
v
Figure 3-21: A fence diagram of the geology at Syferfontein pivot 40
Figure 3-22: Geology and aquifer parameters at Syferfontein 41
Figure 3-23: Groundwater levels for Syferfontein Colliery Pivot 42
Figure 3-24: Durov diagram of the irrigation dam 43
Figure 3-25: Multi-parameter profiling of OBH]. 44
Figure 3-26: Locality plan of Optimum Colliery 45
Figure 3-27: Surface contours in the area of Optimus Pit, indicating the position of the two irrigation
pivots, as well as Lapa Dam 46
Figure 3-28: Floor contours, with flow directions indicated. 47
Figure 4-1: Typical moisture profile during rainstorm or irrigation (after De Marsily, 1986).51
Figure 4-2: Soil water retention curve. Notice that the saturated porosity is equal to the water content
where the matric potential = 0 m (i.e. 0.3 in this graph). In practice, no flow occurs if the matric
potential is smaller than -10 m 54
Figure 4-3: Relationship between moisture content and the unsaturated K-value. Note that the
saturated K-value (i.e. K) is the value where the water content is equal to the saturated porosity (B_J .
................................................................................................................................................. 55
Figure 4-4: Example of K-value for sandy and clayey soil 56
Figure 4-5: Illustration of preferential flow (after Steenhuis et al., 2002) 58
Figure 4-6: Illustration of matrix flow (after Steenhuis et al., 2002) 58
Figure 4-7: Earth worm, a pathway generator (after Steenhuis et al., 2002) 60
Figure 4-8: Comparison between the hydraulic conductivity of the porous medium and the fractured
medium versus the aperture (From Maini and Hocking, 1977) 62
Figure 4-9: Formation of soil water fingers (Source: Steenhuis et al., 2002) 63
Figure 4-10: Thefunnel effect in sandy soil. With a relatively low flow rate raj, afast-moving spout
forms beneath the lower end of an inclined layer of coarse material. Such inequalities in penetration
are minimized when the flow is greater [b}. (After Steenhuis et al., 2002) 64
Figure 4-11: Funnel flow at field site. Blue dye flowing vertically (unsaturated flow) moves laterally
when it encounters a sloping coarse-texture lens. (Source: Soil and Water Lab Website, Cornell
University) 64
Figure 4-12: Pre-conceptual model of the irrigation sites 67
Figure 4-13: Conceptual overview of spoils reactive model.. 74
Figure 5-1: Multi-parameter probe with different sensors 76
Figure 5-2: Graph of profiling down a borehole 76
Figure 5-3: Multi-parameter profiling in WISH 77
Figure 5-4: Specific Depth Sampler (from www.solinst.com) 77
Figure 5-5: Soil sampling methods 79
Figure 5-6: Porous cup lysimeters and vacuum pump 83
vi
Figure 5-7: Installation oftensiometers (left) and replenishment (right) 85
Figure 5-8: Photograph of the different leaching columns used (left) and annular rings (right). 87
Figure 5-9: Idealised depiction of how annular rings minimise the artificialflow afwater along the
soil-wall interface of the lysimeter by redirecting sidewall flow inward (after Corwin, 2000).88
Figure 5-10: An IGS humidity cell, left and an array of cells set up at the IGS, right 90
Figure 6-1: Monitoring borehole at Kleinkopje Pivot 4 97
Figure 6-2: Time graph of the electrical conductivity at Kleinkopje Pivoti 98
Figure 6-3: Time graph of the calcium, sulphate and nitrate at Kleinkopje PivotJ.. 99
Figure 6-4: Time graph of the EC at Pivot 4 boreholes 101
Figure 6-5: Time graph of the Electrical Conductivity at New Vaal Pivot 102
Figure 6-6: PV5 electrical conductivity trend since monitoring began 102
Figure 6-7: Time graph of the Electrical Conductivity at Syferfontein Pivot 104
Figure 6-8: Soil leaching analyses with depth within Pivot 1 106
Figure 6-9: Soil leaching analyses with depth within Pivot 4 107
Figure 6-10: Soi/leaching analyses with depth inside the pivot area 108
Figure 6-11: Soil leaching analyses with depth showing the dam's influence 109
Figure 6-12: Soil leaching analyses with depth outside the pivot area 109
Figure 6-13: Durov diagram of the monitoring holes and piesometers at New Vaal 111
Figure 6-14: Soil leaching analyses within the pivot area 113
Figure 6-15: Soil leaching analyses outside the pivot area 113
Figure 6-16: The porous cup and tensiometer installation (left) and porous cup sampler plus vacuum
pump (right) 115
Figure 6-17: Porous cup installation at Pivot 4 115
Figure 6-18: Sulphate concentrations with depth in the porous cups 116
Figure 6-19: Equivalence plot of various ions at Pivot Major 117
Figure 6-20: Sulphate concentrations liberated from the soil with depth. 117
Figure 6-21: Molar distribution of cations in the soil at Pivot Major 118
Figure 6-22: Sulphate concentrations with depth in the porous cups at Pivot 4 119
Figure 6-23: Molar distribution of cations in the soil at Pivot 4 119
Figure 6-24: Sulphate concentrations with depth in the porous cups 122
Figure 6-25: Equivalence plot of various ions at Syferfontein 123
Figure 6-26: Soi/leaching analysis of sulphate in depth for different positions 123
Figure 6-27: Comparison of soil water and leaching quality 124
Figure 6-28: Molar distribution of cations in the soil at Syferfontein 124
Figure 6-29: Molar distribution of major cations and anions in the soil at New Vaal 126
Figure 6-30: Diffractogram of Bassanite (CaSO.0.5 H20) 131
Figure 6-31: Photograph of the leaching columns 132
vii
Figure 6-32: Graph of sulphate and chloride leaching columns (mg/L) 134
Figure 6-33: Graph of sulphate leaching columns expressed as afraction of the input 134
Figure 6-34: Water and phosphate leachate compared with original chloride in soil 136
Figure 6-35: Tensiometer data for Kleinkopje 1 139
Figure 6-36: Tensiometer datafor Kleinkopje 4 140
Figure 6-37: Tensiometer datafor Syferfontein (tension in mm) 141
Figure 6-38: Texture chart (Soil classification working group -1991; Hillel, 1982) 142
Figure 6-39: Tensiometer data for Kleinkopje 1 with estimated volumetric water content (tension in
mm) 146
Figure 6-40: Tensiometer data for Kleinkopje 4 with estimated volumetric water content .. 146
Figure 6-41: Gypsum precipitation calculations at Kleinkopje pivots (After Van der Westhuizen, 2005)
............................................................................................................................................... 150
Figure 6-42: Leaching drums with different types of soils and gypsum concentrations 151
Figure 6-43: Time graph of the gypsum saturation of the different leachates 152
Figure 6-44: Leaching rate of sulphate from the leaching drums 153
Figure 6-45: EC of the leachate from the irrigated soil 154
Figure 6-46: Ca, Mg and SO. qualities of the leachate over the period of leaching. 155
Figure 6-47: Graphs of the decrease in concentration over time at KK1 158
Figure 6-48: Graphs of the decrease in concentration over time at KK4 159
Figure 6-49: Graphs of the decrease in concentration over time at Syferfontein 160
Figure 6-50: Graphs of the decrease in concentration over time at New Vaal 161
Figure 6-51: Increase in sulphate concentration in well-constructed boreholes 169
Figure 6-52: Clay percentages against the diffusiveflux 171
Figure 6-53: Graph of clay percentage versus unsaturated K. 172
Figure 7-1: EC of all cells over the testing period. 176
Figure 7-2: pH of all cells over the testing period. 177
Figure 7-3: EC and pH of spoil humidity cell over time (deionised water addition) 177
Figure 7-4: EC and pH of spoil humidity eel! over time (mine water addition) 178
Figure 7-5: EC and pH of spoil humidity eel! over time (recycled water addition) 178
Figure 7-6: EC and pH of spoil humidity eel! over time (cell flooded) 179
Figure 7-7: SO. of al! cells over the testing period. 179
Figure 7-8: Ca and Mg over the testing period 180
Figure 7-9: Alkalinity decrease as acidification occurs 180
Figure 7-10: Depletion ofNP and pH over time 181
Figure 7-11: Depletion ofNP vs. %S remaining. 181
Figure 7-12: Sulphate vs. saturation ofgypsum 182
Figure 7-13: Decrease in calcite saturation vs. pH 182
viii
Figure 7-14: Calcite saturation vs. pH 183
Figure 7-15: EC of non-acidic cells over time 184
Figure 7-16: pH of non-acidic cells over time 184
Figure 7-17: EC and pH of spoils humidity cell over time (deionised water addition) 185
Figure 7-18: EC and pH of spoils humidity cell over time (mine water addition) 185
Figure 7-19: EC and pH of spoils humidity cell over time (mine water addition) 186
Figure 7-20: EC and pH of spoils humidity cell over time (recycled water addition) 186
Figure 7-21: EC and pH of spoils humidity cell over time (cells flooded) 187
Figure 7-22: Sulphate concentrationsfor the non-acid-generating spoils humidity cells 188
Figure 7-23: Ca and Mg valuesfor non-acid cells 188
Figure 7-24: NP depletion in non-acid-generating spoils over time 189
Figure 7-25: pH and NP depletion 189
Figure 7-26: Sulphate concentrations and gypsum saturation over time 190
Figure 7-27: Dolomite and calcite saturation vs. pH over the testing period. 190
Figure 7-28: EC variations in columns over time 193
Figure 7-29: SO. variation in columns over time 193
Figure 7-30: Ca variation in columns over time 194
Figure 7-31: Mg variation in columns over time 194
Figure 7-32: Gypsum saturation in leaching columns over time 195
Figure 7-33: Sulphate concentrations and gypsum saturation in columns 195
Figure 7-34: Calcite saturation in columns 196
Figure 7-35: pH recovery in columns over time 197
Figure 7-36: EC in columns over time 198
Figure 7-37: Sulphate concentrations 198
Figure 7-38: Pit in the spoils, illustrating the heterogeneity of the spoils 199
Figure 7-39: Installation of leaching columns 199
Figure 7-40: Installed leaching columns at Optimum pivot 200
Figure 7-41: Sulphate values from field columns over time (mg/L) 200
Figure 7-42: pH infield columns over time 201
Figure 7-43: Gypsum saturation infield columns 201
Figure 7-44: Comparison offield and laboratory sulphate values 202
Figure 7-45: Optimum spoils tensiometer data 203
Figure 7-46: Expected concentration profile over time (using the average mineral assemblage
determined by this study and an aquifer transmissivity of1 m2/day) 205
Figure 7-47: Expected pH profile over time 205
Figure 7-48: Molar concentrations throughout the reaction path. 207
Figure 7-49: Concentrations of major ions over time 208
ix
Figure 7-50: Associate mineral saturations 209
Figure 7-51: Latter part of reaction sequence mineral saturation 209
Figure 7-52: pH variation 210
Figure 7-53: Decrease in reactive minerals over time 210
Figure 7-54: Concentrations of major ions over time 211
Figure 7-55: pH response showing reaction sequence effects over extended time 211
Figure 7-56: Decrease in reactive minerals over time 212
Figure 7-57: Concentrations of major ions over time with reduced recharge 212
Figure 7-58: Associate mineral saturations with reduced recharge 213
Figure 7-59: pH profile with reduced recharge 213
Figure 7-60: The influence of O, concentration on relative activity ofT.ferrooxidans (Jaynes et al.,
1984) 214
Figure 7-61: Decrease in concentrations at lower oxidation rate 215
Figure 7-62: pH recovery ifpyrite rate is decreased by an order 215
Figure 7-63: Mineral saturations at low pyrite oxidation rate (flooding conditions) 216
Figure 7-64: Aqueous concentration in low NP conditions showing an increase in sulphate and
therefore in TDS 216
Figure 7-65: RapidpH-drop 217
Figure 7-66: Mineral saturation showing depletion of carbonates associated with acidification. 217
Figure 7-67: Concentrations of major ions over time with excess NP 218
Figure 7-68: pH profile with excess NP 218
Figure 7-69: Calcite dolomite and pyrite depletion in high NP scenario 219
Figure 7-70: Mineral saturation over time for high NP scenario 220
Figure 7-71: Concentration profile infresh inundated spoils 220
Figure 7-72: Neutral pH infresh inundated spoils 221
Figure 7-73: Mineral saturation where fresh spoils are inundated. 221
Figure 7-74: Concentrations of major ions over time with evaporated Optimum water 223
Figure 7-75: Concentrations of major ions over time with Optimum column water 224
Figure 7-76: Decrease in reactive minerals over time 224
Figure 7-77: pH profile at Optimum over time 225
Figure 7-78: Mineral saturation of the Optimum 2m column water 225
Figure 7-79: Concentrations of major ions over time with SWB simulated water 226
Figure 8-1: Detailed model of Kleinkopje 1 233
Figure 8-2: Detailed model of Kleinkopje 4 236
Figure 8-3: Detai/ed model of Syferfontein 239
Figure 8-4: Deatai/ed model of New Vaal 242
Figure 9-1: Rationalefor defining site-specific criteria. 245
x
Figure 9-2: Data input required to determine impact and define criteria 246
Figure 9-3: Terrain morphological units for soil Ba 19 as an example 249
Figure 9-4: Soil types in study area 249
Figure 9-5: GIM1 input screen with clickable soils map to link to soil properties 262
Figure 9-6: Example of input screen from G1MI. 263
Figure 9-7: Water Quality and Environmental Data in GIM! 264
Figure 9-8: Report of the assessment 270
List of Tables
Table 2-1: Water quality changes over the coalfields (mg/L) 17
Table 2-2: Irrigation water qualities (February 2004) 18
Table 3-1: (a) Average water quality over 4 years of the New Vleishafi dam at Pivot1 (from
Annandale et al., 2006) with statistics on the water in table (b) 29
Table 3-2: (a) Water quality of the Irrigation dam at Pivot4, with statistics on the water in table (b).
................................................................................................................................................. 29
Table 3-3: Average water quality for the Irrigation dam 595 36
Table 3-4: Water quality of the irrigation dam 43
Table 4-1: Key processes in opencast mining that affect water quality (Usher, 2003) 73
Table 5-1: Summary of all the experiments 96
Table 6-1: Water quality of the groundwater at the Kleinkopje Pivot I (November 2005) 98
Table 6-2: Water quality of the groundwater at Pivot 4 (November 2004) 100
Table 6-3: Quality of the groundwater at the New Vaal Pivot in mg/L (March 2005) lOl
Table 6-4: Water quality of the groundwater at the Syferfontein Pivot 103
Table 6-5: Soil properties at Kleinkopje Pivots (%). 106
Table 6-6: Soil properties at New Vaal (%) 108
Table 6-7: Chemical analysis afwater in the piezometers 110
Table 6-8: Down-the-hole profiling of EC (mS/m) 110
Table 6-9: Soil characteristics at Syferfontein Pivot 111
Table 6-10: Analysis from porous cup sampling at Kleinkopje 116
Table 6-11: XRF analysis of the Kleinkopje soils 120
Table 6-12: XRD analysis ofKleinkopje soils 121
Table 6-13: Proportional relationship between water qualities at 2 and 3.5 m depths 122
Table 6-14: XRF analyses at Syferfontein 125
xi
Table 6-15: Clay content in the soil at Kleinkopje Pivot 4 132
Table 6-16: pH and chloride values of soil at Kleinkopje Pivot 4 (background values) 133
Table 6-17: Composition of the leaching water 133
Table 6-18: Water and potassium-orthophosphate extraction values of sulphate 136
Table 6-19: Kleinkopje 1 (Bainsvlei I Clovelly soil): 144
Table 6-20: Kleinkopje 4 (Hulton soil): 145
Table 6-21: Kleinkopje 1 estimated moisture content: 145
Table 6-22: Kleinkopje 4 estimated moisture content: 145
Table 6-23: Average estimated moisture content with depth 147
Table 6-24: Salt balance calculations at the irrigation sites 148
Table 6-25: Percentage salt in layers in depth through the profile down to water level 149
Table 6-26: Daily calculated ETvalues at the Highveld and New Vaal 156
Table 6-27: Conductivity valuesfor differentflow zones through the profile at KK1 (with the flux in
m/d) 158
Table 6-28: Conductivity valuesfor differentflow zones through the profile at KK4 (with the flux in
mld) 160
Table 6-29: Conductivity values for differentflow zones through the profile at Syferfontein (with the
flux in mld) 161
Table 6-30: Conductivity values for different flow zones through the profile at New Vaal.. 162
Table 6-31: Chloride values with depth at the irrigation sites (mgil) 165
Table 6-32: Percentage flux that contributes to the chloride mixing process to give groundwater Cl
concentration 167
Table 6-33: Actual fluxes (m3Idlm2) at the irrigation sites 167
Table 6-34: % Flux that reaches groundwater 168
Table 6-35: Sulphate values at irrigation sites (values are tabled downwards in time from the start of
irrigation until 2006) 169
Table 6-36: Ks and unsaturated K-valuesfor different clay values (silt = 20%) 172
Table 7-1: ABA results from spoil columns 192
Table 7-2: Initial leachate results from columns 192
Table 7-3: ABA results from the spoil columns 197
Table 7-4: Optimum irrigation water quality (before evaporation). 222
Table 7-5: Optimum irrigation water quality (after evaporation) 222
Table 7-6: Optimum 2m column water quality (after evapotranspiration) 222
Table 7-7: Water quality simulated by SWB (water quality input before simulation by SWB).223
Table 7-8: Water quality simulated by SWB (water quality input after simulation by SWB).223
Table 8-1: Summary of all the experiments 243
Table 9-1: Soil suitability rating guide for well-drained soils 251
xii
Table 9-2: Soil suitability rating guide for moderately drained soils 251
Table 9-3: Distribution ofsoils suitable for irrigation in different land types of the area 252
Table 9-4: Minimum monitoring requirements at various types of waste management facilities
(DWAF, 2005) 258
xiii
1 INTRODUCTION
1.1 Overview
South Africa is a country blessed with an abundance of minerals, and is the fourth largest
coal producer in the world. Coal is the main source of energy; the Energy Information
Administration (EIA, 2000) stated that in 1997, 68% of the energy consumption of South
Africa was coal-based.
Coal resources in South Africa are contained in 19 coalfields (Erasmus et af, 1981). The
most important coalfields in the Mpumalanga province are the Springs-Witbank Coalfield and
the Highveld (Eastern Transvaal) Coalfield. A map of the coalfields in South Africa is shown
in Figure 1-1.
• PolgictClblUb "-
~~ \.
)
,-..___./ ,,'"<il" 1I
I
SPRINGnOK~
I
-> .'."m"'/ ;' r
$WAZIl.AND
""
NAME OF COAL-FIELD
LIMPOPO
WATERBERG
SOUTPA:-ISBERG
rAFURl
SPRINGROK FLATS
WESTERN AREA
SPRINGS-WITBANK
IWMATIf'OORT
O.F.S.-VIERFONTETN
OLD SPRINGFIELD
VEREENIGING,SASOLBURG
SOUTH RAND
IIIGI-IVELO
EASTERN TRANSVAAL
kmO~_ ..lOO KUP RIVER
UTRECIIT
VRYHEID
ZUl.Ul.AND
MOLTENO-INDWE
Figure 1-1: The South African Coalfields, after Erasmus et al. (1981).
1
The aerial extent and the period of mining operations that altered the hydrogeological
environment of the coalfields impact on the operational life of the mines for several years
after mining has ceased.
Large volumes of acid mine drainage produced in the Mpumalanga coalfields are mostly
neutralised by seepage through neutralising strata, or artificially by the addition of lime. Mine
drainage is, therefore, often saturated with gypsum (CaS04.2H20). When released into the
surface water environment, the associated high salinities are frequently responsible for
unacceptable water quality degradation.
South Africa is a water-scarce country, and therefore a significant impact on the environment
has been forecast, as early as 1983 (Funke, 1983). However, in recent years, many mines
have been striving to ensure that the negative impacts of their mining operations are kept
within acceptable limits, also in terms of the water qualities.
1.2 The problem of mine closure
After mineable coal has been removed, mines are left to fill with water and seepage takes
place into adjacent, polluting aquifers and rivers.
The National Water Act requires mines to separate clean and dirty water and to contain, and
treat if necessary dirty water, so as to prevent impacts on the surface and groundwater
resources. The Minerals Act of 1991 was the first that forced mining operations towards
sustainable land management instead of just beautifying disturbances. This Act, Act 50 of
1991, stated that the mine remained the property of the owner until a closure certificate had
been issued. The law allowed authorities to gain insight into and control mining activities that
could adversely affect the water environment. This law was substituted by the Mineral and
Petroleum Resources Development Act, Act 28 of 2002. This Act states that no closure
certificate may be issued unless the management of potential pollution to water resources
has been addressed. The MPRDA requires all mines to make financial provision for closure,
and now goes as far as making sure mines have the necessary funding to cover current
liabilities (in the event of bankruptcy).
The government has turned down many applications for mine closure, because of the
concern of possible uncontrolled pollution of water resources in the vicinity of the mines. The
government's reluctance to grant closure certificates has forced the industry to investigate
other ways of managing the water pollution problem. One of the ways forward is to look at
natural passive options, e.g. filling up of mines with water to decrease the ingress of oxygen
and intermine flow. The Institute for Groundwater Studies at the University of the Free State
has done a large amount of research in this regard over the years (Hodgson and Krantz,
2
1998; Hodgson et al, 2001; Grobbelaar, 2001; Hough, 2002; Havenga, 2002 and Vermeulen,
2003).
1.3 Irrigation with gypsiferous water
Irrigation provides a novel approach to the utilization and disposal of gypsum-rich water. The
potential for use of gypsiferous mine water for crop irrigation in South Africa was first
evaluated by Du Plessis (1983). Research at scales ranging from experimental to semi-
commercial over a period of more than ten years has demonstrated the potential to
successfully use this water for the irrigation of a range of crops and significantly reduce the
salt load emanating from mine drainage, by precipitating gypsum within the soil. Jovanovic et
al. (1998) proved that irrigation with this water should not present a soil salinity or crop
production problem within the relatively short time period of three years, provided that careful
fertilisation management is applied. An evaluation of the feasibility and sustainability of this
practice from agricultural and environmental perspectives was the subject of a WRC project
by Annandale et al (2006). The researchers simulated the long-term soil-water and salt
balance, as well as the impact to soil and groundwater resources. The journal Water SA
(Jovanovic et el, 2002) stated that the disposal of mine wastewater is a worldwide problem
wherever operating coal and goldmines, as well as closed underground and open-cast
workings are found (pulles et al., 1995).
The type of water emanating from mines depends largely on the geological properties of the
coal and other geological material with which waters come into contact. The concentrations
of salts and other constituents frequently render such waters unsuitable for direct discharge
to the river systems, except in periods of high rainfall when an adequate dilution capacity is
present and controlled release permitted. DWAF permits mines to discharge polluted mine
water as part of the Controlled Release Scheme. The objective is to allow mines to release
some of their excess water, but in such a way that the rivers and dams, as well as aquatic
organisms, are not negatively impacted upon.
As large volumes of water in underground and opencast collieries are not utilised, the use of
available water resources iun a water-scarce country like South Africa should be optimised.
Gypsiferous mine water can thus be regarded either as one of the greatest problems
associated with mining, or as a potential asset.
Large amounts of wastewater could be possibly available to the farming community for the
irrigation of highly productive soils nearby the coalfields of the Mpumalanga Province where
available water resources are scarce. The concentration of the qypslferous mine water
through evapotranspiration, thereby precipitating gypsum in the soil profile, may reduce
3
environmental pollution by eliminating these salts from the water system (Annandale et aI.,
2002). The aluminium toxicity, often occurring in these soils, could be reduced by irrigation
with gypsiferous mine water (Barnard et a/., 1998). A surplus water problem of the coal-
mining industry could also be significantly reduced by the irrigation option with qypslferous
mine water during dry periods; this could release the aluminum toxicity into river systems
during rainy periods.
Irrigation with mine water is thus currently one of the potential uses of excess mine water.
However, there is continuing concern from especially the regulatory authorities regarding the
long-term impact that larger-scale implementation of this practice may have on water-
resource (specifically groundwater) quality and quantity. The sections, which follow, will
address these concerns.
1.4 Objectives
Sufficient detail does not exist regarding the subsurface behaviour and long-term impact of
the use of mine water for irrigation. The thesis provides the following research studies
needed to fill that gap. The research investigated:
o The determination of hydraulic interaction of irrigated mine water with the underlying
aquifers
o The assessment of the effect of irrigation on water quality and quantity at rehabilitated
opencast colliery spoils and on virgin areas.
o The determination of salt migration and attenuation from irrigated areas under natural
and spoils conditions.
o The establishment of criteria for site selection/operation, monitoring, determination of
impacts and mitigation methods for mine water irrigation areas as part of integrated
mine water management.
1.5 Method of Investigation
Irrigation with mine water has been ongoing for several years at the sites as indicated in the
next paragraph. The sites of the studied areas have not yielded significant indications of
deterioration of groundwater quality monitoring. Methodologies for monitoring and predicting
impacts therefore had to be adapted to provide meaningful estimates of expected salt loads
for the long term and for more intensive irrigation. Thus, detailed groundwater monitoring
throughout the duration of the project was an important objective.
4
Research was done at five irrigation sites (Figure 1-2):
• Two pivots at Kleinkopje Colliery near Witbank, i.e. Kleinkopje 1 (Major) and
Kleinkopje 4
• Syferfontein Colliery between Kriel and Trichardt
• Optimum Colliery (rehabilitated opencast) between Witbank and Hendrina
• New Vaal Colliery near Vereeniging
In Figure 1-3 the position of the collieries in Mpumalanga where the research was conducted,
can be seen.
j_ _j_
-2700000
-3OOIXXJO
-33OCXXXJ
-3600000
-39OCXXJO '----,---___ ,---- r
-1200000 -900000 -6OCOOJ -3OO:XXl o
Figure 1-2: South African Map with the towns near to the irrigation sites.
5
-2880000
-2910000 I
-2940000
LEGEND
-2970000 ' _Towns
I_NRDoaadms .
Rtvere
I j
T
-30000 0 30000 60000 90000 120000 150000
Figure 1-3: Colliery map of Mpumalanga, tYlustrating the collieries with irrigation sites
investigated in this research. (Map courtesy of Grobbelaar et aI., 2002).
To achieve the project aims, the following methodologies were applied:
1. Determination of hydraulic interaction of irrigated mine water with the underlying aquifers.
• Seasonal water level measurement within and down gradient of irrigation pivots
• Soil moisture and salt migration measurement using aspects such as porous cups,
augur drilling, and borehole drilling in the spoils/aquifer.
• Pumping tests and point dilution tests
2. Assessment of the effect of irrigation on the hydrology and water quality at opencast spoils
and on virgin areas.
• Installation of porous cups/lysimeters to collect moisture at various depths in soil and
spoil profiles
• Installation of water interception systems (columns) within unirrigated spoils to
measure water quality and volumes response over time.
• Installation of monitoring boreholes within the areas of irrigation (virgin soil and spoils)
and in adjacent non-irrigated areas.
6
• Trenching in the irrigated areas (spoils and natural) to determine hydrochemical and
hydraulic interactions through direct observation.
• Leaching columns and humidity cells
• Simplified geochemical modelling
3. Determination of salt migration and attenuation from irrigated areas under virgin and
spoils conditions.
• Obtaining soil and overburden samples through direct access methods such as
trenches, drilling and auguring. The profiles obtained in this way were tested through
techniques such as water leaching/extraction and mineralogical determination by
means of XRF/XRD.
• Laboratory column leachate experiments to observe variation in attenuation or salt
migration for different media or conditions.
• Installation of monitoring boreholes within areas of irrigation and in adjacent non-
irrigated spoils.
• Groundwater monitoring results
4. The establishment of criteria for site selection/operation, monitoring, determination of
impacts and mitigation methods for mine water irrigation areas. The lessons learnt are
synthesised into clear guidelines regarding the site selection, groundwater monitoring, and
impact determination for irrigating with mine water in South African conditions. This provides
a suggested best practice for mine water irrigation in South Africa.
1.6 Thesis Structure
The thesis is structured as follows:
• Chapter 1 is an introduction.
• Chapter 2 deals with coal mine water irrigation.
• Chapter 3 provides a general site description.
• Chapter 4 is a discussion 0 ydraulic characteristics and hydrochemical impacts
of irrigation on soil and spoils, and an initial conceptual model of the interaction
between irrigation and water resources.
• Chapter 5 describes the methodologies used in the investigation
• Chapter 6 gives a detailed characterisation of virgin soil irrigation sites
7
o Chapter 7 deals with irrigation at opencast collieries (rehabilitated coal spoils).
o Chapter 8 describes the consolidated models of mine water irrigation for virgin soil
irrigation
o Chapter 9 deals with operation and monitoring guidelines and provides details of a
tool (GIMI) to facilitate irrigation site selection.
o Chapter 10 gives the overall conclusions, and deals with feasibility and
recommendations.
o The Appendices (on the attached CD) contain all the data gathered during the
investigation, calculation sheets, and chapters on:
1. Coal mining in South Africa
2. Irrigation in agriculture - an overview
3. Gypsum in soils.
8
2 IRRIGATION WITH COAL MINE WATER
2.1 Introduction
The mining industry in South Africa has historically been plagued by the accumulation of
excess water (see Appendix A for detail on coal mining and water related issues). When the
mining void is filled, whether opencast or underground, the result is often that water
eventually decants at the surface, or seeps along the strata. As the workings fill up, mine
water is forced out at the lowest interconnection between the mine and the surface. South
Africa is a relatively dry country, and seepage towards, or decant into the rivers, can have a
significant impact on the overall state of the rivers and underground aquifers. This is
particularly relevant in the coal mining industry. Stream water monitoring shows that the
water quality has deteriorated over the last 20 years (Grobbelaar et a/., 2000).
For the sake of current and future generations it is neccessary to safeguard the purity and
quantity against pollution by irresponsible mineral development operations.
The rehabilitation of new mine sites should ideally be planned before starting any mining,
including a future water management plan. This plan should also include the management of
the mine water and groundwater after mine closure. In particular, the data should include
physical, chemical, hydrological and geotechnical properties of the ore (coal), and adjacent
country rocks to develop an EMPR (Taylor, 2000). This is also necessary for operational
mines to ensure that the best result is achieved and that the community and regulatory
authorities are satisfied.
While there have been improvements in mining practices in recent years, it still has
significant negative impacts on the environment, which includes the disturbance of
groundwater during mine construction. These impacts depend on a variety of factors, such as
the sensitivity of the local terrain, the composition of minerals being mined, the type of
technology employed, the skill, knowledge and environmental commitment of the company,
and finally, the ability to monitor and enforce compliance with environmental regulations.
Water pollution from mine waste rock and tailings may need to be managed for decades, if
not centuries, after closure.
One of the problems is that mining has become increasingly mechanised and therefore able
to handle more rock and ore material than ever before. Consequently, mine waste and
therefore also polluted mine water, has increased enormously. As mine technologies are
developed to make it more profitable to mine low-grade ore, even more waste will be
9
generated in the future. This trend requires the mining industry to adopt and consistently
apply practices that minimise the environmental impacts of this waste production.
Once a mine is in operation, water protection must remain the highest goal of the company,
even if it means reduced mineral productivity. Adopting this common-sense ethic is the only
way we can ensure that the golden dreams of mining do not turn into the nightmare of
poisoned streams (www.miningwatch.org). In the right place, and with conscientious
companies, new technologies and good planning, many of the potential impacts are
avoidable.
Not all mine waters are bad news. Some mine waters do have a positive impact on the water
environment. Banks et a!. (1996) state that a considerable number of mine waters are of a
good quality and can thus be used for potable supplies. Furthermore, a substantial amount of
mine water is of good enough quality to be utilised in industries and for irrigation (Reddy et
a!., 2000). Careful management is however necessary to avoid over-utilising such water
(Younger, 1993a; 1998b). If the collieries are excessively dewatered, it can result in:
o The decrease of stream flows and wetlands
o Lowering of the water table near water supplies and irrigation wells
o Land subsidence
2.2 Water volumes available for irrigation
Previous work (Hodgson and Krantz, 1998; Grobbelaar eta!., 2000; Hodgson eta!., 2001 and
Vermeulen, 2003) has stressed that the likely time-scales should not be ignored. The time for
the workings to fill with water, for example, is relatively short. The rise in water levels in
opencast mining is in the order of 1 m per annum. Depending on the pit geometry, decanting
usually occurs within ten years after mining has ceased. Depending on the recharge
coefficient, underground mining could take longer to fill/decant. Many of the operating
underground mines are already 20 - 40% filled with water. Often at closure, only small areas
of these mines remain to be filled by water influx. Decanting at many locations is therefore
predicted within the next 40 - 60 years. Some seepage into the weathered zone will also
occur, but due to the relatively low hydraulic conductivity of the weathered strata, the
seepage component will be relatively small. However, in many of the deeper mines where the
cavities fill up, no decanting will occur due to the absence of any significant piesometric head
differences in the groundwater regimes.
10
Grobbelaar et al. (2000) provided projections of future water volumes to decant from the coal
mines. In total, about 360 ML/d will eventually decant from all the coal mines in combination.
On a catchment basis, this amounts to (in ML/d):
Wilge/Klip Olifants Klein Olifants Vaal Komati
23 170 45 120 2
This demonstrates the anticipated future scale of the water that the industry will have to deal
with. However, an investigation currently undertaken by Hodgson for the coal mining
fraternity indicates that more than half of the available water will be treated for human
consumption, and will not be available for irrigation.
To put these figures in perspective and determine the volumes to be available for irrigation,
the following figures can be considered:
Opencast: The void space in spoils is 20%, which implies that 2000 m3 of water will be in
storage per hectare for every meter of depth in the spoils. The average annual rainfall for the
Mpumalanga region is 680 - 750 mm.
Approximately 75 000 ha of opencast will eventually be rehabilitated in the Mpumalanga
region, according to the Kraai van Niekerk report (Schoeman, 2001). If it is assumed that the
recharge to the rehabilitated opencasts will be in the order of 20% of MAP (Vermeulen,
2003), and that the annual irrigation water for 1 ha is 5 000 rrr', then in excess of 20000 ha
could be irrigated for a single crop. This assumes that all the water can be collected and that
sufficient appropriate areas are available for irrigation.
Underground: The average void space for bord-and-pillar mining is 65%. At an average
mining height of 3m, 19500 m3 of water will be in storage when filled for every hectare mined.
Calculated according the average rainfall, shallow mines with a recharge of 5-10% of MAP
will annually have an amount of 350 - 700 m3 /ha available annually for irrigation. This means
that every 7-14 ha will provide enough water to irrigate one hectare. These values are also
applicable to stooped areas. For deep underground bord-and-pillar areas with a recharge of
4%, 18 hectares will provide enough water for one hectare of irrigation.
11
, I
I'
I'
2.3 Composition of coal mine waters
2.3.1 Introduction
Large amounts of wastewater could possibly be made available to the farming community
and utilised for the irrigation of highly productive soils in the coal-fields of the Mpumalanga
Province in South Africa. Concentrating the gypsiferous soil solution through
evapotranspiration, and precipitating gypsum in the soil profile, may reduce environmental
pollution as these salts are removed from the water system (Annandale et ei, 2002). The
surplus water problem for the coal-mining industry could also be significantly reduced by the
irrigation option (mine water utilisation during dry spells), which could complement the
controlled release into river systems during rainy spells.
2.3.2 Acid Mine Drainage
Acid Mine Drainage is a widespread phenomenon in the mining industry worldwide, and also
affects the water quality of the collieries in South Africa. With the new mine closure law
applicable in South Africa, it is important to find ways of curbing, or at least slowing down the
process of water acidification, in order to attain mine closure certificates. Industry, labour,
government and environmentalists unanimously agree on the issue of AMD as the number
one environmental problem facing the mining industry (Domville et a/., 1994).
In the coal mining industry the pyrite in the coal is roughly three times as high as in the
surrounding rock (Usher, 2002-personal communication). This result in the generation of
sulphate, heavy metals and acidity, and can have numerous environmental consequences
(Usher et aI., 2003), for the following reasons:
o It devastates fish and aquatic habitat
o It is virtually impossible to reverse with existing technology
o Once started, can costs millions to treat
o Is very complex to control and treat (Lawrence and Day, 1997)
According to Seif (2002), as early as 1978, many variations of AMD passive treatment
systems have been studied by numerous organizations on the laboratory bench-testing level.
During the last 15 years, passive treatment systems have been implemented on full-scale
sites throughout the USA with promising results. The concept behind passive treatment is to
allow the naturally occurring chemical and biological reactions that aid in AMD treatment to
occur.
12
It is therefore of the utmost importance for the mining industry to know the
characteristics/capacity of waste rock, overburden, pit walls, pit floor and tailings, to forecast if it
will produce Acid Mine Drainage or not (Usher et al., 2003).
AMD is primarily a function of the geology, geohydrology and mining technology employed at
the mine site (Seif, 2002). AMD is the result of a series of complex geochemical and
microbial reactions that occur when water comes in contact with pyrite in coal, refuse or the
overburden of a mining operation. AMD occurs by the oxidation of the sulphide minerals
(mostly pyrite) in the country rock (coal and surrounding rocks) as a result of exposure to
moisture and oxygen. This results in the generation of sulphates, metals and acidity. Pyrite
(FeS2) is the most important sulphide found in South African coal mines. When exposed to
water and oxygen, pyrite can react to form sulphuric acid (H2S04). The resulting water is
usually high in acidity and dissolved metals such as lead, copper, zinc, arsenic, selenium,
mercury and cadmium (Hadley and Snow, 1974). These heavy metals are fatal, because they
prevent the energy production that is essential for life, by combining with -SH groups
(Stevenson, 1997). Certain naturally occurring bacteria can significantly increase the
oxidation rate. The contaminated water is often reddish brown in color, indicating high levels
of oxidised iron (Hadley and Snow, 1974). Bright orange-colored water and stained rocks are
usually telltale signs of AMD (Usher et a/., 2001). The dissolved metals remain in solution
until the pH rises to a level where the iron oxidizes and precipitation occurs. The precipitates
can be harmful to aquatic life, as the clumps reduce the amount of light that can penetrate the
water, affecting photosynthesis and the vision of animals. When the precipitate settles on the
streambed, it smothers the bottom-dwellers and their food resources.
There are four commonly accepted chemical reactions that represent the chemistry of pyrite
weathering to form AMD. An overall summary reaction is as follows (Seif, 2002):
l
Pyrite + Oxygen + Water -----11> "Yellowboy" + Sulphuric acid
The first reaction in the weathering of pyrite includes the oxidation of pyrite by oxygen. Sulfur
is oxidized to sulphate and ferrous iron is released. This reaction generates two moles of
acidity for each mole of pyrite oxidised.
1). 2 FeS2+ 7 02 + 2 H20 ----l:2 Fe2+ + 4 sol- +4 H+
Pyrite + Oxygen + Water ---I> Ferrous iron + Sulphate + Acidity
13
The second reaction involves the conversion of ferrous iron to ferric iron. This process
consumes one mole of acidity. Certain bacteria increase the rate of oxidation from ferrous to
ferric iron. This reaction is pH dependent with the reaction proceeding slowly under acidic
conditions (pH 2-3) without the presence of bacteria, and several orders of magnitude faster
at pH values near 5. The oxidation for pyrite by Fe3+(aq) is 2 - 3 orders of magnitude faster
than that by oxygen (Appelo and Postma, 1993), and is more rapid than the oxidation of
dissolved Fe2+to Fe3+. This process rapidly consumes all Fe3+ and pyrite oxidation will cease
unless Fe3+ is replenished in the oxidation process of Fe2+by oxygen. It is thus important to
know the oxidation rate for ferrous to ferric iron according to Equation 2. Numerous studies
on the ferrous iron oxidation rate show that, under acid conditions, the rate becomes very
slow (in the pH range of 3, the half life is in the order of 100 days). Such rates are
considerably slower than the oxidation rate of pyrite by Fe3+ (aq). Therefore this reaction
would be referred to as the "rate limiting step" in the overall acid-generating sequence, were
it not for the catalytic effect of bacteria.
2). 4 Fe2++ O2 + 4 H+ ----e>4 Fe3++ 2 H20
Ferrous iron + Oxygen + Acidity ---i> Ferric iron + Water
The third possible reaction is the hydrolysis of iron. Hydrolysis is a reaction that splits the
water molecule. Three moles of acidity are generated as a byproduct. Many metals are
capable of undergoing hydrolysis. The formation of ferric precipitate (solid) is pH dependant.
Solids form at a pH-value above about 3.5, whereas little or no solids will precipitate below
this value.
Ferric iron + Water -----i> Ferric Hydroxide (yellow boy) + Acidity
The fourth reaction is the oxidation of additional pyrite by ferric iron. The ferric iron is
generated in Reaction Steps 1 and 2. This is the cyclic and self-propagating part of the
overall reaction and takes place very rapidly. The cycle continues until either the ferric iron or
pyrite is depleted. Note that in this reaction iron, and not oxygen, is the oxidizing agent.
4). FeS2 + 14 Fe3++ 8 H20 ----i> 15 Fe2 + 2 SO/- + 16 H+
Pyrite + Ferric iron + water -----i> Ferrous iron + Sulphate + Acidity
14
Although these equations yield a correct picture, in that oxygen is the ultimate driving force
for the oxidation of pyrite, with the final products being an insoluble form of oxidized iron and
an aqueous sulphuric acid solution, this is an oversimplification of the problem (Azzie, 2002).
However, the equation do not explain the geochemical mechanisms or rates and it does not
reflect the slow oxidation of aqueous ferrous iron in acid solutions that often results in high
ferrous concentrations in acid mine waters. Furthermore, factors such as microbial catalysts,
neutralization reactions, sorption reactions, and climatic effects have an important influence
on pyrite weathering, but are not considered in the above equations (Nordstrom and Alpers,
1999a). The problem is multi-faceted and it must be emphasized that Acid Mine Drainage
occurs within a complex system for which several factors need to be considered by the
disciplines of inorganic and organic chemistry, geology and mineralogy, hydrology and
microbiology. During the initiation of pyrite oxidation, these are complex chemical and
microbiological processes occurring in microenvironments (Williams et al., 1982). The
existence of these environments is well illustrated by the formation of Jarosite
(KFe3(S04)2(OH)6), a mineral that can only form under acid conditions, but has been found
by Carson et el, (1982) in soil waters of near neutral pH-value.
Micro-organisms are abundant in natural waters containing AMD, and in many cases they are
the only life form under such conditions (Azzie, 2002). As early as 1919, Powell and Parr
suggested that bacteria might catalyze pyrite oxidation. Iron oxidizing bacteria can accelerate
the rate of Fe2+ oxidation by a factor of 106. The two acidophyllic bacteria, Ttuobeciuus
thiooxidans and Thiobaci//us terrooxidens, play an important role in the microbial degradation
of pyrite, and thus are an important factor in producing acid mine waters. The nutritional
requirements of Thiobaci//us are ubiquitous. Nitrogen and carbon dioxide are available in the
atmosphere, sulphur is available in the mine waters, and only phosphorus is needed. These
bacteria have several adaptive techniques that also permit them to tolerate low pH, and high
metal concentrations (Tuovinen et al, 1971).
Thiobaci//us thiooxidans oxidises elemental sulphur but not iron, T. ferrooxidans, oxidizes
both iron and sulphur compounds. A third species Leptospuiltum ferrooxidans, metabolically
behaves like T. ferrooxidans although it has a different morphology, which was first described
by Markosyan (Azzie, 2002). Other bacteria that play a role in these processes are
Thiomicrospira and Suttotoous.
Stages in the development of AMD.
Under natural conditions, the availability of oxygen acts as a limiting factor for oxidation to
occur. Rainwater, the natural source of recharge into the aquifers and into the mines,
contains an average of 8 mg/L of oxygen, which is available for reactions with natural
substances in the soil, such as organic material and sulphur-bearing minerals, including
15
pyrite (FeS2), chalcopyrite (FeS.CuS) and pyrrhotite (FeS). This oxidation process can yield
at most an equivalent amount of sulphate, which is 32 mg/L (Drever, 1997).
A complex combination of organic and inorganic processes and reactions are involved in
AMD. To produce severe AMD where the pH drops below 3, the sulphide minerals should
create an optimum environment for rapid oxidation in the system. This oxidation must
continue for a sufficiently long time to exhaust all of the neutralizing potential of the rock. The
potential of a sulphide-bearing rock to generate acid, is therefore strongly related to the
amount of alkaline material present in the rock (Usher et aI., 2003). When this sulphide-
bearing rock is exposed to moisture and oxygen, sulphide oxidation and acid generation
start. Calcium-based carbonates in the surrounding rock will immediately neutralise the
acidity, and neutral to alkaline conditions will be maintained in the water that passes over the
rock (Broughton et et., 1992). As this neutralizing agent becomes depleted, the water will
start to acidify. Consequently, the pH decreases, which in turn creates favorable conditions
for further acid generation.
As this process progresses, the rate of acid generation accelerates, and the pH progressively
decreases in a step-like manner (Figure 2-1). Each plateau of relatively steady pH represents
the dissolution of a neutralising mineral that becomes soluble at each specific pH. Eventually,
if the acid generation process removes all the neutralizing potential from the surrounding
rock, the pH will drop below 3, and AMD will become severe.
These stages can last for indefinite periods of time even centuries, until the sulphide minerals
are completely oxidised and the rock becomes inert, or until special waste management and
AMD control actions are taken (Durkin and Herrmann, 1996).
pH Evolution of AMD (after Morin. 1983)
Calcite
AI-OH
Fe·OH
Jarosite
Time
Figure 2-1. Stages in pH evolution as a result of different buffering minerals (after Morin,
1983 and Usher et aI., 2001).
16
2.3.3 Gypsiferous Waters
The general trend in the Mpumalanga and Kwazulu-Natal Coalfields is that the southern and
western coalfield waters are richer in sodium water than the northern and eastern areas. The
table below illustrates the average to highest values for a number of mines in these
coalfields. The western and southern coalfields i.e. Kilbarchan/Roy Point, ZAC and
Syferfontein, have a much higher sodium content than the eastern collieries, but a much
lower calcium and magnesium content in general. This implies that for this research project,
Syferfontein was irrigated with sodium-sulphate water (Table 2-2), whilst the irrigation water
of the Kleinkopje pivots were calcium - sulphate rich (gypsiferous type). The New Vaal
irrigation water is also calcium-sulphate rich, but at a far lower concentration than those of
Kleinkopje. The opencast waters at Optimum are also calcium - sulphate rich (gypsiferous
type).
Table 2-1: Water quality changes over the coalfields (mg/L).
Mine Area Ca Mg Na S04
Roy PointiKilbarchan South 324-377 81-291 411-1482 1841-4863
ZAC South 62-298 30-220 300-900 200-1423
Syferfontein South 35-41 84-91 664-879 1220-1232
Khutala North 82-416 42-174 32-103 360-2119
Optimum Northeast 197-534 77-355 83-152 1249-2750
Ermelo East 170-571 120-360 26-50 600-2805
New Largo Northwest 163-473 27-84 15-66 885-1860
Schoongezicht North 275-555 105-223 52-71 1199-2685
TNC Central 345-367 184-208 54-73 865-1634
Minnaar North 220-543 15-555 14-45 1230-1760
Kleinkopje Pivot 1 North 553-751 372-473 104-138 2917-3447
Kleinkopje Pivot 4 North 503-697 195-278 37-65 1935-2650
New Vaal West 68-122 24-94 16-87 172-737
In South African coalfields, co-existing carbonates such as calcite and dolomite can
neutralise acidity (Usher, 2003). Alternatively, the acidity can be neutralised by lime addition.
From the overall reaction of calcite as buffering mineral, it is evident that calcium and
sulphate will increase in concentration:
FeS2 + 2CaC03 + 3, 7502 + 1, 5H20 ¢> Fe(OH) 3 + 2S0/- + 2Ca2+ + 2C02
This increase in Ca2+ and SO/- can only occur to a certain extent, to a point where the
aqueous solubility of these ions is limited by the solubility of gypsum (CaS04.2H20). Using
the PHREEQC geochemical model (Parkhurst and Appello, 1999), the saturation state of the
neutralised mine water used to irrigate at Kleinkopje's Pivot N01 's was determined (Figure
2-2). The results show that the gypsum approached saturation (SI=O) for most of the values.
17
The implication of this is that when irrigation takes place, a small amount of evaporation,
together with the selective uptake of essential nutrients, will result in gypsum precipitation.
Gypsum saturation over time
0.'
0.2
0.'
iii 0 } ~.
~.L~/~( __-.<J.' ~~v I
.<J.2 --
.<J.'
Aug-99 Mar.()() Oet.QO Apr-01 Nov-ot May-02 Dec-02 Jun-03 Jan-04 Aug..Q4 Feb-05
Figure 2-2: Gypsum saturation (SI) in neutralised mine water used to irrigate at Kleinkopje
Pivot1.
Gypsum is a partially soluble salt. The concentration of the gypsiferous soil solution through
crop evapotranspiration precipitates gypsum in the soil section and may remove it from the
water system (see Table 2-2 for irrigation water quality at the different pivots investigated),
reducing potential pollution in the process (Annandale et al, 2002).
Table 2-2: Irrigation water qualities (February 2004).
SiteName ph EC Ca MQ Na K Alkalinity Cl S04 AI Fe Mn
mS/m mg/L mg/L mg/L mg/L as mg/L CaC03 mg/L mg/L mg/L mg/L mg/L
Kleinkopje Pivot 1 5.41 306 697 278 55 14 12 12 2453 0.03 0.01 1.52
Kleinkopje Pivot 4 8.31 505 751 473 138 41 114 94 3447 0 0.06 0.42
Syferfontein 8.84 381 35 91 789 10 621 30 1220 0.08 0.05 0
New Vaal 7.57 166 188 69 138 9 246 17 694 0.01 0.02 0.05
Optimum 7.41 311 438 339 113 39 297 36 2440 0.01 0.3 0.01
To further investigate this, Usher (2003) made use of PHREEQC to analise almost 8000
samples from over the entirety of the Witbank and Highveld Coalfields to ascertain the
degree of saturation with gypsum (Figure 2-3). This figure shows the saturation state,
represented by the saturation index (SI), which describes the state of any solution relative to
the mineral being evaluated (Appelo and Postma, 1993). The maximum solubility of each
mineral at equilibrium is given by its theoretical solubility (the Ksp value). The measured
concentrations in solution are used to determine whether this condition is met.
A value of 1700 mg/L was used by Usher (2003) as a lower cut-off to ascertain solubility
controls. From the 1071 samples in the obtained data set that exceeded 1700mg/L S04,
almost half had SI values approaching equilibrium (a value of 0.1 was used as a cut-off).
Therefore, although not all the values in the higher ranges are at, or approaching equilibrium,
18
indications are that gypsum saturation is an important control on the upper limit of sulphate
values that can be obtained.
III
-~r-------------------+---.-..-..-. ----------------------------~
~~----------------------------------------------------~
o
~( '0
Figure 2-3: Saturation state of gypsum plotted against sulphate concentration (after Usher,
2003).
This observation by Usher (2003) has several implications:
• Determination of gypsum saturation should form part of any hydrogeochemical
assessment in the coalfields.
• Calcium values should be measured on a regular basis so that these assessments
can be made.
• The observed sulphate generation rates (Hodgson and Krantz, 1998) between 5
kg/ha/day and 1O-kg/ha/day sulphate should be reported with an indication of possible
gypsum saturation.
The last mentioned implication could be significant for the often generalised assessments,
e.g. that rehabilitated opencast pits yield between 5 kg/ha/day and 10 kg/ha/day of sulphate
(Hodgson and Krantz, 1998). This rate appears to be fairly constant, despite differences in
degree of spoils saturation, age of spoils, or regional impacts. It is therefore possible that the
observed rates, as derived from flow and concentration calibration, reflect a maximum
sulphate concentration with a consistent recharge. The maximum sulphate concentration is
19
determined by the gypsum solubility. In subsequent reports, Hodgson suggested that gypsum
precipitation will playa role under low flow conditions.
Values from the water at Optimum Colliery clearly indicate the validity of this observation. In
Figure 2-4, the sulphate concentration for the decant water over time, upon which previous
determinations regarding the sulphate generation rate were based, is shown together with
the relative saturation of gypsum.
The SI values indicate that the sulphate values are limited by gypsum precipitation, with
values close to saturation as soon as concentrations rise. This could imply that the true
sulphate generation rates are greater than those derived from flow and concentration
measurements (Van Tonder et a/., 2003).
~ a, _ •. 1
... ,--------------------------------,
•.i +----------------------'r----------I
;l.' +----------tt----±t---------------I
;l.' +-----------f-----i!I---------------I
....+.- ~ __.J:'__ --I
..... -'-- ----l
" .. n"'...t'6 e,D)!:""-(\'" ~",I~'" (11,"''' t. .. ,..a~ u ....·-"O.:- _Itt ......... !' _ "41- ;
GoYJoIot.I,u1' ~I
Figure 2-4: Sulphate concentration and gypsum saturation for decant water from Optimum
Colliery.
The effect of secondary mineral formation is thus an important control on water quality in
South African collieries.
20
2.4 Perception of the final fate of salts
On irrigated lands, irrigation water is the primary source of salts. At all the collieries except
Syferfontein, this water is gypsiferous. In the case of Syferfontein, the water is sodium
bicarbonate rich. The key concern relating to irrigating with these waters, focuses on what
happens to the salts in the water.
The general belief is that the answer is simple-the salts go where the water goes, and they
accumulate in soils where evapotranspiration (combined evaporation from soils and
transpiration by plants) exceeds combined precipitation and irrigation (Cordy and Bouwer,
1999). Thellier et al. (1990) state that irrigation with saline water may result in the
accumulation of dissolved mineral salts in the soil solution, in return resulting in increased
soil salinity. The water returning to the atmosphere by evapotranspiration is essentially
distilled water, leaving behind the salts in surficial soils and in the root zones of plants. To
avoid salinity damage and possibly killing plants or crops, the salts brought in by irrigation
water must be leached from the root zone by applying more irrigation water than can be
evaporated. The leaching results in sustainable crop production and plant growth. If the
amount of water leaching through the soil is too low to remove salts, the salt content in the
soil increases and crop yields may decrease. In such situations, the soil is said to be salt-
affected (Hoffman, 2004)
Generally, the leached water or deep percolation water continues to move downward through
the saprolite soil of the unsaturated zone until it reaches the groundwater (saturated zone),
and must eventually affect groundwater quality. Groundwater salinisation is more likely if the
irrigation water itself is saline or soluble salts are present in the unsaturated zone (Suarez,
1989).
Proper irrigation management can slow down or stop salinisation. If crop fields are
excessively watered or if large amounts of water are allowed to seep away from canals, the
groundwater surface under the field and under the region can be raised (Sommerfeldt and
Chang, 1980). If the groundwater surface rises to within 3 m (depending on soil type) of the
surface, the upper portion of the saturated zone will experience evaporation, which will cause
dissolved salts in the groundwater to become concentrated in the soil profile (Szabolcs,
1986). In arid areas, irrigation can cause the aquifer recharge rate to experience a forty-fold
increase even with well-planned water conservation practices. The ability of the soil and
bedrock to absorb and transport this increase without raising the water table will allow salts
leaching through the soil profile to be carried away. If the soil has layers that possess low
hydraulic conductivity, a perched water table could be created over the regional aquifer. If
drainage is poor due to bedrock, soil or topography the water table will rise and water and
21
salts will eventually reach and be concentrated in the root zone leading to the destruction of
fertility (Schwartz et aL, 1987).
According to Erie et al. (1982) all the water that is not evaporated or transpired, leaches salts
from the root zone. This water contains almost all of the dissolved salts from the irrigation
water. As an example consider cotton grown in south-central Arizona with an efficient
irrigation system that applies 1.6 m of water per year, of which 1.3 m evaporates or transpires
(Erie et aI., 1982), and about 0.3m leaches salts from the root zone. The leaching process
produces 0.3 m of deep-percolation water per year. This moves down to the groundwater
table; although this 0.3 m of water per year contains almost all of the dissolved salts from 1.6
m of irrigation water. As a result, salt concentrations in the deep-percolation water will be as
much as five times higher (Bouwer, 1990) than those of the original irrigation water. This
research will determine the validity of all of these beliefs, and whether some of the salts are
retained in the soil, if so at which quantities, and whether over-irrigation will leach these salts
into the groundwater.
22
3 GENERAL DESCRIPTION OF IRRIGATION SITES
3.1 Introduction
Five different irrigation sites were selected during research by the Dept. of Soil Science and
Agronomy of the University of Pretoria to predict the environmental impact and sustainability
of irrigation with gypsiferous mine water. These sites were used for the current research on
the impact of irrigation with coal mine water on groundwater resources. Four of the sites are
situated in the Mpumalanga province, and one (New Vaal) in the Free State province as
stated in Section 1.5 and illustrated in Figure 1-2 and Figure 1-3.
3.2 Kleinkopje Colliery
Kleinkopje Colliery is part of Anglo Coal and is situated outside Witbank (Figure 3-1). Two of
the irrigation sites are monitored i.e. Pivot 1(Major) and Pivot 4. They were selected because
the soil at Pivot 4 is fairly homogeneous (Hutton soil), whilst that at Pivot 1 varies with depth
(Bainsvlei and Clovelly soils).
-21173000
-2117_
-2117_
-21176000
-21171000
-21179000
----,-
17000 18000
Figure 3-1: Topographic map of the K/einkopje area, indicating the position of Pivot 1and
Pivot4.
Irrigation at Pivot 1 started in 1999 and at Pivot 4 in 2001, and is currently still in operation.
The crops planted were maize, wheat and potatoes.
23
3.2.1 Surface Hydrology
The annual average rainfall (MAP) for the greater catchment area B11G (368 ha) is 698 mm
(SA Weather Service - Rainfall station Witbank no. 0515412-2). The mean annual
evaporation (MAE) for the area is 1600 mm, and the mean annual runoff (MAR) 35.8 mm
(Midgley et a/., 1994) under natural conditions.
3.2.2 Geology
3.2.2.1 Regional geology
Kleinkopje Colliery lies in the Springs-Witbank Coalfield and is flanked by the Highveld and
Ermelo Coalfields. The coalfield extends over a distance of approximately 180 km from the
Brakpan/Springs area in the west to Belfast in the east, and about 40 km in a north-south
direction. It is currently the most important coalfield in the country.
3.2.2.2 Site -specific geology and hydrogeology
• Locally the strata encountered in Pivot1 consist largely of soils, sandstone and coal
below this.
• No detail of the geology at Pivot 4 is available. However, trenches dug showed that the
Hutton soil is deeper than the current water level and that no consolidated rock affects
the downward movement of irrigation water towards the aquifer.
A fence diagram in Figure 3-2 provides a detailed cross-section of the strata indicating the
different layers.
LitholOgy
Figure 3-2: Fence diagram of geology at Pivot 1, drawn through the boreholes.
24
At Pivot 1 the hydraulic conductivity values (obtained from point dilution tests) in the
weathered zone below can be as high as 140 mid in specific zones or preferred pathways
(Figure 3-3). This results in a flow zone above the solid rock. Pumping tests done at all the
boreholes at this site indicated conductivity values for the aquifer itself as 0.001 mid. At Pivot
4 the highest conductivity values obtained in the weathered zone is 120 mid (Figure 3-4), with
the same values in the solid rock as at Pivot 1.
Borehole Log - KK1 Testhole
Depth]m] LocallW - X 2401300 Y: 2874788 00 1.. 1500 00
Uthology Geology Darcy [Velocity (m/d)1 K.'08lue[(mfd)]
000 - 0 40 SOil Black s(Ml, 4 67% clay r~ 0.40
040 - 0 eo SOil ealns'IAelfClo-.e!ly sOIl, 46Z% clay ~i
0.80· 1 00 SOil BainsoJeilClo-.elly 5011, 2.03% clay
1.00·2.50 SOil BalnloJei/Clo\elly 8011
2.50·2.00 SOil e.lnI'lAeIIClowlly sOIl, 18.58% clay
2.80 - 4 00 SOil; Bains'IAeIIClo-.elly sOIl, 9.24% clay
400 - 5 60 SOil Bains\Ae./Clo-.elly 8011, 9% clay
'0 560·1500 SANDSTONE White, medium to fine grained
"
'2
14
'5
Figure 3-3: Geology and aquifer parameters at Pivot1.
Borehole Log - KK4 Testhole
Depth [ml Locality. X 19338,00 Y: 28n675.00 1. 000
Lithology Geology K--.elue [(mId)] 120
~ 0,00· 0 60 SOil' Hutton ecil, '4% clay I 1 ------,
0,60·2.50 SOIL Huttoo soil, 18.5% clay
2.50·300 SOIL Hulton 5011,27% clay
300·350SQIL HlAtOl1IOII, 23 ,,*, clay
~
350·6,50 SOIL. Hulon ao~. 23.22% clay
'0
13.50 - 25.00 SANDSTONE White sandstone
20
25
Figure 3-4: Geology and aquifer parameters at Pivot 4.
25
3.2.3 Positioning of the monitoring boreholes
Pivot1.
At this pivot, six of the ten boreholes were monitored regularly for water levels (Figure 3-5),
but only five are used for hydrochemical profiling and water quality. The other boreholes were
damaged due to agricultural activities and are inaccessible by samplers and probes.
L _~_~_
-2874500
-2874600
-2874700
-2874800 I P•V1BH3
-2874900 l
-2875000
-2875100
-2875200
-2875300
I
23700 23800 23900 24000 24100 24200 24300 24400
Figure 3-5: Position of the monitoring boreholes at Pivot1.
Pivot4.
At this pivot six boreholes were monitored regularly for water levels, profiling and water
quality (Figure 3-6).
26
-2877000
-2877200
-2877400
-2877600 •
-2877800
-2878000
18800 19000 19200 19400 19600 19800
Figure 3-6: Position of the monitoring boreholes at Pivot 4.
3.2.3.1 Water levels
Pivot 1:
Figure 3-7 illustrates the groundwater levels at Pivot1. The water levels, as expected at an
irrigation scheme, are very shallow.
• The time variation of the water levels indicates that the water levels fluctuate in
correspondence with the seasons. Typically, the water level drops during the mid- to
late winter months, and then recovers again during mid- to late summer.
• The rise in the water level during the early months of 2005 is due to high rainfall and
excessive irrigation. Flooding was so great in this period that some of the boreholes
(PV1 BH-5, 6 and on occasion 7) could not be sampled at times.
• PV1 BH-5 has a deeper water level than the other boreholes at the site. This borehole
is drilled into the mine, and drains the groundwater in its vicinity into the mining cavity,
resulting in the lower water level. This is a very important observation, as this
influences the salt balance calculations in Chapter 6.
27
Water Level Depth
Water Level Depth [ml
1.00~-
~PV1BH1
...,._ PV1BH2
-+- PV1BH3
- PV1BH3S
....... PV18Ho1l
~ PV1BH5
_ PV1BH3
....... PV1BH7
__ PV1BH3
..... PV18l-S
Time
Figure 3-7: Water levels at Pivot1.
Pivot 4:
Figure 3-8 illustrates the groundwater levels at Pivot 4.
• The time variation of the water levels indicates that the water levels fluctuate in
correspondence to the seasons. However, during early 2006 there was a rise in
water levels due to the very high rainfall.
• Borehole PV4BH-1 and 2 have been destroyed since early 2005.
Water Level Depth
Water Level Depth [ml
1.00
2.00
3.00
- PV4BH1
4.00 ....... PV4BH2
....... PV4BK3
__ P\l4BH4
5.00
_PV4BI-6
_PV4BI-6
6.00
7.00
8.00
9.00 t-
2001 12002 2003 200' 2005 2006
Time
Figure 3-8: Water levels at Pivot 4.
28
3.2.3.2 Irrigation Water Quality
Pivot 1:
Irrigation is from water of the New Vleishaft dam adjacent to the pivot. This water is pumped
from underground in the dewatering process ahead of mining. Table 3-1 depicts the average
water quality of the irrigation water.
Table 3-1: (a) Average waterqualityover4 years of the New Vleishaftdam at Pivot1 (from
Annandale et aI., 2006) with statistics on the water in table (bj.
(a)
(b)
pH EC Calcium Sulphate
Minimum value 5.2 294 493 1935
Maximum value 6.2 408 697 2620
Standard deviation 0.43 45 87 277
The high TOS value is important. This is the total salinity that is being applied on the soil
during irrigation. The pH value of 5.5 ensures that the metals stay in solution.
Pivot 4:
Irrigation is from a pan adjacent to the pivot. Water in the dam originates from the colliery.
The water is Ca/Mg sulphate-rich, which is typical of water from Kleinkopje Colliery. The
chemical modelling program PHREEQC indicated that the water is in equilibrium with
gypsum, and saturated with dolomite and calcite. Table 3-2 depicts the average water quality
of the irrigation dam. As is the case at Pivot 1, the TOS is very high.
Table 3-2: (a) Water quality of the Irrigation dam at Pivot4, with statistics on the water in table
(bj.
(a)
29
(b)
pH EC Calcium Sulphate
Minimum value 6.4 271 293 1270
Maximum value 8.6 505 818 4392
Standard deviation 0.61 56 122 617
3.2.3.3 Multi-parameter Profiling Pivot1 (see Appendix D for Monitoring data).
In-situ measurements of water qualities were done at five localities at Pivot1.
• There is a slight increase in EC in the down-gradient boreholes BH5 and BH2 at 10m,
which may be the result of the rock composition at that depth (Figure 3-9).
• There is no variation in any parameters in the higher lying boreholes BH6 and BH7.
Borehole Log - PV1BH2
Depth rml locatily-X 2378488 V 287476393 Z.156000
Lithology Geology o ~ _::_g~ 0 ORP 360 42 PH 7 8 2~CONO [mS/{"~ 164 Temp (CJ'S 0
1 I
l
Figure 3-9: Multiparameter profiling at PV1 BH2.
3.2.3.4 Multi-parameter Profiling Pivot 4. (see Appendix D for Monitoring data).
In-situ measurements of water qualities were done at six localities at Pivot 4.
• No significant variations occur with depth in any of the boreholes.
• EC values are very low.
3.3 New Vaal Colliery
The New Vaal pivot is situated next to the Vaal River, close to the town of Vereeniging
(Figure 3-10). Monitoring at this site started early 2001. Maize, wheat, beans and potatoes
were planted at this site. Irrigation has ceased at this pivot in 2005 due to problems with
panding at the surface, resulting in very low yields of crops.
30
·2952000
·2954000
·2956000
t'c~,,--- ._.p ....
\~,}
!
·2958000
r , r r- I
88000 90000 92000 94000 96000 98000
Figure 3-10: Topographic map of the New Vaal area.
3.3.1 Surface Hydrology
3.3.1.1 Rainfall
The annual rainfall (MAP) for the greater catchment area C22F (444 ha) is 643 mm (SA
Weather Service - Rainfall station Vereeniging no. 0438784-3). The mean annual
evaporation (MAE) for the area is 1650 mm, and the mean annual runoff (MAR) 24.3 mm
(Midgley eta!., 1990).
3.3.1.2 Topography
The surface slope of the irrigation area is flat with a very slight dip towards the west and
towards the river. The surface contours in Figure 3-11 illustrate this very clearly. Very little
run-off will occur in this sandy area. All surface flow and expected groundwater flow is in a
westerly direction towards the Vaal River.
31
Figure 3-11: Surface contours for the New Vaal area.
3.3.2 Geology
3.3.2.1 Regional geology
New Vaal Colliery lies in the Sasolburg-Vereeniging Coalfield. The statigraphy of the coalfield
is typical of the coal-bearing margins of the Karoo Sequence.
The succession consists of pre-Karoo rocks (dolomites of the Chuniespoort Group of the
Transvaal Sequence) overlain by the Dwyka Formation (2-15 m thickness), followed by the
Ecca Group sediments, of which the Vryheid Formation is the coal-bearing horizon. In some
places the lava of the Ventersdorp Supergroup underlies the coal. The latter is present over
the whole area and consists mainly of sandstone, shale and coal of varying thickness. An
alluvium (sand) layer is present along the Vaal River. Dolerite intrusions in the form of dykes
and sills are present over the entire coalfield and cause structural complications
3.3.2.2 Site-specific geology
Locally the encountered strata consist largely of soils, siltstones/shale and coal.
Figure 3-12 illustrates an integrated geological model for the underlying lithologies at the New
Vaal pivot, based on the logs from the existing boreholes. A fence diagram in Figure 3-12
provides a more detailed cut through the model indicating the consistency of the different
layers.
32
Lithology
E
SOL
Sft.E
COAL
CLAY
Figure 3-12: Log illustration of geology at New Vaal pivot.
Top
Lrthology
1----E
Figure 3-13: Fence diagram of geology at New Vaal pivot.
The clay layer, illustrated in Figure 3-13, occurs between 2.4m and 4m. This layer has a clay
content of 24%, and greatly influences on the drainage system of the pivot area.
3.3.3 Installation of monitoring boreholes
Five monitoring boreholes were drilled at this site. These were spaced as indicated in Figure
3-14. These five boreholes were sited according to expected water flow. Boreholes NVPV1
and NVPV2 were placed for the purpose of obtaining background values. Boreholes NVPV3
and NVPV4 were located at the expected down-gradient side of the pivot, in order to obtain
33
any irrigation water information, while Borehole NVPV5 was located in the centre of the pivot
to obtain further water quality information on its expected path to the river.
All the boreholes were drilled to a depth of approximately 26 m. This depth was based on the
on-site drilling, ensuring that the borehole was drilled approximately 10m into the geology
below the alluvial system (Figure 3-14).
Figure 3-14: Position of the monitoring boreholes at the New Vaal Pivot.
3.3.4 Water levels
Figure 3-15 illustrates the groundwater levels at the New Vaal Pivot. The water levels are a
little deeper than most other irrigation schemes. It was originally thought that this could be
attributed to the high conductivity in the soils, resulting in rapid drainage towards the river. It
appears, however, to be related to a clay layer at 2.4m - 4m that prevents most of the
irrigation water from migrating downwards. This statement is supported by the water level of
a short piezometer installed into the clay layer. The water level in the piesometer is less than
2m indicating the panding of water above the clay layers identified during soil analysis (see
Section 6.3.2). This was also observed when trenches were dug at this site. Over the course
of the irrigation at New Vaal, areas of water logging were experienced due to the panding in
the soil.
34
Water Level Depth
Water Level Depth [m)
0.0
5.0 g
PV •
..... PV5
....... Short plezo
10.0
-
712003 712004 7~
Time
FIgure 3-15: Water levels at New Vaal Pivot.
The water levels at New Vaal Pivot have varied very little over time, although a general
decline in water levels is evident. The slight variations in water levels can be attributed to
seasonal variation and associated rainfall. The water levels are relatively deep and different
activities were undertaken to determine the reason for this, which will be discussed in chapter
6.
3.3.5 Irrigation water analysis
Irrigation occurs from Dam 595, adjacent to the irrigation site. Water in the dam originates
from the colliery, as illustrated in the Durov diagram, showing the strong sulphate anion
dominance, and with no specific cation dominance of the water in this dam (Figure 3-16).
Mg
Durov Diagram
~\
-~ Na+K.__.~.--S04 •• ~· ~·l
'I.
. ~~~_j• EC (10-1000).. i•\L
~~ pH (6"1
~II~ • 1
Figure 3-16: Durov diagram of the Imqetion Dam 595.
35
The illustration in the Durov diagram (Figure 3-16) and the water quality tables below depict
calcium/magnesium/sodium sulphate water. Table 3-3 depicts the average water quality of
the water sampled at the irrigation dam, with the standard deviation of 37 for EC. The quality
of the water is of far better quality than at the Kleinkopje sites.
Table 3-3: A verage water quality for the Irrigation dam 595.
SiteName pH EC TOS Ca Mg Na K PAlk
mS/m mg/L mg/L mg/L mg/L mg/L mQ/L
Oam595 7.5 125 904 86 31 124 5 29
SiteName MAlk Cl 504 N03(N) P04 AI Fe Mn
mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
Oam595 242 44 410 0 2 0.16 0.30 0.08
3.3.6 Multi-parameter profiling (see Appendix D for monitoring data).
• The water quality in the groundwater was profiled with a multi-parameter probe. Multi-
parameter logs ensure that water qualities are measured as they are in-situ, without
disturbing the water column in the boreholes through sampling. Without such probing,
sampling would be less precise and associated with greater uncertainity with regard
to water quality variation.
In-situ measurements of water qualities were done at five localities in and around the pivot
area at New Vaal. One such profile is illustrated in Figure 3-17.
Borehole Log - PV1
Depth 1ml Locality -X: 92640.40 Y: 2955310.09 Z: 1425.00
Lithology Geology
10
15
20
25
30
35
_j
Figure 3-17" Multi-parameter profiling of PV1.
36
• PV1 is a background value with typical unpolluted aquifer parameters. The EC is low
(46 mS/m) and the pH neutral.
• The measured values indicate clearly that the groundwater flow paths from the pivot
do not intersect borehole PV4, as this borehole is unpolluted. A more westerly flow
direction towards the Vaal River from the pivot is expected.
• PV5 is situated in the centre of the pivot. The measured values indicate clearly that
the mine water used for irrigation has not impacted significantly on the water quality
on the aquifer yet.
• The EC values in boreholes PV2 (above 300 mS/m) and PV3 are elevated (± 90-120
mS/m) compared to the previous three boreholes. These increases were influenced
by spillage of water from the irrigation dam.
3.4 Syferfontein Colliery
Syferfontein Colliery is part of Sasol Mining and is situated between the towns of Trichardt
and Kriel in Mpumalanga (Figure 3-18).
T -.-- -----,--- -r-
15000 17500 20000 22500
Figure 3-18: Regional setting of Syferfontein irrigation slte.
Monitoring on a quarterly basis was performed till August 2004, when a conveyor that was
constructed through part of the site (the grey strip in Figure 3-19), destroyed most of the
monitoring holes. New boreholes OBH6 shallow and deep, 7 OBH and OBH 8 were
37
constructed. Irrigation commenced again In October 2005. Since then a half circle was
recommissioned. Only perennial grasses are being cultivated at this site.
3.4.1 Surface Hydrology
3.4.1.1 Rainfall
The annual rainfall (MAP) for the area at a 50% percentile is 708 mm (SA Weather Service -
Rainfall Station: Secunda no. 047830-3). The mean annual evaporation (MAE) for the area is
1550 mm, and the mean annual runoff (MAR) 50 mm (Midgley et aI., 1990).
3.4.1.2 Topography
The surface slope through the irrigation area is to the north at an average gradient of 1:30.
Boreholes were sited in such a way that any groundwater migration from the pivot would be
intercepted by these boreholes (Figure 3-19). Borehole siting was done with consideration of
the dyke positions and flow directions.
Borehole OBH3 was intended to monitor background water quality. The remaining boreholes
are situated in the northeastern corner of the pivot adjacent to the test quadrant being
researched by the University of Pretoria. An existing borehole, REGM-86, is also close
enough to serve as a monitoring borehole and has therefore been included in the monitoring
system.
Trenches were dug in 2001 to prevent the runoff water from the pivot area entering the
boreholes from surface, thereby preventing skewing of the results.
38
2920000
.OBH8
2920200
2920400
.CoreS
.OBH6S
OBHGO
2920600
.CoreG
LEGEND
2920800
_ Dyke-eW
_ Dyke-NS
Conveye
2921000
23200 23400 23600 23800 24000
Figure 3-19: Pivot site showing the positions of the monitoring boreholes (with the positions
of the dykes, provided by the mine's geologist).
3.4.2 Geology
Syferfontein lies in the Highveld Coalfield and is flanked by the Springs-Witbank and Ermelo
Coalfields. The Karoo Sediments at Syferfontein comprise the Ecca Group (including the coal
seams) and the Dwyka Formation. The total thickness of the Karoo Sediments ranges from
15 - 120 m. Regionally, the Ecca Sediments consist of sandstone, siltstone, shale,
interbedded siltstone, mudstone and coal. Dolerite dykes and undulating sills are present.
The geology at Syferfontein is relatively homogeneous with the weathered sediments up to
10m deep. The strata mainly consist of 2 - 6 m soil, followed by dolerite to 10m, with a thin
relatively impermeable layer of shale below the weathered zone. The shale is underlain by
fresh sandstone with a white appearance. At the depth of 32-35 m the sandstone is again
underlain by shale. Figure 3-20 illustrates an integrated geological model for the underlying
lithologies at the Syferfontein pivot, based on logs from the existing boreholes. A fence
diagram in Figure 3-21 provides a more detailed section of the model indicating the
consistency of the different layers.
39
E
lithology
SOil
Sandstone
Dolerite
Si~stone
Shale
COAL
CLAY
Limestone
Mudstone
Tillite
Granite
Figure 3-20: A model diagram of the geology at Syferfontein pivot, viewed from the south.
Top
Lnhology
Figure 3-21:A fence diagram of the geology at Syferfontein pivot.
40
3.4.3 Geohydrology at Syferfontein Pivot
From the water strikes encountered in the drilling and from subsequent slug testing of the
boreholes, it is clear that the underlying sediments have got very poor water-bearing
properties. Generally the boreholes yielded little water during drilling, with minute quantities
encountered above the shale layer lying at around 35 m.
Slug testing indicated that the sediments show a very poor response and a low hydraulic
conductivity (0.001 mid) is inferred. In certain boreholes moisture (as opposed to definite
water strikes) was associated with the shallower shale layer at approximately 15- 20 m.
These shale layers can therefore be considered as barriers to vertical flow. Any water moving
vertically from the irrigation should be impeded by this deeper lying shale layer. All the
shallower layers show some signs of weathering and groundwater will be able to move
vertically and laterally in these layers. This has been confirmed by point dilution tests
performed, indicating conductivity values of as high as 50 mid above the solid rock.
The low yields and aquifer parameters as obtained in the testing suggest that groundwater
will migrate slowly from the pivot. Any degradation in quality due to the irrigation and fertilizer
application will consequently be a largely localised phenomenon.
Borehole Log - OBH6D
Depth (mi tcceutv . X: 23413.65 Y: 2920500.78 Z:0,00
Lithology Geology o Darcy [Velocity (mId)) 0.12 0 K-value ((mld)!
o ,. -- 500.00·0.80 CLAY: Arcadia, 63.49% clay T I
r1,;-·. '·: ....... .. _ 0.80 - 2.44 CLAY: Yellowish. 40.7% clay r I
f-.-. 2.44 - 3.50 CLAY: Yellowish. shafv. 31.58% clay I
3.50 - 4.65 SOil: vetlowish, shaly. 26.04% clay
4.65 - 7.10 DOLERITE: Clayey, weathered on contact with shal~
F=~--=--;-6.00 - 7.70SHALE: Dark, clayey
6.70 - 8.80 SANDSTONE: Grey, fine grained I
to I-
8.80 - 15.00 SANDSTONE: Yellowish brown
15
n
20 I-
15.00·30.00 SANDSTONE: Greyish Yklile, medium grained
25 I-
30 ~=:=!lIL
Figure 3-22: Geology and aquifer parameters at Syferfontein.
41
3.4.4 Water levels
Figure 3-23 illustrates the groundwater levels at the Syferfontein Pivot. The water levels, as
expected at an irrigation scheme, are very shallow (3-5m below surface).
• All the water levels except those of OBH3 show a similar downward trend.
• The construction of the run-off trenches will also play part in the lower water levels.
Water Level Depth
Water Level Depth [mj
2.00
3.00
4.00
... OBH1 l
......OBH2
..... OBH3
5.00 OBH4
.... OBH5
... OBH6D
6.00 ..... OBH6S
-OBH7
.... OBH8
[ .... REGM86
7.00
8.00 j
9.00
2001 2002 2003 2004 2005
L Time
Figure 3-23: Groundwater levels for Syferfontein Colliery Pivot.
Of importance is the fact that the irrigation activities do not appear to have significantly
affected the observed water levels, since these are at or below the levels measured when the
boreholes were drilled in 2001.The dykes running though the area appear to be more
important in delineating the different water level zones. It is important to note, however, that
the two boreholes with deeper water levels are the upgradient boreholes OBH3 and REGM
86, which could both be considered to be less likely to be influenced by irrigation.
3.4.5 Irrigation Water Quality
Irrigation occurs from a dam adjacent to the pivot. Water in the dam originates from the
colliery. The characteristics of this water are illustrated in the Durov diagram in Figure 3-24.
42
/ Mg
/
\ Durov Diagram
Ca
Na+K
T.I>Jk
•• •• •
504<. -
CI+N03
EC (100-1000)
•• pH (8-10) f'Urate as nitrogen
• t\O Value
• N:l3-N > 20.00
• N:l3-N c 20.00
• N:l3-N < 10.00
• N:l3-N c 6.00
Figure 3-24- Durov diagram of the irrigation dam
Table 3-4 depicts the average water quality of the water sampled at the irrigation dam with
the standard deviation of 65 for EC. The water differs from that at the other irrigation sites in
that it is sodium-rich.
Table 3-4- Water quality of the irrigation dam
3_4 _6 Groundwater Quality
3-4.6.1 Multi-parameter profiling (see Appendix D for monitoring data).
In-situ measurements of water qualities were done at six localities in and around the pivot
area at Syferfontein. As an example the in situ characterization of OBH1 is shown below in
Figure 3-25.
43
Borehole Log - OBH1
Depth [ml Locality - X: 23667,68 Y; 2920238.78 Z. 1612 00
Lithology Geology Coonst~~t~o~ 00 [mgII] 8 60 ORP 220 7 0 PH
-r--r-r-n TTi
(
\
10 \
\
15
20 c-=-=...:::~1900-21 OOSAN05TGjE Grey
~~=~====~~=~2100-2TOOSHALE Bleek
25 E====~
i=====~ r
30
2700-3"OOSAN05TCJIIE Grwy
35
Figure 3-25: Multi-parameter profiling of OBH1.
From Figure 3-25 it can be seen that there is very little variation in the different parameters
with depth. Temperature increases slightly with depth. Dissolved oxygen shows the expected
decrease of DO with depth as oxygen diffusion with depth decreases. The distinct shift in
specific conductivity and pH in OBH1 during February 2002 (blue line) was the result of run-
off water entering the borehole. The profiling indicates a two-year period before the water
quality returned back to normal after the cut-off trenches were constructed.
3_5 Optimum Colliery
3.5.1 Introduction
Optimum Colliery is an opencast mine situated on the Mpumalanga Highveld between
Middelburg and Hendrina, in the catchments of Woestalleen East, Woestalleen West and
Zevenfontein Spruit, all tributaries of the Klein Olifants River, which discharges in the
Middelburg Dam, upstream of the Loskop Dam (Figure 3-26).
The Optimum Colliery can be subdivided into four units: Optimus, Zevenfontein, Bothashoek
and Pullenshope. The eastern Woestalleenspruit runs through the oldest pit, Optimus, and all
the run-off from the pit ends up in the Fanies Dam and Lapa Dam (Hodgson, 1992). From
there it flows through the Woestalleen Spruit and the Klein Olifants River to enter the
44
Middelburg Dam. The water in the pit is derived from groundwater, and recharged through
the spoil from coal slurry disposal and surface run-off .
.: - - ol' "" ....
.... ... -.
,,
,
(
,(,
,,,,,
(
,,
,,
,, , Olif ats Cal chrrent,
_ Towrs
,,
Figure 3-26: Locality plan of Optimum Colliery.
3.5.2 Geology
The area forms part of the Witbank Coalfield, which contains almost a fifth of South Africa's
coal reserves (Cairncross et a/., 1990). The Karoo sandstones, mudstones, siltstones and
coal seams of the Vryheid Formation are underlain by the Dwyka in the vicinity of the study
area. Dolerite in the form of the Ogies Dyke strikes east to west through the study area and a
small sill can be found on the eastern side of the Pullenshope section. Faults are rare in this
area, but fractures are commonly found in the sandstone and coal (Krantz, 1993).
Mining results in the total topography disturbance. The mining depth varies from close to the
surface to about 60m.
3.5.3 Topography
The pre-mining topography was reasonably flat, with pans situated on top of hills. The area has a
mean elevation of 1600m above sea level. The drainage pattern is dendritic with all streams
flowing into the Klein Olifants River in the north.
45
Figure 3-27: Surface contours in the area of Optimus Pit, indicating the position of the two
irrigation pivots, as well as Lapa Dam.
3.5.4 Geohydrology
The two pivots are situated on the rehabilitated spoils of Optimus Pit, as indicated in Figure
3-27. The irrigation site is situated on a topographic high, with the flow directions towards the
Lapa dam. The arrows in Figure 3-28 indicate the flow direction of the water in the spoils
towards decanting in the Lapa Dam.
The decant value at the Optimus pit is at 1572 mamsi, where the water will stabilise in the pit.
The floor of the pit below the irrigation site is higher than the decant point, resulting in all
water in this area draining towards the Lapa Dam. The decant line is depicted in red in Figure
3-27. Therefore no borehole could be sampled in the vicinity of the irrigation, as the floor is
always dry.
46
Figure 3-28: Floor contours, with flow directions indicated.
Irrigation water is pumped from the Lapa Dam. All excess water from irrigation will flow back
through the spoils towards the decant level, and eventually end up in the Lapa Dam again.
This will thus result in the circulation of irrigation water.
3.6 Summary
Each irrigation site differs from the others in either soil type, water quality and composition or
mining methods. Therefore the approach regarding research at each site had to differ, as will
be discussed in detail in the next chapters.
47
4 A REVIEW OF THE HYDRAULIC CHARACTERISTICS AND
HYDROCHEMICAL IMPACTS OF IRRIGATION ON SOil AND
SPOILS
The main difference among the different research sites is the irrigation water quality and the
type of soil on which the irrigation occurs. Of importance to the mining industry is the
availability of soil for irrigation. Both virgin soil (in the natural state without any interference
from mining activity) and rehabilitated spoils are important for future large-scale irrigation.
The hydraulic and hydrochemical properties of these two options differ vastly.
4.1 Virgin soil irrigation
4.1.1 The Unsaturated Zone (Vadose zone)
The unsaturated zone (vadose zone) is the portion of the subsurface above the ground water
table. It contains, at least some of the time, air as well as water in the pores. Its thickness can
range from zero meters to hundreds of meters, as is common in arid regions.
Unsaturated flow represents an interesting challenge, namely the non-linear relation between
degree of saturation, pressure and permeability.
Processes in the unsaturated zone.
o Water storage, plant nutrients, and other substances are present in the unsaturated
zone (USGS, 2001). The unsaturated zone is not always considered a major storage
component of the hydrologic cycle because it holds only a tiny fraction of the earth's
fresh water, which is usually difficult to extract. It is however of great importance for
storing water and nutrients in ways that are vital to the biosphere.
o Transmission of water and other substances occurs in the unsaturated zone. It is held
by some hydrologic viewpoints as a zone that to a large degree controls the
transmission of water to aquifers, as well as to the land surface, to water on the
surface, and to the atmosphere. It may be a controlling factor in the amount of water
that recharges an aquifer, or it may yield information that permits this replenishment
to be quantified.
o The unsaturated zone contains natural and human-induced activity. Its constituents
do not passively reside in place or pass through to the water table. The unsaturated
zone experiences physical phenomena such as thermodynamic interactions, transport
processes of various kinds, and chemical reactions. There are also chemical
reactions involving both natural and artificial substances. There is the biological
activity of plant roots, rodents, worms, microbiota, and other organisms. As a zone of
48
human activity, it is used allover the earth for the cultivation of plants, construction of
buildings, and disposal of waste.
Water moves downwards through porous soils and sediments and, under favourable
conditions, may preserve a record of weathering processes, climatic variations (in the Cl or
isotopic signature), or human activities such as agriculture (N03) and acidification (H+). This
record can be considered as the output from the soil zone and may reflect the properties or
change in properties of soils.
The frequency of measurement suggested is 5-10 year intervals to confirm the movement of
solutes towards the water table. Relatively homogeneous sediments are required where flow
takes place uniformly (diffusive flow). In macro-pore media, possible by-pass flow (preferred
pathways) needs to be taken into account, as when some contaminant travels relatively
rapidly to the water table along fissures, root channels, mouse and mole holes, cracks in the
soil, etc. The quality of groundwater reflects inputs from the atmosphere, from soil and water-
rock reactions (weathering), as well as from pollutant sources such as mining, land
clearance, agriculture, acid precipitation, domestic and industrial wastes. The relatively slow
movement of water through the ground means that residence times in groundwater are
generally orders of magnitude longer than in surface waters. Dispersion, reaction and mixing
ensure that the addition of small amounts of contaminants is commonly difficult to detect.
Below the water table, groundwater is not generally a good archive of past changes because
of dispersion of the input signal.
4.1.2 Unsaturated flow (matrix flow or diffusive flow)
Even with the complexity of soil water flow, diffusive flow of water through soil will still occur.
The rate of soil water flow is variable. Compacted soil layers that are so compact that they
stop root growth will transmit water at slower rates, because of the smaller pore sizes.
Irrigation water management must be tailored to account for rooting depths. If a compacted
layer stops root growth, but not water movement, water and nutrients may move below the
compacted zone and never be used by the crop.
In summary, waters flow through smaller pores partly due to gravitational forces, but also due
to soil wetness or dryness. Water flows from wet to drier soil, but it flows more slowly in dry
soil. The moisture content is dependent on the porosity and the permeability of the soil (de
Marsily, 1986). Below a certain point the water content does not increase and the soil is said
to be saturated, with all the pores filled. This water belongs to the water table aquifer.
49
Darcy's law and the concepts of hydraulic head and hydraulic conductivity have been
developed with respect to a saturated porous medium where all the voids are filled with
water.
Q=KAAh/L
Where in the simplest form:
Q = Total discharge
AIht/L = Hydraulic gradient i
K = Conductivity
A = cross-sectional area to flow
[Darcy's law is a simple mathematical statement which neatly summarises several familiar
properties that groundwater flowing in aquifers exhibits (Wikipedia, the free encyclopedia,
2006), including:
o If there is no pressure gradient over a distance, no flow occurs
o It there is a pressure gradient, flow will occur from high pressure towards low pressure
(opposite the direction of increasing gradient - hence the negative sign in Darcy's law
o The greater the pressure gradient (through the same formation material), the greater
the discharge rate
o The discharge rate of fluid will often be different through different formation materials
(or even through the same material if in a different direction)- even if the same
pressure gradient exists in both cases].
However, soils are seldom saturated, especially near the surface. The voids are only
partially filled with water, and the remainder of the pore space contains air. The flow under
such conditions is called unsaturated flow. The water is subjected to the forces of gravity in
the saturated zone, and to the capillary forces in the unsaturated zone. Water falling on the
soil surface begins by moistening the upper few centimetres, and does not cause an
immediate significant vertical flow. As rain or irrigation causes the water content to increase,
the water spreads downward and moistens a deeper zone (Figure 4-1). If the wetting at
surface continues long enough, the moistening will be progressively greater and eventually
will cause infiltration into the aquifer.
50
7 Soil Surface
sa:.
oCl)
WL
~----------------------M-o-istureContent
Figure 4-1: Typical moisture profile during rainstorm or irrigation (after De Marsily, 1986).
There are two characteristics that are critical to water movement, i.e. water content and
hydraulic conductivity:
1. The most basic measure of the water in an unsaturated medium is water content e or
wetness, defined as the volume of water per bulk volume of the medium. Water is held in an
unsaturated medium by forces whose effect is expressed in terms of the energy state or
pressure of the water. Various types of pressure may be relevant in unsaturated hydrology,
but one called the matric pressure or matric potential (arising from the interaction of water
with the rigid matrix) is of unique interest, as it substantially influences the chief transport
processes. Matric pressure is the pressure of the water in a pore of the medium relative to
the pressure of the air. When a medium is unsaturated, the water is generally at lower
pressure than the air, so the matric pressure is negative. The infiltration of water into the
unsaturated zone can be considered mathematically in terms of both a gravity potential Z,
and a moisture potential W (Childs, 1967). The moisture potential is at negative pressure
due to the soil-water attraction. The moisture 10000
potential increases with decreasing amounts of
1000
soil moisture. Depending upon the soil-moisture
content, either the moisture potential or the
~ 100
II.
gravity potential dominates. At moisture contents .:!.
Z
close to the specific retention, the gravity potential 10
is greater; but when the soil is very dry, the
moisture potential may be several orders of
magnitude greater than the gravity potential 0,2 04 06 0,6
(Fetter, 2001).
LJ
SLOE. r !::IN
y
51
.
(I QS 31 3 X
Greater water content goes with greater matric pressure. Zero matric pressure is associated
with high (saturated or nearly saturated) water content. As matric pressure decreases the
water content decreases, but in a way that is nonlinear and hysteretic. The relation between
soil moisture potential (matric pressure) and moisture content, called a moisture retention
curve (right), is a characteristic of a porous medium that depends on the nature of its pores.
This relation strongly influences the movement of water and other substances in unsaturated
media.
2. The hydraulic conductivity, a measure of how easily water moves through the medium for a
given driving force, is a second characteristic that is critical to water movement. The hydraulic
conductivity has a highly sensitive and nonlinear dependence on the water content. Usually it
is assumed that the flow rate of water is equal to the hydraulic conductivity times the driving
force (typically gravity and pressure differences). This relation is known as Darcy's law. When
the flow is steady, Darcy's law may suffice in itself for quantifying the flow. When the soil is
not saturated, soil moisture flows downwards by gravity flow through interconnected pores
that are filled with water and, to a lesser extent, as a film along particle surfaces in pores that
also contain air. With increasing water content, more pores fill, and the rate of downward
water movement increases. Darcy's law is valid for flow in the unsaturated zone, although the
unsaturated hydraulic conductivity, K(Bv} is not a constant. The unsaturated hydraulic
conductivity is a function of the volumetric water content, Bv. As Bv increases, so does K(Bv}.
The value of the moisture potential, lj!, is also a function of Bv, often ranging over many
orders of magnitude. The relationship between moisture potential and volumetric water
content is determined experimentally for a given soil. The results are graphed as a soil-water
retention curve. The total potential <I> in unsaturated flow is the sum of the moisture potential
lj!(Bv} and the elevation head, Z:
<I> = lj!(Bv) + z
The water table is commonly regarded as the boundary between the unsaturated and
saturated zone, but a saturated capillary fringe often exists above the water table. The water
table is best defined as the surface on which the fluid pressure p in the pores of a porous
medium is exactly atmospheric (Freeze and Cherry, 1979) or p=O, This implies that lj! = 0
and since h = lj!+ z, the hydraulic head at any point on the water table must be equal to the
elevation z of the water table at that point.
The relationship of unsaturated hydraulic conductivity and volumetric water content can be
determined experimentally. A sample of rock is placed in a container. The water content is
kept constant, and the rate at which water moves through the soil is measured. The value of
K(Bv} can be determined by Darcy's law.
52
Darcy's law in the unsaturated zone is given by:
dH
qz = K(H) dz
Where:
K (H) = unsaturated zone K, depending on moisture [water] content
e = water content (degree of saturation)
H = total Head (rn)
dH/dz = soil water gradient
H= h+z
h = matric potential (rn) - or matric suction head
z = depth below surface (rn)
Darcy's law was developed for saturated flow in porous media. To this Richards applied a
continuity requirement suggested by Buckingham, and obtained a general partial differential
equation describing water movement in unsaturated non-swelling soils. The transient state
form of this flow equation is commonly known as Richards' equation and is a highly dynamic
phenomenon. The equation may be represented quantitatively by a combination of Darcy's
law and the continuity equation (Buckingham-Darcy flux equation). The Richards' equation
(1931) combines both these laws in one formula:
Thus:
And therefore:
qz = dhK(H)[- + 1]
- dz
Because of the hysteresis effect of K(H), K(B) is used instead which shows much less
hysteresis:
dh
qz = K(8)[ dz + 1]
53
with:
K = the hydraulic conductivity (mis);
h = the pressure (matric) potential (m), i.e. soil water pressure head);
8= the volumetric water content (m3/m \
z = the gravitational potential or height above a reference level (m), i.e. the vertical distance
taken positive upwards.
(For unsaturated conditions h is negative).
The use of measured or estimated hydraulic properties with formulations such as Darcy's law
and Richards' equation can quantify the movement of water (matrix flow) in the unsaturated
zone. The flow rate of water is often very directly of interest, for example in estimating how
fast water moves down to the water table. It is also critical in the transport of contaminants,
whether they are dissolved in the water or moving in a non-aqueous liquid or solid form. The
usual first step in assessing the rate of contaminants spreading in the subsurface, is to
assess the flow rate of water that to some degree moves the contaminant along with it.
Using of this formula requires the knowledge of two properties of the medium: the
unsaturated hydraulic conductivity, and the differential water capacity, which can be directly
calculated from the water retention curve.
For a wetting front (infiltration) the equation reduces to qz=K" (because the gradient becomes
1 very quickly at saturation). To solve this equation, K(fJ) and h must be known. The relation
between fJand h is called the soil water retention curve, as discussed earlier in the section.
- -160-140 - <>.§.
ns -120 -
+l -100 <>-c:Q) <> <>c0 -80.. -60 <>
.->;<: -40-ns
::2!: -20 -
o -
0 0.1 0.2 0.3 0.4
Water content
Figure 4-2: Soil water retention curve. Notice that the saturated porosity is equal to the water
content where the metric potential = 0 m (i.e. 0.3 in this graph). In practice, no flow occurs If
the metric potential is smaller than -10m.
54
At saturation, the most conductive soils are those in which large and continuous pores
constitute most of the overall pore volume (e.q. sand), while the least conductive are the soils
in which the pore volume consists of numerous micropores (e.g. clay).The relation of K and
suction (or wetness) can be derived from the well-known equation of Van Genuchten and can
be illustrated in a very simple form (Botha, 1996) as:
K(B) = Ks( B - Br )5/2
Bs + Br)
Where:
er= residual moisture content
es = moisture content
K(e) = unsaturated K
Ks = saturated K
The unsaturated K for a specific moisture content and saturated K-value can thus be
determined from a graph (Figure 4-3). The saturated K used can be estimated from the clay
content of the soil (Kirchner et et., 1991). This is discussed in more detail in Section 6.6.6 and
illustrated in Appendix 0 Calculations "Unsaturated K", contained on the DVD.
Unsaturated K-value
0.1
c-o 0.01~ 0.001
0.0001 -
0.00001 .-l...------------------___~
o 0.1 0.2 0.3 0.4
Moisture content (6)
Figure 4-3: Relationship between moisture content and the unsaturated K-value. Note that
the saturated K-value (i.e. Ks) is the value where the water content is equal to the saturated
poroslty (85).
55
10 -~------------~--~
1 -
--- Sandy soil
0.1 -
--- Clayey soil
0.01 -
0.00 1 -I---_._--,.-----,-----..,...---~
o 0.1 0.2 0.3 0.4
Moiture content
Figure 4-4: Example of K-value for sandy and clayey solt.
One important consideration in unsaturated flow is that at low volumetric content, the
relations that hold true in saturated flow may be invalid (Fetter, 2001). The best example is
the fact that for coarse materials such as sand and gravel, the pores are large and drain
quickly. At lower volumetric moisture contents, there may be very few saturated pores. On
the other hand, most of the pores in finer-grained soils may still be saturated. In a soil with
large pores, these pores quickly empty and become nonconductive as suction develops, thus
steeply decreasing the initially high conductivity. In a soil with small pores, on the other hand,
many of the pores retain and conduct water even at appreciable suction, so that the hydraulic
conductivity does not decrease as steeply and may actually be greater than that of a soil with
large pores subjected to the same suction. Since the soil in the field is largely unsaturated, it
often happens that flow is more appreciable and persists longer in clayey than in sandy soils.
For this reason, the occurrence of a layer of sand in a fine-textured profile, far from
enhancing flow, may actually impede unsaturated water movement until water accumulates
above the sand and suction decreases sufficiently for water to enter the large pores. This
simple principle is all too often misunderstood (Hillel, 1982).
Thus, at lower values of 8v the unsaturated hydraulic conductivity of clay may be greater
than that of sand. A layer of sand in a fine-textured, unsaturated soil may retard the
downward movement of infiltrating water owing to its low unsaturated hydraulic conductivity.
The ratio of surface area to volume of a particle is inversely proportional to its radius. The
surface area of grains exposed to pores in a fine-grained porous medium is thus larger than
in a coarse-grained porous medium. A fine-grained porous medium may therefore contain
more fluid than a coarse-grained porous medium, but will also adsorb a much larger volume
of fluid. This explains why a clayey formation, with its fine-grained matrix, forms a poor
aquifer, although it may contain larger quantities of water than a corresponding coarse-
56
grained sand aquifer. Less force will thus be needed to remove water from a pore with a large
radius than from one with a small radius. Pores with large radii, therefore, will drain before
pores with smaller radii in unsaturated flow.
4.1.3 Preferential flow (macropore flux)
4.1.3.1 Introduction
It is usually assumed on a macroscopic scale, water flow in soils obeys Darcy's law. This
assumption is not valid in structured or aggregated soils in which varying flow velocities occur
on a macroscopic scale (Hutson, 1983). Water applied to the soil surface initially moves
rapidly through macropores and other pathways, but more slowly through finer matrix pores.
By Poiseuilles law, the total flow rate of water through a capillary tube is proportional to the
fourth power of the radius, while the flow rate per unit cross-sectional area of the tube is
proportional to the square of the radius. A 1-mm-radius pore will thus conduct the same
amount of water as 10 000 pores with a radius of 0.1 mm each. There is thus a short-
circuiting effect that causes the water to be transmitted rapidly into the profile along larger
pores, cracks and other channels, leaving the fine matrix relatively dry. (Darcian models will
be most applicable to horizons that have a mieropore system. All water will enter and leave
the horizon vertically. Any lateral flow will thus invalidate such Darcy models).
Preferential flow refers to the uneven and often rapid movement of water and solutes through
porous media (typically soil) and fractured media, characterised by regions of enhanced flux,
so that a small fraction of media participates in most of the flow, allowing much faster
transport of a range of contaminants, including pesticides, nutrients, trace metals, and
manurial pathogens (Steenhuis et aI., 2002). This creates significant consequences for
groundwater quality (Figure 4-5). High pesticide concentrations which are associated with
preferential flow phenomena, were initially described by Lawes et al. (1982). During field
drainage experiments, the authors distinguished between preferential and matrix flows
(diffisuve, piston like), and pointed out that the relative importance of the two kinds of
drainage depends on soil type and rainfall intensity. Matrix flow is the relative slow and even
movement of water and solutes through the soil while sampling all pore spaces, obeying the
convective-dispersion theory, which assumes that water follows an average, flow path
through soil (Figure 4-6). These two types of flow affected solute transport differently.
57
rain
istribution layer
tile line
Figure 4-5' Illustration of preferential flow (after Steenhuis et aI., 2002)
rain
distribution la yer
matrix WithOut~)
preferential \
flow paths )'
tile line
Figure 4-6: Illustration of matrix flow (after Steenhuis et aI., 2002).
Due to its rapid movement, preferential flow allows much faster contaminant transport,
creates significant consequences for groundwater quality and has direct impacts on drinking
water and human health, animal waste management, nutrient and pesticide management,
and watershed management. In the mining sector, the phenomenon of preferential flow has
been used in designing the best configuration for tailings left behind by mining operations. By
including clean material at strategic places where preferential flow is directed, the toxicity of
58
leachates can be greatly reduced. Because of the rapid movement of the water, the dilution
factor is also much larger, and concentrations measured in the water will be greatly reduced.
Subsurface preferential flow, in which small areas of the subsurface carry large portions of
the flow, has been observed in a variety of soils (Hill and Parlange, 1972; Beven and
Germann, 1982; Glass et ai., 1987, 1989; Baker and Hillel, 1990; Ritsema and Dekker, 1994
and Johnson et al, 1983). As preferential flow provides a mechanism to bypass most of the
porous media, the effects include enhanced solute transport, less filtering and adsorbing of
contaminants in the soil, and fast travel to groundwater or flow lines. Preferential flow can
result from macropores and other inhomogeneities (Beven and Germann, 1982), but it has
also been observed in uniform soils with uniform rainfall and low application rates. Sharma
and Hughes (1985) studied the groundwater recharge in Western Australia and concluded
that more than 50% of recharge occurs through preferred pathways bypassing the soil profile.
If this is the case for sandy coastal aquifers, it can also be valid for the porous soils at the
research sites.
Kirchner et al (1991) studied recharge in the Dewetsdorp aquifer in the central parts of South
Africa, using 20 neutron moisture probes at 13 sites with different types of soil to a depth of
1.7m. The Zero Flux Plain (ZFP)-method, based on locating the point of zero hydraulic
gradient, and thus the zero reference flux in the soil profile, summing up the changes in water
content below that point, was used. It was found that even during the 1988 floods in the
central parts of South Africa a zero flux plain still existed. This implies that there was no water
movement through that plain. The only explanation for this behavior of the unsaturated zone
is that recharge has occurred along preferred pathways by bypassing the soil matrix.
linfiltration took place through the soil matrix of the upper (30 cm) of soil. Measurable
downward fluxes occur, until a relatively heavy clay layer is reached, from where the water
movement is predominantly horizontal. At specific locations the clayey layer may have wide
cracks, leaving a pathway for water to travel along. Dieleman and De Ridder (1963) found
similar results when studying the recharge near Lake Chad in Central Africa. In a similar
study at another test site at De Aar in the Karoo by Kirchner et al (1991) tests yielded a
saturated K-value (Ks) of 1*10-4 mId for the first seven days after the start of the experiment,
and for the next 14 days no measurable amount of water flow occurred due to swelling of the
clay. However, after the flood, some movement through the matrix did occur. (The
determination of the unsaturated K-value is an integral part of the calculations of the Darcy
fluxes in the unsaturated zone). The unsaturated K value is related to the saturated K-value
measured (4.3*10-3 mId) and was calculated at 7*10-6 mId. With the total recharge, measured
at 5% in the aquifer during the 30 days after the flood, this low value (7*10-6 mId) could only
be representative of the soil matrix, and strengthens the belief that recharge had also taken
place along preferred pathways. At this site it was also found that downward infiltration
59
occurred until a clayey layer was reached, from where the movement of the soil water is
predominantly lateral. Wood et al. (1997) found that at the Southern High Planes of Texas
and New Mexico macro pore flux is between 60% and 80% of total recharge flux.
Preferential flow (also referred to as rnaeropore flow by Wood et al. (1997) and J. Wagenet of
Cornell University, USA) in soil may result from biological activity (e.g. root channels, worm
holes, meerkat or rodent holes, etc.), geological tectonic forces (e.g. subsurface erosion,
cracks and fractures), geological pressure release forces (bedding plane fractures) or
agrotechnical (manmade) practices (e.g. ploughing, bores and wells). Along these paths
solutes and contaminants are transmitted to groundwater at much higher velocities than
those predicted by Darcy's equation. Surface cracks and channels that bypass the root zone
are also responsible for rapid water and chemical transport through the unsaturated zone.
Figure 4-7' Earth worm, a pathway generator (after Steenhuis et aI., 2002).
4.1.3.2 Types of flow.
Soil texture is important in the type of flow (Steenhuis et aI., 2002):
1. Highly conductive rnaeropore pathways form and persist in structured soils. These
macropores are the result of a variety of soil-forming factors such as flow through non-
capillary cracks or channels within a profile. In clay and loam soils, for example, areas of
relatively low permeability are riddled with channels consisting of cracks partially filled with
sand and small stones, as well as passages formed by roots and earthworms. When it rains,
water infiltrating the ground follows these channels in preference to the surrounding matrix,
60
whose small pores are penetrated comparatively slowly. Flow is often fairly uniform through
the uppermost soil horizon (typically the plow layer for current or former agricultural soils),
which functions as a "distribution zone", distributing the flow to macropores in the subsoil.
Different channels sometimes cross each other, resulting in the mixing of water and solutes.
Flow in macropores and channels can occur with little or no interaction with the surrounding
soil matrix. On the other hand, matrix flow is the comparatively slower movement of water
and solutes percolating though soil pores. The relative importance of the two forms of
percolation is dependent on soil type and rainfall intensity. For example, well-structured soils
consisting of clay and loam mixes typically experience low permeability rates. In such soils,
less than 1% of the pore volume consists of cracks and subsurface channels. However,
during rain events, water infiltrating from the soil surface often flows through these channels
in reference to the surrounding soil matrix, whose small pores are penetrated comparatively
slowly. Even though these channels consist of a relatively small percentage of the total pore
volume, they may be responsible for the bulk of moisture and solute transport after an
infiltration event. Preferential flow may be initiated well below soil-water saturation (Steenhuis
et et., 2002). Fracture flow below the water table is thus also a case of preferential flow (van
Tonder, 2006). Maini and Hocking (1977) give the equivalence between the hydraulic
conductivity in a fractured medium and that of a porous medium. For example, the flow
through a 100m-thick cross-section of a porous medium with a hydraulic conductivity of 10-7
mis could also be the result of a single fracture with an opening no wider than 0.2mm in a
fractured rock medium with an impervious rock matrix. This shows the immense importance
of the flow of one single fracture (or preferential pathway) even when it is not very wide.
61
~
~
~ 10-~
~
"-
.nsl
"-
0
~
::l
t:
nl
~C.
10-4
10 100 1000
Thickness of a continuous medium equivalent to one fracture (m)
Figure 4-8: Comparison between the hydraulic conductivity of the porous medium and the
fractured medium versus the aperture (From Maini and Hocking, 1977).
2. Instability in the wetting front leads water to find its way down through coarse soils in a
number of channels called Fingers. This fingering phenomenon also occurs in homogeneous
soils. Hill and Parlange (1972) were the first to document preferential flow in homogeneous
sand at low infiltration rates. They have shown that water finds its way down through sandy
soils in a number of channels called Fingers. A poorly conducting layer of topsoil at the
surface produces a wetting-front instability. Gravity drives the instability, while surface
tension has a contrary, stabilizing effect. What occurs is analogous to water dripping from a
sponge; the pull of gravity is opposed by surface tension, which makes the drops increase in
size before they fall. In the case of soils, this balance of forces determines the diameter of the
fingers. In sand this is quite homogeneous, the fingers are nearly vertical and do not merge
with one another. If the sand is less homogeneous; however, the fingers deviate from a
strictly vertical path, and can merge. Finger type preferential flow also creates a non-uniform
moisture profile in the soil. The soil within the initial preferential flow paths will remain wetter
than the soil outside of the fingers. The persistence or horizontal spreading of this non-
uniform water content will affect subsequent infiltration events. Glass et al. (1989) found that
in laboratory sandy soils, non-uniform water contents remained after two weeks of continuous
infiltration. Subsequent infiltrations followed the old finger paths, resulting in persistent
preferential flow. Conversely, if the soil is uniformly wet, Liu et al. (1994) found that the finger
widths are much greater, while Diment and Watson (1985) and Lu et al. (1994) found that
62
preferential flow might even be completely inhibited. At irrigation sites the type of soil will
determine if the soils will wet uniformly.
Figure 4-9: Formation of soil water fingers (Source: Steenhuis et aI., 2002).
3. Funnel flow occurs when sloping geological layers cause pore water to flow laterally,
accumulating at a low region (Steenhuis et al., 2002). If the underlying region is coarser,
finger flow may also occur. The way in which soil is layered makes a big difference in how
water and solutes find paths down to ground water. The way in which these flow paths merge
depends on inhomogeneities in the soil, and sloping structural interfaces have a considerable
effect on the degree of merging and rate of flow. Soil deposited by ocean currents, rivers,
lakes, or glaciers consist of strata formed at different times and under different conditions,
causing them to differ in composition and permeability. Various experiments have shown that
at low flow rates, coarse layers act as funnels, collecting water from a broad area and
channeling it through a small number of drainage fingers. At higher flow rates, these coarse
layers begin to leak, and the funnel effect becomes less significant (Figure 4-10).
63
b
, • j t II , + i tt •• + • + + • i , I + • li!!! 11111[[1111[1
~\
llV' r LIJLIIJl]JJl1 1111
Figure 4-10: The funnel effect in sandy soil. With a relatively low flow rate [a], a fast-moving
spout forms beneath the lower end of an inclined layer of coarse material. Such inequalities
in penetration are minimized when the flow is greater [bj (After Steenhuis et aI., 2002).
One might expect that soil water in contact with the coarse layer might travel straight through
the layer. However, under unsaturated flow conditions, greater pressure is required to push
moisture into large pores from the small pores (Steenhuis et aI., 2002). Therefore, moisture
tends to flow over the layer rather than through it. At low flow rates, the horizontal component
exceeds the vertical component until the edge of the layer is reached where moisture then
funnels through. Since moisture is forced to take this narrow route, the concentration of
moisture and solutes may once again bypass much of the soil matrix and be directed to the
groundwater. The diversion of flow by layering is significant, because even in areas where
groundwater monitoring is conducted routinely, pollution may be missed entirely if the well is
placed in an incorrect location (see discussion in Section 6.6.5).
Figure 4-11: Funnel flow at field site. Blue dye flowing vertically (unsaturated flow) moves
laterally when it encounters a sloping coarse-texture lens. (Source: Soil and Water Lab
Website, Cornell University).
64
4.1 .4 The hydrochemical effects of irrigation with saline water on the soil.
Any water used for irrigation carries ions in solution. Depositing this water on fields in the
form of irrigation can affect the concentration of salts in croplands. If these salts become
excessively concentrated, it can lead to salinisation (Steffen and Savina, 1996). Salinisation
can reduce yields in its earliest stages and eventually lead to the destruction of fertility in the
soil. This is referred to by Steffen and Savina as "The Secondary Salinisation Process":
Salinisation has a direct effect on both plant growth and the structure of the soil. If the soil is
saline a plant will have to expend energy to bring water into its cells because it is forced to
work against osmotic potential. The cation exchange capacity (CEC) affects the stability of
colloid-sized particles in the soil. The cation's positive charge will be attracted to the negative
charge found on clay particles, which make up most of the colloid fraction. Divalent cations
(Ca, Mg) will allow the colloidal particle to group together closely enough so that Van Der
Wahls forces will cause the clays to flocculate, or form stable aggregates. Sodic soils, whose
CEC is dominated by mono-valent sodium cations, will tend to be dispersed and not form
stable aggregates. Sodium is mono-valent and can not pull the colloid particles close enough
together for the short range Van Der Wahls forces to act. Sodic soils will tend to have a dark,
organic appearance due to the dispersion of clay and organic particles while saline soils will
tend to have a white crusty surface due to the precipitation of salts (Lax et aI., 1994). Both
these effects will lead to decreased permeability and hydraulic conductivity.
4.1.5 Prior-conceptual model for irrigation at virgin soil sites.
A new mindset and paradigm for studies that may involve groundwater has been suggested
by LeGrand and Rosen (2000), the objectives are (1) obtaining optimal value from existing
information, (2) reaching a high plateau of knowledge as a basis for further study and (3)
providing an early perspective to be explained to involved stakeholders.
Specialised hydrogeologists rarely use experience, generalisations and qualitative linguistic
modeling to obtain full synergistic value that will maximise understanding. The application of
systematic reasoning is refered to as prior conceptual modeling explanation (PCME) and
represents an initial high grade, synergistic analyses of hydrogeological foreknowledge.
PCME leads to a package of open-ended thoughts and statements that allow for further
study, provisional decisions or definite decisions. If properly prepared and presented at an
early stage, a large part of the total hydrogeological information needed for site studies at an
average site is already available, partly in unrecognised or latent form. With this advanced
background of foreknowledge, adequately developed and displayed, only a fraction of time
and money normally used in site studies may be needed in most cases.
65
Monitoring by the Institute of Groundwater Studies over the years has indicated that very little
salinity applied on surface during irrigation is detected the water table. (See Appendix D for
"Monitoring data"). This indicates that most of the salts are retained in the soil profile. The
general belief is that the salts move according the water currents and accumulate in soils
where evapotranspiration (combined evaporation from soils and transpiration by plants)
exceeds combined precipitation and irrigation (Cordy and Bouwer, 1999). Due to the
composition of the coal mine waters, gypsum precipitation should occur. Annandale et al.
(2002) concluded that there is thermodynamic proof, in the form of calculated Ca and S04
activities, that gypsum is present in the soils irrigated with gypsiferous mine water.
Concentrating the gypsiferous soil solution through crop evapotranspiration precipitates
gypsum in the soil section and therefore removes it from the water system and reducing
potential pollution (Annandale et al, 2002). More clayey layers will retard the vertical flow of
the water, resulting in salt build-up due to precipitation and adsorption.
Hutson (1983) states that water applied to the soil surface initially moves rapidly through
macropores and other pathways, but more slowly through finer matrix pores.
Subsurface preferential flow of water, in which small areas of the subsurface carry large
portions of the flow, has been observed in a variety of soils (Hill and Parlange, 1972). Due to
the nature of the soils at the irrigation sites, preferential flow will occur due to the abundance
of earth worm and rodent holes, as well as cracks in the more clayey soils. According to
Sharma and Hughes (1985) a decrease in chloride values with depth in the unsaturated zone
implicates bimodal flow. Under preferential flow or bypass flow conditions part of the
precipitation water (with relatively low chloride concentrations) will flow directly along these
preferred pathways to below the root zone. This bypass flow component, unaffected by
evapotranspiration, dilutes the soil water that has been concentrated by transpiration,
resulting in relatively low concentrations in the percolation zone with respect to the
concentrations encountered in the root zone. The decrease in chloride values in the
boreholes thus implies preferential flow at the test sites.
Work at numerous sites in the Mpumalanga Coalfields has indicated a saturated weathered
zone of 10m - 15 m deep. Water moves laterally along this zone of higher transmissivity, and
salts that does reach the water table will become diluted in this zone of higher flow. Only a
small part of the salts will move along vertical fractures to the deeper aquifers. Monitoring
may thus portray a false picture of what really reaches the water table if boreholes are
constructed incorrectly, or sampled at incorrect depths.
66
Figure 4-12 presents a concept of what occurs at an irrigation site. Irrigation and rain are
applied, and a large portion of this water evapotranspirates. The rest moves down the soil
profile, with a build-up of salts in the soil, and the conservative ions e.g. chloride moving
through. The concentration of the chloride with depth indicates if bimodal or only diffusive
flow would occur. Once reaching the water table, lateral flow will dilute the water in the
saprolite zone, with some moving further down towards the deeper aquifer. Some will
daylight or seep into streams.
I Are;) 1 Irrignlion (mm)so. ET:O'.g.
1 RT"'"') 11
1 1 1 1
, Runoff
Clay%
~(? SO,· Q! • 1mm)
(7m) SO••
(? m) SO,". Cl •
( 111w."':»'.!'" Clay% 3m
i ~% M~
- - - - - - - - - - - - - - - á-m)So; ':-Q:- - - - - - - r - - - - - - - - - - - -1- - - - -
: .;r~ .......
Vertical +flow % ~.._.
.!l.,;;rn/d. K- +-------~~~re-~~ft~----- 5m
:ê
I
q= A~ _s strerunosr aklng(f)
K= fractures into fractured oquifer (1 m) SO,", 0 " 7m
10m
(1 m) SO, B. IJ' 15m
Figure 4-12: Pre-conceptual model of the irrigation sites.
4.2 Spoils irrigation
Spoils irrigation refers to irrigation on opencast rehabilitated land, which is underlain by
spoils.
4.2.1 Introduction
In opencast pits, the material that remains is extremely heterogeneous, ranging from
boulders down to very fine particles. This provides a highly reactive area when compared to
underground mining, and if large quantities of pyrite remain in the spoils, high sulphate
generation rates can be obtained. The recharge rate is far higher than that for underground
mining; thus dilution is an important mechanism in such a system. Localised areas of acidity
(hotspots) will occur, which serve as initiators of large-scale pyrite oxidation. The decant
67
elevation in these pits is of prime importance; where the elevation is at a point where most of
the spoils are submerged, the water quality is expected to improve over time. Where largely
unsaturated conditions subsist, the likelihood of acidification and poor quality water with high
sulphate increases.
Coal strip-mining was introduced in the Mpumalanga Coalfields in 1971, and became
widespread during the mid-1970s. During the 1980s the awareness of changes in the
hydrological properties of soil became widespread. (The mining method involves the
complete removal of the overburden above the coal in adjacent strips). Following the removal
of the coal, the material from an adjacent strip is dumped into the void and graded to form a
new surface topography. Usable soil materials stripped ahead of mining are then replaced on
the new surface. Because of the large volumes involved, heavy machines are required, which
exert a compactive force on the soil over which they travel. Various soil amelioration and
revegetation operations then follow to complete the rehabilitation process). In an effort to
improve rehabilitation and quantify impacts, large mining companies conducted a number of
studies (Tanner, 1993; Viljoen, 1993; Van der Merwe, 1993).
In order to understand these processes better, a research initiative by the WRC was
conducted at six sites at Kriel, Middelburg South and Optimum (Schoeman et a/., 2002). The
study was linked to work previously done during 1994 - 1996 at Arnot, Kleinkopje and New
Vaal. The findings include:
o The cover soils consisted of structureless clay loam to sandy clay loam, derived from
orthic, red, yellow-brown and plinthic soil horizons, and ranged in depth from 0.55 to
1.05 m. The pH of the soil cover was low in places (average pHKcL 5.7, 5.5, 5.0,4.8,
4.5 and 4.4 at the six sites respectively), and that of the spoil material was low at
Middelburg South «4.5) but above 6 and 7.1 at the Optimum and Kriel sites
respectively. Pockets of strong acidity do occur in spoils material, but acidity due to
pyrite oxidation was not identified as a major limitation to land use (Usher et a/.,
2003).
o Mineralogically, the fine fractions of the cover soils are strongly dominated by kaolinite
(55-80%). Approximately 10% clay-sized quartz occurs. Water-dispersible to total clay
ratios vary between 0.01 and 0.34, indicating relatively high to intermediate aggregate
stability. The cover soils are susceptible to compaction and hardsetting behaviour.
o Water available for evapotranspiration varied between 45% and more than 100% of
the annual rainfall. Maximum rates of evapotranspiration were in the order of 5 to 8
mm per day. The net gain or loss of soil water during the season was small.
68
• Water loss through deep percolation varied from zero to 40%. This was strongly
affected by rainfall distribution, and also differed between sites. Some spoils are
almost permanently dry below a depth of 1-1.3 m. Internal water drainage occurs by
means of deep percolation from the soil profile to a depth greater than the bottom of
the root zone. Bennie ef al. (1998) describe it as the unidirectional vertical flux q
(mmid) through an arbitrary plane, the bottom of the root zone (WD, m). This is mainly
determined by the hydraulic conductivity (K(e), mm/d), which is a logarithmic function
of the average water content e, so that q = K(e). Regarding the estimation of K and
deep percolation, the image emerging from studies by the various researchers is one
of slowly permeable spoil materials with Ks-values of less than 0.007 mid. Seasonally,
lateral drainage may occur along major slopes, leading to water accumulating in
hollows. Although percolation through the spoil matrix is slow, it may be greatly
influenced by settling cracks.
• The calculated run-off varies between 1.9 and 10.8% of the annual rainfall. The
reconstructed soil profiles can generally be regarded as poorly drained. This causes
certain topographical features, e.g. local hollows, to become water collection sites
through lateral surface as well as subsurface run-off, particularly during wet periods.
Viljoen (1992), in a study of soil formation in spoil material between 1 and 18 years old at
Optimum, observed that dense layers appear to become less obvious over time and that
coarse fragments in the upper soil appear to weather rather rapidly. The clay content of the
upper spoil and signs of wetness increase over time, particularly in low-lying landscape
positions. De Villiers (1992), in a study of nearly 100 profiles, noted that dark, fissile shales
particularly appear to soften and disintegrate quite rapidly. This is consistent with their often
high pyrite content.
4.2.2 Hydraulic Properties af Opencast Collieries
4.2.2.1 Spoils hydraulic properties
Hawkins (2004) investigated the predictability of surface mine spoil hydraulic properties in the
Appalachian Plateau. During surface mining, strata overlying the coal (overburden) are
removed by blasting and excavation. Once the coal is extracted, the fragmented spoil is
dumped and pushed into the pit during backfilling. Initially, material placed on the elevated
spoil ridges and piles tend to be poorly sorted. Larger spoil particles will however roll to the
base of the spoil ridges, creating well-sorted, highly transmissive zones. Extreme
heterogeneity is introduced into the backfill material by the broad range of particle sizes and
differences in particle sorting (Hawkins, 1998).
69
Heterogeneities introduced during mining and reclamation cause spoil to exhibit a dual-flow
groundwater system. Mine backfill contains a substantial percentage, by volume, of large
(macro) voids. The void volume may equal the spoil swell (a volume increase from mining,
which may be as high as 20 to 25%. Porosity values of 16 to 23% have been recorded by
field measurements in mine spoil in the coalfields (Rahn, 1976; Hawkins, 1998). The macro-
voids are within a matrix of unconsolidated material composed of clay-sized «2 microns) to
boulder-sized (>2m) particles. The macro-voids' behaviour is similar to that of a karst aquifer,
capable of storing and transporting large volumes of water. They may exhibit multiple water
tables within the same unit (Caruccio et a/., 1984), whereas spoil material itself behaves as a
highly transmissive unconsolidated porous medium. Mine spoils exhibit an average hydraulic
conductivity of two orders of magnitude greater (geometric mean) than the adjacent unmined
strata (Hawkins, 1998).
The water level and conductivity forecasting are based on the factors of mine backfill age,
spoils lithology and thickness, dip of the pit floor and distance to the closest unmined strata
(highwall). Spoil age is important as a parameter, because spoil properties evolve with time.
In aquifer testing, mine spoil behaves as an unconfined aquifer under most circumstances. A
summary of aquifer testing illustrates a wide range of K-values, and ranges over seven
orders of magnitude from 3.5x10-4 to 1.73x1 03 mld. The smaller K-values associated with
mines <31 months old are attributed to the poor interconnection of voids and pore spaces of
the relatively freshly acclaimed spoil in which the water table may take up to 24 months to
rebound. On the other hand, small K-values at mines >100 months old appear to be caused
by spoil compaction, rock weathering, piping and settlement that decrease the void and pore
space as these sites age. However, no reasonable correlation for K could be approximated
by regression analysis for age or any other factors.
Greater saturated thicknesses were associated with higher sandstone content (>35%). This
relationship is related to the higher rates of recharge associated with blocky sandstones.
Analyses indicate that the prediction of saturated thickness is directly dependent on the
sandstone/shale ratio, age of the spoil, total spoil thickness, and the distance to the highwall.
This relationship is strongest for sites <60 months old, but is still somewhat viable for sites>
60 months in age. The decreased predictability at the older sites appears to result from
differing degrees of spoil compaction, piping, settling, and rock weathering commonly
observed between mine sites.
Even though it seems as if the findings of the different researchers differ, there is truth in all
the statements. From personal experience the author has seen spoils that drain fast, as well
as spoils that drains very slowly. Much depends on the type of spoils, the age and the clay
70
content in the weathered rock, as well as the capping material and the degree of compaction
by construction vehicles.
Sracek et al. (2004) investigated the geochemical characterisation of Acid Mine Drainage
from a waste rock pile at Mine Doyon in Québec, Canada. Pore water in the unsaturated
zone was sampled from suction Iysimeters and with piezometers in underlying saturated
rocks. The investigation revealed strong temporal (dry period vs. recharge period), and
spatial (slope vs. central region of pile) variability in the formation of Acid Mine Drainage. The
main secondary minerals observed were gypsum and jarosite.
The AMD production in mine tailings has been studied extensively, for example by Dubrovsky
et al. (1985) and Blowes et al. (1991). In contrast, much less is known about AMD generation
from waste rock. Several waste rock studies have evaluated temperature and gas
concentration profiles (Lefebvre et aI., 2001 and Kuo and Ritchie, 1999). These studies
revealed the significant role of air convection in the unsaturated zone of very reactive waste
rock piles. However, the geochemical environment within waste rock piles is less well known.
Until recently, the focus in waste rock geochemical studies has been on the quality of
drainage water flowing from the waste rock base (Morin et al., 1994). This type of monitoring
represents an integral part of the acid drainage produced in various parts of the dump over
space and time. However, this information does not provide a clear picture of geochemical
processes within the pile (Ritchie, 1994). There have been attempts to obtain data on water
chemistry within the unsaturated zone of waste rock piles (Shafer et aI., 1994). They were not
very successful, however, because of the low water content and correspondingly high suction
values characteristic of semi-arid climate conditions in the northwest USA. The acquisition of
geochemical data from the unsaturated zone of a waste rock pile remains a topic of on-going
research (for example, Smith etal., 1995 and Stockwell etal., 2001).
The approach that Sracek et al. (2004) use for the assessment of water quality within waste
rock deposits was based on sampling leachate from both unsaturated and saturated zones
within the pile combined with transport and geochemical modelling, and solid phase analysis.
4.2.2.2 Hydrogeology of coal mine tailings
Tom and Blowes (1999) studied the hydrogeology of tailings impoundment. They stated that,
because flowing pore water in tailings can transport dissolved constituents from the enclosed
impoundment into the surrounding environment, an understanding of the flow system is
critical to accurately assess the environmental impact of tailings disposal. In addition, long-
term water-treatment costs are affected by the amount and quality of the effluent from the
tailings. Important aspects of the groundwater flow system within the tailings and surrounding
natural geological materials are the direction and rate of groundwater movement and the
71
distribution of recharge and discharge areas. Delineating recharge areas is important,
because geochemical processes such as sulfide oxidation may degrade the quality of the
recharging pore water, ultimately affecting the quality of the pore water throughout the
impoundment. When the locations of recharge areas are known, steps may be taken during
mine decommissioning to protect these areas from the effects of sulfide oxidation.
To characterise the pore water flow system within the tailings, the hydraulic conductivity (1<),
hydraulic head and the water table elevation at piezometer nests located within the tailings
impoundment were measured.
Results of the study indicated that hydraulic conductivity measurements with the Hvorslev
method (Hvorslev, 1951) range between 8x10-5 to 8x1 0-2 mid. The values obtained using the
constant-head permeameter were consistent with this range, with fractured samples
displaying K values of up to one order of magnitude greater than the matrix (10-4 vs. 10-5
mid). The narrow distribution of Kwithin the tailings was a reflection of the uniform grain-size
distribution. There was a trend toward decreasing Kwith increasing depth, which may be due
to consolidation of the tailings. The lowest Kvalues «1 x 10-8 mis) were obtained from the
deepest piezometers, thought to be installed within the underlying silt and clay (AI et aI.,
1994b).
4.2.2.3 Stratification and mineralogy of spoils versus hydrochemistry
Some mining types have a set of conditions that promote poor quality water, while others
ameliorate the effects of pyrite oxidation. In all mining types, however, the local mineralogical
conditions provide the most important driver of water quality. Newbrough and Gammons
(2002) emphasise the importance of understanding the relationship between mine drainage
chemistry and the mineralogy of the surrounding rocks, and how these relationships can vary
on a district scale.
Usher (2003) indicates the validity of this observation by means of trenches dug in the spoils.
In underground mining, this is no less important. A case study at an underground
compartment shows that the roof sediments have a higher probability of acidification than the
coal seam itself, implying that the high extraction areas pose the greatest risk. The balance
between the increased recharge, with the addition of alkalinity and faster inundation, against
the higher surface area and lower net neutralising potential will determine the final water
quality.
Table 4-1 summarises some of the key issues from a hydrogeochemical perspective:
72
Table 4-1: Key processes in opencast mining that affect water quality (Usher, 2003).
Rehabilitated Opencast
Recharge 15%-25%
Dilution can occur
Relative Surface Area High
Faster reactions; high salt loads
Porosity 25%
Flooding can occur to prevent excessive salt loads
Coal Removal 90%
If coal seams are the most likely to acidify, it reduces the risk
Decant Likely
Outflow can occur with small amount of material flooded
4.2.2.4 Conceptual model for irrigation on spoils
The way in which the spoils heaps react to irrigation is an important consideration for
correctly including the influences within a reactive model. Major aspects are schematically
indicated in Figure 4-13
1. All the spoils above the flooded or final decant elevation will remain reactive for
extended periods after closure. Pyrite oxidation will continue and, provided that
enough calcareous carbonate material occurs, concentrations will not exceed in the
order of 5000- 6000 mg/1. If acidity occurs, this value can be greatly increased.
Recharge onto rehabilitated spoils will be in the order of 15-20%.
2. In the flooded spoils, oxidation reactions will decrease in most cases. James (1996)
shows this to be the case in laboratory tests on South African spoils, while research
by Vermeulen (2003) shows this to hold for spoils at Transvaal Navigational Colliery
(TNC) and Schoongeschigt Colliery. Van Tonder et al. (2003) and others have shown
that the inflow and outflow volumes in the pit are in the same order. The transmissivity
of the spoils is extremely high, resulting in a flat water table controlled by the decant
elevation.
73
Figure 4-13: Conceptual overview of spoils reactive model
3. If irrigation is added to the normal recharge, the irrigated areas will have an additional
salt load and provide an additional source of moisture to sustain AMD processes in
the unflooded zone. This additional water may also provide a medium of transport for
the precipitated salts in the unflooded zone.
4. In general, all rehabilitated open cast pits will decant at the lowest level at which the
spoils intersect the topography. This decant will constantly add salt load to the
system in the long term, as long as water infiltrates the spoils. It is considered unlikely
that, on the scale of more than 75 000 ha, engineered covers will be largely
employed. This constant salt load addition must therefore be included in the
assessment of any comparative management.
74
5 METHODOLOGY
5.1 Introduction
Irrigation with mine water has been ongoing at the different pivots (as described in Chapter 3)
for periods of three to nine years. Monitoring the groundwater quality at these sites has not
detected significant deterioration of groundwater quality. The focused and detailed monitoring
of the groundwater throughout the duration of the project is one of the pillars of this project,
and may even be extended beyond the duration of this work.
To achieve the aims of this study, the methodologies that are outlined in this chapter were
applied. Detail of each approach is given in the different sections of this chapter.
5.2 Groundwater monitoring
Boreholes were drilled at all the irrigation sites, and placed at such positions that the water
levels and water quality, inside and outside the irrigation areas (for background values), could
be measured.
The boreholes were drilled by means of air percussion method. The boreholes were drilled to
a diameter of 165 mm. All the boreholes were cased as a precautionary measure against
collapse.
The monitoring investigation included:
o Investigation of the local geology
o Water level measurements and interpretation.
o Multi-parameter chemical profiling of the water quality in the boreholes.
o Sampling and chemical analyses of water in the boreholes
5.2.1 Geochemical profiling
The boreholes were profiled with an YSI multi-parameter Sonde probe (Figure 5-1)
connected to a laptop computer. Multi-parameter logs ensure that water qualities are
measured, as they are in-situ, without disturbing the water column in the boreholes through
sampling. The advantages of such probing are as follows:
o In-situ measuring of temperature, DO, ORP, pH and conductivity.
o The position of fractures can be identified.
75
• Stratification of the water column can be recognized.
• The data can be used for hydrochemical modelling.
.•~ - -~ - --- ----'~,..~,... --... ,,-- . - ~
\~ .........-
!~.. ~'-' .:- -.ói • ~ ~ ::.. • ., 0>' _
Figure 5-1: Multi-parameter probe with different sensors.
An example of probing results is shown in Figure 5-2. These data can be related to
geohydrological programs such as WISH for interpretation, as illustrated in Figure 5-3. The
stratigraphic nature of the water column is clearly shown in Figure 5-2.
BH1.DAT
,.'
09:40 09~41 09!41 09'42 09:42
04/05/04 04/05/04 04/05/04 04/05/04 04/05/04
OateTime(M/OIY)
Figure 5-2: Graph of profiling down a borehole.
76
Borehole Log - PV1
Depth rml Locality. X: 92640.40 Y: 2955310.09 Z: 1425.00
Lithology Geology ORP 250 48 PH 6.8 SPCOND1mS{'l'd 14 Temp ICJ 22
10
15
20 \
25
30
35
Figure 5-3: Multi-parameter profiling in WISH.
In piesometers such as those installed at New Vaal a TLC (temperature, water level,
electrical conductivity) meter was used to measure the electrical conductivity and
temperature in depth.
5.2.2 Chemical analysis
Following the water quality probing in the boreholes, sampling was done with a sophisticated
pressurized stainless steel depth sampler (Figure 5-4). Depending on the result from the
profiling, sampling depths were determined.
Figure 5-4: Specific Depth Sampler (from www.solinst.comJ
77
It is essential that sampling should always be at exactly the same depth in order to get a
uniform picture of the true water quality. Sampling at different depths can result in incorrect
concentrations. In order to prevent heavy metals in the water samples to precipitate, part of it
was filtered and acidified to a pH of below 3 immediately after sampling. After every sample
all the equipment was washed with deionized water to ensure that cross-contamination is
ruled out.
All analyses were done at the Institute for Groundwater Studies at the University of the Free
State. The following methods were used:
Anions: A filtered water sample was injected into a stream of carbonate-bicarbonate eluent
and passed through a series of ion exchangers. The anions of interest were separated on the
basis of their relative affinities for a low capacity, strongly basic anion exchanger (guard and
separator columns). The system used is the Dionex DX-120 ion chromatograph and the
columns used are lonpac AG 14 and AS14 columns.
Cations: Dissolved metals were determined in filtered and acidified samples. An ICP source
consists of a flowing stream of argon gas. A sample aerosol is generated in an appropriate
nebulizer and spray chamber and is carried into the plasma subjecting the constituent atoms
to temperatures of 8 - 10 000 K. The high temperature of the plasma efficiently excites atomic
emission. The light emission is registered and the concentrations of the metals determined.
5.3 Soil Sampling
Obtaining soil and rock samples is a vital component of any drainage prediction methodology
and relies on adequate representative sampling. Sampling was conducted at the various
irrigation areas by means of (Figure 5-5):
o Core drilling
o Auger drilling and
o Test pits / trenches
o Samples were taken inside as well as outside the irrigation areas to obtain
background values. These samples were analysed at distinct lithological horizons to
determine clay content as well as the ion / mineral content.
o An excavator was used to dig the inspection pits in virgin soil and spoils to depths of
approximately 4m. From these, the stratigraphic nature of the soil, salient features
and the hydraulic features could be studied. In-situ samples were taken, which stored
in airtight bags.
78
Figure 5-5· Soil sampling methods.
5.4 Soil Analysis
5.4.1 Soil Properties
Sampled soil was analysed at the Glen Agricultural College near Bloemfontein to determine
the soil texture and clay content. For this determination, a 7 Fraction Texture Analysis was
used as follows:
Weighed 10 g of soil, add 25 ern" of calgon, and dispersed the sample for 10 minutes.
Washed dispersed sample in a 1 litre cylinder and stirred for 1 minute. Pipetted 25 ern" of the
suspension (pipette at the 1 min reading mark) into weighed beakers. After 4 min: pipetted
25 crrr' of the suspension (pipette at the 4 min reading mark) into weighed beakers. After six
hours: pipetted 25 crrr' of the suspension (pipette at the 6-hour reading mark) into weighed
beakers. Placed beakers in an oven at 105°C and dried until constant mass. It was cooled in
a desiccator before weighing the beakers. From this the silt and clay fractions and sieve
sand for sand fractions could be calculated which was important in the soil moisture
calculations.
79
5.4.2 Soil chemical analysis
All chemical analyses were done at the Institute for Groundwater Studies. All samples were
milled to pass a 250-llm sieve and dried in the oven at 45 DC (ISO 11048).
Normally all leaching methods in soil sciences were developed to determine total plant-
available sulphate (Tabatabai, 1982 and Dejager, 2005). De Jager developed a method to
determine the dissolved and exchangeable/adsorbed sulphate in the soil due to irrigation
activities, which was used to analyse the soil:
A suspension of 5g soil and 50 ml of water was shaken for four hours and left to suspend for
16 hours. It was then filtered with Whatman No1 filter paper after which the extract was
analysed with ion chromatography and ICP to determine the precipitated part of the sulphate
and other ions. The soil residue from the filter was then dried, and a suspension of 5g soil
and 50 ml potassium-ortophosphate was treated in the same way to determine the adsorped
sulphate and other ions.
5.4.3 XRDandXRF
All samples were sent to the Geology Department at the University of the Free State for X-ray
diffraction analysis to determine the presence of gypsum and clay minerals. The mineral
composition of a sample will ultimately determine the contribution of different species in the
system as a whole (Usher et aI., 2003). The analyses were carried out with a Siemens D5000
Diffractometer with a CuKa radiation at 30kV and 40mA. Diffraction scans of the X-ray
intensity pattern against a 28 angle were viewed and interpreted using DIFFRAC-AT V.3.0.
software together with the JCPDS Mineral Powder Diffraction File Data Book (1980a) and
JCPDS Mineral Powder Diffraction File Search Manual (1980b). X-ray diffraction from
crystalline solids occurs as a result of the interaction of X-rays with the electron charge
distribution in the crystal lattice. The diffraction peaks are individually considered to be the
result of the diffraction of the incident X-ray beam of wavelength A from crystal lattice planes,
having Miller indices hkl and spacing dhkl. Plotting the measured intensity against the 28
angle (or corresponding crystal d-spacing) produces a diffractogram from which the minerals
in the sample can be characterised (Wainerdi and Uken, 1971).
X-ray Fluorescence (XRF) is the key technique for characterizing the element composition of
rocks and soil. XRF is generally used for analysis of bulk solids, and can be used to routinely
analyse for almost any element heavier than neon including non-metals. XRF analyses have
the additional benefit that it does not depend on putting a sample into solution, thus
eliminating problems with insoluble residues.
80
5.5 Soil water sampling (Porous cup Iysimeters)
The unsaturated zone is the portion of the subsurface above the groundwater table. It
contains air as well as water in the pores. Its thickness can range from 0 meters, as when a
lake or marsh is at the surface, to hundreds of meters, as is common in arid regions (GCRIO
Website http://www.gcrio.org/geo/grndwt.html).Atirrigation sites the elevated perched water
table will normally vary between 2m and 10m.
Soil Water Samplers, also referred to as suction Iysimeters or porous cups, are used to
collect water samples from soil profile. The samplers are installed at the desired depth and
left in the soil, allowing periodic sampling to occur with minimal disturbance of the soil. The
samplers consist of a porous ceramic cup and a sample collection tube. A vacuum pump is
used to create a vacuum in the sampler, which draws water from the soil matrix, which
passes through the ceramic cup and into the sampler. The water sample can then be
extracted from the collection tube and taken to the lab for chemical analysis.
In 1961, Skaling and Wagner, of the University of Missouri, fashioned the first suction
Iysimeter (Soil Report Newsletter of Soilmoisture Equipment Corp, 2002). The suction
Iysimeter was a cylindrical device consisting of a porous ceramic cup (to withdraw soil pore
water); a body tube to act as a reservoir; and a simple stopper assembly with a single hole for
creating a vacuum and retrieving the sample. These early suction Iysimeters allowed pore
water to be extracted from unsaturated soils near the surface. Subsequent changes to the
suction Iysimeter added a "pressure" port to the stopper assembly and the
"pressure/vacuum" Iysimeter was born. The new pressure/vacuum Iysimeters extend the use
of Iysimeters to greater depths or even remote locations. Recently Iysimeters have been
extensively used to evaluate solute transport models (Corwin et al., 1992; Klein et ei, 1997
and Butler et aI., 1992), to monitor the fate and mobility of soil contaminants (Burne et aI.,
1994; Kelly et el, 1998 and Corwin et al., 1994) and for evapotranspiration studies (Prueger
etal.,1997).
Unsaturated soils and the vadose zone are composed of a mixture of soil particles, water
held on the surface of soil particles and in small capillary spaces between the particles and
interconnecting air passages that are open to the atmosphere at the soil surface.
Extraction of a pore liquid sample requires a hydrophylic porous material with numerous pore
channels such as ceramic to transport soil water fluids without alteration or leaking, in order
to establish and maintain intimate hydraulic contact between interstitial pore liquid and the
sampler chamber. Because of its surface tension characteristics, water in contact with a
ceramic cup will completely fill its pores whenever the air pressure differential across the
ceramic does not exceed a critical value related to pore size called the "air entry value" or
"bubbling pressure". Volumetric porosities of the interface material should be greater than
81
10% and pores small enough to sustain air pressures "or bubbling pressures" greater than 10
psi Ceramic soil water samplers have a maximum pore size of 2.5 microns and can maintain
a vacuum in excess of 0.95 bar.
Generally there are four possible materials considered for the interface: porous plastic films,
porous plastic shapes, sintered metals and ceramics. The first three materials have
significant drawbacks in terms of functionality. Delicate plastic films tear easily and cannot be
well supported. Porous plastic shapes have large non-uniform pores and require special
surface treatments to become hydrophylic. Sintered metals have high exchange capacities,
will frequently oxidise, and again have non-uniform pore sizes. Ceramics are historically
applied and is known for their long-term use, with alumina and porcelains preferred for their
inert and tough characteristics.
Today's suction Iysimeters and pressure/vacuum Iysimeters rely on a surface of uniform
wetted pores that act as contact points that draw fluids from soils, through interface pore
channels into the reservoir of the sampler (lysimeter). If the pores of the sampler's interface
material are either not small enough in size or not uniform, the water links between the soil
and the sampler break. Air instead of water then enters the sampler under vacuum. In dry
conditions, water in the pores of the ceramic will move out of the cup, into the surrounding
soil, and the vacuum in the sampler will be relieved. Since the unsaturated hydraulic
conductivity of soil decreases exponentially with increasing dryness, liquid phase movement
virtually ceases in dry soils. When a wetting front pattern approaches, the soil moisture
tension will decrease. As the soil moisture tension falls below about 2 bars, a vacuum can be
maintained in the soil water sampler. At a soil moisture tension of about .65 bars, a pore
liquid sample will be drawn into the sampler.
The photograph in Figure 5-6 illustrates what the porous cup Iysimeters built for this research
look like. PVC tubing and fittings were bonded using a mix of epoxy bonds. Two pipes were
connected to the tubing, with one placed into the ceramic cup. The latter is used to extract
the water with. The water was easily extracted by pressurising the sampler through the one
pipe and thus forcing the water out through the deeper pipe in the ceramic cup.
82
Figure 5-6: Porous cup /ysimeters and vacuum pump.
The soil water samplers were installed in diamond drilling holes. The ceramic cups were
bedded in very fine-grained silica flour so that the porous ceramic was in intimate hydraulic
contact with the soil. These samplers were installed every meter down to 4 meters, bedded in
the silica. The hole was filled halfway with the original soil, up to the next sampler. A cement
seal layer was then applied, and soil was entered on top of this again until the next sampler
was reached. The top of the hole was also sealed with a layer of cement. The cement layers
ensured that water would not move vertically down the borehole, but rather downwards in the
undisturbed soil.
Whenever a sample was needed, the sampler was placed under suction for three days, after
which the water was sampled using very low pressure.
5.6 Tensiometry
Soil water tension, soil water suction, or soil water potential are all terms describing the
energy status of soil water. Soil water potential is a measure of the amount of energy with
which water is held in the soil. A soil water characteristic or water release curve shows the
relation between soil water content and soil water tension.
Tensiometers have been used for many years to measure soil water tension in the field.
Tensiometers are water-filled tubes with hollow ceramic tips attached at one end and a
vacuum gauge (or mercury manometer) and airtight seal at the other end. The device is
installed in the soil with the ceramic tip in close contact with the soil at the desired depth. The
water in the tensiometer eventually arrives at pressure equilibrium with the surrounding soil
through the ceramic tip. Water is extracted into the soil through the ceramic tip, creating
83
tension in the closed tube. As the soil is re-wetted, the tension gradient reduces, causing
water to flow into the ceramic tip. As the soil goes through wetting and drying cycles, tension
readings can be taken.
Most commercially available tensiometers use a vacuum gauge to read the tension created
and have a scale from 0 to 100 centibars (one bar or 100 centibars of pressure or tension is
equal to 14.7 psi). The practical operating range is from 0 to 75 centibars. If the water column
is intact, a zero reading indicates saturated soil conditions. Readings of around 10 centibars
(eb) correspond to the field capacity for coarse-textured soils, while readings of around 30 eb
can approximate field capacity for some finer-textured soils. The upper limit of 75 eb
corresponds to as much as 90% depletion of total available water for the coarse-textured
soils, but is only about 30% of the depletion for silt loam, clay loams, and other fine-textured
soils. This limits the practical use of tensiometers for coarse-textured soils or to high
frequency irrigation where soil water content is maintained at a high level.
Tensiometer readings may be used as indicators of soil water and also of the need for
irrigation. When instruments installed at shallower depths of the root zone reach a certain
reading, they can be used to determine the time to start irrigation, based on soil texture and
crop type. Similarly, instruments at deeper depths of the root zone may be used to indicate
when adequate water has been applied. It can also be used to determine the soil water
volumes (and movement) deeper down towards the water level. Careful installation and
maintenance of tensiometers are required for reliable results. The ceramic tip must be in
intimate and complete contact with the soil. During this study holes were drilled with a core-
drilling rig to install the tensiometers in. A slurry of silica flour and original soil was introduced
into the hole at the desired depth into which the tensiometer tip was inserted. The hole was
filled with some of the original soil, sealed with cement halfway between two tensiometers
and the process was repeated with the next tensiometer in the same fashion as with the
porous cup Iysimeters. A few days are required for the tensiometer to arrive at equilibrium
with the surrounding soil.
For this research, Lorentz and Pretorius from the School of Bioresources Engineering and
Environmental Hydrology at the University of Kwazulu-Natal designed custom-made
tensiometers to measure tension as deep as 6 m, and depending on the depth of the water
table, tensiometers were installed at every meter down the drilled hole (Figure 5-7). The
operation of these tensiometers is described in Manual for the operation of Tensiometer
Nests using the HOBO Logger System (Thornton-Dibb and Lorentz,2002).
A four-channel logger was developed to record signals from pressure transducers attached to
tensiometers and submerged in groundwater wells. The automatic tensiometers house the
water phase in a hydraulic hose, which is protected using PVC conduit. The components are
84
all modular, which allows for convenient assembly in the field to lengths determined in-situ.
Tests indicated no apparent temperature effects on the electronics that would cause
observable signal fluctuations, and that the diurnal fluctuations in automatic tensiometer
signal appears to be the result of expansion and contraction of air entrapped in the
tensiometer. Hence a rigorous maintenance program, ensuring that air pockets are removed
from the apparatus at regular intervals, must accompany the use of automated tensiometers
(Figure 5-7).
Figure 5-7" Installation oftensiometers (left) and replenishment (right).
5.7 Leaching columns
Leaching column tests were undertaken to determine the kinetic behaviour of waste rock,
soil, ore or tailings stored on the surface and exposed to atmospheric weathering (sub-aerial
storage), or stored under water (sub-aqueous storage). In either case the aim is to monitor
water (leachate) quality with time by cyclic (daily, weekly or monthly) sampling. Unlike
humidity cell procedure (see Sectiion 5.8), there is little, if any, standardisation of column
testwork procedure, allowing considerable flexibility. This flexibility permits column operations
to be highly site- or material- specific with regard to material particle size and size range
(which for waste rock, soil, or drill core is usually greater than that used for humidity cell tests,
but still less than that of site conditions), sample mass, water infiltration or flow rate, and
degree of oxygenation.
Different leaching column tests were performed to investigate various aspects of the
movement of water, and the accumulation and dilution of salts in the soil and spoils of various
irrigation areas, including the following:
85
o Leaching of acidic and non-acidic spoils with deionised water, irrigation water and
irrigation return water. Three duplicate sets of 2-m columns with a diameter of 160
mm were packed with raw spoils obtained directly from the mines. After thorough
mixing, equivalent amounts of spoils were loaded sequentially into each column by
hand. The columns were loaded until each contained 50 kg of material. Initially the
cells were leached with 5 litres of water to completely inundate the column and leach
out the accumulated oxidation products.
o In situ columns were installed at the spoils irrigation site at Optimum. These columns
were 215 mm in diameter. Three columns with lengths 2m, 4m and 4.5m were
installed and filled with spoils in the exact sequence of collection from the pit in which
these columns were placed. These columns include a dead end at the bottom, which
was filled with fine-grained washed gravel where the water could assemble. A small
piezometer pipe was installed in each column to measure water levels and extract
water.
o Leaching of unpolluted background soils with irrigation water (25 pore volume cycles)
with gypsiferous waters in order to determine precipitation and adsorption capabilities
of the soils. The columns were 65mm in diameter and 2.5 m in length.
o Leaching of irrigated soil columns (sampled down to the water table) with rainwater.
This was done to determine the time required to flush polluted soils, at which rates
and tempos. The columns were 65 mm in diameter and 3.5 m in length, analogous to
the real distance to the water table. These columns were not scaled down.
o Short, wide drums were used for gypsum leaching. Different gypsum concentration
scenarios were run for the different types of soils at Kleinkopje Pivot Major (Bainsvlei
and Clovelly soils -lower clay content, 5-10%) and Pivot 4 (Hutton soil - higher clay
content, average 18%), Syferfontein Pivot (Arcadia soil-high clay content, in the order
of 60%) and New Vaal Pivot (sandy soil). For the Kleinkopje Pivots three scenarios
were simulated i.e. 10 ton gypsum / ha, 20 ton gypsum / ha and 40 ton gypsum / ha,
but for Syferfontein and New Vaal only one scenario each i.e. 20 ton gypsum / ha.
During each leaching event 75 mm of rainwater was applied to each drum, and the
water analysed. The soil was then dried out under natural conditions. This process
was repeated eight times. After each drying cycle, the soil was loosened as if
cultivated by the farmer. This was done to simulate realistic agricultural conditions.
The photographs in Figure 5-8 show the different column designs used in the experiments as
well as annular rings that were installed to prevent sidewall flow.
86
Figure 5-8: Photograph of the different leaching columns used (left) and annular rings (right).
A common criticism of many soil Iysimeter designs has been the existence of artificial flow
paths along the soil-wall interface (Saffigna et al., 1977 and Corwin, 2000). This artificial flow
is referred to as sidewall flow. Sidewall flow in a soil Iysimeter indicates the artificial
channelling of water due to the separation of the soil from the Iysimeter wall creating
airspace. These airspaces serve as artificial flow paths that permit the rapid flow of water and
the transport of solutes, which is a condition that is not representative of the field. Because of
this anomalous flow behaviour along the column walls, the past use of Iysimeters to evaluate
solute transport models and the mobility of contaminants through soil has been questionable.
A simple Iysimeter design modification utilising annular rings to divert sidewall flow near the
soil surface into the soil column to minimize the occurrence of sidewall flow along the
remainder of the column's length, was evaluated by Corwin. A chloride-tracer experiment was
used to evaluate the effectiveness of annular rings in minimising sidewall flow in a mesoscale
soil Iysimeter (0.6 m in diameter and 1.83 m in height). The tracer-experiment data showed
that even though sidewall flow may not have been completely eliminated it was reduced to an
undetectable level based on chloride distributions and time domain reflectometry
measurements. However, a delicate balance exists between minimizing sidewall flow and
significantly altering the natural water-flow dynamics when using annular rings. The simple
design modification provides a means of using a disturbed column of soil to evaluate models
of solute transport, and to study preferential flow and contaminant mobility without concern
for spurious data due to artificial flow along the soil-wall interface of the Iysimeter.
87
Figure 5-9: Idealised depiction of how annular rings minimise the artificial flow of water along
the sou-wet! interface of the /ysimeter by redirecting sidewall flow inward (after Corwin, 2000).
Aside from the use of annular rings, another aspect that is crucial to minimising sidewall flow
is the way in which the soil is packed into the Iysimeter or column, particularly at the soil-wall
interface. Precautions were taken to create a column of soil that correlates as closely as
possible with naturally occurring preferential flow paths. Care was taken to add very small
increments of soil into the column that were compacted uniformly with applications of water
rather than using the physical pressure of pounding. This was done in an effort to pack the
column uniformly. After each increment of soil was added, a series of wetting-and-drying
cycles were used to compact the soil. These cycles, with the purpose to promote particle
settling and compaction of soil particles, has been a common approach for packing
Iysimeters (Shih and Rosen, 1985 and Bowman, 1988), and used in this research.
5.8 Humidity cells
Kinetic tests are intended to mimic the geochemical processes at mining sites, usually at an
accelerated rate. These tests require more time and are considerably more expensive than
static tests (US EPA, 1994).
Acid-base accounting procedures are used as a screening process to categorise materials
into potentially acid generating, potentially non-acid generating and uncertain groups. For
material where the potential for acid generation is uncertain, kinetic test work is performed in
an attempt to define acid generation characteristics. The term kinetic is used to describe a
88
group of test work procedures wherein the acid generation (and metal solubilisation and
transport) characteristics of a sample are measured with respect to time (Mills, 1998).
Internationally, humidity cells have become the most popular devices for conducting kinetic
tests. locally the usage has, unfortunately, been very limited. Cell designs can vary in the
material of construction, geometry and size. A typical cell is constructed of plexiglas with the
dimensions 10 cm diameter by 20 cm in length, and has a nominal capacity for 1 kg of rock.
The rock sample is typically crushed to grain sizes of less than 6 mm in diameter and placed
on a perforated plate to permit the flow of air through the bed of rock. The cell can be
provided with a bubbler tube containing water, attached to a tight-fitting lid, through which the
exiting air is passed. The bubbler provides a visual check that air is flowing through the cell,
and allows the operator to achieve a semi-quantitative balancing of airflow through a bank of
humidity cells. Dry or humidified air is supplied to the underside of the perforated plate. The
temperature of the water in the humidifier should preferably be maintained slightly above
ambient to ensure a good supply of humidified air. leachate, usually distilled water, is
periodically added through the lid of the cell. The mode of addition can vary. In some test
programs, the water, typically 250 ml to 500 ml is added slowly over several hours
(percolation leaching). In this method, the valve at the bottom of the cell is open to allow free
draining. In other programs, sufficient water can be added to completely submerge the
sample for a period of time, before the bottom valve is opened to allow draining (Usher,
2003). Humidity cells that are constructed as illustrated in Figure 5-10 are currently used at
the Institute for Groundwater Studies (IGS).
Humidity cell tests are performed to evaluate the long-term acid-producing potential of mine
waste rock, tailings or spent ore. The test simulates the accelerated weathering of the
sample. This is done by passing moist air followed by dry air through the sample chamber;
moist air for three days, followed by dry air for three days, and distilled water on the seventh
day. The leachate is analysed for pH, conductivity, sulphate, acidity and dissolved alkalinity.
This one-week cycle is typically run for 20 weeks (Usher, 2003). Samples are usually
selected to represent the various lithologies at the mine, along with a range of ABA and/or
NP/AP values. In this way, humidity cell results can be used to identify or confirm the AMD
risk associated with a range of ABA values. It is normal for data to be quite erratic over the
first few cycles, before consistent results are obtained. This is due to the removal of readily
soluble components from prior oxidation and weathering. It is not unusual for humidity cell
tests to continue for several months, or even more than one or two years (lawrence and Day,
1997).
89
Figure 5-10: An IGS humidity cell, left and an array of cells set up at the IGS, right
(after Usher, 2003).
Different weekly methodologies for this research were used. Two identical sets of non-acid
generating rock were used for the first experiment that ran over 20 weeks (obtained from
Optimum Colliery). This was followed with two sets of acid-generating rock for the second
experiment (obtained from Middelburg Mines).
• Cells 1 and 4 were leached each week by the addition of mine water.
• Cells 2 and 5 were leached each week by the leachate from that column, made up to
500 ml by mine water.
• Cells 3 and 6 were leached each week by the addition of deionised water.
• Two cells, not connected to the humidity cell system, were filled with irrigation water
and also closed at the top. The rock in the cells was totally submerged, and the water
quality was only measured every seven weeks to keep as little air as possible from
entering the system. This was done to simulate "flooded" background pit conditions.
5.9 Point Dilution Tests
The point dilution test is a simple method to find the groundwater velocity. Since high
velocities are related to highly conductive units, the test is used to locate fractures and other
conductive features at less time and cost than the geophysical methods. The test requires
careful data measurement and interpretation, but will not indicate groundwater flow
directions.
90
The use of artificial tracers involves the injection of tracers in the subsurface and their
detection down gradient in monitoring wells (Freeze and Cherry, 1979; Riemann, 2002 and
Van Tonder et aL, 2002b). Taking the effect of dispersion into account, the groundwater flow
rate and direction can thereby be evaluated. However, this experiment needs more funding
and is time intensive. An easy and inexpensive alternative is the use of point dilution tests to
estimate groundwater velocity (Lamontagne et al, 2002 and Drost et aL, 1968). A single well
point dilution test method aims to relate the observed rate of tracer dilution in a borehole, or a
segment isolated in a borehole, to the Darcy velocity in the aquifer. The groundwater flow
gradually depletes a tracer introduced to the well to produce a concentration-time relationship
from which the Darcy velocity and horizontal hydraulic conductivity can be evaluated.
The aim of the point dilution experiment can be summarised in the following three points:
o Delineation of hydrogeologic profiles within a borehole
o Estimation of Darcy velocity
o Estimation of horizontal hydraulic conductivity.
The experiments were performed in boreholes at each irrigation site. The procedures were
the same in all the boreholes (Gebrekristros, 2006). The background EC-value was firstly
measured. After estimating the volume of water in the borehole section, NaCI was used as a
tracer. Changes in concentration were measured in terms of electrical conductivity (EC).
From experimental results, the concentration of NaCl, in mg/L, was found to be four times the
value of the EC (mS/m). This means that every change in concentration of NaCI (in mg/L)
resulted in a linear increase in the EC (in mS/m). The amount of tracer solution introduced
should be high enough to be distinguished from the background EC value in the borehole. If
the concentration is too high however, it may create a density gradient and the vertical
movement of the solution may occur. A 100% increase of EC i.e. doubling the initial
concentration is usually sufficient, although the increase may vary from 50 to 200%
depending on the groundwater flow velocity. If the velocity is expected to be high, then the
influence of solution movement by density gradient would be insignificant and a higher
concentration of tracer solution could be applied.
The pump inlet is positioned at the bottom of the borehole while the injection pipe outlet is at
the top, just below the water level. A flow cell on the surface connects the abstraction and
injection pipes. The flow cell has an opening through which the tracer is injected and the EC
probe could be inserted for measuring inside the flow cell. Water is abstracted and flows
through the flow cell and injection pipe, and finally back into the borehole. The tracer solution
must be inserted uniformly in a single well volume. The insertion of the tracer is very
important. Improper insertion results in an error in the estimated velocity value.
91
The water level in the aquifer must remain static and steady-state flow must be maintained
throughout the test. If the experiment is continued while the water level is lower than the
static position, the dilution rate will be higher until the borehole recovers to its static position
and Darcy velocity will be overestimated.
After the release of the tracer, the circulation is continued for at least one well volume so that
the solution will mix completely. During the borehole-mixing process, activities that may result
in the increase of the rate at which the tracer moves out of the well, termed as dispersion by
Lamontagne et al. (2002), should be avoided. Tracer dispersion results in the dilution of the
tracer at higher rates at an early stage. This would occur because some of the tracer initially
displaced upstream by dispersion would later re-enter the well by advection. Turbulent flow
due to a high pumping rate could result in dispersion; therefore, the optimal circulation
velocity may be a compromise between greater mixing and less dispersion.
After the concentration/EC had stabilised, indicating complete mixing, the pump is switched
off and the reading of EC values proceed at specific depths within the borehole. The
measurement continued until the solution is diluted to about 20% of its initial value.
Before the evaluation of the velocity, the measured EC value was normalised as follows:
Where
C(t) = normalised EC
Cb = background EC value
Ci = measured EC value
Co = Initial EC at time to
The Darcy velocity for the different sections in the borehole can be evaluated from the
measured concentration values using the Drost and Neumaier (1974) equation as follows:
q = v--Inc(t)
aAt
Where:
q = Darcy velocity
V = volume of fluid contained in the test section
A = cross sectional area normal to the direction of flow
a = borehole distortion factor (assumed to be equal to 2)
t = time when EC value was Ci
92
From these values, the flow zones at depth could be identified, and the flux as well as the
conductivity for the flow zones could be calculated.
5.10 Slug tests
In a slug test, a small volume of water is suddenly removed from the borehole after which the
rise of the water level in the borehole is measured. Alternatively, a small slug of water is
poured into the borehole and the subsequent fall of the water level is measured (Kruseman
and De Ridder, 1994). This is very useful to determine the transmissivity of the upper soil
layers at the irrigation site where no water level occurs. If these values are compared to those
of the layers below where conductivity values were obtained by the point dilution method, the
conductivity values for the whole profile can be established.
In South Africa slug tests are conducted for the following two reasons:
o To estimate the hydraulic conductivity (K) of the aquifer in the vicinity of the borehole
(Van Tonder and Vermeulen, 2005) and
o To obtain a first estimate of the yield of a borehole (Vivier et al, 1995).
As we are only interested in the hydraulic conductivity (K) of the aquifer in the vicinity of the
borehole, only the applicable theory will be dicussed. Usually the Cooper method (Cooper et
al, 1967) or the Bouwer and Rice method (1976) is used to estimate the K-value (or T-value
in the case of the Cooper method).
In the FC program, developed by Van Tonder et ai, 2002, the Bouwer and Rice method
(1976) was applied to calculate these conductivity values. The method applies to unconfined
conditions and is based on the Thiem equation. The Bouwer and Rice equation reads:
Where:
re= radius of the unscreened part of the borehole where the head is rising
rw= horizontal distance from the borehole centre to the undisturbed aquifer
Re= Radial distance over which the difference in head hO is dissipated in the flow system of
the aquifer
d = length of the borehole screen or open section of the borehole
ho= head in the borehole at time=O
h,= head in the borehole at time t
93
The estimated K-value of Bouwer and Rice is dependent on the thickness open to flow d
which in this case refers to the different layers in the soil.
At all the irrigation sites a number of boreholes were augured to a depth just above the water
level. The boreholes were saturated with water to minimise the unsaturated flow conditions
discussed in Chapter 4 (water was poured down the augured holes for a number of times
over a period of two days). They were then filled with water, and the recession time was
measured, after which the hydraulic conductivity values were calculated for the different
zones.
5.11 Pumping Tests
Pumping tests are the most important experiments for aquifer investigation in the
groundwater industry (Van Tonder et aI., 2002). They are the only method that provides
simultaneous information on the hydraulic behaviour of the well (borehole), the reservoir and
the reservoir boundaries, which are essential for efficient aquifer and wellfield management.
The general objectives of a pumping test are:
o A better understanding of the aquifer system,
o Quantification of its hydraulic characteristics and properties
The interpretation of pumping test data is based on mathematical models that relate the
drawdown response to the discharge of the pumped well. The results obtained from this short
duration test are then used to estimate the borehole performance over a period of many
months (or even years). The mathematical model could be solved by the application of
analytical or numerical techniques (Van Tonder et aI., 2002). If the aim is setting up a 3D
numerical model, a number of piezometers must be installed (to measure pressure heads
and vertical K-values for each layer).
Every borehole in a fractured-rock aquifer reacts differently, and therefore it is of no use to
give one general recipe on how to conduct and analyse pumping test data. If the objective of
the pumping test is to estimate aquifer parameters that are to be used in e.g. a numerical
management model, the constant rate test is most important and is set as minimum
requirement for parameter estimation. One of the most important factors of a constant rate
test in this case is selecting the abstraction rate during the test. The yield must be chosen in
such a way that no main water-yielding structures will be dewatered during the test.
94
The FC (flow characteristic method) software was used (Van Tonder et ai, 2002) and
includes the following procedures to deal with the different types of flows and aquifers in
South Africa:
o Porous aquifer solutions (Theis, Cooper-Jacob I and II and Hantush methods and
also a solution for water-table aquifers).
o Step drawdown and multirate analyses.
o Fractal pumping test analysis (Barker's Generalised Radial Flow Model).
o Slug test analysis (Bouwer and Rice method).
o Different diagnostic plots for flow regime identification (e.g. derivatives, second
derivatives, log-log (Theis)-plot, lin-log (Cooper-Jacob)-plot, square root of time plot,
fourth root of time plot, spherical and recovery plot).
A number of pumping tests were performed at each pivot site to determine the transmissivity
values. These values could then be compared with the values of the other flow zones higher
up in the profile, in order to determine the flow paths of the water over the total depth of the
aquifer measured. Very low pumping rates were used in all the boreholes « 0.1 LIs). The
data were analysed with the FC program, and transmissivity values calculated.
In Table 5-1 a summary of all the experiments and the sites where they were conducted, are
portrayed.
95
Table 5-1: Summary of all the experiments.
Site Descri I>tion
Kleinkopje 1 (M ajor) and 4
Syferfont ein New
Vaal Borehole monitoring
Kleinkopje 1 (M ajor) and 4
Syferfont ein New
Vaal Pits and soil sampling
Kleinkopje 1 (M ajor) and 4
Syferfont ein XRD
Kleinkopje 1 (M ajor) and 4
Syferfont ein
OQt:imum XRF
Kleinkopje 1 (M ajor) and 4
Syferfont ein
Optimum Porous cups
Kleinkopje 1 (M ajor) and 4
Syferfontein
Optimum Tensi om ete rs
Kleinkopje 1 (M ajor) and 4
Syferfontein
Optimum Point Dilution Tests
Kleinkopje 1 (M ajor) and 4
Syferfontein
Optimum Falling head tests
Kleinkopje 1 (M aior) and 4
Syferfont ein
Optimum Pum ping tests
Laboratory Leaching colum ns -spoils
Leaching colum ns -
Laboratory adsorption
Leaching colum ns - saft
Laboratory leaching
Laboratory Leachinq drum s
Optim um -lab Hum idity cells
96
6 DETAilED CHARACTERISATION OF VIRGIN SOil IRRIGATION
SITES
6.1 Introduction
• In this chapter, the groundwater quality is discussed at various sites, together with
related results and conclusions.
• This will be followed by a discussion of the soil properties, the soil moisture and
gypsum precipitation and adsoption in the profile.
• An assessment of the hydraulic behaviour and chemistry will be investigated,
including soil moisture, a salt balance, salts leaching through the profile, as well as flow
determination. The influence of clay on these properties will also be discussed.
6.2 Groundwater monitoring
Detailed groundwater monitoring has been ongoing at the irrigation sites since the inception
of the irrigation project by the University of Pretoria in 1999. (The complete data set of all the
monitoring data at the different sites is portrayed in Appendix D). As described in Chapter 3,
boreholes were drilled inside and outside the pivot areas for monitoring of the groundwater.
Figure 6-1: Monitoring borehole at Kleinkopje Pivot 4.
97
6.2.1 Kleinkopje Colliery
6.2.1.1 Groundwater quality analysis Pivot 1
The analysis of the sampling done during November 2005 is illustrated in
Table 6-1.
• According to these analyses, the water quality of all the boreholes sampled inside the
pivot area is acceptable water, except for elevated nitrate in PV1 BH2.
Table 6-1: Water quality of the groundwater at the Kleinkopje Pivot 1 (November 2005).
SiteName pH EC Ca Mg Na K PAlk MAlk
PV1BH2 5.69 73.5 51.5 36.2 30.4 9.2 0 13.6
PV1BH3 6.45 3.6 0.6 0.2 7.8 2.0 0 15.4
PV1BH6 6.43 5.36 1.5 0.8 7.4 2.1 0 18
PV1BH7 6.42 6.25 3.6 1.5 7.9 6.5 0 17.5
SiteName F Cl N02(N) Br N03(N) P04 S04 Br
PV1BH2 <0.01 41.1 0 0.1 <0.01 197.0 <0.04
PV1BH3 <0.01 1 0 ~"'l!! <0.01 1.2 <0.04
PV1BH6 <0.01 1.5 0 0.02 1.52 <0.01 0.9 <0.04
PV1BH7 <0.01 4.8 0 0.06 3.07 <0.01 1.8 <0.04
Electrical conductivity
EC[mS/m] SANS 241 :2005 (', " 150, 370)
120
100
......P.V1BH1
__ PV1BH2
80 r_ PV1BH3__ PV1BH3S
__ PV18H4
60
_=PV~1BH7~J ::
__ PV1BHB
__ PV1BH9
40
20
0 ,
1997 19'99 2005
Time
Figure 6-2: Time graph of the electrical conductivity at Kleinkopje Pivot 1.
98
• The overall water quality trend of all the boreholes is sideways (Figure 6-2), indicating
that over the period of irrigation, water quality has not deteriorated within the
underlying aquifer. The only exception is PV1BH2, where a rise in salinity is evident.
l
__ P111BH1
Sulphate as S04 [rrg/L] __ P111BH2
500
__ P111BH3
400 _ P111BH4
300 ....... P111BH5
__ P111BH5
200
-- P111BH7
100 __ P111BH8
__ P111BH9
calcium as ca [ ]
tOO
80
60
40
20
~997
Time
Figure 6-3: Time graph of the calcium, sulphate and nitrate at Kleinkopje Pivot1.
• The majority of the groundwater constituents showed no increase over time, as
illustrated in the time graph of calcium, sulphate and nitrate in Figure 6-3. This
suggests that these macro constituents do not reach the groundwater. The exception
is PV1 BH-2, which deteriorated drastically in water quality since 2003. As this
borehole is down gradient from the expected flow direction, the vast increase in
dissolved solids is cause for concern. The water is probably being transported from
elsewhere in the pivot area by the bedding-plane fracture situated at approximately
10m towards PV1 BH-2. This situation must be closely monitored to minimise regional
impacts.
• The steep increase in nitrate in PV1 BH2 since 2001 correlates with the
increase in dissolved solids. The values are currently an order higher than in
the other boreholes. As this borehole is down gradient from the expected flow
direction this vast increase in nitrate is cause for concern.
Conclusion: The salinity in the monitoring boreholes is surprisingly low for an irrigation site,
given that the TOS off the irrigation water is nearly 4000mg/L. This indicates that the salts are
attenuated in the soil, but because PV1BH2 shows an increase in salinity over time, the other
boreholes are probably constructed wrongly (see Section 6.6.5.3 for a discussion on this
issue).
99
6.2.1.2 Groundwater Quality analysis Pivot 4
The analysis of the sampling done during November 2004 is illustrated in Table 6-2.
According to the analysis, the water quality of all the boreholes sampled is ideal except for
PV4BH2, which has elevated nitrate content. Unfortunately this was the last time all the
boreholes could be sampled. Since then two of the boreholes were destroyed.
Table 6-2: Water quality of the groundwater at Pivot 4 (November 2004).
5iteName pH EC Ca Mg Na K PAlk MAlk
PV4BH1 5.16 13.7 5.8 5.0 4.3 7.14 0 2
PV4BH2 6.89 73.5 49.9 43.6 36.8 5.41 0 95
PV4BH3 5.2 6.6 1.8 1.4 4.8 2.04 0 1
PV4BH4 5.25 5.1 1.6 1.0 5.2 0.83 0 1
PV4BH6 5.11 9.16 3.5 2.4 7.7 2.10 0 3
5iteName Cl 504 N03-N F N02-N NH4 P04 Br
PV4BH1 4.2 0.4 13.75 <0.01 <0.01 <0.1 <0.1 <0.04
PV4BH2 50.3 240.0 0.04 0.03 <0.01 <0.1 <0.1 0.12
PV4BH3 4.9 4.0 2.46 <0.01 <0.01 <0.1 <0.1 <0.04
PV4BH4 5.3 4.1 2.66 <0.01 <0.01 <0.1 <0.1 <004
PV4BH6 4.1 0.4 9.28 <0.01 <0.01 <0.1 <0.1 <0.04
• The overall water quality trend of all the boreholes indicated relative stability,
indicating that water quality did not deteriorated within the underlying aquifer over the
initial period of irrigation (Figure 6-4). The increase in the EC in PV4BH2 during 2004
is the result of a broken-down pivot that caused water logging. The borehole is
downgradient and water infiltrates via the broken casing into the borehole. The EC
should return to the original values over time (there was already a slight improvement
during the next sampling run). Unfortunately this borehole has been destroyed, and
cannot currently be monitored.
• Since 2006 the electrical conductivity of boreholes inside the irrigation area indicated
a relatively sharp increase.
Conclusion: The exceptionally high rainfall during the rainfall season may be the cause of
leaching of salts from the soil profile. Only continued monitoring at this site will indicate if this
trend will continue.
100
Electrical conductivity
EC [mS/m) SANS 241 200S (',' 1S0, 370)
120
100
80 1_ PV46H1 '
-r Pv'4BH2
..,..._ PV4BH3
_ Pv'4BH4
60 __ PV4Bt-fi j
....... P\I4BH5
40
20
0
1999 2001
Time
Figure 6-4' Time graph of the EG at Pivot 4 boreholes.
6.2.2 New Vaal Colliery
6.2.2.1 Groundwater Quality Analysis (see Appendix D for Monitoring data).
According to the analysis of the sampling done, the water quality of all the boreholes
sampled, except PV2, indicates very little pollution. PV2 contains completely different water.
The reason for this is that the water originates from the dam overflowing. Unfortunately this
borehole became blocked and could not be sampled since 2005. Regarding the 2005
sampling (Figure 6-3), it is clear that the water in the shallow piesometers. installed above the
clay layer, is highly polluted. Evapotranspiration resulted in an increase in sailinity, and can
clearly be seen in the salinity of the water in the piesometers, which is much higher than that
of the irrigation dam.
Table 6-3: Quality of the groundwater at the New Vaal Pivot in mg/L (March 2005).
"I 'I -, I (,I / ) ) 101
Electrical conductivity
ECI! SANS 241'2005 (-, -, 150, 370)
400
300
~l
~ P\I3 I
- PV4
200 ~PV5 ;1
I~~piezo
~dpiezo
100
o
7rNJ1 7= 7f2003 7rNJ4
Time
Agure 6-5' Time graph of the Electrical Conductivity at New Vaal Pivot.
This indicates that, over time, the water quality did not deteriorated significantly within the
underlying aquifer because of irrigation. The reason may be dilution due to the high lateral
flow of water through the sandy layers above the clay layer and high natural recharge of the
sands. The clay layer previously discussed also limits downward migration and therefore only
gradual contamination of the aquifer occurs from slow leakage through this layer.
It is insightful to focus on Borehole PV5 in the pivot area as far as trends are concerned
(Figure 6-6).
Electrical conductivity
EC (mS/m! SA dnnkmg water- humans (Class 0 70 Class L 150 Class II 370)
70
60
50
40
30
20
10
0
7aaD1 712002 712003 712004 712005
Time __j
Figure 6-6: PV5 electrical conductivity trend since monitoring began.
102
Although the water quality in this borehole is acceptable, the results from the analysis show
that there is a steady and consistent rise in the electrical conductivity in the borehole. This
indicates that the irrigation water is slowly leaching through the clay layers and starting to
influence the groundwater quality directly below the pivot. The trend in this borehole is cause
for some concern and future phases will need to investigate this further if irrigation is to
continue. At the current increase rate, irrigation will start to cause marked deterioration in
values beyond the trigger values for this site within a few years.
6.2.3 Syferfontein Colliery
6.2.3.1 Groundwater Quality analysis
The analysis of the sampling done during June 2006 (the last time sampling was possible) is
illustrated in Table 6-4. According to this analysis, the water quality of all the boreholes
sampled excluding OBH6S is acceptable, except for Nitrate. OBH6S is a shallow borehole
and shows the effect of the irrigation.
Table 6-4: Water quality of the groundwater at the Syferfontein Pivot.
5iteName pH EC Ca Mg Na K MAlk Cl 504 N03-N Br
m5/m mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l
OBH1 7.55 99 86 49 55 1.7 414 102 4 1.8 0.3
OBH2 7.81 95 76 46 66 2.0 415 41 37 11.4 0.6
OBH3 7.96 88 83 64 35 1.9 327 17 175 3.8 0.1
OBH6D 7.60 99 89 48 69 0.8 315 13 177 . J!Od 0.3
OBH65 7.80 189 198 'JU >Jm6 1.7 481 28 1715 7.4 0.2
OBH7 7.46 87 105 62 158 5.9 383 24 432 4.6 1.6
OBH8 7.61 90 70 38 77 1.5 344 24 65 10.8 0.7
The electrical conductivity of all the boreholes is illustrated in Figure 6-7. The deeper new
borehole (OBH6D) inside the irrigation area shows no variation or increase in electrical
conductivity from the other boreholes outside.
103
Electrical conductivity
EC [J SANS 241 :2005 (-, -, 150, 370)
400
300 ~
--OBH1
... OBH2
... OBH3
..-OBH4
OBH5
--OBH6D
200 ... OBH6S
... OBH7
-+- OBH8
... REGM86
~
100 l
o 2001 2002 2003 2004 2005 2006
Time
Figure 6-7: Time graph of the Electrical Conductivity at Syferfontein Pivot.
OBH 1 has shown a vast improvement since the construction of run-off trenches in 2003 and
was similar in water quality to the other boreholes when monitoring ceased.The overall water
quality trend of all the boreholes is sideways, indicating that over the period of irrigation water
quality has not deteriorated within the deeper aquifer. However, the shallow aquifer has
deteriorated.
6.2.4 Overall conclusions from the groundwater monitoring at the pivots
• The water levels at all the pivots have varied very little over time. The water levels at
New Vaal are deeper than most other irrigation schemes. This can be attributed to the
high conductivity in the soils above the clay layers, resulting in faster drainage
towards the river. The water levels at the other irrigation sites are very shallow due to
the increased local recharge.
• The overall water quality trend in the deeper aquifer indicates no significant water
quality deterioration over the monitoring period. The exception is Borehole PV5 at
New Vaal within the pivot area itself which shows signs of consistent water quality
degradation, with increases in most of the constituents.
• The irrigation water has, as yet, shown a minimal impact on the groundwater. Some
exceptions occur, but none of the boreholes outside the pivot areas show any
meaningful change in water quality from leaching from the irrigated area. This implies
104
a very slow movement of the constituents in the irrigation water, and that these are
attenuated by different mechanisms between the surface and the aquifers, often in the
clay layers. The amount leaching through the soil is small enough to be easily diluted
by the regional groundwater flow. It thus appears that the soil type and morphology
plays a very important role in the vertical movement of the irrigation water
constituents.
o It is unknown how long this accumulation of salts in the upper layers can continue
before leaching into the underlying aquifers, but in the short to medium term, the
evidence from groundwater monitoring shows that irrigation with mine water does not
hold significant threats for the regional groundwater quality.
The irrigation water has, as yet, shown only a minimal impact on the groundwater. It is
important to consider reasons that the irrigation water is not represented in the groundwater.
The hydraulic and attenuation factors preventing the salts in the mine water used for irrigation
from being mobilised down the soil profile and into the aquifer are important considerations in
this process. The soil composition and associated sorption and hydraulic properties may be
informative. The amount leaching through from the pivots is small enough to be easily diluted
by the regional groundwater flow. The larger concentrations of nitrate are due the agricultural
activity at the pivot. Of concern is the fact that the relatively mobile constituents (e.g. nitrate)
are able to move through the soil profile into the groundwater and results in quality
degradation.
6.3 Soil analysis
To determine the soil composition and texture, which is important in understanding the
downward flux of the irrigation water, trenches were dug and core drilling was done in order
to sample the profiles at the different sites with depth. The soil was also analysed to
determine the salt balance in the profiles with depth.
6.3.1 Kleinkopje
To determine the soil composition and texture, two trenches were dug for inspection and soil
property analysed. The soil at Pivot 4 is fairly homogeneous (Hutton soil), whilst that at Pivot
1 varies with depth (Bainsvlei and Clovelly soils). The distinctive geological layers were
selected, sampled and soil properties were determined in terms of relative clay, sand, silt etc.
content. The results are shown in Table 6-5 and indicate clayey layers in the soils.
105
Table 6-5' Soil properties at Kleinkopje Pivots (%).
Pivot Name Depth (m) Sand Coarse silt Fine silt Clay
1 0-0.4 79.19 5.18 9.34 4.67
1 0.4-0.8 78.39 9.74 4.62 4.62
1 0.8-1.0 73.74 10.30 13.93 0.00
1 1.0-2.5 61.11 16.46 9.29 18.58
1 2.5-2.8 66.33 15.37 9.24 9.24
4 0-0.6 72.50 9.20 4.60 13.79
4 0.6-1.0 64.32 11.75 4.62 18.48
4 1.0-2.0 56.06 10.80 13.93 18.58
4 2.0-3.0 50.50 15.29 4.60 27.59
4 3-3.2 53.27 19.49 4.62 23.10
4 3.2-3.5 59.60 9.79 4.64 23.22
Three core boreholes were drilled at each of the pivots, two inside the pivot areas and one
outside for background values. The core samples were analysed for the major cations and
anions at the IGS laboratory. Figure 6-8 and Figure 6-9 shows the leaching of water-soluble
constituents for two cores inside the pivot areas.
~_ ..
Borehole Log - KK1-inside
Depth (mj Locality. X: 5.00 Y: -5.00 Z: 0.00 I
I
---------------
Figure 6-8: Soil leaching analyses with depth within Pivot 1.
106
Borehole Log - KK4-inside
Oepth[m) Locality. X: 7.00 Y: ·7.00 Z: 0.00
,3..5yO.0.5OSon. Hl.(IonllOd,2
--------- 1
Figure 6-9: SOIl leaching analyses with depth within Pivot 4.
From these figures it is clear that there is currently a build-up of salts in the soils above and
within the clayey layers, and that a limited proportion of the associated salts move through
into the groundwater.
6.3.2 New Vaal Colliery
Different activities were undertaken to determine the reason for the relatively deep water
levels:
1. Two trenches were dug inside the pivot area, and inspected. Water seepage is very
shallow, and occurred approximately 2.4m below surface.
The distinctive geological layers were selected, sampled and soil properties were determined
in terms of relative clay, sand silt etc content at Glen Agricultural College. Table 6-6 confirms
the clay layer below 2.4 m, which is also a reason for the water logging at this pivot.
107
Table 6-6: Soil properties at New Vaal (%).
Depth (m) Sand Coarse silt Fine silt Clay
0-0.6 94.50 4.60 1.01 3.59
0.6-0.75 94.95 4.64 0.70 3.95
0.75-0.85 96.02 4.57 1.05 3.52
0.85-1.0 91.46 6.13 0.65 3.97
1.0-2.1 93.47 4.62 1.39 3.23
2.1-2.2 96.48 4.62 0.92 3.70
2.2-2.4 89.34 7.21 1.40 3.27
2.4-4.0 42.27 15.25 18.96 23.70
8.0-9.0 57.14 15.10 14.07 14.07
2. Three auger boreholes were drilled at the pivot, one inside the pivot area and one
outside for background values. Another one was drilled between the irrigation dam
and the pivot to determine the leaching properties from the irrigation dam towards the
pivot, and to determine the influence of the dam on the water logging problems. The
auger samples were analysed for the major cations and anions at the IGS laboratory.
The figures below show the leaching of water-soluble constituents for the auger hole
inside the pivot, with a high concentration of constituents in the clay layer.
Borehole Log - NV-inside
Depth imi Locality - X: 9.00 Y: -9.00 1:. 0.00
I
3.9)·100SOHO cs.,.111-
I- ...;.
,
,
&,OJ-iIXlMNO CI.~1iI\, 14'1tocl,y,
- 1,",,~r'I"'1I.1;/'IIo1Ul'Od..
900-lo.OOClAV $ardy~
10 -
-
Figure 6-10: Soil leaching analyses with depth inside the pivot area.
108
Borehole Log - NV-dam influence
Qepth(mJ locality.Xl1.00 Y:·l1,OO Z:O.OO
K(mo'klll 20
L- _j
Figure 6-11: Soil leaching analyses with depth showing the dam s influence.
Borehole Log - NV-outside
Depth(m) Loca!ity. X: 10,00 Y: -10.00 Z: 0.00
EC(mSlm]
50
~ J
----------------------------j--------
r-------_I-
Figure 6-12: SoH leaching analyses with depth outside the pivot area.
These graphs (Figure 6-10 to Figure 6-12) show that there is a significant clay layer below
2.4 m in the pivot area (as also depicted in Table 6-6) and slightly deeper towards the dam.
This layer impedes downward migration of water and associated salts, which is associated
with the observed ponding in the pivot area.
109
3. Two piezometers were installed close to PV5 to study the influence of the clay layer
on the irrigation waters.
Table 6-7' Chemical analysis of water in the piezometers.
pH EC Ca Mg Na K PAlk MAlk
m5/m mg/L mg/L mg/L mg/L mg/L mg/L
8.33 232 118.55 33.54 407.63 20.59 4.00 140
F Cl N02(N) Br N03(N) P04 504 B
mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
0.8 126 0.09 0.44 7.96 0 968 1.321
A chemical profile of the EC down the hole in PV5 (near the centre of the pivot), and the
piezometer hole adjacent to PV5, also shows vast differences in water quality (Table 6-8).
The very high EC values in the piezometer hole compared to the boreholes also indicate that
water does not migrate easily downwards through the clay layer at 2.4 m.
Table 6-8: Down-the-hole profiling of EG (mS/m).
Depth PV5 Piezo
1.47 148.7
2 242.2
12.78 17.3
13.5 17.9
14 37.3
15 39.7
20 37.9
25 37.9
28.5 35.9
Conclusion: There is a significant clay layer below 2.4 m in the pivot area and slightly deeper
towards the dam. This layer impedes downward migration of water and associated salts,
which give rise to the deeper water levels that are uncommonly deep for an irrigation site.
The installed shallow piezometers support these findings since the water quality measured in
these indicated elevated concentrations consistent with evaporated mine water that could not
migrate downwards (Table 6-7). The Durov diagram in Figure 6-13 illustrates these
differences in water quality between the monitoring boreholes (except PV2) and the
piesometers very clearly.
110
Durov Diagram
/,... \~
Ca
T.PJ~
•..l.A.• "".......804 ./-10 .....".-,. .... - ~• PV2• PV3• • PV4 J• PV5
• Short piezo
~olidpiezo
Cl - ..... -- - ..-EC [m8fm] (10-1000)
.."r.J
• l. • pH (6-9)
o • •
-
Figure 6-13: Durov diagram of the monitoring holes and piesometers at New Vaal.
6.3.3 Syferfontein Colliery
In order to determine why the aquifer below and adjacent to the pivot area is not polluted, soil
tests were performed.
• Core drilling was carried out at various sites inside and outside the irrigation area.
From the core the distinctive geological layers were selected and soil properties were
determined in terms of relative clay, sand silt content at Glen Agricultural College. The
results are shown in Table 6-9.
Table 6-9: Soil characteristics at Syferfontein Pivot.
Depth (m) Sand Coarse silt Fine silt Clay
0-0.8 28.04 8.99 0.10 63.49
0.8-1.44 na na na na
1.44-2.44 44.19 16.27 0.40 40.70
2.44-3.5 62.11 8.42 0.40 31.58
3.5-4.65 67.71 10.02 0.30 26.04
4.65-5.46 82.50 10.00 0.40 10.00
The soils are of the Arcadia form with a high clay content. They vary in thickness from 300
mm to 900 mm. Two layers are visible. The upper layer of 200 mm has a strong, fine,
angular blocky structure. The horizon 200+ mm has a strong, coarse, angular blocky
structure. It lies on weathered dolerite saprolite. The colour of the soil is black. It is very high
in clay. Slickensides occur deeper than 400 mm. A few distinct mottles occur at the bottom
111
of the soil layer and above the saprolite. Some soft lime concretions and planar
accumulations of calcite occur deep in the saprolite (>1500 mm).
The morphology is typical of well-developed Arcadias. It is high in swelling smectite clay
(See Section6.4.3.1). Smectite clay is physically very active. It swells when wet and shrinks
when dry. After rain it is swollen and only a few isolated macropores may exist. Water
movement is restricted to micro pore flow. Water infiltration in Arcadia soils can be lower
than 10 mm/h (Ie Roux, 2004 during site inspection). In these Arcadias the infiltration rate
and saturated hydraulic conductivity in the soil layer may be 1 mm/ho This is 175 mm in a
week. All additional rain will accumulate in pools and run off.
The underlying saprolite is coarser and micro pore flow is restricted. This results in the
accumulation of water in the bottom of the soil layer and the formation of mottles. A typical
Highveld thunderstorm may react slightly differently. After the dry winter, the soil is dry and
cracks form. The cracks extend as deep as the saprolite. During the first downpour the
water penetrates the cracks and drains through the saprolite quite quickly. It moistens the
underlying weathering rocks and initiates a new incidence of chemical weathering. This
process results in calcite formation. The limited amount of calcite is an indication that
weathering is slow or leaching is high. The low saturated hydraulic conductivity of the soil
restricts drainage and therefore the later alternative is unlikely.
o Leaching tests were also performed with deionised water through the cores drilled
inside and outside the pivot area. The leaching analyses on these samples can be
graphically represented (Figure 6-14 and Figure 6-15) to show the layers where
enrichment in soluble salts occurs. From these it is clear that the shallow layer under
the irrigated area has values that are far higher than those in the lower layers and
outside the pivot. The clays in the shallow zones appear to prevent migration and
attenuate the salts strongly in the upper profile.
112
Borehole Log - Syferfontein1
Depth Iml Lccellty- X: 1.00 Y: ·1.00 z: 0.00
Figure 6-14: Soil leaching analyses within the pivot area.
Borehole Log - Syferfontein3
Depth [ml localrty-X 300 Y -300 Z 000
SOoI [mgJkQ1 N03(N) lm"1 N.I[IfIO*gJ ieo e K[mgllq;j1
lOS [mg.4lg]
300 1800
0.0 r ,roy .. 0 . l
0.5
10
15
20
2.5
3.0
3.5
4.0
4.5
50
Figure 6-15' Soli leaching analyses outside the pivot area.
113
6.3.4 C:onclusions
From the soil analysis it is clear that most of the salts are captured in the upper one or two
meters of the soil profile. The salts move along the profile in the soil water. The data indicates
that the clay layers, which playa major role in the vertical flux of the water, therefore also
have an influence on the salt distribution through the soil profiles. In order to determine what
the salinity of the soil water through the different profiles is, soilwater sampling was
performed at the different sites.
6.4 Unsaturated soil sampling
6.4.1 Introduction
If gypsum accumulates in the upper meter of the irrigated soils, as predicted by Annandale et
al. (2002), there would still be downward movement of Mg and 804, given that Mg will be
more exchangeable than Ca in the upper part of the soil profile and that the concentration of
804 in the irrigation water is twice that of Ca. Furthermore, the solution that percolates from
the base of a gypsiferous zone is saturated with respect to gypsum (under equilibrium
conditions), and therefor soluble Ca must also be present in leachate from the zone
(Campbell, 2001). Even if irrigation is managed so that leaching is negligible during the dry
season, it is apparent from the data collected that significant movement of the solutes added
by irrigation would take place during the rainy season. This means that in the long-term
almost all of the Mg, at least half of the 804, and some of the Ca added by irrigation must
enter the groundwater system. If sustainable gypsum precipitation does not take place, then
close to all the added Mg, 804 and Ca will move to the local aquifer.
Porous cups were installed at the depths of 4, 3, 2 and 1m respectively (Figure 6-17 and
Figure 6-17) at the different irrigation sites. This is to determine the movement of the water
and the dissolved solids in the unsaturated zone, and will thus give a good indication of the
ion movement, adsorption and ion exchange of the different ions in the soil. From the
saturation index values in Table 6-10 it is clear that the water is saturated with respect to
gypsum in the first meter of the soil at both sites, indicating the precipitation of gypsum.
Deeper down the soil water is undersaturated with respect to gypsum and should not
precipitate.
114
6.4.2 Kleinkopje
Four porous cups and tensiometers were installed at the depths of 4, 3, 2 and 1m
respectively at Kleinkopje Pivot Major and Pivot 4. Seven sampling runs were done, of which
the data (November 2004) are included in Table 6-10.
Figure 6-16: The porous cup and tensiometer installation (left) and porous cup sampler plus
vacuum pump (right).
Figure 6-17: Porous cup installation at Pivot 4.
According to the soil water data in Table 6-10 the water is saturated with gypsum in the first
meter at both sites, indicating gypsum precipitation.
115
Table 6-10: Analysis from porous cup sampling at Kleinkopje.
Pivot No Depth (m pH EC Ca Mg Na K MAlk F Cl Br N03 N S04 SI Gypsum
mS/m mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
1 1 6.63 329 488 328 74 25 150 0.16 53 144.5 32.5 2160 -0.11
1 2 7.34 201 353 106 57 11 37 0.18 42 0.2 27.2 1217 -0.3
1 3 8.17 185 372 66 58 11 39 0.17 23 0.2 27.2 1108 -0.28
1 4 7.16 80 78 54 33 5 100 0.30 25 0.2 6.1 309 -1.22
4 1 6.97 1411 4029 1111 74 0.17 130 1.3 62.6 8732 0.42
4 2 8.25 138 91 128 38 5 245 0.08 110 4.5 466 -1.13
4 3 6.99 129 178 62 49 3 70 0.08 100 0.3 42.2 422 -0.84
4 4 7.07 166 89 58 52 9 97 0.21 33 0.3 12.7 350 -1.14
6.4.2.1 Kleinkopje Pivot 1 (Major)
The data from the porous cups monitoring provides some insight into the reasons for the
apparent lack of salinity increase in the aquifer (Figure 6-18). The monitoring data show a
steady decrease in sulphate concentrations with depth. This suggests that the majority of the
salts are currently contained within the uppermost portions of the soil profile.
Kleinkopje Pivot 1(Porous Cups)
Sulphate (mg/L)
0
-+-Jan-05
1 _Nov-04
:[2 -+-Feb-05
..sg..
:.: ~May-05- 3 __'_Aug-05
c -+-Oct-05
4 -I-Nov-05
5
0 500 1000 1500 2000 2500 3000
Figure 6-18: Sulphate concentrations with depth in the porous cups
If the relationship between cation qualities in the soil water for two meters depth and three
meters are compared, it is clear that calcium decreases much more than the other cations,
indicating gypsum precipitation. In Figure 6-19 the equivalent concentrations of the ions are
shown.
116
Kleinkopje 1 Jan 2005 -Calcium
-Magnesium
r -SodiumConcentration(eqw) -Chloride
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 -Sulphate
0.5 1--------------'------------1
1.5
_ 2
.5.
-= 2.5
Co
2.l 3
3.5
4
Figure 6-19: Equivalence plot of various ions at Pivot Major.
To verify this observation, leaching tests were done on representative samples obtained with
depth in the soil profiles. In these tests, an excess of KCI was added to each soil sample and
the liberated ions in the supernatant analysed (Figure 6-20 and Figure 6-21). The results of
the soil tests suggest that the porous cup results are consistent with the trapped pore water,
precipitates and adsorped ions at each of these levels.
Kleinkopje Pivot 1Soil
Sulphate (mg/kg)
o 1000 2000 3000 4000 5000 6000
o ~---L---------------------~~------------_.
0.5 -1--- r...6----___::=====---=----=-~ .~..-..--.1
Ê 1.5 /
-a 2~ 2.351~l !-
3.5 1/
4 ~-- --
4.5 --'---------------------------'
Figure 6-20: Sulphate concentrations liberated from the soil with depth.
117
Molar distribution of KCI soluble cations on soils- Kleinkopje Pivot Major
10.0000 ,-------------------------,
1.0000 -~ ~
o. 00 • 0.5~~<_-~1,~,__00_-_-_- __ 50_0 2=.00=0 ---=~2.500 3. 00-----=: -+-ca]
0.1000 -
Jl K I=~
0.0100
0.0010 L-- _.J
Depth
Figure 6-21: Mo/ar distribution of cations in the soli at Pivot Major.
Figure 6-21 indicates that calcium decreases with depth, while magnesium and sodium are
probably affected by the clay layer just below one meter. Potassium also decreases with
depth.
One of the implications is that there are hydraulic and attenuation factors preventing the salts
in the mine water used for irrigation from being mobilised down the soil profile and into the
aquifer. The soil composition and associated sorption and hydraulic properties may be
informative. The typical composition determined by standard soil composition tests is
provided in Table 6-5.
This shows a marked increase in clay content below the depth of 1m.
6.4.2.2 Kleinkopje Pivot 4
At this pivot the porous cups were not as successful as at Pivot Major, and less water was
retrieved. However, the same trend can generally be seen as at Pivot Major (Figure 6-22).
118
Kleinkopje Pivot 4 (PorousCups)
Sulphate (mgil)
0 2000 4000 6000 8000 10000
0 ,
- 1 • -+-Nov-04.s2 _Jan-OS______Feb-OS
..s.:.:..g. 3 ~lIt1ay-05
c ---'JIE- Aug-OS
4 ___ Nov-OS
5
Figure 6-22: Sulphate concentrations with depth in the porous cups at Pivot 4.
Molar distribution of KCI soluble cations In soils - Kleinkopje Pivot 4
10.0000,----------------------
10000
o OSO' 100 1.SO 2.00 250 300 350
'. -+-K
01000 -.' Mg
........ Na
0.0100
0.001'-0--------- -----'
Depth
Figure 6-23' Molar distribution of cations in the soil at Pivot 4.
The soil leaching tests (Figure 6-23) indicate that the magnesium initially increases, and then
decreases. This may be the result of sorption at the clay layer at 3m depth. Sodium and
potassium level out at 3m, but do not decrease in value.
The results to date point to several potentially important findings for the wider application of
mine water irrigation. It would appear that where the soils are richer in clay content, there is a
significant attenuation of salts in the shallower zones. The groundwater monitoring results
indicate that this attenuation makes mine water irrigation a viable option in the short to
medium term where gypsum-rich waters are used. The fact that the data collected from the
porous cups, the soil salinities and the groundwater quality indicates the same trends is
119
encouraging. Analysis of the tensiometer data over time, continued groundwater and soil
water monitoring and detailed analysis of the soil characteristics in terms of hydraulic and
mass transport properties at each site, should allow the development of accurate conceptual
models of the interaction between irrigation and the underlying soils and aquifers.
6.4.2.3 XRD/XRF Analysis
Sixteen samples were sent to the Geology Department of the University of the Free State for
analysis (eight from each pivot), especially for gypsum and clay determination, with depth in
the soils.
Table 6-11: XRF analysis of the Kleinkopje soils.
NO. Depth Si02 Ti02 AI203 Fe202 MnO MgO CaO Na20 K20 P205 H20- LOl TOT
I
P1-1 0-0.2 84.56 0.49 4.86 2.71 0.04 0.13 0.53 0.12 0.33 0.06 0.9 5.05 99.78
P1-2 0.2-0.4 85.07 0.42 4.63 2.54 0.03 0.18 0.41 0.12 0.3 0.04 0.84 4.58 99.16
P1-3 0.4-0.8 86.13 0.45 4.71 2.59 0.03 0.12 0.15 0.12 0.3 0.04 0.67 3.79 99.1
P1-4 0.8-1.0 85.69 0.53 5.62 3.48 0.02 0.11 0.08 0.13 0.33 0.04 0.6 3.1 99.73
P1-5 1.0-1.5 73.27 0.59 8.82 11.09 0.1 0.16 0.07 0.12 0.49 0.06 1.03 4.53 100.3
P1-6 1.5-2.0 74.02 0.73 10.36 8.14 0.05 0.22 0.07 0.12 0.66 0.05 1.23 4.42 100.1
P1-7 2.0-2.5 76.81 0.77 10.24 5.51 0.04 0.21 0.08 0.13 0.82 0.03 1.01 3.98 99.63
P1-8 2.5-2.8 78.91 0.58 8.11 5.73 0.2 0.18 0.07 0.15 1.18 0.03 0.95 3.12 99.21
P4-1 0-0.3 86.11 0.49 5.4 2.72 0.02 0.12 0.17 0.13 0.16 0.05 0.52 3.14 99.03
P4-2 0.3-0.6 83.69 0.6 6.83 3.55 0.03 0.14 0.08 0.12 0.2 0.04 0.79 3.17 99.24
P4-3 0.6-1.0 82.15 0.67 7.68 3.84 0.03 0.18 0.08 0.12 0.23 0.05 0.85 3.43 99.31
P4-4 1.0-1.25 79.78 0.78 9.13 4.46 0.02 0.22 0.08 0.13 0.24 0.05 0.97 3.87 99.73
P4-5 1.25-1.5 80.06 0.77 9.07 4.76 0.04 0.2 0.07 0.12 0.24 0.05 0.98 3.91 100.3
P4-6 1.5-1.75 76.63 0.84 10.66 4.89 0.04 0.2 0.09 0.13 0.27 0.04 0.9 4.35 99.04
P4-8 2.0-2.5 76.37 0.84 10.89 5.4 0.04 0.09 0.06 0.11 0.3 0.05 0.77 4.49 99.41
I
It is evident from Table 6-11 that there are indications that gypsum precipitation should occur
since Ca values are far higher in the shallow portions, while Mg values are more evenly
distributed. This suggests that the distribution is chemically controlled rather than by surface
interactions since these two ions should have similar sorptive capacities. This is supported by
the Na values, which are evenly distributed with depth.
The XRD analysis in Table 6-12 the clay to be kaolinite, which theoretically, should adsorp
sulphate (Mott, 1981).
120
Table 6-12: XRD analysis of Kleinkopje soils.
NO. Quartz Orthoclase Hematite Plagioclase Kaolinite Gypsum Montmorillinite
P1-1 XX <x x <x
P1-2 XX <x x <x
P1-3 XX <x x <x
P1-4 XX <x x <x
P1-5 XX x x <x
P1-6 XX x x <x
P1-7 XX <x x x <x
P1-8 XX <x x x <x
P4-1 XX
P4-2 XX <x
P4-3 XX <x
P4-4 XX x
P4-5 XX x
P4-6 XX x
P4-8 XX x
DominantXX >40%
Major X 10 - 40%Accessory <x 1 - 2%
Minor x 2 - 10% Rare «x <1%
6.4.3 Syferfontein
Four porous cups and tensiometers were installed at the depths of 4.8, 3.5, 2 and 1m
respectively. This is to determine the movement of the water and the dissolved solids in the
unsaturated zone. Seven sampling runs were done. The results with depth are illustrated in
Figure 6-24. Figure 6-24 illustrates the soil leaching analysis of sulphate with depth for the
different positions within and outside the pivot area.
121
Syferfontein Pivot(porous Cups)
Sulphate (mgil)
0 500 1000 1500 2000
o -
1 -+-Jan-05
- _Nov-04- 2-E ----.- Feb-05.cc.. 3 ---?f- May-05Q)
c ~Au9-054
--+-Oct-05
5 ---
~-+-Nov-05
6
Figure 6-24: Sulphate concentrations with depth in the porous cups
If the relationship between cation qualities in the soil water for depths of 2 meters and 3.5
meters are compared, as indicated in Table 6-13 below, it is clear that there is sorption and
ion exchange between magnesium and potassium because the potassium has increased in
depth while magnesium decreased. The reason why sodium does not exchange is because
the sodium content of the irrigation water is already very high. In Figure 6-25 the equivalent
concentrations of the major ions are shown for the February 2005 sampling. The sodium
sulphate water results in high sodium and sulphate content in the upper layers of soil, which
decreases with depth. The conservative chloride anion moves through the different layers to
the same degree, without any precipitation or adsorption.
Table 6-13: Proportional relationship between water qualities at 2 and 3.5 m depths.
Ca Mg Na K
3.40 25.38 3.91 0.20
122
-Calcium
Syferfontein Feb 2005
-MagneSiUm
-Sodium
Concentration (eqw) -Chloride
0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 [-Sulphate
0.5 r-----'------'------'--------'--------------j
1.0
1.5 l-f-~----',
2.0 1-1---1--
:g 2.5
ii 3.0 I-Y'~--__= ---_===_"""""-
Q.
~ 3.5 M=--
4.0 ""--~--~
4.5
5.01----
5.5 '------------------------'
Figure 6-25: Equivalence plot of various ions at Syferfontein.
Syferfontein Pivot Soil
Sulphate (mg/kg)
0 2000 4000 6000 8000 10000
0.0
1.0 x
~
...... 2.0 r-
3.0 ,_ -+-BH1-inside.§.
:; 4.0 --- BH2-inside
~c. 5.0 --.- BH3-outside
6.0 __ ~BH4-outside-
7.0
8.0
Figure 6-26: Soil leaching analysis of sulphate in depth for different positions.
To verify this observation, leaching tests were done on representative samples obtained with
depth in the soil profiles. In these tests, an excess of KCI was added to each soil sample and
the liberated ions in the supernatant analysed. Analysis of the soil indicates, as is the case
with the Kleinkopje pivots, high sulphate concentrations in the upper one meter of soil.
123
Syferfontein Pivot
Sulphate (mgIL)
~r-~10 ==10~0 ~~1000~~10~000~r=.l2 - - Porous cups-NO\.Q004
-Ê 3- -Soil BH1-inside-5 4 -.- Soil BH2-inside
og. 5 ----*"- Porous cup-Jan2005
6 +-I'_'_---
~__._ Porous cup-Feb2005
7
8
Figure 6-27 Comparison of soil water and leaching quality.
From the comparisons in Figure 6-27, it is clear that the clays play an important role in
retaining the salts at the upper few layers of the profile. A comparison between the soil
leaching and porous cup data supports this hypothesis.
Molar distribution of KCI soluble cations in soils - Syferfontein
10.0000 ,----------------------,
1.0000 --
c
0 6
;:l
~c
Q)
uc 0.1000 ____ Mg
0
u....
<II -. --*-Na"0
:!
0.0100 <,
0.0010 _L...__ -'
Depth
Figure 6-28: Molar distribution of cations in the soil at Syferfontein.
From the molar distribution of the cations (Figure 6-28) it can be concluded that all the
cations move through the clayey Arcadia soils at the same degree, with sodium obviously at
higher molar concentrations because of the composition of the irrigation water. There is much
less magnesium in the water than calcium or potassium.
124
6.4.3.1 XRD/XRF Analysis
Eight samples were sent to the Geology Department of the University of the Free State for
analysis, especially for gypsum and clay determination, with depth in the soils.
Table 6-14: XRF analyses at Syferfontein.
t\O. ~ Si02 Ti02 Al203 Fe202 M10 NgO CaO Na20 K20 P205 ~ la Tar S(flPITt J
SYF1.1 0-0.2 00.25 1.59 11.76 9.16 0.24 1.54 233 0.00 0.78 0.07 3.57 6.84 00.00 3400
SYF1.2 0.2-0.4 63.24 1.51 11.2 8.43 0.23 1.37 213 0.93 0.78 0.07 3.57 6.84 100.3 831
SYF1.3 0.4-0.6 63.46 1.53 11.51 8.00 0.23 1.43 200 0.00 0.78 0.00 200 6.57 100.3 1446
SYF1.4 0.6-0.8 53.00 1.51 1249 11.55 0.3 292 4.35 0.85 0.77 0.13 3.9 6.73 00.16 E63
SYF1.5 0.B-1.0 48.39 1.93 13.37 10.69 0.18 281 8.55 0.94 0.79 0.18 3.93 7.27 00.00 245
SYF1.6 1.0-1.2 52.23 1.73 13.74 11.28 0.21 249 5.57 1.42 0.87 0.24 4.83 4.81 00.42 >100
SYF1.7 1.2-1.44 43.52 0.9 9.37 26.31 0.45 231 3.00 0.84 0.00 0.16 4.28 7.77 00.93 >100
SYF1.8 1.44-1.6 53.32 0.77 15.84 7.71 0.00 1.65 0.95 0.45 273 0.00 5.26 5.25 00.00
NO. Quartz Orthoclase Hemathite Plagioclase Kaolinite Gypsum Montmorillinite
SYF1.1 X <x XX
SYF1.2 X <x XX
SYF1.3 X <x XX
SYF1.4 X <x XX
SYF1.5 X <x XX
SYF1.6 X <x XX
SYF1.7 X <x XX
SYF1.8 X <x XX
Dominant XX >40%
Major X 10 - 40%
Minor x 2 -10%
Accessory <x 1 - 2%
Rare «x <1%
The sulphur values obtained from the XRF are of interest since they indicate a strong
attenuation in the shallowest layers of the profile. This is consistent with the leaching tests
and porous cup values. According to the XRD analysis the clay is montmorrilonite.
125
6.4.4 New Vaal
No porous cups were installed in New Vaal as piezometers were installed in the irrigation
area.
The molar concentration comparison of the cation analysis in depth in the soil indicates that:
• Calcium and magnesium concentrations increase drastically in the clay layer at
2.4m depth. Even the potassium, which has a far lower concentration in the water,
increased in the clay layer.
• The sodium shows a drastic drop in concentration at 2.4m depth, but recovers
again below the clay layer. Contrary to calcium magnesium and sulphate, it also
shows a decrease in concentration with depth.
• Chloride and sulphate were included in Figure 6-29 and follow the same pattern
as all the major cations except sodium.
Molar distribution of KCIsoluble cations in soils- New Vaal
0.7000 .,.-------------------------------,
0.6000
0.5000
---+-Ca
_K
0.4000
_Mg
__ Na
I_-CSl04
0.0000 1--I~:::::_~"..I11==::::;::::---------:=====-----==.--
O. 0 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10 00
-0.1000 L.,_ ---'
Depth
A'gure 6-29: Molar distribution of major cations and anions in the soil at New Vaal
6.4.5 Conclusions
There are hydraulic and attenuation factors preventing the salts in the mine water used for
irrigation from being mobilised down the soil profile and into the aquifer.
• Calcium decreases significantly more than the other cations, indication gypsum
precipitation.
• Sorption and ion exchange occurs between magnesium and potassium.
126
• The clays play an important role in retaining the salts to the upper few layers of the
profile.
The next section will deal with gypsum precipitation, and determine how much of the sulphate
adsorbs in the soi/.
6.5 Gypsum precipitation from the CaS04-rich water
6.5.1 Introduction
Whether gypsum is present or not in the irrigated fields has been a point of debate since the
project started. Campbell (2001) concluded that gypsum was not present at Pivot One
(Major) or Pivot Four. This conclusion was based on the apparent disequilibrium of saturated
pastes with respect to gypsum, which were found to have concentrations below where
precipitation is expected.
The simulation of gypsum precipitation and dissolution is based on the calculations of the
activities of Ca2 2+, Mg +, SO/· and CaS04 according to the Debye-HOckel theory. Gypsum
precipitates when the ion activity product of Ca2+ and SO/· exceeds the gypsum solubility
product (2.63 x 10.5 at 25°C according to Drever, 1997). According to Campbell (2001) there
are shortcomings in the simulations of gypsum solubility at the irrigation sites:
• The Mg S04 ion pair was not taken into account. In the presence of sufficient Mg in
solution, this ion pair will playa comparable role to that of CaS04 in reducing the
activity of SO/· and enhancing gypsum solubility.
• Ion exchange and adsorption were not taken into account.
Campbell stated that he could not detect or confirm any gypsum with the XRD-method in the
soils at the Kleinkopje pivots. He also stated that calculations of the gypsum saturation
indices of saturated paste extracts from soils in the irrigated fields demonstrated their
undersaturation with respect to gypsum. This suggests that only very small amounts of
gypsum, if any, are present, although Van den Ende (1991) shows that saturated paste
extracts from soil can be undersaturated with respect to gypsum, while the in-situ soil solution
is oversaturated, and that free gypsum will thus be present in the soil. Campbell also
indicates data suggesting that the adsorption of calcium and sulphate ions is the favoured
mechanism to remove these ions from solution. It is possible that, with continued irrigation,
the adsorption sites will reach equilibrium with a soil solution, which is sufficiently
overaturated with respect to gypsum, and that precipitation may then occur. It is possible that
127
this precipitation already occurs during the irrigation seasons, but that intense periods of
rainfall would tend to dissolve a large amount of this gypsum.
However, Annandale et al. (2006) stated that "according to Bohn and Bohn (1996), the strict
convention of saturation indices that log lAP < 0, which means that a solution is
r;
undersaturated with respect to a certain solid phase, can often not be applied in
environmental chemistry. Most solubility products in the literature emanate from minerals in
their standard state i.e. pure, well crystalline minerals. Minerals in the soil are not pure end
member phases, but are often poor or micro-crystalline (Essington 2004). Over-or
undersaturation of a soil solution could simply mean that the mineral under investigation is
not in its standard state. According to Bohn & Bohn, (1996) a log lAP ratio between -1 and 1
Ksp
can be considered as confirmation that the soil solution is saturated with respect to the
specific mineral. A ratio lower than -1 indicates that the solution is either undersaturated with
respect to the pure mineral, or possibly that the solid phase is a minor constituent of another
solid. Differences within this range can be attributed to uncertainty in the lAP measurements,
as well as uncertainties of the reported Ksp in the literature. In the case of gypsum, four
different Ksp'swere obtained. The standard state Kspfor gypsum is reported by Nordstrom et
al (1990) to be 10-4 58 10-4 64. , . according to Lindsay (1979), 10-4 62. is reported by Essington
(2004) and the default Kspfor gypsum in the Minteq. v.4 database file is 10-4 61. .
In the study by Annandale et al. (2006), the four solubility products obtained from the
literature were introduced into PHREEQC, and the saturation of the soil solutions with
respect to gypsum, was evaluated using all of the above reported solubility products. A
saturation range is therefore obtained, which is based on the inherent uncertainty of the true
standard state Kspof gypsum. Using the Kspreported by Nordstrom et al (1990) in saturation
index calculations therefore gives the upper limit of the saturation range, while the Ksp
reported by Lindsay (1979) gives the lower limit of the saturation range. The minerals
kaolinite, goethite, hematite, gibbsite and quartz have been identified by X-ray diffraction in
the soils of Pivot 1 (Major) and Pivot 4 (Campbell, 2001). To obtain ion activity- and solution
speciation approximations as close as possible to those in the soil solution, the above-
mentioned solid phases were introduced in PHREEQC and equilibrated with the saturated
pastes. The saturation range for all depths at Pivot 1 at Kleinkopje was between -0.25 and
0.04 and between -0.09 and 0.04 for Pivot 4. At the time Campbell collected his samples at
Pivot 4 (1999) the solutions were outside the proposed range of -1 to 1 proposed by Bohn
and Bohn (1996), and were indeed undersaturated with respect to gypsum. Since 2000, the
soil solution saturation with respect to gypsum has increased, and gypsum has accumulated
in the soil profile.
128
Annandale et al. (2006) stated that saturated paste extracts are generally accepted as an
approximation of the soil solution composition, ion activity and solution speciation. However,
Campbell (2001) pointed out that the gravimetric water content of the samples collected is
generally far from that of a saturated paste extract. The increase in the water content of the
samples will definitely result in the dissolution of gypsum and an underestimation of the solid
phase gypsum. Calculations with PHREEQC showed that for a specific sample the volume
increase could result in the dissolution of approximately 10.1 mmol '1, using the default Ksp
for gypsum in the Minteq. v.4 database file (10.4.61). The concentration of the soil solutions to
field moisture levels increased the gypsum saturation from 0.00 to 0.64.
According to Annandale et al. (2006), gypsum quantification of the irrigated soils was initially
based on the difference between total extractable Ca and S04 and a certain Ca and S04
sorption capacity. The assumption is that before gypsum precipitation will occur, a certain
sorption capacity must be filled. However, this approach was abandoned because the
determination of a certain sorption capacity of Ca and S04 in the laboratory is difficult to
extrapolate to the field for various reasons:
o The sorption capacity of a soil is a function of pH, ionic strength and ion composition
of the soil solution (Rietra et a" 1999). The experimentally obtained sorption capacity
is therefore a conditional sorption capacity, and not necessarily true for all conditions.
Concentration the soil solution through drying and dilution of the soil solution as a
result of rain will influence this sorption capacity. The solution, with which the soil is
equilibrated, should be similar to the water quality of the irrigation water, which means
that soil samples should be equilibrated with neutralised acid mine drainage (NAMD).
o Surface precipitation can occur in the double layer despite the fact that the bulk
solution is undersaturated with respect to the pure mineral. The NAMD is saturated
with respect to gypsum, which basically means that it is impossible to determine
whether the decrease in Ca and S04 in solution after equilibration was the result of
precipitation or adsorption.
A thermodynamically more sound, and analytically easier approach was therefore chosen by
Annandale et al. (2006). The saturation range was determined for the saturated pastes
extracts. If the saturation range indicated that the solution was saturated with respect to
gypsum water extractable Ca was taken as gypsum. A lesser error is involved when water
extractable Ca is used to calculate gypsum because the water extractability of exchangeable
Ca is lower than that of S04. More than 50 % of the Ca immobilised in the soil was
exchangeable at both Pivot 1 and Pivot 4. Total Ca immobilised in the soil profile at Pivot 1
was 23.9 mol m-2 of which 11.4 mol m-2 was gypsum. At Pivot 4 the total Ca immobilised was
25.9 mol m-2 of which 10.9 mol m-2was gypsum. The amount of gypsum precipitated at Pivot
129
Major was equivalent to 19.6 tans I ha and that at Pivot4 to 18.9 tans I ha. (This is in line with
the findings of this research).
Annandale et al. (2006) concluded that there is thermodynamic proof, in the form of
calculated Ca and S04 activities, that gypsum is present in the soils irrigated with NAMD. The
saturation ranges were the same as those reported for natural gypsiferous soils in the
literature. The authors stated that over the last four years gypsum accumulated in the soils to
such an extent that the solutions were saturated with gypsum.
6.5.2 Determination of the presence of gypsum in irrigated soils
A number of soil samples were collected at shallow depths at Kleinkopje Pivot Major and
Pivot 4. Some of the samples were scraped from surface where white precipitate was in
abundance, usually in lower lying areas. These samples were sent to the Geology
Department at the University of the Free State for X-ray diffraction analysis to determine the
presence of gypsum.
No gypsum was detected in any of the samples, even those that consisted mostly of
precipitate scraped from surface. Gypsum has hkl Miller indices of 1111.21, 1111, 462, 1111,
460, 2101 and 424, which did not correspond with any of the diffractograms obtained.
However, further research has revealed that bassanite, the hydrous form, CaS04 0.5H20, or
calcium sulfate hemihydrate, is present in most of the samples. The hkl Miller indices of 824,
741, 1002 and 628 corresponded accurately with those of the diffractogram of bassanite
(Figure 6-30). Gypsum can be transformed into bassanite upon heating. It reaches a level
corresponding to the semi-hydrate bassanite at 70-90°C. However, dehydration starts at
40°C. Heat build-up definitely occurs between the densely populated plants in the irrigation
area, and probably dehydrates the gypsum sufficiently to transform it into bassanite in the
upper soils. This may be the reason why Campbell (2001) also did not detect any gypsum in
the soils. Precipitated CaS04 is thus definitely present in the soils, although it exists in its
dehydrated capacity as bassanite.
130
900
800
700
BOO
BOO
...
200
Il,
100 , I ! 1, I I,
I .' \ J I \' \ i !
~4i~","!,,~j ~_ .. ~""_" __ .}y ~~I .. I I'V: \I'~J~ ..(.}•\.lN\ ........,........"... ......~~/_·t\..I. ~",__""V.N\..) ........_. .
13 21 a ~ ~ ~ ~ n
File: a:\bassanit.raw
Figure 6-30: Diffractogram of Bessenite (CaS04 0.5 H20).
6.5.3 Sulphate adsorption:
Finely divided clay minerals and organic matter have electrical charges on their surfaces and
tend to adsorb ions from solution onto their charges. Most clay minerals, due to their
molecular structure, have a permanently negative surface charge, which allows for the
substitution of metal cations of different charge so that clay will have a net deficit of positive
charge in their structure (McBride, 1994). However, kaolinite (AbSi205(OH)4) has little or no
permanent negative surface charge. The surface charge is positive at a low pH and negative
at a higher pH, thus possessing CEC at a high pH and AEC at a low pH. According to Matt
(1981). kaolinite has a tendency to adsorb sulphate ions on the positively charged surfaces of
the hydroxides of Fe and AI, and is enhanced by low pH conditions. Sulphate may form
irreversible inner-sphere complexes on these surfaces. Kaolinite reversibly adsorbs sulphate
at its positively-charged edge sites. Sulphate adsorption is most favored at a pH of 2, and
decreases with a rising pH (Matt, 1981), becoming negligible above a pH of 6 (Tabatabai,
1982).
Low levels of S04 adsorption are observed in topsoils relative to deeper horizons (Tabatabai,
1982), probably because of the abundance of negatively charged organic matter in topsoils.
This is believed to be the result of electrostatic repulsion, which causes anions in solution to
be concentrated in the centres of soil pores, where the flow velocity of the soil is the greatest
(Leij and van Genuchten, 1999).
The literature provides evidence of a synergistic enhancement of S04 sorpion by Ca
adsorption (Ajwa and Tabatabai, 1995 and Davis and Burgaa, 1995). A comparison of
extractable Ca and S04 suggests that there may be co-sorption of calcium and sulphate in
131
the irrigated soils of Pivot 1 and Pivot 4 (Campbell, 2001) where kaolinite is the major clay
(see Table 6-12).
Leaching column experiments to determine the adsorption capacity of sulphate:
In order to determine the adsorption capacity of sulphate in the soils at Kleinkopje Pivot 4,
experiments were conducted where field conditions were simulated in leaching columns at
the laboratory of the Institute for Groundwater Studies. A four-meter soil profile (down to the
water table) was taken at Kleinkopje Pivot 4 and proportionally packed into three-meter
columns (Figure 6-31).
Figure 6-31: Photograph of the leaching columns.
The soil used in the leaching columns was analyzed at the Glen Agricultural College near
Bloemfontein to determine its clay content, and XRD was done at the Geology Department of
the University of the Free State (See Section 5.4).
Table 6-15' Clay content in the soil at Kleinkopje Pivot 4.
Depth (m) Sand Coarse silt Fine silt Clay
0-0.6 72.50 9.20 4.60 13.79
0.6-1.0 64.32 11.75 4.62 18.48
1.0-2.0 56.06 10.80 13.93 18.58
2.0-3.0 50.50 15.29 4.60 27.59
3-3.2 53.27 19.49 4.62 23.10
3.2-3.5 59.60 9.79 4.64 23.22
132
Leaching tests done on the soil outside the pivot areas by the IGS laboratory indicate a pH of
mostly below 5, and also confirms the presence of chloride in the soil (Table 6-16). The acidic
conditions in the soil will also enhance sulphate adsorption (Tabatabai, 1982).
Table 6-16: pH and chloride values of soil at Kleinkopje Pivot 4 (background values).
Depth ph Cl
m mg/kg
0-0.3 4.67 33.6
0.3-0.6 4.42 22.6
0.6-1.0 4.50 55.7
1.0-1.25 4.74 91.1
1.25-1.5 5.19 141.4
1.5-1.75 4.91 102.1
1.75-2.0 4.50 20.5
2.0-2.5 4.53 16.8
2.5-3.0 4.54 16.8
3.0-3.2 5.06 16.7
3.2-3.5 4.85 16.3
3.5-4.0 5.29 50.3
The pore volume of the soil in the columns was obtained by filling the soil profle in the
columns and three different scenarios were run, each with different water qualities. The
composition of the water was such that the electrical conductivities corresponded with those
of the irrigation water at Kleinkopje 4 (555 mS/m). The NaCI was added to the leaching
waters of Columns 1 and 3 to ensure ion exchange. (Ion exchange is a reversible chemical
reaction wherein an ion [an atom or molecule that has lost or gained an electron and thus
acquired an electrical charge] from solution is exchanged for a similarly charged ion attached
to an immobile solid particle. Na can exchange with Ca and Mg). The composition of the
waters is tabled in Table 6-17. Columns 1 and 3 were leached with sodium-sulphate water
with different NaCI concentrations. Column 2 was leached with pure NaCI water.
Table 6-17: Composition of the leaching water.
Na2S04/25L NaCI/25L EC (mS/m)
Colomn 1 55g 35g 555
Colomn 2 90 g 659
Colomn 3 83g 7g 545
Each column was leached with twenty-five pore volumes of the different solutions. After each
pore volume was drained, the leachate was analysed for sulphate and chloride. From the
time graph in Figure 6-32 the following can be concluded:
133
S04/CL Leaching
3100.0
2600.0 -+-1 Low 804 Cone
__ 3 High 804 Cone
CJ
c: 2100.0 -.k- 2 Chloride
0
u ~ 1 Background
...J 1600.0 -+-2 Background
~ 3 Background
0 1100.0
Cl) ..vi -1-1 Chloride111+ 1 ---+1+1 ~ II -3Chloride
600.0
__ ••• • •••• -- __ - ww • ._
100.0
Figure 6-32: Graph of sulphate and chloride leaching columns (mg/L).
1. There is a definite increase in chloride in Column 2 (the one leached with NaCI). A
possible explanation is the large pore volume of the column. As such, the initial
amounts of the sodium chloride solution acts as an elutant in the column and over
time the proportion of the leaching solution will increase in the leachate.
2. There is a definite decrease in sulphate in both Columns 1 and 3. This may be due to
some adsorption in the column (SeeTabie 6-18). However, the change in these
values is not significant, as indicated by plotting the percentage of the original input
sulphate against time (Figure 6-33).
3. The dip in the sulphate values for the five leaching events in the middle cannot be
explained.
504 leaching %
'00
~
'50
'0
~ 801---
~
... 70-+---
60 ---ti-='%s~-3% so4_[-
Date
Figure 6-33: Graph of sulphate leaching columns expressed as a fraction of the input.
134
After the last leaching procedure, the columns dried out over a period of a month, after which
they were cut lengthwise. The soil was sampled at each layer to correspond with the clay
analysis. These were then analysed to determine sulphate adsorption. Normally all the
leaching methods were developed to determine total plant -available sulphate (Tabatatai,
1982 and deJaqer, 2005). De Jager developed a method to determine the dissolved and
exchangeable/adsorbed sulphate in the soil as a result of irrigation activities, which was used
to analyse the soil with:
A suspension of 5g soil and 50 ml of water was shaken for 4 hours and left to settle for 16
hours. It was filtered with Whatman N01 filter paper. The extract was then analyzed by ion
chromatography to determine the precipitated part of the sulphate. The soil residue from the
filter was dried, and a suspension of 5g soil and 50 ml potassium-orthophosphate as then
treated the same way.
The results in Table 6-18 indicate:
o An average sulphate adsorption of 24.9% of the total extracted sulphate. The rest
(water extracted) is precipitation. This means that not all sulphate will be leached by
excessive irrigation or rain. It must be noted, however, that desorption can occur and
that the 25% sorbed sulphate will not be irreversibly fixed in the soil profile.
o What is also significant from these data is that the highest adsorption values occured
in the layers where the original soil chloride values (see Table 6-16, and also Table
6-18) were high. The graphs are plotted in Figure 6-34. There is no direct correlation
between the clay content of the soil and the adsorption values, but the increased
values are definitely a function of the hydraulic properties of the soil. The layer with
increased clay content retards the downward movement of the water at 2.5m, and this
result in the increased water content above 2.5m. As a result of this, the "natural"
chloride values are elevated, andthis also results in increased precipitation and
adsorption of ions from the irrigation water.
o As was found by Campbell (2001), the phosphate-extractable sulphate data appeared
to be independent of pH (Table 6-16). However, this can only be determined from
sorption experiments in which the soils are equilibrated with sulphate solutions at
various pH-values.
135
Table 6-18: Water and potassium-orthophosphate extraction values of sulphate.
Unpolluted soil (background values) Leachate values
Depth Cl in soil Clay Water KP04 Adsorption
m mg/kg % mg/kg mg/kg %
0.3 33.6 13.79 303.50 84.65 21.8
0.6 22.6 13.79 156.50 83.16 34.7
0.8 55.7 18.48 404.50 148.91 26.9
1 55.7 18.48 436.60 163.18 27.2
1.2 91.1 18.58 518.40 239.88 31.6
1.5 141.4 18.58 526.40 215.52 29.0
1.8 102.1 18.58 474.20 173.80 26.8
2 20.5 18.58 412.70 96.49 19.0
2.5 16.8 27.59 469.90 99.93 17.5
3 16.8 23.1 575.30 135.10 19.0
3.5 16.3 23.22 598.10 154.14 20.5
--
H20/KP04 Leaching ~
I
Leaching (mg/kg)
0 200 400 600 800 I I'
t
0
0.5 ~ .. '.
1 t""'t~ [-+-waterl t
Ê 1.5 --KP04~~
.J::. 2 ......---coo"'I +' -*- Chloride
C-
cCII 2.5 ~ --EC J
3 -
3.5 ::.-- -
4
I
-_ _jL
Figure 6-34: Water and phosphate leachate compared wlfh original chloride in sou.
6.6 Assessment of the hydraulic behaviour and chemistry in the unsaturated zone
at the virgin soil irrigation sites
In order to determine the hydraulic behaviour, salt balances and attenuation, and the
movement of the salts at the various irrigation sites, the results of the various experiments
that have been performed on site and in the laboratory will be analysed and discussed in the
following sections. This includes the soil analysis, porous cup analysis, monitoring analysis,
tensiometry, leaching columns, pumping tests, slug tests and point dilution tests.This will be
discussed in the following sequence:
• Moisture content at the irrigation sites.
136
• Salt balance through the profile.
• Leaching possibilities of the salts after irrigation has ceased.
• Aquifer parameters and the determination of flow paths.
• Determination of unsaturated / vertical flow through the profile and subsequent salt
attenuation.
6.6.1 Measurement of soil moisture potential
The moisture content of a fully saturated porous medium must be clearly equal to the porosity
of the medium (Botha, 1996). In unsaturated flow, this value of 8, also known as the
saturated moisture content, is conventionally denoted by the symbol 85. The moisture content
and water saturation of a porous medium must thus satisfy the inequalities:
o S I} S e ( I}), and 0 ·Ss; S I
Where: e is porosity.
Total soil moisture potential is often thought of as the sum of matric and osmotic potentials
and is useful for characterising the energy status of soil water with respect to plant water
uptake. The sum of the matric and gravitational (elevation) heads is generally called the
hydraulic head (or hydraulic potential), which is useful in evaluating the directions and
magnitudes of the water-moving forces throughout the soil profile. Methods are available for
measuring matric potential as well as total soil moisture potential, separately or together
(Black, 1965). Matric potential in the field is measured with a tensiometer, whereas in the
laboratory use is often made of tension plates and of air pressure extraction cells.
Tensiometer: If left in the soil for a long period of time a tensiometer follows the changes in
the matric suction of soil water. As soil moisture is depleted by drainage or plant uptake, or as
it is replenished by rainfall or irrigation, corresponding readings on the gauge occur. Suction
measurements by tensiometry are generally limited to matric suction values below 1
atmosphere. This is a result of the vacuum gauge or manometer measuring a partial vacuum
relative to the external atmospheric pressure, as well as of the general failure of water
columns in macroscopic systems to withstand tensions exceeding 1 bar. In practice the
useful limit of most tensiometers is at about 0.8 bar. As described in Section 5.6 the School of
Bioresources Engineering and Environmental Hydrology at the University of Kwazulu-Natal
designed special tensiometers for this project to measure tension as deep as 5 m, and
depending on the depth of the water table, tensiometers were installed at every meter down
drilled core holes at the different irrigation sites, except New Vaal (which is too sandy for
137
good ceramic contact). Results vary in the degree of success, with the tensiometers at
Kleinkopje 4 being the most successful.
The tensiometer data can be interpreted as follows:
o The higher the tension values above 0 bar, the drier the soil.
o At 0 the soil is saturated.
o Below 0 ponding may occur.
K/einkoPie 1:
At Kleinkopje 1 (Figure 6-35) the one-meter tensiometer initially malfunctioned. However,
when auger holes were drilled during the summer and the winter, a visible difference in
moisture content could be observed, with the summer period being wetter. At 2m, the
tensiometer indicates wet conditions during the summer rain period and drying-out conditions
during the winter irrigation period. During September there was again an increase in moisture
when the area was flooded due to a malfunctioning pivot. The 3 m tensiometer is very close
to the water level and indicates very wet conditions (even ponding conditions). The 4 m
tensiometer is below the water level and indicates wet conditions similar to the 3m one.
From these data, it is clear that even at a site with such a shallow water level there is a visible
decrease in soil moisture during winter when effective irrigation is being practiced. This will
result in salts retained in the upper layers of the soil, as indicated by the soil analyses and the
soil water analyses.
138
Kleinkopje1
- 2500 402000 35
-~ 1500 30 :E2C 1000 25 .£
.0e-n 500 20 ~15 'B.
ii 0 10 .~Cl..
I- -500 5
-1000 0
/ qj''? 'I)'? ~'? ,?'? ;p'? ~'? .11lD!im(1..~ ~~ ~~ ~ ·21lD!im"- ~ ~ ~~ '"I,~ 'l>.if'I\! 631lD!im
Time x41lD!im• Pil! CPf1im)1
Figure 6-35' Tensiometer data for K/einkopje 1.
K/einkopje 4:
At Kleinkopje 4 the results from the tensiometers were much better, with all four tensiometers
indicating good results.
Initially the one-meter tensiometer did not work, but after excessive rain contact was
achieved. Since then the layer remained fairly wet. The two-meter tensiometer indicates an
increase in moisture during the rainy season, and started to dry out slightly during winter. The
three-meter tensiometer was installed at the increased clay layer and indicates ponding at
this level after the high rainfall events of the summer. This corresponds with observations
while digging the observation pits. At the end of July all the tensiometers showed an increase
in moisture, with the three-meter one again indicating ponding. It could not be confirmed, but
it is most probably because of an increase in irrigation during this time just after the winter
crop was planted. At four m the effect of the clay layer can be seen clearly. This layer dries
out at a steady rate during winter. This is because of the retarding effect the clay layer has on
the downward movement of the water, but also because the water leaks into the aquifer. The
leaking will occur because the conductivity value is larger below the clay layer than in the
clay itself (See section 6.6.4 on aquifer parameters).
139
The site shows the importance of the clay layers in the downward movement of the soil
water. This will also have an effect on the salts being attenuated in the soil layers above the
saprolite zone.
Klein kopje 4 Tens iometers
~[]][]
3SJ[]
3[]][]+-----------------------------------------~~~4_----15~~
I 2SJ[] s2[]][]+- -L ~~~~--------~~~~--------~----~~ ~
C lSJ[],I r--------------#----~~~~~------------~~--~~.~1[]][]
SJ[]
~ ~~~==~~~~~~pi~;;~~==='i0 =~15
I- [] ._-----,l 10 :::
-&lIl a...
-1[]][]
-lSJ[] .'UWrmr
i'~ .:AlWnnr
,~r, "~,~/ II.:IIWnrn:.:.tJW"",
Date '1Iwjll04u,
Figure 6-36: Tensiometer data for K/einkopje 4.
Svferfontein:
Shortly after the tensiometers at Syferfontein were installed in June 2004 the pivot broke
down. All the results until March 2005, when the pivot came into operation again, were thus
discarded. The soils have a very high clay content at Syferfontein and the tensiometers did
not produce satisfactory results. The two-meter one did not function at all. The only feasible
conclusion from this is that the topsoil (with a clay content of 60%) also dried out slowly
during the winter.
140
Syferfontein Tensiometers
6000~----------------------------------------------~
co --4BClOmm
cUI --35l10mm
Il)
I- --1000mm
-3000..1-.-. ___'
Date
Figure 6-37' Tensiometer data for Syferfontein (tension in mm).
6.6.1.1 Calculation of soil moisture
The pressure potential of water in soil in relation to water content may be portrayed by means
of retention curves. The relationship between water content and pressure potential is not
unique due to the hysteresis effect in the filling and emptying of soil pores. Changes in pore
geometry and water content will occur during swelling of clays. Retention curves determined
under transient and static conditions may differ (Hutson, 1983). In practice, these effects are
usually ignored and a unique non-hysteretic relationship between water content and pressure
potential is assumed. Retention curves are determined in the laboratory by measuring the
water content of soil cores at increasing pneumatic potentials. These tests are time-
consuming and laborious, prompting the search for simpler methods of estimating retentivity
from South African soil types.
Hutson (1983) set up regression equations for most of the soils in South Africa, based on
texture. Soil texture refers to the size of mineral particles or the relative amounts of sand silt
and clay, ranging in size range from fine to coarse. Fine-textured soils contain more clay size
particles and have relatively high porosity but the pores are small and often discontinuous. In
contrast, coarse textured soils contain sand-sized particles and have more porosity but
bigger pores that are connected. The proportion of sand, silt and clay particles in the soil
determine whether a soil is classified as sandy, silty or clayey. Clayey soil has the smallest
soil particles, and many small pore spaces. Soils with a high number of clay particles have a
very high water-holding capacity and are very fine-textured, making them feel smooth and
sticky (like soap) when wet.
141
In South Africa the International Soil Science Society particle size classification is used, with
the addition of a medium sand fraction. The limits are:
o Clay < 0.002mm
o Silt 0.002 - 0.02mm
o Fine sand 0.02 - 0.2mm
o Medium sand 0.2 - 0.5mm
o Coarse sand 0.5 - 2mm
A balanced soil (loam) contains about 40% sand, 40% silt and 20% clay and is preferred for
growing crops. The soil particle diameter of clay is < 0.002mm, that of silt between 0.002mm
and 0.05mm, sand 0.05mm to 2mm, with gravel >2mm).
Textural classes were defined by Loxton (1961). The textural triangle was adapted from that
of the USDA Soil Survey (Hutson, 1983). The textural triangle converts the relative
percentage contribution of each separate (by weight) into a textural class or name such as
loam, clay, silty clay loam and so on. For example: a soil which contains 45% sand, 25% clay
and 30% silt is a loam. Note that silt is determined by difference: % silt = 100% - (45% sand +
25% clay). (see Figure 6-33.)
SAND GRADE CHART TEXTURE CHART
010: Values are e.pressed 90 coarse und 2.0 - 0.5 mm eend 2.0 -0.05 mm
as porcenlages of medium sand 0.5 - O.25mm slit 0.05 -O.OO2'mm
the total sand. fine+¥ery fino lionel 0,f--*-~20 clay c:O.002mm
0.25 - 0.05mm
" COARSE SAND
Figure 6-38: Texture chart (Soil classification working group -1991,' Hillel, 1982).
Clay mineralogy plays an important role in determining the physical properties of soil
(Hutson, 1983). Clays increase the total surface area and are usually the cause for swelling
or shrinking. Highly exchangeable sodium levels can lead to extreme swelling and particle
dispersion. The structure of most soils changes to some extent during wetting and drying
cycles. The extent of this change (or structural stability) depends on the composition of the
142
soil material. Sandy soils in South Africa generally have a high structural stability during
wetting and drying, while the stability of soils containing swelling clays or high exchangeable
sodium levels have low structural stability. Air-ta-water permeability ratio (AWR)
measurements were introduced by Reeve (1953) as a method of evaluating structural
stability. A completely stable porous material has an AWR of 1 while 20 roughly indicate the
instability threshold in soils. These high AWR values may be caused by swelling of the soil.
A large number of published regression equations relate moisture retention to particle size
distribution and bulk density (Petersen et a!., 1968; Rivers and Ship, 1978; Hall et a!., 1977;
MacVicar eta!., 1977 and Motram eta!., 1981).
Soil is composed of solid particles (minerals and organic matter) and pore space (air and
water). The characteristics of the soil are determined by the size, distribution and shape of
the solid particles in addition to the size and number of pore spaces.
Bulk density is the weight of a given volume of soil, which includes the pore spaces. It can be
easily measured by gently pressing a small cylinder into the soil, removing the core and
weighing after drying to remove the water contained in the pore space. Average values would
be 1.3 -1-75 g per ern". Coarse textured soils will usually have a higher bulk density because
they have less pore space than fine textured soils. Bulk density is an important property of
soils since it affects how easily plant roots can penetrate the soil when they propagate. Real
Density is the weight of a given volume of the soil solids only. It would be equivalent to the
average density of the soil minerals and the organic matter.
According to Hutson (1983), particle size distribution, organic matter content and bulk density
accounts for between 60% and 90% of moisture retention. He combined data from several
sources in South Africa to derive equations relating water content at several potentials with
different particle size criteria and bulk densities. In this research project, where moisture
content data down to the water level was needed, and as it is very difficult to sample intact
soil cores to that depth, this method of Hutson is very practical and convenient.
As pressure potential (lJ!) decreases, the proportion of the variance of water content
(0) accounted for by regression increases since moisture retention increases. Of the
independent variables included in the regression, clay content accounts for the greatest
proportion of variance in 0 and lJ!. Silt content increases in importance as potential increases
but bulk density has a negligible influence until saturation is approached. At saturation water
content is related solely to bulk density.
Factors that have an influence on soil water retention are particle shape, bulk density, clay
mineralogy, organic matter content, structure and degree of aggregation. According to
Hutson (1983) some of these factors defy precise quantitative description, and consequently
143
their effect on retentivity may be assessed in qualitative terms only. Using data from a wide
spectrum of physical environments he developed a model, in equation form that is based on
clay and silt percentages together with bulk densities, which may be applied to South African
soils.
This is needed to calculate the salt balance in the soil. The soil texture analyses done on the
Kleinkopje and Syferfontein soils (sand, silt and clay content were compared to these
regression equations) were derived for stable soils by Hutson (1983) with an air water ratio of
10-100 (Hensley, 2006). The tensiometer data at the irrigation pivots were then used to
calculate the soil moisture at the different depths. The water content is calculated as follows:
~: 8.10 = 0.0558 + 0.0037CI + 0.0055Si + 0.0303Db, with the assumption that 8Áj1s is
linear between Vpand 8.10 on the moisture retention curve.
Eg2: 8.30 = 0.0150 + 0.00384CI + 0.00572Si + 0.0463Db, with the assumption that 8Áj1s is
linear between 8.10 and 8.30 on the moisture retention curve. This equation is being used
where the tension from the tensiometer data is above 1000mm.
Where:
8= volumetric water content (m3/m 3)
lj! = moisture potential
Vp= pore volume
Db= bulk density of soil
The saturated volumetric content of the soil was derived from the bulk density of the soils,
using the formula (Hensley, 2006):
es = 1- (Db/2.65)
From Eq1 and Eq2 the moisture content can be calculated for the pivots in relation to the soil
composition obtained from the soil analysis done at Glen (Table 6-19 and Table 6-20
Table 6-19: Kleinkopje 1 (Bainsvlei / Clovellv soil):
Eq1: El.,o
Estimated 8v% Estimated Eq2: Estimated
Depth Db at qrs= 1000mm % 8Vpat 8v% at tjJs=
(mm) Si Clay (rnq/rn" water saturation 3000mm water
1000 25.8 18.6 1.65 31.7 37.7
2000 25.8 18.6 1.7 31.8 35.8 28.6
3000 24.6 9.2 1.75 27.8 34
4000 Below the water table.
144
Only the 2000mm depth graph rose above 1000mm tension, therefore the 8-30 equation was
used to calculte the moisture values at 3000mm tension.
Table 6-20: Kleinkopje 4 (Hutton soil):
Eq1: E:)_1O
Estimated 8v% Estimated Eq2: Estimated
Depth Db at \jJs= 1000mm % 8Vpat 8v% at \jJs=
(mm) Si Clay (mg/m3) water saturation 3000mm water
1000 25 19 1.65 31.4 37.7
2000 20 28 1.7 32.1 35.8
3000 24 23 1.75 32.6 34
4000 14 23 1.75 27.1 34 23.4
(Only at a depth of 4000mm, which lies between 8-10 and 8-30, Ws> 1000mm).
From these calculations the moisture content at different depths and moisture potential
values are extrapolated and estimated in Table 6-21 and Table 6-22 (Hensley, 2006) and
displayed in Figure 6-39 and Figure 6-40:
Table 6-21: Kleinkopje 1estimated moisture content:
Depth
(mm) Estimated 8v% at different \jJsvalues (mm water)
Saturation 500 1000 1500 2000 2500 3000
1000 37.7 34.7 31.7
2000 35.8 33.8 31.8 31 30.2 29.4 28.6
3000 34 30.9 27.8
Table 6-22: Kleinkopje 4 estimated moisture content:
Depth
(mm) Estimated 8v% at different \jJsvalues (mm water)
Saturation 500 1000 1500 2000 2500 3000
1000 37.7 34.6 31.4
2000 35.8 34 32.1
3000 34 33.3 32.6
4000 34 30.6 27.1 26.2 25.3 24.3 23.4
145
Values were estimated for every 500mm of tension in order to calculate more accurate
averages for the different depths (1000mm to 4000mm) to correspond with the porous cup
water quality data needed to calculate a salt balance,
Kleinkopje1 Tensiometers
2500 40
2000
C 1500 30 .1000mm
0 1000 ·2OOOmm
tIJ 20 .3000mmC
Cl> 500 x 4000mm
I-- 0 o Precipttation10
~500
-1000 0
-d'~ ~<)~r§''t" $l~ >'l~ >'l~ <)~ >'l~
Pf'O éf?; ~<:/. ..)-~ ~:_..;:; ~e<:/. ~o~ <;l
"ii &f '),~ \:i ');: ~
Date
Figure 6-39: Tensiometer data for Kleinkopje 1with estimated volumetric water content
(tension in mm).
Kleinkopje 4 Tensiometers
4000 40
3500
23,4%"__"""O 35
3000
-=--~~
2500 ° 30
Ê 2000 ° ~ 25,3% ,,""S ", -~-~ " • 1000 mm0 j~ 0 ~ '_r- ~ 251500 " • ZaOOmmC
0 1000 ':I? 10t. ! ''"---' ':Ii LlOl:-f-- 20 3000 mm'ciii 500 ~ 30,6".~ ..r--.. 4000mm~<01Ol9&ooooooo OO~~ ~3,8% 15
GI 0 _' ",- .,..70 o PreCipitatIOn
t- o ...° ~ 0
,500 ° """'lj '--~." I .--,_ :; /.17''6Z' -..Il.~\.::'?'~ -~ - /;~ 100 :
0 I
,1000 °.. 0 Eo ° o 00 -- 50 0 0 00
,1500 °0 0 o 0 0
~,,\'I.C!J
Date
Figure 6-40: Tensiometer data for Kleinkopje 4 with estimated volumetric water content.
From the moisture content values in the tables above (Table 6-21 and Table 6-22), and the
tensiometer graph average, the volumetric water content (needed for a salt balance) was
calculated at each depth (Table 6-23), The average tension of the tensiometers was
calculated against specific periods of time per tension value, The Syferfontein values are
146
included, but will be less accurate because the tensiometers did not function as well as those
at Kleinkopje.
To obtain these averages per specific depth, the number of days that the tension was for
example saturated, between 250mm and 750 mm was calculated against the 500mm tension
moisture content, the number of days between 750mm and 1250mm against the 1000mm
tension moisture content up to the highest tensions obtained. These values were then
devided between the total period of days to obtain an average for a specific depth.
Table 6-23: A verage estimated moisture content with depth.
Depth 1000mm 2000mm 3000mm 4000mm
Kleinkopje 1 0.32 0.373 0.34 0.34
Kleinkopje 4 0.358 0.339 0.34 0.28
Syferfontein 0.384 0.35 0.274 0.262
It must be kept in mind that the profile above 1m will be less moist than at 1m. Therefore the
conditions above one meter w/~1be much more favourable for gypsum precipitation, and more
will be precipitated in relation to what w/~1be in the soil water deeper down in the profile
where the moist conditions will be less conducive to the precipitation of the salts.
6.6.2 Salt Balance of the profile
One of the critical issues in successful irrigation with the coalmine waters is how much of the
salts are retained in the soil profile over the medium term. A salt balance is used to measure
the amount of salt that is being introduced into the soil, the amount that is being removed,
and the amount currently present within the soil and ground water (Szabolcs, 1986). Salts
accumulate in soils where evapotranspiration (combined evaporation from soils and
transpiration by plants) exceeds combined precipitation, and will follow the water flow in the
event of excessive irrigation and high rainfall events. As saline groundwater approaches the
land surface, plants begin to show signs of salinity damage and may die from salty water in
the root zone and waterlogging, basements may flood, water levels may rise into landfills,
and underground pipes can be damaged.
If a high percentage of the salts precipitate (Annandale et aI., 2006) and also adsorb (for
example 25% of the retained sulphate, see Section 6.5.3) in the soil, irrigation with the mine
waters will not have a significant influence on the groundwater except if these salts leach out
rapidly. (see Section 6.6.3 on leaching).
As explained in the previous section, soil is composed of solid particles (minerals and organic
matter) and pore space (air and water). The characteristics of the soil are determined by the
size, distribution and shape of the solid particles in addition to the size and number of pore
spaces. The soil hydraulic characteristics of a vast range of soil, including that of the irrigation
147
sites, were performed by Lorentz et al. (2001), using the Van Genuchten, Brooks-Corey and
the Campbell parameters. An average of these bulk density values was used in the
calculation of the salt balance. (A spreadsheet with the calculations is in Appendix D).
To calculate the salt balance in the profile, the following data are needed:
o Irrigation quantity and quality (from Annandale et aI., 2006).
o Bulk density of the soil (Lorentz et aI., 2001).
o Volume of water in the soil pores (as discussed in Section 6.6.1).
o Quality of soil salinity (soil leaching data). This is the total salt, which includes
precipitated salts and salt in the soil water.
o Quality of soil water (porous cup and water sampling data).
To calculate the salts retained in the soil, the following calculations was made:
Bulk density * thickness of the layer *sulphate concentration in the soil
To calculate the percentage salts retained in the soil:
Sulphate retained / total sulphate applied over the period of irrigation
To calculate the sulphate retained in the soil water:
Sulphate concentration in soil water * thickness of layer *moisture content at that depth
From these calculations the following results were obtained per hectare of irrigation:
Table 6-24: Salt balance calculations at the irrigation sites.
% Total sulphate % of Total sulphate % of Salts
applied through applied through retained that Sulphate in
irrigation retained irrigation in soil are in soil Sulphate in soil water
Site in vadoze zone water water soil (ton/ha) (ton/ha)
Kleinkopje 1 67 56 83 56 47
Kleinkopje 4 77 60 77 48 37
Syferfontein 82 64 78 42 33
New Vaal 86 30 35 18 6
Because the calculations are so sensitive to the input parameters, they cannot be regarded
as exact figures, but rather in a range of ± 10 % accuracy. The value at Kleinkopje 1 is lower
148
than the others, because a borehole near the test site drains into the mine underneath, and a
large amount of salts in the upper shallow layers was flushed down to the deeper systems
(personally observed). On average, between 75% and 80% of the salts are retained in the
pore water and the rest precipitates and is sorbed (Table 6-24). Hydraulic, adsorption and
attenuation factors prevent the salts in irrigation mine water from being directly and rapidly
mobilised down the deeper soil profile and into the aquifer. New Vaal is the exception where
a large amount of the salts is retained in the clayey soils deeper down the profile. In Table
6-25 the percentage salt per layer (in depth) in relation to the total salt in the profile is
depicted. From this table it is clear that most of the salts in the clayey soil of Syferfontein are
retained in the upper 0.6 m. At Kleinkopje 4 the salts have migrated a little deeper, with most
remaining at 1-2 m in the moist area above the clayey layer.
Table 6-25· Percentage salt in layers in depth through the profile down to water level.
Kleinkopje 1 Kleinkopje 4 New Vaal Syferfontein
Thickness (m) % salt in Iay_er Thickness % salt in layer Thickness (m) % salt in layer Thickness (m) % salt in layer
0,2 31,5 0,3 15 0,3 0,6 0,3 63
0,2 27,1 0,3 9 0,3 0,2 0,3 25
0.4 10,0 0.4 20 0.4 0,7 0.4 9
0,2 4,1 1 52 1 11 1 0,7
0,5 8,9 0,5 0,1 0,2 2,6 1 0,7
0,5 5,3 0,5 2,8 0,2 5,0 1 1,2
0,5 4,5 1 0,1 0,6 15 1
1 8,2 0,5 1,1 1 14
0,5 0,3 3,5 40
1 11
What will happen when irrigation cease at the different sites? How long will it take for the
salts to leach out with rain, and at what rate? These questions will be addressed in the next
section.
6.6.3 Leaching of the salts in the profile
Campbell (2001) states that associations between sulphate retention maxima and zones
representing past or present regions of hydrological transition suggests that the accumulation
of sulphate in any particular region of the profile is significantly influenced by seasonal
variations in the soils' water content and water flow in the soils. In particular, the influx of
rainwater during each wet season appears to move the bulk of sulphate present in the upper
part of the profile to the top of the saturated zone. The high-intensity, short-duration
rainstorms results in the rapid movement of significant volumes of rainwater through the
unsaturated zone in these sandy soils. Since sulphate adsorption, especially on kaolinite
edge sites, tends to be reversible (Matt, 1981) rainwater passing through the unsaturated
zone would cause sulphate to desorb and be transported down to the water table. More
stagnant water flow conditions at the water table will cause a net accumulation of dissolved
sulphate.
149
The leaching requirement is the amount of water that must travel through the soil and into the
groundwater in order to leach the desired amount of salts. In general one volume unit of fairly
low salinity irrigation water can leach 80% of salts from the same volume of soil. This is only
a general rule and significant variation is related to the different soil types (Abrol et al., 1988).
Irrigation water must serve two purposes:
• Some of the water will be used by plants (transpiration) and some will be lost to
unavoidable evaporation. Both processes will raise the salt concentration in the soil.
• The second purpose of irrigation water is to leach the excess salt, created by
evapotranspiration, and maintain the salt balance in this way.
However, in the case of irrigation with colliery water, the water quality is such that it will not
leach the salts in the soil, but instead result in the accumulation of more salts. The only way
to leach the salts is with heavy rainfall events. In order to simulate (and determine) the effect
of rainfall events on the leaching of gypsum from the soil, leaching drums were set up in the
laboratory (Figure 6-42). The gypsum content in the soil was based on the work done by
Annandale et al. (2006). According to these calculations approximately 20 tons gypsum / ha
will precipitate in the upper layer of soil at the Kleinkopje pivots over 15 years. These
simulations and measurements are illustrated in Figure 6-41 (Van der Westhuizen, 2005).
Gypsum(ton h~)
0.0 5.0 10.0 15. 20.0 25.0 30.0 35.0 40.0
0-20
20-40
Depth (em)
40-60
60-80
80-100
100-120
Figure 6-41: Gypsum precipitation calculations at Kleinkopje pivots (After Van der
Westhuizen, 2005)
150
Figure 6-42: Leaching drums with different types of sous and gypsum concentrations.
Different gypsum concentration scenarios were run for the different types of soils at
Kleinkopje Pivot 1 (Bainsvlei and Clovelly soils -lower clay content, 5-10%) and Pivot 4
(Hutton soil - higher clay content, average 18%), Syferfontein Pivot (Arcadia soil -high clay
content, in the order of 60%) and New Vaal Pivot (sandy soil). For the Kleinkopje pivots three
scenarios were simulated i.e. 10 ton gypsum / ha, 20 ton gypsum / ha and 40 ton gypsum /
ha, but for Syferfontein and New Vaal only one scenario each was simulated i.e. 20 ton
gypsum / ha. Artificial rainfall used was of the same quality that has been measured by Van
der Westhuizen (2005). The water had a TDS of 80mg/l (EC of 14 mS/m) with a NaCl value
of 40mg/l and a CaC03 value of 40mg/L.
During each leaching event, 75 mm of rainwater was applied to each drum, and the water
analysed. The soil was then dried out under natural conditions, assisted by wind. This
process was repeated eight times. After each drying cycle, the soil was loosened as if
cultivated by the farmer. This was done to simulate real agricultural conditions.
151
Calculations with PHREEQC indicate that the leachate is saturated or over saturated with
respect to gypsum (Figure 6-43). The exception was during the second last leaching event
(12 December 2005) when all the drums were not completely dried out. The moist soil
resulted in the leachate being under saturated with respect to gypsum.
Gypsum saturation of leachate over time
0.25
0.2
0.15
E 0.1 --.-syter
~ 0.05
0 I-+--NV&c. 0.05 ___ KK4-1Oton
CJ) -0.1 ~KK4-2Oton
0.15
-0.2 ___..._KK4-4Oton
0.25 _._ KK1-2Oton
-0.3
-t- KK1-4Oton
f0':J \\::J':J \\::J':J \\::J':J \\::J':J \\::J':J \\::J':J \\::J'O
,,\::J\\::J(;) ,,\::J\,,\::J ,,\::J\1-'"]; ",,\\::J'?> ",,\1-'"]; ,,1-\\::J" ,,']).1-\::J \::J1-\\::J" -KK1-1Oton
Date
Figure 6-43: Time graph of the gypsum saturation of the different /eachates.
Increase in Ca2+ and SO/- can only occur up to a point, where the aqueous solubility of these
ions becomes limited by the solubility of gypsum. The sulphate concentrations in the leachate
were all in this range. This means that the maximum quantity of gypsum leached from the soil
is bound by the gypsum saturation of the water moving through the soil. Calculating the
sulphate (1800-2000 mg/L) and calcium (600-700 mg/L) concentrations in the leachate
indicates that, if 20 tons gypsum / ha precipitated in the upper layer of soil, at least 40 - 50
heavy rainfall events (75mm) will be needed to leach the precipitated salts from the soil.
152
-- - -
Drum leaching - Sulphate
:e::nJ
g 6000
c: 5000 - r-+- KK110ton
~0-... 4000 --ns <; __ KK120ton 1c: 3000 - -.- KK1 40tonCl)
u 2000 ~c: •..... ~ --*- KK4 10ton0
U 1000 " __..._ KK4 20ton
'It
0 0 -+- KK4 40ton
C/)
~\:)'0 ~'0 ~'0 b~f5 b~ ~f5 f'J'f:><:O~,,<o r:,fl
R> y\SR> -+-NV 2Oton
\:)OJ~ ,,\:J -,~\:J -,~'l,; ,,"V" ~\:) ~."\:) \:) \:)"V" - Syfer 20ton
DATE
-- _j
Figure 6-44: Leaching rate of sulphate from the leaching drums.
According to Figure 6-44 the soils at 8yferfontein will leach faster than the others. High Na in
the 8yferfontein water causes high ionic strength, and due to these effects the activity
coefficient of the ions decreases, resulting in the higher solubility of Ca and 804 before
gypsum equilibrium is reached.
It must be noted that this represents the minimum rainfall events needed, as the conditions
in the experiment are optimal for leaching.
• Evapotranspiration, which will decrease the volume water moving through the soil,
has not been taken into account during the experiment.
• The leaching is also a direct function of the moisture content of the soils. The drier the
soil, the more water is needed to flush the salts.
• Recharge, and the subsequent vertical movement of water, are bound by the amount
of rain. It was assumed that very little recharge would occur in events of less than 15-
20mm. Only high rainfall events will thus effectively leach the salts.
It should be noted, however, that even if gypsum accumulates in the upper meter of the
irrigated soils, as predicted by Annandale et al. (1999), there would still be downward
movement of Mg, 804 and other salts, given that Mg will be more exchangeable than Ca in
the upper part of the soil profile and that the concentration of 804 in the irrigation water is
twice that of Ca. Furthermore, the solution that percolates from the base of a gypsiferous
zone is saturated with respect to gypsum (under equilibrium conditions), so soluble Ca must
also be present in leachate from the zone, i.e. if Ca is not zero. Even if irrigation is managed
so that leaching is negligible during the dry season, it is apparent from the data that the
significant movement of the solutes caused by irrigation will take place during the rainy
season. This means that in the longterm almost all of the Mg, Na, K and Cl, at least half of the
153
S04, and some of the Ca added by irrigation must enter the groundwater system. If
sustainable gypsum precipitation does not take place, then nearly all the added Mg, S04 and
Ca will move te;>the local aquifer.
In order to form an idea of the time it will take to leach the gypsum retained in the soil after
irrigation has stopped, leaching columns were set up with soil from the irrigation sites. The
soil was augured down to water level depth at Kleinkopje 1 and 4, and was then placed in 4m
leaching columns in exactly the same sequence as it was taken from the profile. The same
quality "artificial" rainwater that was used in the previous experiment was then used to leach
the soils. The EC was measured after each leaching event and the experiment continued
until the leachate quality approached that of the elluent (artificial rainwater) (Figure 6-45).
Leaching of Irrigation soil
r
800~--------------------------------~
700 I ~
600
Ê 500+~\-
ï -+-KKD400 '-\ ---KK4
1rl 300 ~ -00 Water
200
100f-
L o 10 20 30 40 50Leaching events (125mm/cycle)
FIgure 6-45' EG of the leachate from the irrigated soil.
Under these optimal conditions the columns indicated that at least 30 leaching events of 125
mm each (3750 mm) are needed to decrease the EC down to 38 and 34 mS/m respectively.
To decrease it to 16 ms/m (nearly the same as that of the rainfall) 42 leaching events (5250
mm) are needed.
All the important macro elements were analysed for after selected events. The leaching
trend, especially for sulphate, is the same as that of the EC (Figure 6-46).
154
Leaching of irrigated soil - water quality
7000
6000
::J 5000en i__KK1-S04
.§. I __ KK4-S04
.c: 4000 -.._ KK1-Ca0, ~KK4-Ca
~.. 3000oe I=:=::~:
0
0 l-KK4-Na
2000
10 15 20 25 30 35 40
Leaching cycles (125mm/event)
Figure 6-46: Ca, Mg and S04 qualities of the leachate over the period of leaching.
To compare the volume of rain needed to the time it will take to leach the salts from the soil
profile, a few assumptions had to be made:
• The leaching is also a direct function of the moisture content of the soils. The
drier the soil, the more water is needed to flush the salts.
• Recharge and the subsequent vertical movement of water, are bound by the
amount of rain. Very little recharge will occur in events of less than 15-20mm. Only
high rainfall events will thus leach the salts. An average of rainfall available for
leaching (if this assumption was upheld) since 1929 was calculated and relates to
338mm/a if daily evapotranspiration is also subtracted (see ET Calcs in Appendix
D). It is also assumed that all the flow is diffusive flux (which will leach the salts
uniformly) and not macropare (preferred pathway) flux, which will only leach areas
immediately surrounding the preferred flow areas.
• Daily evapotranspiration has to be compared to the rainfall in order to determine
how much water will pass the root zone and will be available for leaching. From the
ET figures for maize, wheat and pastures, obtained by Green (1985), it is assumed
that the evapotranspiration is the same throughout the crop period, which is not
correct.
The total ET (mm) for these crops at the Highveld is as follows:
155
Crop ET Time span Total days
Maize 550 15 Dec - 14 April 119
Wheat 570 15 June-13 Nov 150
Pastures 1300 Jan -Dec 365
And for New Vaal (ET in mm):
Crop ET Time span Total days
Maize 730 15 Dec - 14 April 119'
Wheat 645 15 June-13 Nov 150
Lucerne 1500 Jan -Dec 365
The evaporation in the hot summer months (also the rainy season) will be much
higher than during winter. The assumption was made that there is a linear
relationship between temperature and ET (average temperatures obtained from the
SA Weather Bureau). A spreadsheet on all these calculations can be seen in ET
Calcs in Appendix D. The average daily evapotranspiration figures are tabled in
Table 6-26.
Table 6-26' Daily calculated ET values at the Highveld and New Vaal.
HIGHVELD NEW VAAL
Maize Month Daily ET Pastures Month Daily ET Maize Month Daily ET pastures Month Daily ET
December 4.7 January 4.1 December 6.2 January 4.9
January 4.9 Febnuary 4.0 January 6.5 February 4.7
February 4.7 March 4.0 Febnuary 6.2 March 4.5
March 4.7 April 3.5 March 6.0 April 3.9
lApril 4.2 May 3.2 April 5.2 May 3.6
l\!_heat May Daily ET June 2.7 Wheat May Daily ET June 3.0
June 3.1 Julv 2.7 June 3.5 Julv 3.2
July 3.1 August 3.2 July 3.5 August 3.6
August 3.6 September 3.7 August 4.1 September 4.3
September 4.2 October 3.8 September 4.7 October 4.5
October 4.3 November 3.8 October 4.9 November 4.5
November 4.5 December 4.0 November 4.9 December 4.7
With all these assumptions in mind, it will take at least 15.5 years (5250 mm/ 338 mm/a
available on average after small rainfall events and evapotranspiration were subtracted from
annual rainfall) to leach the soil profile at the Highveld with rain after irrigation has stopped.
At New Vaal it will take 44 years if the same crop (maize and wheat) is planted on the same
type of soils as in the Highveld, because the ET is much higher and the rainfall is less (only
630 mm average against the 720 mm at the Highveld). However, the measured gypsum load
in the sands at New Vaal is only 20 ton/ha against the 48 - 56 ton/ha at the Kleinkopje sites
and 42 ton at Syferfontein (see spreadsheet on salt balance in Appendix D.). From the salt
leaching experiments and the illustration of sulphate leaching from the drums in the time
series graph in Figure 6-44, it is clear that the salts in the New Vaal sand profile will leach at
156
the same rate as the Highveld soils at Kleinkopje. It can thus be assumed that the New Vaal
soil profile will also leach in 44*(20/56) years, i.e. 16-18 years.
To test these values, they can be compared with a vertical K-value calculation. If assumed
that porosity is 35%, there will be 0.525 m3 per m2 of water for the first 1.5m of the profile,
where most salts are currently retained. If the vertical K is 1*10.4 mld for diffusive flow, it will
take 0.525 / 1*10.4 days for the water with the salts to move through the profile, which will be
14 years.
It must be kept in mind that these leaching figures are derived from the absolute optimal
conditions, with 100% diffusive flux through the soil profiles. In reality it will take much longer
(and with a non-linear decrease in gypsum) because of the rnaeropore flow. According to
Table 6-34 in the following section 25% of the water applied through irrigation will reach the
water level. This will mean that under the same conditions (wet irrigation conditions) and with
an annual rainfall of 750 mm, it will take 30 years to leach the gypsum. The conditions in the
soil will however be drier under normal cultivating conditions, and recharge will maximum be
5-6% (Usher et aI., 2006). Then it will take more than a hundred years to leach the sulphate
from the soil. In the following section (Section 6.6.4) ca/culations wHIshow that most of the
recharge through the profile is along preferred pathways.
6.6.4 Identification of the aquifer parameters and flow zones through interpretation of point
dilution, slug and pumping test data (See Sections 59 to 5 11).
6.6.4.1 Kleinkopje 1:
This borehole had very low salinity water with a background value almost below the detection
limit of the EC-measuring instrument. As a result, only 20g of NaCI (circulated in the volume
of water of the borehole) was enough to increase the EC of the tested section by about 250%
during the point dilution test. The water level was at 2.45 m and the test conductive to a depth
of 12m.The point dilution test clearly indicates a high flow zone at 6 m, because of the dilution
that occurs at this level (Figure 6-47), and also one at 8-9m.
157
Dilution of salt concentration
0
2
- 4 -+--20min l-E ___ 30min.c -.-60min~ 6c. -*-90min J
cQ) 8 ___"_120min_'_480min
10 --+-Backgro und
12
0 50 100 150
EC (mS/m)
Figure 6-47' Graphs of the decrease in concentration over time at KK1.
The values in Table 6-27 indicate the conductivity values of zones of higher flow (preferred
pathways or fractures), and not of the matrix (or aquifer) as a whole for this section of the
profile. The high flow zone at 6 m confirms what has been encountered while digging the test
pit at this site when high flow was encountered at a depth of 4m. The values above the water
level were obtained from the falling head tests. The transmissivity value of the aquifer, as
determined by the numerous pumping tests performed at this site, ranges from 0.2 - 0.3 m2/d,
which relates to matrix K-values of <1mid.
Table 6-27' Conductivity values for different now zones through the profile at KK1 (with the
nux in mld).
Depth (m) K (mld) Flux q
1.5 0.05 0.001
3 0.03 0.003
4 127.10 0.318
5 19.36 0.048
6 32.81 0.082
7 21.91 0.055
8 24.17 0.060
9 32.69 0.082
10 25.67 0.064
11 28.95 0.072
12 18.44 0.046
158
6.6.4.2 Kleinkopje 4:
The water level in the borehole was 2.03 m and the profile was tested to a depth of 15m. The
background EC-value of this borehole at the top 5m was almost half of the rest of the
borehole. This may indicate the existence of at least two aquifer systems where by, fresh
water is flowing in the upper section. 250g of NaCI was added to significantly increase the
concentration from the background value during the point dilution test. Three layers with
different velocities could be distinguished from the result. These are: above 5m which is
characterised by a lower velocity, from 7 - 10m which has a higher value, and between 10 -
15m which has an intermediate velocity value (Figure 6-48).
Dilution of salt concentration KK4
0
2
4 r=:::=~0"k9'O""d
g 6 --.-60
o...e.. 8 ~90
C-
G)
C 10 ~120
-200
12 -1-1020
14
I
l16 0 100 200 300 400EC(mS/m)
Figure 6-48: Graphs of the decrease in concentration over time at KK4.
The values in Table 6-28 indicate flow zones with higher conductivity values (preferred
pathways or fractures), and not of the matrix (or aquifer) as a whole for that section of the
profile. Low conductivity values occur in the soil above the water table (obtained from the
falling head tests), basically indicating no flow or very low flow zones in the upper soil. The
relatively low conductivity value of the clay layer at 3m confirms the reason for panding at this
depth, as indicated by the tensiometers. The high flow zone at 7m has been confirmed during
pumping tests when water could be heard and seen running into the borehole at a depth of 6-
7m after the water table level was draw down below this level. The transmissivity value of the
aquifer, as determined by the pumping tests at this site, ranges from 0.2 - 0.3 m2/d, which
relates to an average aquifer K-value of <1mId (See Appendix 0 Pumping tests, to confirm
these values).
159
Table 6-28: Conductivity values for different flow zones through the profile at KK4 (with the
flux in mid).
Depth (m) K (mld) Flux q
1.3 0.06 0.005
2 0.58 0.002
2.5 22.57 0.056
3 4.76 0.012
5 57.44 0.144
7 111.12 0.278
10 29.17 0.073
15 35.20 0.088
6.6.4.3 Syferfontein:
The test was conducted from the water level at 3.36m to a depth of 20m. Nael was added to
increase the solution by roughly 250%. There was not much dilution over time, which
indicates that the groundwater velocity was very low through the whole profile. It took more
than 66 hours for the tracer to be diluted to its initial background concentration. There was
also not much difference in concentration along the profile over time, which indicates that the
groundwater flow velocity is more or less the same throughout the section (Figure 6-49).
Because of the low velocity value, aggravated by the high solution concentration added,
movement of the solution downwards due to density gradient was observed. To interpret the
Darcy velocities, the movement of the tracer by density gradient had to be compensated for.
Dilution of salt concentration Syferfontein
0
5
-+-Background
_30
10 ----; -1-60
Ê ~90
;; 15 ~120
Cl. -+-"IlO
Cl)
0 ---!-240
20 -1)
l-2025 300
30
0 100 200 300 400 500
Concentration
Figure 6-49: Graphs of the decrease in concentration over time at Syferfontein.
The groundwater flow velocity is fairly uniform throughout the borehole, except at 4.5m where
there was an elevation in the conductivity value. This is not clearly illustrated in the graphs in
160
Figure 6-49, but is portrayed in the calculations (obtained from the point dilution test) in Table
6-29. Some of the water that moves down the upper soil profile will move lateral along this
zone, but not as much as at the Kleinkopje sites.
Table 6-29: Conductivity values for different flow zones through the profile at Syferfontein
(with the flux in mid).
Depth (m) K (mld) Fluxq
1 0.086 0.0002
2.5 6.55 0.016
4.5 39.48 0.091
7 6.53 0.016
10 5.64 0.014
15 3.66 0.009
20 2.44 0.006
25 0.95 0.002
6.6.4.4 New Vaal:
There was not much difference in concentration along the profile over time, which indicates
that the groundwater flow velocity is more or less the same throughout the profile in depth.
Unfortunately the flux of the clayey layer between 2.4 mand 4m could not be determined with
this test. (The water level is below the clayey layer and could thus not be determined with the
point dilution test. The auger holes used for the falling head tests were not deep enough due
to the sand caving in).
Dilution of salt concentration
0
2 -
- 4 - • Background Ig 6 ~~ ~60
.. ~ 4ol.:. 8 +- - -,lIE-90
Q.
Q) 10 --- ~ ~.
C ~20
12 - ~ 90
-
14 - --~-+-~
16
0 2 4 6 8 10 12
Concentration
Figure 6-50: Graphs of the decrease in concentration over time at New Vaal.
161
However, the areas immediately above and below the clayey layer shows elevated hydraulic
conductivity values (Table 6-30), indicating preferred flow zones, as observed while digging
the inspection pits.
Table 6-30: Conductivity values for different flow zones through the profile at New Vaal
Depth (m) K (mld) Flux q
1 32 0.320
2.5 50 0.500
4 50 0.501
7 39 0.390
9 13 0.130
14 17 0.177
The lower conductivity values deeper down the profile is due to higher clay content in the soil
deeper down (See Section 6.3.2).
These tests indicate lateral flow zones with higher conductivity (fractures or weathered
zones), which will have an influence on the quality of water moving down towards the water
level and from there into the aquifer. To reach the water table, the water that moved past the
root zone will fol/ow different paths and flow mechanisms to reach it.
6.6.5 Determination off/ow at the Irrigation site
6.6.5.1 Introduction
If the flow through the unsaturated zone is bimodal (i.e. diffuse plus preferential flow), the
chloride content in the soil water of the unsaturated zone wil/ be different from the chloride
content in the groundwater. If the chloride in the groundwater is less than the chloride in the
soil water, preferential flow paths exist (Selaoio, 1998 and Gehrels, 1998). This is the case at
all the researched irrigation sites. The chloride mass balance method is thus ideal to
calculate the flows at these sites.
6.6.5.2 Chloride mass balance and moisture fluxes
The chloride mass balance method is one of the techniques that are often used to estimate
moisture fluxes through the unsaturated zone, thereby enabling the assessment of recharge
to groundwater basins. The chloride mass balance method has been used to evaluate
recharge processes in a wide range of semi-arid environments (Eriksson and Khunakasem,
1969; Allison and Hughes, 1978; Edmunds and Walton, 1980 and Sharma and Hughes,
1985). Of the tracer techniques available, Cl balance techniques appear to be the simplest
and least expensive method for recharge estimation (Allison et el, 1994). No sophisticated
instrumentation or dependence on the measuring of specific runoff events is required, and
162
when used, groundwater estimates of the method is independent of whether recharge is
focused or diffuse (Bazuhair and Wood, 1996). The method is useful because long-term
average precipitation and chloride concentrations in rain and groundwater entail less
uncertainty and are generally less expensive to acquire than physically based parameters
commonly used in analysing recharge via water balance approaches.
Chloride is used for recharge estimation because of its conservative nature and the relative
abundance in precipitation. The application is based on a comparison of the chloride
deposition rate at the soil surface with the concentration in the soil water or groundwater.
The Cl concentration increases relative to the concentration of rainwater as a result of
interception, soil evaporation and/or root water uptake by the vegetation.
The theory is that solutes will be transported downwards through the unsaturated zone at
rates that depend on precipitation amounts. Some of the solutes may be removed from the
soil water by a variety of secondary processes, such as plant root uptake, mineral
precipitation and adsorption. Conversely, some of the solutes may also be released by the
decay of dead plant material, by mineral dissolution, or by desorption (Edmunds et al, 1988).
A prerequisite in the use of a chemical tracer is that only tracers that are not subject to
chemical reactions, and for which there is no net release or storage from the soil matrix on
their way down, are considered.
The chloride balance method, developed by Eriksson and Khunakasem (1969), compares
total chloride deposition at the surface with the soil water concentration below the active root
zone or the concentration in groundwater. Under diffusive or piston flow conditions, Cl
increases in the root zone, until a constant value is reached below the root zone (Allison,
1988). Under steady-state conditions of piston flow, the flux of Cl at the surface is equal to
the flux of Cl below the active root zone, and the conservation of mass leads to (Eriksson and
Khunakasem, 1969):
PCfp =R C/sw
Where:
P = mean annual precipitation
R = recharge (mm/a), and
Cl, = mean Cl concentrations (mmol/I) in precipitation
Cisw = mean Cl concentrations (mmol/I) in soil water
Sharma and Hughes (1985) provide a slightly more complex estimation of recharge. They
introduce a bimodal model including a bypass flow component. If it is assumed that
soilmoisture movement occurs through the soil matrix as well as through preferred pathways,
163
then the total recharge can be evaluated as a two-component system. Soilmoisture transport
in such a flow regime consists of a slow diffusive and quick preferential flow component.
Preferential flow may well extend beyond the root zone. Bypass flow in the root zone may
become more diffuse in the percolation zone, but it is also possible that it persists all the way
down to the water table. If the profiling depth is limited to within the unsaturated zone, only
the diffuse recharge component is estimated.
The basic assumptions of the method are that:
o There is no source of chloride, other than that from precipitation in the soil water or
groundwater.
o The chloride ion behaves conservatively and is thus not adsorbed by nor is it leached
from vegetation, unsaturated zone sediments and/or aquifer formations (Edmunds et
a/.,1988).
o Diffuse and/or preferential flow often needs to be critically examined. Soil moisture
and solutes may be transported through the unsaturated zone by preferred pathways
(macro-pores).
Under preferential flow or bypass flow conditions some of the precipitation water (with
relatively low chloride concentration) will flow directly along the preferred pathways to below
the root zone. This bypass flow component, unaffected by evapotranspiration, dilutes the soil
water that has been concentrated by transpiration, resulting in relatively low concentrations in
the percolation zone with respect to the concentrations encountered in the root zone.
Assuming conservation of mass in such a bimodal flow situation, total recharge (RT) can be
partitioned into two components as (Sharma and Hughes, 1985):
and
where Rsw is recharge attributed to diffuse flow (mm/a); Rpp, is recharge attributed to
preferential flow (mm/a), C/gw, C/sw and Cfpp are the Cl concentrations in mmol/I of
groundwater, soil water and bypass flow, respectively. C/pp, is assumed equal to the Cl
concentration of precipitation. When Cisw is close to Clgw, the total recharge is approached by
the diffuse recharge component. Conversely, when Cisw is higher than Clgw diffuse recharge
is not the only recharge component and bypass flow occurs as well.
164
If the chloride concentration of the preferential flow component (Clpp) is nearly identical to the
chloride concentration of the precipitation, then enrichment of soil moisture solutes by plant
uptake becomes negligible.
For these equations, it is assumed that the two flow components of water are thoroughly
mixed upon groundwater recharge to the aquifer. Rapidly decreasing chloride concentrations
below the root zone may be an indication of preferential flow. If the chloride concentrations
become constant with depth, the moisture flux can then be calculated.
6.6.5.3 Bimodal Flow calculations for the irrigation sites
The chloride values at all the different irrigation sites showed a decrease in chloride values
with depth, indicating bimodal flow (Sharma and Hughes, 1985). The values of the chloride
with depth are illustrated in Table 6-31. These values were obtained from the porous cup
data, the monitoring data as well as from Annandale et ai, 2006. The values were averaged
overtime.
Table 6-31: Chlonde values with depth at the irrigation sites (mg/L).
Cl value of
Weighted irrigation groundwater outside 1m depth 3m depth Cl value
Site water Cl value irrigation area Cl value Cl value atWL
Kleinkopje 4 42 20 130 100 33
Kleinkopje1 13 20 48 42 23
Syferfontein 24 28 72 55 28
New Vaal 34 8 79 130 41
The calculations were derived according to chloride values at 1m Oust below root zone}, at
the 3m level and at the water level (van Tonder, 2006). (These calculations are illustrated in
the spreadsheet named "Irrigation flux" in Appendix D). This implies:
1. Diffusive flow (av) at the root zone level. At this level there is an increase in
chloride concentration due to evapotranspiration.
2. Bimodal (or bypass) flow, i.e. vertical as well as preferred pathway flow (Qsp) at
3m. This component portrays the pollution flux that reaches the water level. The
chloride value has decreased from the value just below the root zone, and can
only be attributed to dilution due to preferred flow, where water that is not
evapotranspirated (water that moves directly through along pathways) mixes with
diffusive flow water.
3. Lateral flow (aL) at water level. This indicates the groundwater component of
water (lower Cl) that flows laterally which plays a role in the mixing fraction of the
salts.
165
The calculations are as follows:
1.
Qv = _C_lirrig * Rain + Ir_r_ig_:a::t:i.o.._n_
Cluz 365days
Where:
Cl,rrig= chloride value in the rain
Clrz= chloride value at 1m, just below the root zone
Rain = total annual rain
Irrigation = total annual irrigation
2.
= Cl .QSp Cl 3m */i"(/'Q v -Cl. a/) *Qj/~II" Cl
3/11
Where:
CI3m = chloride value at 3m depth
3.
Where:
Clwl = chloride value at water level
Clouts/de = chloride value of aquifer outside the pivot area
From these formulas the vertical and the lateral flux components that contribute to the mixing
values of the chloride could be calculated (Table 6-32).
166
Table 6-32: Percentage flux that contributes to the chloride mixing process to give
groundwater Cl concentration.
Site Qv+Qsp (bypass) QL (lateral)
Kleinkopje 4 16.3 83.8
Kleinkopje1 13.6 86.4
Syferfontein 1.8 98.2
New Vaal 27.3 72.7
These values have a major impact on the feasibility of irrigation on these sites at the short to
medium term. The saline water that moves down through the soil profile contributes
approximately 15% or less to the mixing process of the water. This causes dilution and
explains the lower salinity qualities measured down the soil profile (as samples with depth in
the porous cups). The low value obtained at Syferfontein (1.8) may be due to the fact that
irrigation ceased at this site for nearly two years and that the system has not reached steady
state, with the lateral component contributing a lot more than should be the case if irrigation
continued.
The actual fluxes (m3/d/m2) for the three components Qv, aBp and Ql calculated are as follows
(Table 6-33):
Table 6-33: Actual fluxes (M/d/d) at the irrigation sites.
Site Qv (m3/d/m=2K diffusive) Qsp (K bypass) QL
Kleinkopje 4 7.23E-04 2.80E-04 5.17E-03
Kleinkopje1 3.34E-04 5.77E-05 2.48E-03
Syferfontein 7.58E-04 3.07E-04 5.75E-02
New Vaal 9.65E-04 O.OOE+OO 2.57E-03
(Note: The values in both Table 6-32 and Table 6-33 are not fixed values, but should be
seen as ranges of what the different fluxes will be - the calculations are sensitive to various
parameters).
• The bypass component of the flux (OBP) at Syferfontein is larger than at the other sites
(Table 6-33). The reason for this is that the clay normally cracks very quickly near the
surface during the drying out process by the sun and wind, resulting in large cracks,
which serve as conduits for water. During thunderstorms, most of the water infiltrates
quickly before the cracks close up because of swelling. This has been observed
during leaching experiments in the laboratory, and on numerous occasions in the
field. At Syferfontein, the Arcadia clays are not very deep (O.2m -1.0m) and cracks will
extend into the weathered (clayey) dolerite and shale below.
167
From these fluxes the % flux moving down into the groundwater in relation to the water
applied to the pivot sites, as well as the actual volume of water that reaches the groundwater,
were calculated as illustrated in Table 6-34.
o Also portrayed in the table is the K-value derived from these fluxes (av + aSp) at
each of the sites.
o The value for Kleinkopje 1 is small in comparison with the other sites (9%). This is
due to the fact that water is lost due to drainage down the borehole into the mine.
Table 6-34: % Flux that reaches groundwater.
% Flux of total water Total water that
apllied that moves reaches WL
Site down to WL K-value (m3/ha/d)
Kleinkopje 4 24.07 0.001004 10.0
Kleinkopje1 9.39 0.000391 3.9
Syferfontein 25.55 0.001066 10.7
New Vaal 24.47 0.000965 9.7
The salinity increase in the monitoring boreholes does not always reflect the downward
movement of the water. This strengthens the belief that most of the monitoring boreholes are
constructed incorrectly. The boreholes known to be constructed correctly (PV1 BH2 at
Kleinkopje 1, PV4BH2 at Kleinkopje 4 and OBD6D at Syferfontein) show an increase in salt
concentration as portrayed by the downward fluxes, whilst the other boreholes e.g. PV1 BH7
at Kleinkopje 1, PV4BH6 at Kleinkopje 4 and OBH5 at Syferfontein show unrealistically low
values of salts in the aquifer (see
Table 6-35) and is graphically illustrated in Figure 6-51. The borehole at Syferfontein situated
outside the pivot, but next to the dyke (OBH2) running through the pivot, shows elevated
values, which implies that some of the water drains along the dyke. At New Vaal, the high flux
value is due to the sandy nature of the soil. Monitoring (see Section 6.2.2.1) indicates an
increase in sulphate in the Borehole PV5 in the middle of the site. It is thus very important to
know what is being measured at a monitoring site.
168
Table 6-35: Sulphate values at irrigation sites (values are tabled downwards in time from the
start of irrigation until2006).
Kleinkopje 1 Kleinkopje 4 Syferfontein
Time PV1BH7 PV1BH2 PV4BH6 PV4BH2 OBH6D OBH5 OBH2
1 3 3 5 10 126 40 133
2 2 6 1 7 116 50 131
3 3 12 5 2 126 46 138
4 4 18 1 3 131 39 166
5 2 26 1 115 177 39 139
6 3 53 1 55
7 1 89 1 203
8 0 135 5 240
9 0 197
10 2 275
Increase in Salinity
300
250
::::! 200 i-+--PV'BH7-KK1
C)
E _PV1BH2-KK1.......... 150 - _ PV4BI-6-KK4Cl) I~ns ~PV4BH2-KK4..c: 100
Q,
:; ! _OBI-6D-Syfer-+--OBHS-Syfer50 A
~ .... _;z l_OBH2-Syfer
0 ,~ ~ ~- --- ~ • •
-50 0 2 4 6 8 10 112
Time
Agure 6-51: Increase in sulphate concentration in well-constructed boreholes.
6.6.6 The influence of clay in the unsaturated profile on the soil water that reaches the
water level
A ve'Y important aspect of the research at the imgation sites is to determine the influence of
the clay on the downward movement of the soil water, and what the unsaturated hydrulic
conductivity and saturated hydrulic conductivity values will be.
The relation of the unsaturated K and moisture can be derived from the well-known equation
of Van Genuchten and can be illustrated in a very simple form (Botha, 1996) as (also see
Section 4.1.2 on Unsaturated flow):
K(B) = Ks( B - Br )5/2
B. + Br)
169
Where:
er= residual moisture content
es = saturated moisture content
K(e) = unsaturated K
Ks = saturated K
To help geohydrologists obtain an estimated unsaturated K-value that will be applicable to
the deeper layers below the root zone, a spreadsheet was set up to calculate the unsaturated
K-value according to the equation above. To determine the K(O) (unsaturated-K) value for a
specific moisture content (0) derived from a tensiometer reading at an irrigation site (See
Section 6.6.1), the following is thus needed (see spreadsheet "Unsaturated K" in Appendix
D):
1. Residual moisture content
2. Saturated moisture content
3. Saturated K- value
The unsaturated K-value at the irrigation sites will differ from that under normal field
conditions in the sense that the soil will always have a higher moisture content, resulting in
swelling of clays (if present), thus decreasing the flux, and subsequently the unsaturated K-
value.
1. It was decided to use the 8.1500 regression equation (thus at 15 bar matric pressure) from
Hutson (1983) as a default value to determine the residual moisture content, as suggested by
Van Genuchten (1980).
According to the regression equations derived by Hutson (1983) in relation to the soil texture
(sand, silt and clay content) this can be calculated from:
0-1500 =0.0602 + 0.0032CI + 0.0031Si - 0.026Db
2. The saturated volumetric content of the soil can be derived from the bulk density of the
soils, using the formulae (Hensley, 2006):
Os= 1- (Db/2.65)
(Where 2.65 is the average density of soil minerals).
3. Saturated K:
To determine the Ks for different clay percentages at the irrigation sites, the diffusive flux
obtained from the calculations for each site in Section 6.6.5.3 and illustrated in Table
170
6-33was plotted against the clay content at each site (Figure 6-52). From this a relationship
with a correlation of 74% was obtained:
y = 0.00008 x + 0.000971
Where:
x is the clay content and y is the diffusive flux (saturated K-value).
These K-values are very low (10-4 mid) and indicate that vertical movement in the moist
irrigation soils is very slow. This equation was calculated for and will be valid for average clay
contents up to 40%, but with individual layers up to 65%.
Diffusive flux/Clay
0.0012 -
0.001
><
::I 0.0008
lO: e
Q)
-> 0.0006 - y = -0.000008x + 0.000971'jjj 2::I R = 0.738532is 0.0004 -
0.0002 -
o -,
0 10 20 30 40
Clay
------------~----~--~
Figure 6-52: Clay percentages against the diffusive flux.
o Work by Campbell (1985) revealed an equation where clay and silt fractions were
used to calculate the saturated-K. The equation reads:
o Ks = 2* 10-3e -4.26 (ms+mc)
(Where rn, is the silt content and methe clay content in fractions).
o This equation was further developed by van Tonder et al. (2002) for borehole
protection zones in the Bainsvlei area outside Bloemfontein and reads as follows:
Ks = 30*Exp(-0.3*Clay)
This formula is suited to normal field conditions where the soil does not have a high
moisture content, and the prediction of saturated K-values will be too high for irrigation
soils where the high moisture content will cause swelling of the clays, thus reducing the
flux.
171
These three equations were used in calculating the unsaturated K for different clay values
and for a silt value of 20%, which is a representative value for the soils at the irrigation sites,
and is illustrated in Table 6-36. From this it is clear that the values obtained by van Tonder et
al (2002) for the Bainsvlei soils are high for soils with low clay content, whilst for high clay
content the results are low in comparison with the other two methods. The values calculated
from the irrigation sites and those calculated by Campbell are very similar, as can also be
seen from Figure 6-53.
Table 6-36: Ks and unsaturated K-values for different clay values (silt = 20%).
Ks Unsat K
~alnSVlel - Irrigation - t::IalnSVlel- Irrigation -
Clay% van Tonder Vermeulen Campbell van Tonder Vermeulen Campbell
10 1.49E+00 8.91E-04 5.57E-04 3.87E-01 2.31 E-04 1.45E-04
20 7.44E-02 8.11 E-04 3.64E-04 1.17E-02 1.28E-04 5.73E-05
30 3.70E-03 7.31E-04 2.38E-04 3.41 E-04 6.73E-05 2.19E-05
40 1.84E-04 6.51 E-04 1.55E-04 9.37E-06 3.31E-05 7.89E-06
50 9.18E-06 5.71 E-04 1.01 E-04 2.36E-07 1.47E-05 2.61 E-06
60 4.57E-07 4.91 E-04 6.62E-05 5.17E-09 5.56E-06 7.49E-07
Clay versus Unsaturated K
70
60
50
0~ 40
>- -+-lrrigationJ
III
(3 30 l_Campbell
20
10
0
O.OE+OO 5.0E-05 1.0E-04 1.5E-04 2.0E-04 2.5E-04
L Unsaturated K
Figure 6-53: Graph of clay percentage versus unsaturated K
• Another equation by Naney, which is based on the bulk density value of the strata
(Naney et et., 1983) to determine the saturated K-value is:
Ks = 110.16*Db-22.2 mld
(This equation is also included in the spreadsheet "Unsaturated K" in Appendix D).
172
6.7 Conclusion
Geochemical modeling has indicated that the soil water in the top one meter is saturated with
gypsum, and precipitation should occur. In the deeper moist layers the water is still
undersaturated with regard to gypsum. The salt balance calculations revealed that 80% of
the sulphate applied during irrigation since 1999 (2001 at other sites) has been retained in
the soil profile (also as gypsum), most of it in the upper two meters. 80% of this retained
sulphate is in the soil water. After irrigation ceases, it will take more than 15 year to leach the
sulphate from the soil profile under optimal conditions. Due to bimodal flow and low recharge
it can be more than 100 years.
The chloride values at all the different irrigation sites showed a decrease in chloride values
with depth, implicating bimodal flow. Twenty five percent of the total volume of water applied
through irrigation reaches the water level. There is a direct correlation between clay content
of the soil and the diffusive component of the bimodal flow. Due to the moist conditions at
irrigation sites, swelling of the clay results in a very low vertical conductivity value of 10-4 mid.
This correlates well with the findings of Campbell (1985).
Due to lateral flow in the weathered saprolite zone, dilution occurs (lateral flow contributes
approximately 85% towards the dilution process). Water will move along the weathered zone,
and part of the salinity will move deeper down into the fractured aquifer, some will daylight or
seep into adjacent strata. This means that the release of salinity into the groundwater is a
very slow process, which makes irrigation a viable option over the short to medium term.
It is very important to construct the boreholes at irrigation sites correctly. Otherwise a false
picture of what really happens regarding the increase in salinity in the underlying aquifer
emerges.
173
7 IRRIGATION AT OPENCAST COLLIERIES (SPOilS)
7.1 Scenarios considered for spoils irrigation
The range of potential interactions in spoils is highly variable; the fact that the on-site
observations will not provide a complete picture, several methods can be carried out by
applying a series of laboratory investigations was undertaken, including the following:
o Leaching columns of spoil material (spoils is waste material from the mine and can be
potentially acid generating and non-acid generating)
o Humidity cells of spoil material (potentially acid-generating and non-acid generating
scenarios were run)
o Columns within the spoils
The different scenarios considered by these tests therefore included:
o Irrigation with high quality water
o Irrigation with mine water
o Irrigation with re-circulated mine water
o Inundation of spoils through over-irrigation or irrigation where water could not
drain away.
In addition to these alternatives, geochemical modelling was used to illustrate factors
of fundamental importance.
7.2 Laboratory testing of spoils irrigation
7.2.1 Humiditycells
Humidity cell tests are performed to evaluate the long-term acid-producing potential of mine
waste rock, tailings or residual ore. The test simulates the accelerated weathering of the
sample by passing moist air through the sample chamber, followed by dry air; moist air is
used for three days, followed by dry air for three days, and distilled water on the seventh day.
The leachate is analysed for pH, conductivity, sulphate, acidity and dissolved alkalinity. This
one-week cycle is typically run for 20 weeks. Samples are usually selected to represent the
various lithologies at the mine, along with a range of ABA and/or NP/AP values. Therefore,
174
humidity cell results can be used to identify or confirm the AMD risk associated with a range
of ABA values.
7.2.1.1 Methodology
A degree of standardisation has been attempted in the USA, with the adoption of the ASTM
standard methodology. In Designation: D5744-96 Standard Test Method for Accelerated
Weathering of Solid Materials using a Modified Humidity Cell (ASTM, 1996), a cell 203mm
(8.0 in.) in height by 102 mm (4.0 in.) diameter is specified for material crushed to 100%
passing 6.3 mm (crushed core or waste rock and coarse tailings).
In a typical test, a seven-day cycle is employed: three days of dry air, three days of humid air,
followed by leaching on the seventh day. The next cycle is started on the 8th day of the test by
carrying out the humidity cell tests as long as possible (as discussed in detail in Chapter 5). It
is normal for data to be quite erratic over the first few cycles, before consistent results are
obtained. This is due to the removal of readily soluble components from prior oxidation and
weathering. It is not unusual for humidity cell tests to continue for several months, or even
more than one or two years (Lawrence and Day, 1997).
The different weekly methodologies were as follows:
o Cells 1 and 4 were leached each week by the addition of mine water (first order
analogue to irrigation)
oCelis 2 and 5 were leached each week by the leachate from the columns themselves,
made up to 500 ml by mine water (to illustrate the effect of recirculation)
o Cells 3 and 6 were leached each week by the addition of deionised water (Response
of spoils under natural conditions, which forms the baseline for comparison)
o An additional two cells were filled with water and sealed off to similute flooded
opencast pit conditions.
7.2.1.2 Results
The key objectives of laboratory kinetic tests are (a) long-term stable reaction rates under
kinetic conditions and (b) depletion times for acid-generating, acid-neutralising, and metal-
leaching minerals. Test interpretations therefore focus on calculating these values.
Recommended equations by Morin and Hutt (1997) are listed in Appendix D.
175
7.2. 1.2. 1 Acid-generating spoils
The humidity cells were run for the standard ASTM 20-week period to determine reaction
rates and sequences in the potentially acid-generating spoils (obtained from Middelburg
Mines). Duplicate cells for the scenarios of clean water addition, mine water irrigation,
recycling and flooding were run.
Electrical conductivity
EC [mS/m]
1000
800
... Fklodedl
...,._Mine1
600 ..... Recyde1
_Deiori!l!d1
_ Flocded 2
_Mine2
400 ...,._ Recyde1
... DEiori9:l2
200
1212005
Time
Figure 7-1: EG of all cells over the testing period
Figure 7-1 shows that the EC changes over time. The differences observed are very
important, as they show the impact on concentrations of leachate under different regimes,
and illustrate that the inundation would result in a stable chemistry. Clean water irrigation will
give the lowest concentrations, mine water irrigation will give relatively stable values, and
recycling causes the progressive degradation of water quality to very high salt loads, while
the flooded values show a constant concentration.
Figure 7-2 shows the pH changes over time. The progressive acidification in all the cells
except the flooded ones is important in terms of expected long-term behaviour.
176
pH
pH[J
8
seceert
6 [
_".. Mine1
~ Recyde1
_DEiorise:f1
5 __ Flowed 2
...... Mine2
...... Recydel
...... Oelori:!HJ2
3 J
~12005 1012005 11/2005 1212005 112006
Time
Figure 7-2: pH of all cells over the testing period
The pH and EC response of each scenario is given by Figure 7-3 to Figure 7-6.
350
Electrical conductivity [mS/m)
300
250
200
150
100 1:~EiOrise:f1
L_D90rise:f2
pH [J
6
2
9/2005 ...1.0r.2005~ --1-:-112"00:-:'"5--- "1212005 112006
Time
Figure 7-3: EG and pH of spoil humidity cell over time (deionised water addition).
177
550
Electrical conducti~1y ImS/mJ
500
450 .:
400
350 JT
pH IJ I
7.0
6.0
5.0
4.0
1012005 1112005 1212005 112006
Time
Figure 7-4: EG and pH of spoil humidity cell over time (mine water addition).
1a00
I Electrical conductivity [mS/m]
800 I
700 I
1--- Recydel
....r Rec:yde1
pH IJ
6.0
5.0
4.0
10/2005 1112005 1/2006
Time
Figure 7-5· EG and pH of spoil humidity cell over time (recycled water addition).
178
450.,---
Electrical conducti~ty[mSlm]
400
350
1======- _Aooded1300 -.-Aooded2
pHil
7.8
7.6
7.4
7.2
1012005 ~ 11t2005 112006
Time
Figure 7-6: EG and pH of spoil humidity cell over time (cell flooded)
The sulphate response is of even greater interest and shows that, when the recycled water
contains significant ions other than Ca and S04, the sulphate can increase to very high
values over time. The other ions increase the ionic strength, thereby resulting in the activity
coefficients decreasing so that the equilibrium point for gypsum saturation becomes very high
(see also Figure 7-12).
Sulphate
S04 [mgil]
__ Flocxled1
-....Mine1
~REJ;yde1
OêOr19;!d1
_FIoa:JEd 2
_Mine2
ËRIlCyde1
Oaor19:d2 J
~12005
Time
Figure 7-7: S04 of all cells over the testing period.
179
2000
Magnesium (mgil)
1500
1000
500 .... FIoooed1
...... Mine1
..... Recyde1
_ Deiori9:d1
.... Floooed 2
.... Mine2
600 .... Recyde1
!..... DeioriSEd2
500
400
300
200
10~1200s "12005 12/2005 112006
Time
Figure 7-8: Ca and Mg over the testing period
As acidification occurs, the alkalinity of the waters decreases, as expected (Figure 7-9).
Total alkalinity
T.A1k [mgil)
100
__ Floooed1
....... Mlne1
1 ..... Recyde1
50 _Oeiori!J3dl
_Mlne2
_.. Recyde1
_ OEiorised2
~12005 112006
Time
Figure 7-9: Alkalinity decrease as acidification occurs.
The pH-drop is expected to correlate with the consumption of the available neutralising
potential. This is shown on Figure 7-10. It is also clear from Figure 7-11 that the available
sulphide outlasts the available NP by far, and if these cells are a good analogue for field
conditions, long-term acidification can be expected. It must be noted that the standard
equations for determining depletions are not valid where the effluent is mine water or
180
recycled leachate, and that the reported values are approximations based on well-
constrained assumptions.
100
Rom.lnlng_NP _ (%J I l
50
.... Floooe::l1
__ Mine'
_.,_ Rec:yde1
_Deior'isej1
..... FIocx:le::l2
pHIJ .... Mlne2
_... Recycle1
..... Dêori9:d2
~05 11...2005 1212005 1/2006
Tirre
Figure 7-10: Depletion of NP and pH over time.
Remaining_ %S
Remaining_%S [ I
100
so
80
70
60 ~
50 ..
912005 1012005 1112005 1212005 112006
Time
Figure 7-11: Depletion of NP vs. %S remaining.
The saturation indices for different parameters were determined by PHREEQC. It is clear that
the upper sulphate and calcium concentrations are limited by gypsum saturation. This is also
clear in considering the calcium and magnesium values given in Figure 7-8. It is clear that, in
the recycled cell, the Mg keeps increasing, but Ca remains relatively constant due to gypsum
saturation.
181
0.0
0.1
0.2
0.3
0.5
0.6 I .... Minel
........Recyd et
;---====----==_ ..... Daorissj10.7 J Mine2Sulphate Img/LI ...... Recyde1
...... DEioris:rj2
8000 1
::I::
0912005-------1-0I2~005 '112005 1212005 112006
Time
Figure 7-12: Sulphate vs. saturation of gypsum.
The calcite saturation in Figure 7-13 and Figure 7-14 as a function of the pH shows that the
calcite is rapidly and consistently below saturation. As the pH drops, calcite becomes
insufficient to buffer the generated acidity and remains undersaturated as the pH drops.
o -
10
15
.... Mine1
-r Recyde1
..... DEioni!B:I1
20
_ Mjne2
pH II l -.. Recyde1~Deori91:12
6
1012005 1112005 1212005 11:2006
Time
Agure 7-13: Decrease in calcite saturation vs. pH
182
SI Calcite VS pH
.:JI...",.....
•." .• ••
5.5 6.5 • 7.5 8 .... • j.5 95• •• • • •
-, ·r~l•'y• .,1• • ... • .
• ... ~••
-2 •
.• ...•.•
-3 •
• •
•
_4L_ ~
pH
Figure 7-14: Calcite saturation vs. pH
7.2 1.22 Non-acid-generating spoils
The humidity cells were intially run for the standard ASTM 20-week period to determine
reaction rates and sequences in the potentially non-acid-generating spoils (obtained from
Optimus Pit). After consideration, the test period was extended for an additional six weeks to
investigate the possibility of significant changes. Duplicate humidity cells were also run for
the scenarios of clean water addition, mine water irrigation, recycling and flooding.
Figure 7-15 shows the EC changes over time. The observed differences were compared to
acid-generating spoils, as they show the impact of the different regimes more clearly and
illustrate that inundation would result in a stable chemistry. Clean water irrigation give very
low concentrations, while mine water irrigation will again give relatively stable values, and
recycling causes an even more marked degradation of water quality, which would result in
extremely high values compared to clean water irrigation.
Figure 7-16 shows the pH changes over time. The stable pH, showing increases over time
after which it stabilises, is in stark contrast with the potentially acid-generating spoil cells.
183
Electrical conductivity
EC (mSlm]
700
600
r_ FloOOl
1 -e- MrlEMEier3
500
..... Recyde3
Deiaised3
-MIlfMEier1
400 ~ Recyde1
~ Oeicrlsad1
_ Fkxx!2
300
200 l:~~~ J""'lP Fl0003- D;X1nun"'l-er
100
o ,~ ,
312005 412005 512005 612005 712005 8/2005
Time
Figure 7-15' EG of non-acidic cells over time.
pH
pH[]
9.0
8.5
~Flocxl1
..... MineNcie!3
__ Recyde3
8.0
_"'Deiori~3
___ Mina.va:erl
...... Recyd.1
7.5 ....... Deiorisa:l1
_Flocxl2
~MineNaeJ2
7.0 ... Recyde2
.... Deiori5ed2
_Flocxl3
6.5
6.0
312005 412005 512005 612005 712005 8i2005
Time
Figure 7-16: pH of non-eddie cells over time
The pH and EC responses of each scenario are given by Figure 7-17 to Figure 7-21.
184
80
70 Electric conductivity !mS/m]
60
50
40
30
20
10
.... Deiorised3
----- ,.- -r Deiorised1
I ..... oecnsecz
pH!]
8.5
8.0
7.5
7.0
6.5
6.03/2005 412005 512005 612005 712005 812005
Time
Figure 7-17' EG and pH of SpOlYS humidity cell over time (deionised water addition).
1:0
Electricalconductivity [mSlmJ
380.0
370.0
360.0 .
350.0
340.0
330.0
320.0
310.0 ~ Mine1 1
I ....... Mlne2pH [J
8.13 I
7.87 i
7.601
3/2005 512005 612005
Time
FIgure 7-18: EG and pH of spohs humidity' cell over time (mine water addition).
185
400
Electrical conductivity [mS/m]
350
300
I_M""""113 --
250
pH[]
8.1
7.9
412005 5/2005 612005 712005 812005
L~/2005 _
Time
Figure 7-19: EG and pH of spoils humidity cell over time (mine water addition).
700
Electrical conductivity [mS/m]
600
500
400
300
_ Recyde3
200 -=l===========:_ ....... Recyde1
-.. Recyde2
pH[]
8.5
8.0
7.5
7.0
4/2005 5/2005 6/2005 712005 8(,1005
Time
Figure 7-20: EG and pH of spoils humidity cell over time (recycled water addition).
186
400
Electrical conductivity [mS/m]
350
300
250t=~~------~--~==~======~======~==--~
pH[]
8.2
7.9
7.7
~4/:20OS-~ 512005 6/2005 712005
Time
Figure 7-21: EG and pH of spoils humIdity cell over time (cells flooded).
From closer scrutiny of the different scenarios it is clear that, where the spoils are non-acid
generating, irrigation with mine water causes a severe deterioration in quality, compared to
clean water irrigation. Recycling the water results in a ten-fold increase of acid generation
over clean water irrigation, as opposed to a factor of three to four for the acid-generating
spoils.
The sulphate concentrations for the different scenarios are shown on Figure 7-22. The sharp
contrast between the eventual sulphate in these cells and the acid-generating cells is worth
highlighting. The flooded cells and mine water cells have similar values, showing that, where
the spoils do not generate significant sulphate, no other ions are released, and if throughflow
rates are sufficiently high, the effect of secondary minerals is such that sulphate values will
be in almost the same order as the input quality. The upper limits for each value are two to
three times less than for the acid cells.
The calcium and magnesium values show the stabilising effect of gypsum solubility on
calcium and the steady increase in magnesium. The values recorded for recycling, for
example, are two to three times lower than for the acid-generating cells.
187
Sulphate as S04
504 [mgil]
6000
5000
4000
-r Mnewater1
3000 ... Mnewater2~~~~~~_q:,tirrumrT>fnwaterj -+- Mnewater3- RecyCledl
_ Recycled2
2000 .... Recycled3
1000
..,,_.
o • 1;;::1 • • • • • .~.
312005 412005 512005 612005 712005 BI2OO5
Time
Figure 7-22: Sulphate concentrations for the non-acid-generating spoils humidity ce/Is.
800 r
fv1agnesium [mgA...]
600 I
400 ~~5I~~==~*=~~~~~~~~;:~~;;=;;;~ ~A~1
_M~er3
200 -.- Recyde3
200 1
100
~12005 412005 Sf-WOS 6/2005 712005 812005
Time
Figure 7-23: Ca and Mg values for non-acid ce/Is.
The depletion of the neutralising potential in these cells is far lower, as expected, with almost
none of the available NP depleted during the period of testing (Figure 7-24). It is important to
note that the depletion occurs over the extended 26-week period rather than the standard 20-
week ASTM testing period.
The increase in pH over the testing period supports this, and indicates the importance of
placing spoils irrigation only on non-acid-generating spoils.
188
Remaining_NP _
Remaining_NP _ [%]
100
95 ~e"cyde3eiori~3-~'
Mine.vaer1
...... Recydel
~ Deiorisaj1
-r Mine.vae12
...... Recyde2
90
~Oeiori~2
85
312005 412005 512005 612005 7áoo5 812005
Time
Figure 7-24: NP depletion in non-acid-generating spoils over time.
9.0 ~
pH [1
8.5
8.0
7.5
7.0 ...... MineNaer3
___ Recyde3
6.5 ....... Deorised3
~ Min9Mi:er1
6.0 ...... Recyde1
.... DEiori9:d1
Remaining_NP _ [%1
...... Mlne..véiel2
...... Recyde2
....... DeoriSfd2
95
90
4/2005 512005
Time
Figure 7-25' pH and NP depletion.
The saturation state of these waters was determined using PHREEQC. This shows that, over
the testing period, the gypsum saturation maintains sulphate and calcium concentrations at
relatively low values compared to the acidic cells. It is important to note that the mine water is
gypsum-saturated, as the spoils are initially. This is maintained with all mine water (Figure
7-26).
189
6000
Sulphate [mgil) I
5000
4000
3000 ~~:=:=J
2000 .... Mroe.oa8"3
_ R"",de3
1000 ..... De
r
... hl • • • • • • • • • • • • • • • • • • • • • • ..... ::la"i:sede3 ~
... Oecnsed1
_...Mr19.'\BI:e12
_ Recyde2
1.0
2.0
3.0-
612005
Time
Figure 7-26: Sulphate concentrations and gypsum saturation over time
The saturation of carbonates shows that, in contrast to the acidic cells, calcite and dolomite
are at or above saturation throughout the reactions (Figure 7-27). The fact that these values
remain at or above equilibrium shows that the system is not under stress from an acidification
perspective, even for the extended testing period.
Dolomite and Calcite Saturation vs pH
... •
5.. :<:1'-;.:.~:-
-5 .... • :t!.'~....••• I.,.,., .. .,
.... . • si Calcite l
• .. • ••• • Si_Do'Omit~
·10
•..~'"• • •
• •·15
•
\
·20·L-----------------------------------------------~
pH
Figure 7-27: Dolomite and calcite saturation vs. pH over the testing period.
190
7.2.2 Leachingcolumns
7.2.2.1 Methodology
Three duplicate sets of 2 m columns with a diameter of 160 mm were packed with raw spoils
obtained directly from the mines. After thorough mixing, equivalent amounts of spoils were
sequentially loaded into each column by hand.
The columns were loaded until each contained 50 kg of material. Initially the cells were
leached with 5 litres of water to completely inundate the column and leach out the
accumulated oxidation products.
The different weekly methodologies were as follows:
o Columns 1 and 4 were leached by the addition of 1 litre of mine water each week.
o Columns 2 and 5 were leached by the the leachate from that column, made up to 1
litre by mine water each week.
o Columns 3 and 6 were leached by the addition of 1 litre of deionised water each
week.
The EC and pH of each column were measured weekly with full analyses of the leachate in
each column every four weeks.
It is very important to note that the water to spoils ratio in the columns is far lower than in the
humidity cells. The reaction progress is therefore much slower and the secondary mineral
influence becomes more pronounced (Morin and Hutt, 2001). These columns are, however, a
better real-time analogue to the field, since the weterspolts ratio and rates of water addition
were adapted to mimic field conditions.
7.2.2.2 Results
7.22.2 1 Initial data (Potentially non-acidic spoils)
Acid-base accounting was performed on the spoils obtained from the Optimum site to obtain
an indication of the acid-generating potential of the spoils. Results are given in the Table 7-1
below:
191
Table 7-1: ABA results from spoil columns.
Initial pH Acid Potential
Final pH (Open) NP NNP (Open)
kg/t CaC03 kg/t CaC03 kg/t CaC03
7.66 3.50 1.84 10.49 8.65
7.33 2.76 3.51 4.76 1.26
7.56 3.41 13.09 14.43 1.34
7.67 4.14 4.86 16.41 11.56
From the results, it is clear that, under open conditions such as those in the laboratory, the
spoils would be expected to have some potential to acidify, but this potential is relatively low.
Furthermore, according to accepted criteria of ABA interpretation, the spoils fall into the
uncertain zone, where ABA alone cannot give a high-confidence assessment of the acid
generation from the spoils. The result is that the initial leachate has very high concentrations
of different ions, particularly sulphate and the associated Ca and Mg from buffering by calcite
and dolomite.
Initial leachate results are given in Table 7-2.
Table 7-2: Initial leachate results from columns.
No. Date pH EC Ca Mg Na K MAlk Cl 804
WEEK1
LC1 29-Mar-05 7.24 657 560 1237 104 55.8 464 238 5581
LC2 29-Mar-05 7.46 588 551 1249 91 51.2 296 232 5632
LC3 29-Mar-05 7.83 505 486 1078 80 42.1 278 55.4 5227
LC4 29-Mar-05 7.21 664 557 1291 89 49.8 329 237 5678
LC5 29-Mar-05 7.76 660 662 1238 87 54.7 214 247 5856
LC6 29-Mar-05 7.28 659 576 1329 81 48.7 322 208 5740
The mine water used to "irrigate" the spoils was obtained from the Lapa Dam at Optimum,
with the quality indicated by the table below:
192
EC2981 K I MAlk I26.3 231
7.2.2.22 Results
The EC and sulphate variations over time revealed a very interesting phenomenon, namely
that, where clean water irrigation is employed, the values decrease over time. The reason for
this lies in the fact that the spoils were oxidised prior to the experiment. Since they represent
the results of the in-situ spoils at Optimum, it can be expected that oxidation occurred over
time in these spoils and the secondary products from oxidation reactions and the subsequent
neutralisation would be contained in the spoils.
EC Lab
700 ,-----------------------------------,
600 -+-LC1
500 __ LC2
E
(ij 400 ___ LC3
.§.
o 300 ~LC4w 200 -lIE-LC5
100 ___ LC6
o L- -- ----~
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
.):-'1><: ~'1><: :,;.,.<f .):-'1>~ _...'.1>~ -::,v~ .':l .':l ve;, 0fQ~ 0fQ'Q pC}
'\)0;/ ~. ,,'I:) '\)'1:) 0,.'I:)"'~: ~ tV "eo)?' '\)~. 0,.~' ,,~
Date
Figure 7-28: EG variations in columns over time.
Leaching columns Sulphate -+-Mjnewate~
___ M ine return water1
-6- Deio nized water1
6500 ,------------------------------------1 __Minewater2
;g: _""'_M ine return water25500 __ Deionized water2
E -Optimum column4m
C 4500
:f8! 3500
+'
~ 2500
(s,) 1500-
o
500
<:)" <:)"
l <,_# ,,"
,,<:)" ,,"
9:f~<' c!'~
... ~'I>'"'" ·v"'"<:l ')9;:f ... <:)"i
Date
Figure 7-29: S04 variation in columns over time.
193
These findings are mirrored in the results of the Ca and Mg provided in Figure 7-30 and
Figure 7-31
-- --+- Mine water 1
Leaching columns-Calcium _____ Mine water return
- ...... Deia nized water1
~MjneVvater2
700 ~M ine water return_'_Oeio nizedy.gter2
650 ,~ --t--Optimum columJ 600 -Optimum columnClE 550 [I:0 500 __ . "'" - -;-lC'G 450lo.I: 400 .~
uCII 350 - -I: ~-)0.
o0 300 -_ , -250
200
Date
Figure 7-30: Ca variation in columns over time.
I leach ing columns- Mag nesium ~::: :::;~,"m111-.- Deionized water1- 1500 -r--------------------~ ~ Minewater2.-..J 1300 , ~ Minewater return2Cl)
-E l=:*- Deionizedwater21100C
0 900
'+
.n.
i
...s.. 700
C
Cl)
o 500
C
0 300
U
100
Date
Figure 7-31: Mg variation in columns over time.
The gypsum saturation over the same period is of interest and shown below:
194
Gypsum saturation in leaching columns
0.2,-----------------------------------------------------------------,
0.1
O·
-0.1
--L1
E -0.2 --L2
sV"I --L3
(ii -0.3 I=~:J
-0.4
-0.5
-0.6
-0.7 L- ---'
29-Mar-05 05-Apr-05 03-May-05 31-May-05 28-Jun-05 26-Jul-05 29-Aug·05 20-Sep-05
Date
Figure 7-32: Gypsum saturation in leaching columns over time.
It can be seen from the above that the deionised cells become undersaturated as secondary
products are flushed out, but where leachate is recirculated or mine water used, the values
remain close to saturation.
0.2 r-------------------~----------------------------------------,7OOO
0.1
6000
-+ -+- I _-+-
29-Mar 31~ -2S"'un- 2e!!Ji:r 2~g -Z(}!Sep
_ 5000
-0.1
·0.2
4000 1~. ~Lo~
-0.3 3000 -L6_S04
-+1 -LS_S04
-0.4
2000
-0.5
1000
-0.6
-0.7
Figure 7-33: Sulphate concentrations and gypsum saturation in columns.
195
As can be expected in these columns, the carbonates are not stressed in the reaction period
and calcite and dolomite are at or above expected equilibrium.
Calcite saturation in leaching columns
29-Mar-05 05-Apr-05 03-May-05 31-May-05 28-Jun-05 2B-Jul-05 29·Aug-05 20-Sep-05
Date
Figure 7-34: Calcite saturation in columns.
7.2.2_3 Potentially acidic spoils
The column tests were repeated with potentially acid-generating spoils. All the caveats
outlined for the previous columns hold in this case as well. The most important are the ratios
between the rock and water and the resultant limited reactivity of the system. Columns are
excellent real-life analogues, but are limited, as they cannot provide the entire reaction
sequence within the usual test period.
Acid-base accounting was performed on the spoils used in the leaching columns to obtain an
indication of their acid-generating potential. Results are given in Table 7-3.
From the results, it is clear that, under open conditions such as those in the laboratory, these
spoils are expected to acidify.
196
Table 7-3: ABA results from the spoil columns.
Acid Potensial
Initial pH Final pH (Open) NNP (Open)
kg/t CaC03 kg/t CaC03
7.62 2.700 76.796 16.22
7.43 2.580 32.806 -20.34
7.84 2.150 58.377 -22.15
7.45 1.790 57.209 -37.52
A few parameters are shown to contrast with the equivalent humidity cells. In the pH
variation, neutral conditions are established and maintained in the short term (Figure 7-35).
This indicates that, although the potential for acidification exists, this is not immediately
manifested. This is borne out by the results of geochemical reaction modelling in Section
7.2.4.
pH - Acidic spoils l
7 -
6.8
6.6 =~ -+-LC1
6.4 --LC2
6.2
::I: I--.-LC3
C. 6 ~LC4
5.8 - --.-LC5
5.6 -+-LC6
5.4
5.2
5
05-0ct-05 24-Nov-05 13-Jan-06 04-Mar-06
I
Date _j
Figure 7-35' pH recovery in columns over time.
The EC (Figure 7-36) and sulphate (Figure 7-37) shows that, as in the previous columns,
there are initial high concentrations that decrease as the accumulated reaction products are
flushed out of the system. This fact alone makes columns less than ideal for understanding
what would happen in the very long term during spoils irrigation. It does indicate, however,
that the initial salt loads in such a system would be very high, with stabilisation occurring in
time.
197
EC- Acidic spoils
1400
-
eE.ns FL~~
U I:=LC2w ---.- LC3
O~-------r----------------,,------~ ----*- LC4~
-+-LC5
~<:;)<o ~ ,;:><0 ~ ,;:><0 ,;:><0 ,;:>CO ,;:>CO ,;:>co f.,.,;:>co f.,.,;:>CO
ei ~o 1"~o eP 0-~ \§? x~ ~'?j ~'?j
I • LC6
,,<00 <:;)bt "bt<::J <:;)":J''!, '),.0;[ ,,'1: <:;)bt '"lY
Date
Figure 7-36: EG in columns over time.
Sulphate - Acidic spoils
16000
14000
4
-::J 12000 I • LC~.Csl 10000 ! ----- LC2
..C...l 8000 --
-.-- LC3
CU ----*- LC4
.c::
c. 6000-::s 4000 -- ~..... ~ -...-- .. -+-LC5CJ) ~~ ~.. ____ LC62000
0
10/05/05 11/24/05 01/13/06 03/04/06
Date
Figure 7-37: Sulphate concentrations.
7.2.3 Field leaching columns
A deep trench was dug in the spoils (Figure 7-38) at one of the pivots at Optimum to install
leaching columns of 250 cm in diameter. Three columns were buried vertically in the trench,
one at 5m, one at 4m and one at 2m. The columns were sealed at the bottom, and filled with
30cm of silica gravel to prohibit the leachate water in the bottom with the spoils reacting
chemically. A small tube with a perforated bottom was installed in each of the columns for
sampling purposes (Figure 7-39 and Figure 7-40).
198
The columns were then filled with spoils in the same sequence as they occur in the pit from
which they were dug, and compacted. Unfortunately the 5 m column leaked, and only the 4 m
and 2 m were sampled over time.
Two 4m experimental columns were installed in a hanger at the University of the Free State,
also filled with spoils from the pit, and with small taps at the bottom to drain water. The aim
was to leach irrigation water through the one, and de-ionised water through the other after the
spoils have compacted, in order to determine the hydrochemical and hydraulic interactions
through direct observation and measurements.
Figure 7-38: Pit in the spoils, il/ustrating the heterogeneity of the spoils.
A'gure 7-39: Instal/ation of leaching columns.
199
Figure 7-40: Installed leaching columns at Optimum pivot.
7.2.3.1 Results
The results indicate that very high initial sulphate and salt loads can be expected (Figure
7-41). This was consistent with the findings in the laboratory columns. The differences
between the 2m and 4m depths should be noted and are a result of both the mine water
addition at the surface and, more importantly, the greater pre-irrigation reaction in the shallow
depths where pyrite oxidation can more easily be accomplished due to optimal oxygen and
moisture contents at these shallower depths.
--
Sulphate
S04 IJ
5500 I l
5000
4500
4000
3500 -
412005 612005 812005
Tlrre
Figure 7-41: Sulphate values from field columns over time (mg/L).
200
The pH response was similar to that observed in the laboratory columns, and showed that
neutral conditions will exist until the available NP has been depleted.
pH
7.5
1__CPT2M I
l:_CFT~
sj ~-_j
412005 612005
Time
Figure 7-42: pH in field columns over time.
The gypsum saturation profile over time was consistent with the explanation of pre-existing
oxidation products, and it is clear that the columns were initially oversaturated with gypsum,
which moves to equilibrium over time.
Gypsum Saturation in Field Columns
0.1,-----------------------
0.08
0.06
0.0<4
0.02
-__+_-2m4m_l[
05
-0.02
..(I.D4
-0.06
-0.08
-0L.1- --'
Date
Figure 7-43: Gypsum saturation in field columns.
A comparison of the field columns with laboratory data showed a fairly good correlation, but
suggested a large amount of pre-existing reaction products in the 2m column, which
correlates well with the initial values in the laboratory. Due to the lower water ratios, this will
persist for a lengthy period.
201
Figure 7-44: Comparison of field and laboratory sulphate values.
7.2.4 Discussion
The laboratory and field experiment results are vital for understanding different scenarios in
the field. The humidity cell results provide a broad indication in terms of the expected full
reaction sequence, therefore providing the best experiments for the most likely fundamental
interactions. The laboratory columns are useful as they provide a better understanding of the
true rates at which the processes occur and thus provide a closer "real-time" analogue of the
field situation, although the volumes of water added reflect a far longer time of irrigation than
is usually the case in the field. The good comparison with the field situation is encouraging in
this regard.
7.3 Hydraulic response of the spoils
At the Optimum Spoils, tensiometers were installed to measure the soil moisture variation
over time. Unfortunately the deeper tensiometer at 4m malfunctioned and only data from 0.8,
2 and 3 m could be interpreted. From the high rainfall period during late summer it is clear
that the spoils were quite wet, especially at 2m where the tension was close to zero. After the
rain has stopped, and irrigation was applied twice a week at 10mm per cycle (and later at
15mm), the spoils started to dry out and the tensions at the different depths increased,
especially at 3 m. Even the 0.8 m tensiometer started to dry out during July to August but
once again increases in wetness when more irrigation (15mm) was applied. This indicates
that controlled irrigation during wintertime will not enhance increasingly moist conditions,
which is conducive to sulphate generation (See Section 7.4.2.3 for modelling of irrigation with
202
reduced recharge). The drop in tension and the subsequent increase in moisture at the end
of October are attributed to the fact that the pivot broke down above the tensiometer
installation, resulting in flooding in the area for more than a week. Since then, when the
rainfall started again, all three depths remained constant in moisture content, which is the
same as during late summer.
From these data it is clear that during summer time, especially when rainfall is high, there is
an increase in moisture in the spoils. If controlled irrigation were applied during winter, it
would not have a negative effect on the surroundings (winter pastures need a lot less water
to grow effectively). Less irrigation over a larger area will thus be the answer in reducing
excess water at opencast collieries.
Optimum Tensiometer
2000 ~----3-(1-0~0m--m--------------------------~ 40
- 1500 --- 2000mm 35-800mm ~E 1000 .... 30- o precip~8tio-;.._;r ~ - ;;z- ~: ....~_E 500 1--I~'_~~'10~--= ..:._..-- ~ ~ 25c:
0 0 20
1cii -500 15
~ -1000 10
-1500+-~~--~--~-~--~-~-~--~--+ 5
-2000 o 0 0 0J-:;0_;;,jA •,•.....;:.,0 ...o. ~~o ..;.0l.,a, .. ::-. ........ ...l... 0
~,\)":.> ...Q('\)~ '8l,1f> ~If> ..,.".'elf> ~If> ~If> rrlf> o:\)~ N,1f>
~~ &r" &~ \9-"Y \9-",)'" ej\_i'-V ~,\_i'-V ~\:/:P ~CJ'0 'l,.Of~
Date·
Figure 7-45' Optimum spoils tensiometer data.
7.4 Geochemical Modelling
Geochemical modelling can be used for various reasons, such as:
• To characterise and interpret current contaminant load.
• Environmental assessment (source and receiving environment).
• To predict future contaminant concentrations and loads.
• To assess future treatment needs.
• To compare management and decommissioning options.
203
To support this, Lichtner (1996) states: "Computer models can provide, if not a direct
quantitative description, at least a far better qualitative understanding of the geochemical and
physical processes under investigation than might otherwise be possible."
The geochemical model, PHREEOC (Parkhurst and Appelo, 2001) was used to assess the
minerals that exert control over the observed water quality. The results were entered into
PHREEOC, so that an indication of the saturation state of each sample, relative to common
minerals within the South African coalfields, could be determined. The approach was the
widely recognised saturation index determination, wherein the saturation state of each water
is compared to the theoretical solubility of the minerals, as defined by its Kspvalue.
Work in other coal mining areas in South Africa indicates that the carbonate minerals, calcite
and dolomite are important controls on the pH in water, and that secondary minerals such as
gypsum and ion hydroxides are important in controlling the concentrations of these species.
The evaluations for these waters are presented in the sections below.
To obtain a generalised indication of what could occur at a typical opencast spoils site, a
case study from Usher (2003) is discussed. Geochemical modelling was applied, using
PHREEOC to speciate the water and determine saturation indices. At the encountered
~
colliery, gypsum saturation acts as a control on the maximum sulphate conc~ntrations,
whereas calcite and dolomite are the principal buffering agents as far as pH is concerned.
Eary et al. (2003) used a similar approach, in terms of ABA, saturation considerations and
simplified modelling, to characterise the hydrochemistry of a coal mine.
Geochemist's Workbench (Bethke, 1996) was used to obtain a profile of the expected water
quality over time. The data used in this simulation are derived from the average acid and
base potentials for the ABA boreholes and extensive mineralogy on the 73 core borehole
samples. From this detailed mineralogy, the principal minerals are characterised as quartz,
kaolinite, illite and montmorrilinite, with lesser quantities of K-feldspar, calcite, dolomite, pyrite
and Ca-plagiocase minerals. The data used in this simulation are derived from the average
acid and base potentials for the ABA boreholes and detailed mineralogy on the samples. The
values used are: AP: 2.8% pyrite; NP: 2% dolomite; clay (kaolinite with some illite) in excess,
in the presence of oxygen and water. The results demonstrate the expected sequence of
events, but there were no detailed determinations of the reactive surface area. Oxygen is
only fully available on the outside of rocks in the spoil. The rocks are saturated with water on
the inside and pyrite does not oxidise. For oxidation to take place inside a rock, it has to
disintegrate through weathering. This could take many years, depending on specific
circumstances. Based on these principles, the conclusion is that the outside of a spoil
boulder (at excess acid potential and above the water table) could acidify within a matter of
204
less than ten years, while its inside would still be alkaline. Trenches dug into existing spoil as
part of research for a WRC project by Usher et al. (2003) confirm this observation.
Time
Figure 7-46: Expected concentration profile over time (using the average mineral assemblage
determined by this study and an aquifer transmissivity of 1rrt/day).
Time
Figure 7-47' Expected pH profile over time.
It is worthy mentioning that within 12 months of this study, excess mine water flowed from the
control area. The quality of this water was in excess of 3 000 mg/L S04 at a pH of around 2.5.
It is clear that using a properly constrained conceptual model in conjunction with other data
205
can lead to more accurate predictions of water quality, even when key parameters have not
been measured. An important caveat is that the time graph is not specific; the profiles
indicate the expected evolution, but not its timing.
For any particular spoils area, it is possible to generate similar hydrochemical evolution
diagrams, if the correct data are collected.
7.4.1.1 Neutralisation of AMD
The expected mechanism of AMD neutralisation in the spoils is of great importance, since it
will show the effect of irrigation with mine water.
The system attempts to buffer acidity, generated by AMD, as far as possible. In both the
underground and opencast areas, minerals such as calcite and dolomite, which occur within
the Karoo lithogies as cementation material or veiniets (observed in shales, coal and
sandstone) and the alkalinity of the influent water, provide the necessary buffering (Usher,
2003).
Calcite neutralisation reaction (calcite dissolution by sulphuric acid):
H2S04 + CaC03 = CaS04 + H20 + CO2
Dolomite neutralisation reaction (dolomite dissolution by sulphuric acid):
2H2S04 + CaMg(C03) 2 = CaS04 + MgS04 + 2H20 + 2C02
In the opencast, the influent water consists of rainfall and fairly inert groundwaters with
limited alkalinity, while the irrigation mine water may be either devoid of, low, or high in
alkalinity, such as at Syferfontein.
The concentrations obtained, also provide a vital understanding of the expected system
response. Depending on where along the early part of the reaction path the system is when
sampled, completely different ion ratios will be obtained. In the early part of the reaction path,
the system contains a typical Na-HC03 water, showing an enrichment in sulphate relative to
bicarbonate (Figure 7-48). The sodium content remains constant, with the water becoming
progressively more calcic as the reaction proceeds. As the reaction continues in the longer
term, sulphate and calcium will be constrained by gypsum precipitation for as long as calcium
is released by calcite buffering. When this ceases, the system rapidly acidifies and sulphate
concentrations rapidly rise, since no more gypsum can be produced to remove sulphate from
solution (Usher, 2003).
206
Md:ar ratios
UOE'{)l
121lE'{)1
1WE'{)1
.<:0..s. eWE'{)2
~:;
<:
8...... 6WE'{)2
i
'WE'{)2
2WE'{)2
ODOEiOO
0 100 200 an ~ 500 600 700 BOO 900 1000
stap.
Figure 7-48: Molar concentrations throughout the reaction path.
7.4.2 Reaction modelling
Simplified reaction modelling was performed to elucidate the most important factors likely to
be responsible for water quality changes in spoils. Variations in spoils geochemistry and
irrigation were considered.
7.4.2.1 Base case
A relatively simple base case of spoils irrigation using Geochemists Workbench was
modelled. The major reactive minerals are considered to be pyrite as acid generator, and
calcite with a minor proportion of dolomite as carbonates acting as neutralisers. Associated
with these minerals are a large proportion of quartz, clay minerals in the form of illite with
some montmorrilinite, and feldspars (K-felspar as major specie). The expected recharge from
irrigation was added to the system in the form of pure water. Oxygen and carbon dioxide
were initially allowed to equilibrate with the atmosphere and it was assumed that a free
exchange of gases could occur in the system.
For the first simulation, a slight excess of pyrite is assumed, with calcite and dolomite making
up an almost equivalent mass of neutralising capacity.
Assumptions have to be made regarding the reaction surface areas. Consequently the values
for concentration should be regarded as illustrative guides of expected behaviour. Time
periods should also be regarded as relative and not absolute.
207
The expected water concentrations over time (Figure 7-49), the associated mineral saturation
variation for the carbonates and gypsum (Figure 7-50 and Figure 7-51), the depletion of
reactive minerals (Figure 7-52), and the associated pH-response are provided (Figure 7-53).
This is the baseline against which all the variations are considered and the comparison to
analogous laboratory simulations must be viewed with these data. The fluid variation shows
relatively high values of 804, Ca and Mg, as seen in the field and in the cells, but also a
decrease over time as the reactive minerals become depleted. The Ca and 804 values are
furthermore limited by the saturation of gypsum in particular.
~ 1500
'0>
-E
cCl)
Q)
c
Eo 1000 -
ou
]2
::::l
;;:::
Q)
Eo
CJ)
I
O~~~I~-H~~:O~;WI~_K_+~~~~aC~+:~I::===_=_~_~_~~_~__J
o 5000 1e4 15000 2e4 25000 3e4 35000
Time (days)
Figure 7-49: Concentrations of major ions over time.
208
~.-.. -5-a-
0)
0
::::..
+
+ro -10
o
~
c
~
ë -15
0
.~
-....:Jro
Cl) -20
10 20 30 40 50 60 70 80 90 100
Figure 7-50: Associate mineral saturations.
0 ,
ICalcite Gypsum
~te
S2' -5
a-- l
0)
.
+-
Q--
+ro -10
o
~
c
:2
ë -15
0
~-.....:Jro
Cl) -20
-25~--~----~----~--~----~--~----~----~--~----~
50 60 70 80 90 100
Time (years)
Figure 7-51: Latter part of reaction sequence mineral saturation.
209
7il-----_j~
I
0..
5 10 15 20 25 30
Time (years)
Figure 7-52: pH variation.
From the pH variation, it appears that acidification is unlikely. However, a consideration of the
relative depletion of pyrite against the carbonates (Figure 7-53) suggests that acidification
may occur. For this reason, the model was rerun as described in the following section.
50
45
40
<il 35
E
~
.9 30
(f)
~
acl
'Ë
al
E
0
Cl)
o 10 20 30 40 50 60 70 80 90 100
Time (years)
A'gure 7-53: Decrease in reactive minerals over time.
7.4.2.2 Base case run for longer period
In the light of the observed pH response and depletion of the carbonates with relation to
pyrite, it was decided to run the reaction for a longer period oftime. It can be seen that, due
to the elevated recharge from clean water irrigation, the concentrations subside with time.
210
~ 1500
e.sn
CJ)
ë
calo
~ 1000 -
oo
:g
:::J
;;:::
al
Eo
Cf)
50 100 150 200
Time (years)
Figure 7-54: Concentrations of major ions over time.
The pH response is the most important, as it shows the dangers of arbitrarily defining a
reaction period and then baking a decision on such a finding (Figure 7-55). This
demonstrates that acidification would occur due to the persistence of pyrite over the
carbonates, as shown in Figure 7-56.
7
6.5
6
5.5
I
0..
5
4.5
4
3.5
3
50 100 150 200
Time (years)
Figure 7-55: pH response showing reaction sequence effects over extended time.
211
50
45
40
Ii) 35
E
~
.9 30
!/)
(ii rite
IcD
E
Q)
E
0
Cl)
o 50 100 150 200
Time (years)
Figure 7-56: Decrease in reactive minerals over time.
7.4.2.3 Base case with recharge reduced
The base case assumes a certain amount of increased recharge. One of the key factors is to
consider the effect of diminished recharge on the spoils chemistry. The effect of a 50%
decrease in irrigation recharge is illustrated in the figures below (Figure 7-57 to Figure 7-59).
As expected, the concentrations are greatly increased, but not doubled, due to the saturation
effects previously discussed.
50 100 150 200
Time (years)
Figure 7-57 Concentrations of major ions over time with reduced recharge.
212
O~~GYpSum
sz -5 l
(5
oOl=
+
+ro -10
CJ
"3
c
~ -15
co
~
r:o::J
Cl) -20
50 100 150 200
Time (years)
Figure 7-58: Associate mineral saturations with reduced recharge.
One of the key points to highlight is that the recharge reduction has very little effect on the pH
profile. Only if the recharge contains a significant amount of alkalinity or acidity will this
significantly alter the pH profile.
I
a.
50 100 150 200
Time (years)
Figure 7-59: pH profile with reduced recharge.
213
7.4.2.4 Inundation resulting in order to lower pyrite rate
The effect of inundation on the spoils chemistry is also of importance. Here it is assumed that
the initial system is started with the same conditions as previously but with oxygen excluded.
It was shown that pyrite oxidation decreases due to the exclusion of oxygen and the
decrease in activity of Thiobaccilus ferrooxidans when oxygen values are depleted.
The data in Figure 7-60 show that T. ferrooxidans activity ceases when O2 is depleted.
However, bacterial activity resumes even under extremely small increases in O2 content, and
reaches a maximum at approximately 0.01 mole or 1% of O2 (Jaynes et a/., 1984).
1.0
.8
.->--
-
> .,6
u
c:r
Q)
> .4
-0
Q)
~ .2
0
0 .21
O2 Mo le Freet! 0 n
Figure 7-60: The influence of O2 concentration on relative activity of T. ferrooxidans (Jaynes
et a/., 1984).
The effect of this response was simulated and the results indicate that, if the pyrite oxidation
rate or activity is inhibited, the system will dilute (Figure 7-61). Pyrite oxidation and the
resulting acidity, sulphate production, neutralisation and solubility effects are the principal
drivers in the system. When the pyrite effects are reduced, the other minerals also react far
less aggressively.
Equally important is the fact that, under these conditions, the pH will very quickly become
neutral and then increase as the carbonates equilibrate with the available Pco2 (Figure 7-62).
The result is that the system reaches equilibrium with calcite and will remain neutral. The
calcite and dolomite are also not depleted rapidly, since the system is not placed under
stress by the added acidity under full pyrite oxidation kinetics (Figure 7-63).
214
l
0ICt'o---N-a-' '-I0----- IK I20-----30-----4-0----5-0-----eo-----7-0----e-o-------- 100
Time lyears)
Figure 7-61: Decrease in concentrations at lower oxidation rate.
:r:
Q.
7.2 -
7.15 -
o 5000 1e4 15000 2e4 25000 3e4 35000
Time (days)
Figure 7-62: pH recovery If pyrite rate is decreased by an order.
215
-.2 -
z
0
-.4C
gOl Calcite
ocu -.6 -
}
c: ~ypsum
~ -.8
ë
0
~:s -1
ennl
-1.2
-1.4
0 5000 1e4
Time (days)
Figure 7-63: Mineral saturations at low pyrite oxidation rate (flooding conditions).
7.4.2.5 Base case with low NP simulation
As a variation of the base case, a situation was modelled where the pyrite greatly exceeds
the Ca-containing carbonates. It is clear that the concentration of sulphate becomes
significantly higher as does the resultant TOS, and that the pH-drop is rapid. This is shown on
Figure 7-64 and Figure 7-65 respectively.
2500
~.s 2000
.!!l
c
Q)
cog- 1500
o
CJ
'0
'S
"" 1000 -
Q)
E
eon
, I I I I I I I I
50 60 70 80 90 100
Time (years)
Figure 7-64: Aqueous concentration in low NP conditions showing an increase in sulphate
and therefore in TDS.
216
7 -
6.5 -
6 -
:c 5.5
-
0..
5 -
4.5
4
5000 1e4 15000 294 25000 3e4 35000
Time (days)
Figure 7-65' Rapid pH-drop.
Mineral saturation is of interest, as it shows the rapid undersaturation of dolomite and calcite
as conditions become acidic. The gypsum remains relatively close to equilibrium for a long
period of time, but due to the depletion in calcite and dolomite, this becomes undersaturated,
as seen below.
o _idi! G9S§d1iJ
Q'
CJ
gCl
5000 1e4 15000 294 25000 3e4 35000
Time (days)
Figure 7-66' Mineral saturation showing depletion of carbonates associated with sadsïcenon.
217
7.4.2.6 Base case with high NP simulation
Alternatively, NP greatly exceeds the acid potential, as indicated by the pyrite content. This
result in the most stable chemistry due to the fact that sulphate and calcium are limited by
saturation, while calcite and dolomite are not rapidly depleted, giving a stable pH profile over
time (see Figure 7-67 and Figure 7-68).
4500 r- I I I I I I I I I I I I I I I I I I 1_
~ -
4000~ -
~ ~ -
,~r- -
~ r- -
~~r- -
~ r- -
c8. 2500 CSO- -
E ~4 -
8 20001["'------------------- .~..
~ 1500 = =
~ 10~0=0l~-------,~ Ca+-+
500 - -
OjC~'" = 1,Mg=~+=K=+=Na+=I==I =I==I I I I I I Io 10 20 ===~===-~~I ~I~~I ~I H~C;~03~~==30 40 50 60 70 60 90 100
Time (years)
Figure 7-67' Concentrations of major ions over time with excess NP.
8",-,-,-,-,--,-,-.-,--,-,-,-,-,--,-,-.-,--,-
7.9
7.8
J:
Q_
10 20 30 40 50 60 70 80 90 100
Time (years)
Figure 7-68: pH profile with excess NP.
218
The depletion of pyrite against calcite and dolomite is shown below. This behaviour explains
the pH profile, as the NP is always in excess of AP with the result that the pH rises. The
sulphate profile is also explained by this graph, as it indicates the continued production of
sulphate in relation with the presence of pyrite. The concentration in the water is however
controlled by gypsum solubility, which is at equilibrium until the pyrite is depleted, whereafter
it consistently becomes increasingly undersaturated (Figure 7-70) .
..-.
(Jl
Er..o...
..0.....)..
(Jl
~
Q)
c
E
Q)
Eo
Cf)
10
Time (years)
Figure 7-69: Calcite dolomite and pynte depletion in high NP scenario.
The final depletion of pyrite and the consequent undersaturation of gypsum also results in an
increase of dolomite saturation, as more Ca can now solubilise in the water (Figure 7-70).
219
Or, I I Gypsuml
IJ olomite
-.5 -
o 10 20 30 40 50 60 70 80 90 100
Time (years)
Figure 7-70: Mineral saturation over time for high NP scenario.
7.4.2.7 Flooded cellslcolumns
The variation of the flooded system is another scenario to consider similar to that of 7.4.2.4,
as it is assumed to start with relatively unoxidised fresh spoils. The results show both low
generation of salts and neutral pH profile .
. ..".'.4
120 r- -
~ -
C) 100 '- -
.~s - -
<Il
ë 80 ~co; -
<Il
c
0 ~ -
0..
E
0 60-
0
:"Q -
::J
<;:
<Il 40 r- Co' j
E
0
Cl)
M9+~
'"o~D5?7""" : I•I
5 10 15 20 25 30 35 40 45 50
Time (years)
Figure 7-71: Concentration profile in fresh inundated spoils.
220
7
_I 1 1 1 I 1 1 1 1 1 1 1 1 1 1 I I I 1-
6.9 c- -
- -
6.8- -
- -
6.7 r- -
r-- -
6.6 r-
e -a, 6.5 r- -
r-- -
6.4 r-- -
r- -
6.3 r-- -
r- -
6.2 r-- -
r-- -
6.1 r- -
I I I I I I I I I I I I I I I I I I I
0 5 10 15 20 25 30 35 40 45 50
Time (years)
Figure 7-72: Neutral pH in fresh inundated spoils.
It is not surprising that the carbonates are in equilibrium and the gypsum is greatly below
saturation levels.
0
I Gypsum-2
52
ê -4
Ol
0
::::;..
+
+
oCl!
~ -8
C
~
ë -10
.2
n:;l
nl -12
Cf)
-14
-16
5 10 15 20 25 30 35 40 45 50
Time (years)
Figure 7-73: Mineral saturation where fresh spoils are inundated.
221
7.4.2.8 Geochemical modeling with Optimum Colliery water
The final three scenarios consider the result of leachate from irrigation. Three simulations
were run with the following water qualities:
o The Optimum irrigation water (Table 7-4) evaporated by 85% (Table 7-5), as
determined by PREEQC (this is in line with the expected ET at the pivot site). The
simulation after evapotranspiration is illustrated in Figure 7-74.
Table 7-4: Optimum irrigation water quality (before evaporation).
pH EC Ca Mg Na K PAlk
m8/m mg/L mg/L mg/L mg/L mg/L
7.99 309 376 313 112 19.1 0
MAlk F Cl N02(N) Br N03(N) 804
mg/L mg/L mg/L mg/L mg/L mg/L mg/L
185 0.62 34 <0.01 0.16 0.18 2290
Table 7-5· Optimum irrigation water quality (after evaporation).
o The water that was sampled in the 2m column in the spoils at Optimum. (The original
irrigation water quality is the same as in Table 7-4. This water has moved pass the
root zone of the grass planted on top of the column and probably indicates the most
accurate scenario. The high nitrate value is due to fertilizer application. The simulation
after evapotranspiration is illustrated in Figure 7-75.
Table 7-6: Optimum 2m column water quality (after evapotranspiration).
pH EC Ca Mg Na K PAlk
m8/m mg/L mg/L mg/L mg/L mg/L
6.53 496 504 986 29.1 18.6 0
MAlk F Cl N02(N) Br N03(N) 804
mg/L mg/L mg/L mg/L mg/L mg/L mg/L
31 0.2 43.39 <0.01 0.04 17.76 5423
o Water quality from Soil Water Balance program simulations, i.e. water that moved
passed the root zone. The simulation after evapotranspiration is illustrated in Figure
7-79.
222
Table 7-7: Water quality simulated by SWB (water quality input before simulation by SWB).
pH TOS Ca Mg Na K HC03 Cl S04
mg!1 mg!1 mg!1 mg!1 mg!1 mg!1 mg!1 mg!1
6.50 3336 480 265 47 0 580 38 1920
Table 7-8: Water quality simulated by SWB (water quality input after simulation by SWB).
Ca S04 Mg Na K Cl HC03
mg/L mg/L mg/L mg/L mglL mg/L mg/L
216 15853 38 4775 30 3861 347
Optimum evaporation water (as recharged into
spoils)
ti) 9000
~ 8000
,,. • Ca++.§_ 7000 .. • CI-~ 6000 ...HC03-
~ 5000
g_ >~K+4000 f
E 3000 J Yo Mg++
0
o .f2000 • Na+
'1:1
-::l 1000 + 804--LI.. 0 III
0 20 40 60 80 100 120
Time (yea.rs)
Figure 7-74: Concentrations of major ions over time with evaporated Optimum water.
From the simulations it is clear that the water quality measured in the spoils columns seems
to be fairly accurate regarding movement past the root zone. The simulation for the Optimum
2m column water in Figure 7-75 and that of the simulated evaporated water in Figure 7-74
are very similar.
223
Optimum 2m Column as recharge water
- 5000t»
.
--
.lil:
~ 4000
... -+-Ca++~ 3000 _Cl·c41 ----Ir-HC03·
C _"_K+
0
Q.
E 2000
~Mg++
_Na+
o0 --+-S04-·
-a 1000
::s
li.
0 -", ,-- "'~-,-- .----_
0 50 100 150 200 250
Time (years)
Figure 7-75' Concentrations of major ions over time with Optimum column water.
Op,tËmum2m column - Reactirve m~nerall dep,l~et~on
3~-------------------------------------.
_ 2.5~----------------------------------~
o ,--------.1
c;= 2+----.:!!1IIk-----------------------------------l -+- Pyrite
ti, -Calcite
;; 1.5 ---+- Dolomite
~~ j~~~==========::~~~;;~::::=--GypsumI:~ 0.5 =====~
O~~~~~~~====~~~~~~r----~
o 50 100 150 200 250
Time (years)
Figure 7-76: Decrease in reactive minerals over time.
The simulation of the Optimum column water indicates that the sulphate generation only
starts to decrease after 90 years of irrigation (Figure 7-75). Figure 7-76 indicates that calcite
and dolomite will be limited by the saturation of gypsum (Figure 7-78) and deplete within 50
years, togetherwith a decrease in pyrite and precipitation of gypsum. During this time the pH
will drop (Figure 7-77). This demonstrates that acidification would occur due to the
persistence of pyrite over the carbonates. After calcite and dolomite are depleted, gypsum
will start acting as buffer.
224
Optimum 2m columns - pH
8
7
6
5
::cx::. 4
3
2
1
0
0 50 100 150 200 250
Time (years)
Figure 7-77' pH profile at Optimum over time.
I~Optimum 2m column - Mineral saturation5
0-
Cl 0
.-Q -5t::
0 -+-Calcite
;.:-n.
;
l. -10 ___ Dolomite
:::l ~Gypsuml
nl -15
IJ)
.n.l. -20
G>
t::
:ij -25
0 50 100 150 200 250
Time (years)
Figure 7-78: Mineral saturation of the Optimum 2m column water.
The values obtained from the Optimum simulations are much lower than those predicted
based on Soil Water Balance model (SWB) simulated leachate in Figure 7-79. It therefore
seems as if the SWB model predicts values that are too high, and the effect that this has on
the spoils irrigation is that this will increase the concentration at which gypsum equilibrium
occurs.
225
SWB Simulated water as leach ate recharge
20000
18000
F 16000
-= /' • Ca++
.!:- 14000 • C~
~ 12000 /
a " HC03-gl 10000 ! 'x K+
ê' 8000 :t8 /~ x Mg++
~ 6000 • Na+i/ + 804-
1.1- 4000
2000 :/~
o I'
o 20 40 60 80 100 120
Time (Vears)
Figure 7-79: Concentrations of major ions over time with SWB simulated water.
Generally this indicates that mine water irrigation on spoils wil/lead to a further deterioration
of spoils water quality, when compared to rainfal/ infiltration into the spoils.
7.5 Discussion of geochemical response
Laboratory and field studies, together with geochemical model testing, are carried out to
determine the general behaviour of the spoils under various conditions. Each approach has
its own particular advantages and drawbacks, but the cumulative understanding from the
collective use of these techniques allows the construction of feasible conceptual reaction
models and a derivation of expected outcomes from the different applications of mine water
irrigation.
The tests and models indicate the great importance of the reactive nature of the spoils on
which irrigation will occur. The potential for acid generation is a very important indicator of
where to apply mine water irrigation, while it is also clear that in high sulphide areas, the
irrigation will enhance reactions and therefore lead to a higher salt load generation.
The re-use of irrigation in certain systems has both advantages and disadvantages. The
advantages of a "closed" or semi-closed loop means that the salt is not released into the
environment. It is shown that secondary minerals solubility constraints will limit the possible
upper values. It is therefore clear that the nature of the recirculating water is important.
226
Where the mine water used for irrigation is largely gypsyferous, the recirculation scenario is
feasible under the correct conditions (e.g. impermeable floor rocks). Where this water
contains many other elements, such as the sodium bicarbonate waters at Syferfontein, the
solubility limits become far higher due to the increase in ionic strength of the percolating
water. Under these conditions, concentrations will continue rising to very high levels. This
continuous rise is confirmed not only by theoretical considerations and geochemical models,
but also by observations from laboratory tests. The implication is that the nature of the spoils
must be considered in terms of what the likely soluble concentrations of different elements
will be.
Where mine water of relatively consistent quality is used, the changes in longer-term quality
will be far less dramatic. The system reaches an equilibrium state and can continue in this
state for long periods of time. The nature of the mine water is important in terms of the
alkalinity, acidity and total ionic strength. The former two parameters will play a role in
determining the depletion of the neutralisation capacity of the rock, which in turn will
determine the relative reactivity and generated salts. Where mine water containing some
alkalinity is used, this will decrease the salt loads in the long term, provided that the alkalinity
is mostly associated with calcium-containing carbonates. Where this is not the case, the ionic
strength effects of the co-existing dominant cation will result in the benefits of added alkalinity
being offset and concentrations rising steadily in the longer term.
The complexity of the interactions makes it necessary to consider several factors
simultaneously in terms of the feasibility and effects of mine water irrigation on spoils.
227
8 CONSOLIDATED MODELS OF MINE WATER IRRIGATION FOR
VIRGIN SOil IRRIGATION
8.1 Introduction
Irrigation with mine water is currently one of the most promising uses of excess mine water.
However, there is continuing concern, especially from the regulatory authorities regarding the
long-term impact that larger-scale implementation of this practice may have on water
resource (specifically groundwater) quality and quantity. Other options for the use of the
water e.g. the treatment of water for human consumption are being investigated. If this
becomes a reality, irrigation will only be an option at collieries where there is excess water,
thus reducing the area of irrigation drastically.
Opencast: About 75 000 ha of opencast will finally be rehabilitated in the Mpumalanga
region. As the annual water use to irrigate 1 ha is 5 000 rn'', then the excess of 20000 ha
could be irrigated for a single crop.
Underground: For shallow mines every 7-14 ha will provide enough water to irrigate one
hectare. These values are also applicable to stooped areas. For deep underground bord-and-
pillar areas 18 hectares will provide enough water through recharge to the underground
workings for one hectare of irrigation.
Because the overall water quality trend in the deeper aquifer indicates that the water quality
does not show significant deterioration over time the aim of the research was to determine
the feasibility of irrigation with coalmine water, and what the effect will be on the groundwater.
From all the work at the test sites and in the laboratory, detailed models of the water and salt
migration at the different sites could be established.
8.2 Virgin soil model
Based on the earlier chapters, a general model for irrigation sites is as follows:
o Tensiometer data indicates that the soil throughout the profile is high in moisture
content, with the exception of the top 0.5 - 1m. On average the moisture content is
above 30%. The tensiometer data also indicates that the deeper layers dry out
during winter.
o The clay rich layers play an important role in the moisture content, with a build-up of
moisture above these layers. This indicates that clay does playa role in the vertical
flux.
228
o The decrease in chloride values with depth in the profile (obtained from the porous
cup data) indicates that preferential flow occurs at the irrigation sites together with
the diffusive flow. Between 80% and 90% of the water needed for the dilution
resulting in the chloride values at water table comes from lateral flow. (This
calculated lateral flow corresponds with the calculated Darcy flux).
o At least 90% of the water that moves past the root zone reaches the water level.
Due to lateral flow in the higher conductivity zones in the weathered saprolite zone,
water moving past the root zone will have a lower concentration of salts.
o There is a definite relationship between the clay content and the vertical K-value.
The vertical saturated K-value calculated is 10-4 mid on average. This low value is
the result of the swelling of the clays due to the moist conditions below the irrigation
sites. The values obtained from these sites correspond with those obtained by other
researchers. The unsaturated K-value derived from these findings indicates a value
one order smaller than the saturated K, at the infered soil moisture contents
measured from tensiometer data.
o Data from soil analysis with depth through the profile indicates that most of the salt
is contained in the top 2 m of the profile. Chemical modelling of the soil water
indicates saturation of the water with respect to gypsum above one meter, implying
gypsum precipitation. Deeper down the soil water is unsaturated with regard to
gypsum. Approximately 80% of the salts applied over the years of irrigation are
retained. There is thus not a high degree of continuous leaching due to preferential
flowpaths. Data from soil water analysis obtained from the porous cup sampling
indicates that most of these salts occur in the soil water (about 60% of the total salts
applied), and that the balance precipitates in the soil or gets adsorbed. This implies
that over the short to medium term the irrigation with coal mine water does not
influence the aquifers to a great degree. Dissolved salts leach to the aquifers at a
very low rate and are diluted at such a fast rate because of lateral groundwater flux.
As a result low concentrations are detected through borehole sampling.
o When irrigation ceases, it will take more than 15 years under optimal diffusive
conditions to leach the salts out of the profile again. Due to the fact that preferential
flow occurs, it can be more than 100 years. Of the sulphate captured in the soil, 25
% is adsorbed. This means that not all sulphate will be leached by excessive
irrigation or rain.
o Irrigation with gypsiferous water:
229
Gypsum, more than any other single product on earth, uniquely helps soils to be more
productive and fruitful (see Appendix C on Gypsum in soils). It also improves the soil
structure and replaces harmful salts like sodium, which are detrimental to plant growth
and development since they rupture and destroy plant cells. If gypsum is added, this
will lead to a decrease in exchangeable sodium percentage (ESP). The calcium ions
replace the sodium ions on the clay particles, 'stickiness' increases and sodicity
decreases.
According to this research gypsum precipitates in the first meter in the soil profile. Salts
are released very slowly into the groundwater, and the addition of water due to lateral
flow dilutes it to such an extent that very little influence on the groundwater can be
detected.
Recommendation: Irrigating with gypslferous type water is an option over the short to
medium term (10 years).
o Irrigation with sodium water:
However, the high sodium water at Syferfontein influences the ESP. High levels of
sodium make the clay particles less 'sticky', so they do not adhere, or hold together,
very well. Following rain in which the electrolyte concentration is relatively low, the
combination of sodicity and raindrop action causes the clay particles to disperse
instead of remaining in their original arrangement. The disruption of the soil structure,
together with clay dispersion, greatly reduces the soil permeability since the larger
pores are blocked.
Recommendation: Irrigating with sodium-type water is not an option over the medium to
long term (> 10years).
It is recommended that monitoring continue at the sites that are still in operation to study
further trends.
8.3 Detailed site evaluation for virgin soil sites
In this section a detailed description and evaluation of each site will be dealt with. A reference
to each section discussed is as follows:
o Water levels - section 3.3.4
o Groundwater monitoring - section 6.2
230
o Soil analysis - section 6.3
o Unsaturated soil sampling - section 6.4
o Gypsum precipitation - section 6.5
o Measuring moisture potential - section 6.6.1
o Salt balance - section 6.6.2
o Leaching of sulphate - section 6.6.3
o Identification of aquifer parameters - section 6.6.4
o Determination of the flow hydraulics - section 6.6.5
o Influence of clay on vertical conductivity - section 6.6.6
8.3.1 }(Ieinkopje 1
The soil at Pivot 1 varies with depth (Bainsvlei and Clovelly soils) and is underlain by
sandstone and coal. The area below the pivot is undermined. The soil is on average 4m deep
with a 4 - 6m weathered sandstone zone beneath itt. The soil texture at this site varies greatly
with depth and there is a distinctive clay-rich layer at 2m (18.6%). Above this layer the clay
content is <4% and below this layer the clay content is 9%.
The water levels, as expected at an irrigation scheme, are very shallow (approximately
2,8m). The time variation of the water levels indicates that they fluctuate in correspondence
with the seasons. Typically, the water level drops during the mid- to late winter months, and
then recovers again during mid- to late summer.
The total water that is applied to this site is 1500mm annually, of which 700mm is rain. The
irrigation water applied is 800mm with a sulphate content of approximately 2200 mg/L. The
gypsiferous type water results in an increase in the soil salinity, with most of this in the upper
0.7m of the profile. The data for the soil water monitoring from the porous cups shows a
steady decrease in sulphate concentrations with depth (decreasing from 4700 mg/kg at 1m to
260 mg/kg at 3m - this is more than at the other sites. Irrigation at this site commenced a few
years before the other sites, indicating movement down the soil profile, resulting in the larger
concentrations of salts deeper down). Calcium decreases much more than the other cations,
indicating gypsum precipitation in the upper drier soil. However, most of the salts in the moist
layers 0.7 meter and deeper are retained in the soil water. Tensiometer data show a moisture
content of 0.32 at 1mand 0.37 at 2m. The total mass of salts retained in one hectare of
irrigated land is 56 ton, of which 47 tons is in the soil water (84 % of the total salts retained).
231
The decrease in the chloride values down the soil profile confirms preferential flow in the
unsaturated and weathered zones. Under preferential flow or by-pass flow conditions some
of the precipitated water (with relatively low chloride concentration) will flow directly along
these preferred pathways to the water level. This bypass flow component, less affected by
evapotranspiration, dilutes the soil water that is concentrated by transpiration, resulting in
relatively low concentrations in the percolation zone with respect to the concentrations
encountered in the root zone. Sharma and Hughes (1985) introduced a bimodal model
including a bypass flow component. From a variation of this model it was calculated that the
diffusive flux attributes 14% to the mixing process (of diluting the salts) and the lateral flux
86%. Of the total volume of water applied to the site, 9.4 % reaches the groundwater. This
results in a build-up of salts in the profile above the water level, but there will also be an
increase in salinity in the water below the pivot, as has been illustrated with the well
constructed boreholes. This water will move through the weathered zone into the deeper
fractured aquifer along vertical fractures, or laterally. Some may also daylight in streams or
as springs.
Water balance:
o Total water applied /day (irrigation + rain) = 42 m3/d
o Total volume of water per 1 ha that reached aquifer below irrigation pivot = Ov+Osp =
3.9 m3/d
o Lateral flux along 1 ha = Oh = 0.25 m3/d (estimated from Cl mass balance). Darcy
estimate of lateral flux along 1 ha =T*i*L=3*.001 *100 = 0.3m3/d, which is in agreement
with the Cl mass balance estimate.
o 35% of the water that moves past the root zone (after ET) reaches the water level.
This is 9% of the total water applied daily (3.9 m3/ha/d). This value is lower than at the
other pivot site, probably because a large amount of the water drains down to the
undermined area near to the test site.
Salt balance:
o Total salts applied (ton/ha) = 83 ton/ha
o Total salts retained in profile = 56 ton/ha. 67% of the total salts applied to Pivot 4 is
retained in the soil
o Total retained salts in soil water = 47 ton/ha, which is 56% of total salts applied.
A summary of the detailed model at Kleinkopje 1 is illustrated in Figure 8-1.
232
e
,I1~ .4.1..
..-
11----1
---Jt-
u"III -
o
(jJ
E
..t
N
II
t(l o
II
o N"
II...
~
~
.ol
-., ...
Ê
E
~
t:;.
c
c'iiri:
8.3.2 Kleinkopje4
The Hutton soil at Pivot 4 varies with depth and is underlain by sandstone. The soil is on
average 6-7 m deep with a 4 - 6m weathered sandstone zone beneath it. The soil texture at
this site varies greatly with depth and there is a distinctive increase of clay at 3m (27%).
Above this layer, the clay content is 14 - 19% and 23% below it.
The water levels, as expected at an irrigation scheme, are shallow and vary between 3m and
4m. The water levels fluctuate in correspondence with the seasons. Typically, the water level
drops during the mid- to late winter months, and then recovers again during mid- to late
summer.
The total water applied to the site is 1500mm annually, of which 700mm is rain. The irrigation
water amounts to 800mm, with an average sulphate content of 2200 mg/L. The gypsiferous
type water results in an increase in the salt content in the soil, with 96% occurring in the
upper 2m of the profile (48 ton/ha). The data for the soil water monitoring from the porous
cups shows a sharp decrease in sulphate concentrations with depth (from 1600 mg/kg at 1m
to 2 mg/kg at 3m. The sulphate concentration in the soil at 4m, just below the water level, is
59 mg/kg, which confirms that part of the salts reaches the water level, resulting in a build-
up). XRD results indicate that calcium decreases far more than the other cations, indicating
gypsum precipitation in the upper drier soil. Most of the salts is retained in the soil water.
Tensiometer data showed a moisture content of 0.358 at 1mand 0.339 at 2m. The total mass
of salts retained in one hectare of irrigated land is 48 tons (77% of total salts applied), of
which 37 tons are in the soil water (60 % of the total salts retained). The clay layer at 3m
causes ponding in the soil above this depth.
The decrease in chloride values down the soil profile confirms preferential flow in the
unsaturated and weathered zones. From a variation of the Sharma and Hughes model it was
calculated that the diffusive flux attributes 14% to the mixing process (of diluting the salts)
and the lateral flux 86%. Of the total volume of water applied at the site, 24% reaches the
groundwater. This results in a build-up of salts in the profile above the water level, but there
will also be an increase in salinity in the water below the pivot, as has been illustrated with
the well-constructed borehole adjacent to the pivot. This water will move through the
weathered zone into the deeper fractured aquifer along vertical fractures, or laterally
(pumping tests in the boreholes at this site indicates influx of water at 6m). Part of the water
may also daylight in streams or as springs.
Water balance:
• Total water applied /day (irrigation + rain) = 42 m3/d
234
o Total volume of water per 1 ha that reached aquifer below irrigation pivot = Ov+OBp=
10 m3/d
o Lateral flux along 1 ha = Oh = 0.54 m3/d (estimated from Cl mass balance). Darcy
estimate of lateral flux along 1 ha =T*i*L=5*.001 *100 = 0.5 m3/d, which is in
agreement with the Cl mass balance estimate!.
o 88 % of the water that moves past the root zone (after ET) reaches the water level.
This is 24% of the total water applied daily (10 m3/ha/d).
Salt balance:
o Total salts applied (ton/ha) = 62 ton/ha
o Total salts retained in profile = 48 ton/ha. 77% of the total salts applied at the pivot 4
have been retained in the soil.
o Total retained salts in soil water = 37 ton/ha, which is 56% of total salts applied.
A summary of the detailed model at Kleinkopje 4 is illustrated in Figure 8-2.
235
CD
('f)
N
..E.-
,....,
~mn r
cl
(/1
E
'<t
isii
~I o~
__________________ L •..~j~
,... JIIIIo-Jl I ~~I;:
(
oc
I) _. ---_. I -111
' ~~
i5 cJ, > :f
e ~
~ II
II (;
(3 ri'5
II
~
oII
~
cl) .-J..-!
"Ê
E
~
t:::.
c
'iil
0::
---__,.-.-,~~
\-_----~--Il10- -----_. t€t
I!OS uOWlH
8.3.3 Syferfontein
The geology at Syferfontein is relatively homogeneous with the weathered sediments up to
10m deep. The strata mainly consist of 2 - 6 m Arcadia soil, followed by weathered dolerite
and shale to 10m. The shale is underlain by fresh sandstone with a white appearance. At the
depth of 32-35 m the sandstone is underlain by shale. The Arcadia soil at Syferfontein varies
with depth. The soil texture at this site varies greatly with depth, with a high clay content. The
clay % at the top is 63%, decreasing to 31% at the water level of 4m. Below the water level it
decreases to 26%.
The water levels, as expected at an irrigation scheme, are shallow and vary between 3 mand
5 m. The water levels fluctuate in correspondence with the seasons.
The total water applied to this site is 1500mm annually, of which 700mm is rain. The irrigation
water amounts to 800mm with a sulphate content of 1370 mg/L. The sodium sulphate water
results in an increase in the salt content of the soil, with 95% occuring in the upper 1m of the
profile (41ton/ha). The data for the soil water monitoring from the porous cups shows a sharp
decrease in sulphate concentrations with depth (decreasing from 6400 mg/kg at 1m to 17
mg/kg at 3m). Most of the salts are retained in the soil water. Tensiometer data showed a
moisture content of 0.384 at 1mand 0.35 at 2m. The total mass of salts retained in one
hectare of irrigated land is 42 ton (82% of total salts applied), of which 33 tons are in the soil
water (64% of the total salts applied, or 78% of the salts retained).
The decrease in the chloride values down the soil profile confirms preferential flow in the
unsaturated and weathered zones. From a variation of the Sharma and Hughes model it was
calculated that the diffusive flux attributes 2% to the mixing process (of diluting the salts) and
the lateral flux 98%. Of the total volume of water applied at the site, 26% reaches the
groundwater. Most of the salts build-up in the profile above the water level, but there will also
be an increase in salinity in the water below the pivot, as illustrated with the well-constructed
deep borehole inside the pivot. (The sulphate value in the water is 150-175mg/L).
Note: For a long period during the duration of the research project, the pivot was broken
down. Due to flooding directly onto the irrigation site very near to the experimental site,
the lateral flux values obtained may be exaggerated.
Water balance:
o Total water applied /day (irrigation + rain) = 42 m3/d
o Total volume of water per 1 ha that reached aquifer below irrigation pivot = Qv+Qsp =
11 m3/d
237
o Lateral flux along 1 ha = ah = 5 m3/d (estimated from Cl mass balance). Darcy
estimate of lateral flux along 1 ha =T*i*L=7*.005*100 = 3.5 m3/d, which is in
agreement with the Cl mass balance estimate.
o 94% of the water that moves past the root zone ~after ET) reaches the water level.
This is 26% of the total water applied daily (11 m fha/d).
Salt balance:
o Total salts applied (ton/ha) = 51 ton/ha
o Total salts retained in profile = 42 ton/ha. 82% of the total salts applied at the pivot
have been retained in the soil.
o Total retained salts in soil water = 33 ton/ha, which is 64% of total salts applied (or
78% of the salts retained).
A summary of the detailed model at Syferfontein is illustrated in Figure 8-3.
238
en
M
N
e E,
1~ T""" l' ~ ~ c! ~T""" ~T"""
~ ___.. ........ .:~ "It ~ , ~ ,..
j j j , j~ CJ
U U U ,
~.... ,
tw .... ,11 ,," ,~___.... Ol.. ~ 13 "l r' r ,~~';},r'l ffi r- T"""II , IIII It It
___.... "il4 0 0~~~ ([I (fI Iff:I'~t:'"'f E: J5 E E E:{ r'"l ,~
1111 .. '0-'l °..- -,..
____.. I~I J~
... ~ -.-. - ----~---------...
.. (.)
I~ .... .!l,.. i5ii ------...,. Jl
-- :r....J Wj~~u I~ (")eCl T"""II~ ~ 00 ~N
&f ... r--
tB N II 0
II II II 0
~ " 0 0
0 ~
~ T"""cO
....,j.{_:j; pf (.".-') ..- II~ "", ~ II a~ ~II II... " "... ot U'l0 0
Ilo 0 ~ U'lU'l
"'r- ~ t./) IT""".._. _.,. ci' I I~ --._.. ti l:s, ti --Ê
E ~
f'j ~
r.::::..
c 1 ~
'jij .... i1
0::
.... 1,@
/-------~
... .~j.~ ------~ -----...,. g
- ~-_r,~_~ 12LI..,..---
.e ~
~ ~~
~.=.. ~~ l~~~ i~ li ~i ~~
~ d~! ~CC ~~ ~(D diQl,) §.~ ~~~ II II 'cc;~ cII;~II dIflI II il cM00. c;~ ~~ q" ~II cII;~II ecrr;~II
I
8.3.4 New Vaal
At New Vaal the profile is a very sandy soil with a clay layer of 24% above 3m. Above this
layer the clay content is 4%, and 14% below it. Drilling at the site indicated that the sand is
quite clayey deeper down, which will retard the vertical movement of water. The clay layer at
3m results in the retardation of the vertical movement of the water, and causes a build up and
ponding on surface during periods of high irrigation.
The water levels at the irrigation site are deeper than at the other sites. They vary between
10m and 12 m and shows a downward trend over time, which implies that the water drains
away.
The total water that is applied to this site is 1500mm annually, of which 640 mm is rain. The
irrigation water amounts to 800mm with a sulphate content of 450 mg/L. The sulphate-rich
water results in an increase in the salt content of the soil. It is evenly distributed down the soil
profile up to the water level, with some build-up above the clay layer at 3m. In the upper 1m
very little salts are retained, with less than 200kg/ha in the soil. Only 36% of the salt are being
retained in the soil water, indicating less moist conditions than at the other sites, which
implies that the water moves down the soil profile relatively quickly. (The chloride in the
aquifer indicates a recharge of more than 10% for the vicinity). The total mass of salts
retained in one hectare of irrigated land is 18 tons (85% of total salts applied), of which 6.5
tons are in the soil water (30 % of the total salts retained).
In the upper layers an increase in chloride with depth occurs, which implies that no
preferential flow occurs, or that the salts are being retained in the clayey layers. The
decrease in the chloride values down the deeper soil profile confirms lateral flow in the
unsaturated and weathered zones. From a variation of the Sharma and Hughes model it was
calculated that the diffusive flux attributes 27% to the mixing process (of diluting the salts)
and the lateral flux 73%. Of the total volume of water applied to the site, 25% reaches the
groundwater. There is a slow increase in salinity in the water below the pivot, as illustrated
with the borehole inside the pivot. (The sulphate value in the water is 93 mg/L).
Water balance:
o Total water applied /day (irrigation + rain) = 39 m3/d
o Total volume of water per 1 ha that reached aquifer below irrigation pivot = Ov+OBp=
10 m3/d
o Lateral flux along 1 ha = Oh = 0.26 m3/d (estimated from Cl mass balance).
o All the water (100%) that moves past the root zone (after ET) reaches the water level.
240
Salt balance:
o Total salts applied (ton/ha) = 21 ton/ha
o Total salts retained in profile = 18 ton/ha. 85% of the total salts applied at the pivot
have been retained in the soil.
o Total retained salts in soil water = 6.4 ton/ha, which is 30% of total salts applied.
A summary of the detailed model at New Vaal is illustrated in Figure 8-4.
241
N
'<:t
N
...e.... ~
:tt
~ JIll ~
---+- jCJ ~ J
I'.,) "CJ
...
01
~ r-;t :B ~ ;.ë I
N
II II i "I,
0' , • UI
(tl
J o u 'n (",l) U) "f I
J~ II ~E 19 -S ~ ~I I
--T-----------------i~~-,-1--- ...II
___IL ~ ~" Um
... .. ~iii ______ _.. : 11
oc: I ii.s.-J
~IR:>';~ ~~ ';
II 6'-1 a
~ ~II
0 Lf) 1'
Lf) rQ--'
ti!: II ~-:
u:i ("3 0
~ ffi ~
Q IIII Ó I~ II...
Ó (/') Ol .". 0 Ó(/')
(/') ê ( (/')
Lf)
~ Ii 'l
/'')1 l g êr--P1j I."._'Ï. J
r- .r.-._ e..
.... "":-':.... m ]I ~
~ ~
J ...
1
8.4 Summary of Results
In Table 8-1 a summary of all the experiments, the results as well as the implications for the
research are portrayed.
Table 8-1: Summary of all the experiments.
Site Description Results Implications
1. The majority of the groundwater constituents at KK Pivot 1 and 4 and
Syferfonrein have not shown increases over time. This indicates that these 1. II is clear at KK that there is currently a build up of salts in the soils
macro constituents are not reaching the groundwater. Very little, if any, above and within the clayey layers, and that a limited proponion of
Kleinkopje 1 (Major) pollution are detected in the groundwater. the associated salts move through into the groundwater.
2. PV5 within the New Vaal pivot area is showing signs of consistent water
Kleinkopje 4 2, The clay layers in the sands of NV are very imponant in theQuality degradation, with increases in most of the constituents. The retardation of the movement of soil water and constituents down into
Syferfontein piesometers show high salt values in the water above the clay layer. the aquifer, 00 we
New Vaal Borehole monitoring 3. The water levels at aU the sites are elevated. measure correctty? Is the leakage into the aquifer quick enough?
1, There are hydraulic and attenuation factors preventing the salts in
the mine water used for irrigation from being mobilised down the soil
profile and into the aquifer. The soil composition (especially the clay)
KK 1: Sandy soil but shows a marked increase in clay content below the depth and associated sorption and hydraulic properties may be informative,
of 1m (18%). Calcium decreases with depth, while magnesium and sodium are 2, The results point to several potentially imponant findings for the
affected by probably the clay layer just below one meter. Potassium also wider application of mine water irrigation, It would appear that where
decreases with depth, KK4: The magnesium initiatly increases, and then the soils are richer in clay content, there is a significant attenuation of
decreases. This may be the result of sorption at the clay layer at 3m depth, salts in the shallower zones, The groundwater monitoring results
Sodium and potassium level out at 3m, but do not decrease in value, Syfer: indicate that this attenuation makes mine water irrigation a viable
High sulphate concentrations in the upper one meter of soil. It can be option in the short to me~ium term whe.re gYP,s~mrich waters are
concluded that alt the cations move through the clayey Arcadia soils at the used. The clays play an important role In re.talmng the salts to t~e
same degree. with sodium obviously at higher molar concentrations because of uppe~ few layers of the profile. The comp~nson between the SOil.
Kleinkopje 1 (Major) the composition of the irrigation water. The magnesium is a lot less in the water leaching and.porous c.u
p data. supports this. 3. The clay layer In
than calcium and potassium. NV: Calcium and magnesium concentrations th~ s~ndy ~Oll results .In very httle movement of salts pass the laye:.
Kleinkopje 4 increase drasticalty in the clay layer at 2.4m depth. Even the potassium, which This ISconirtned by plesometer measurements. The clays results In
Syferfontein has a far lower concentration in the water, increased in the clay layer. Chloride panding. Soil texture is important in soil moisture calculations.
New Vaal Pits and soil samplinq and sulphate follow the same pattern as all the major cations except sodium.
Heat built-up definitely occurs between the densely populated plants
Kleinkopje 1 (Major) No gypsum could be detected in any of the samples, even those that consisted in the irrigation area, and probably dehydrates the gypsum enough to
mostly of precipitate scraped from surface. It has been revealed that bassanite, be transformed to bassanite in the upper soils (dehydration starts at
Kleinkopje 4 the hydrous form, CaS04 O.5H20, or calcium sulfate hemihydrate, is present in 40°C). Gypsum is thus definitely present in the soils, may it be in a
Syferfontein XRD most of the samples. more dehydrated form of bassanite.
Kleinkopje 1 (Major) KK: There are indications that gypsum precipitation should occur since Ca KK: This would suggest that the distribution is chemically controlled
Kleinkopje 4 values are far higher in the shallow portions. while Mg values are more evenly rather than by surface interactions since these two ions should have
distributed. similar sorptive capacities. This is supported by the Na values, which
Syferfontein Syfer:The sulphur values obtained from the XRF are of interest since they have an even distribution with depth. More gypsum present in upper
Optimum XRF indicate a strong attenuation in the shallowest layers of the profile. layers.
KK 1: It is clear that calcium decreases a 101 more than the other cations in The results of Ihe soil tests suggest that the porous cup results are
depth, indicating gypsum precipitation. Syfer: There is sorption and ion consistent wilh Ihe trapped pore water. precipitates and adsorped
exchange between magnesium and potassium because potassium increased in ions at each of Ihese levels. The fac
depth and magnesium decreased. The reason why sodium does not exchange that the data collected from the porous cups. the soil salinities and
Kleinkopje 1 (Major) is because the sodium content of the irrigation water is already very high. The the groundwater quality indicate the same trends as the soil analysis
Kleinkopje 4 sodium sulphate water results in high sodium and sulphate content in the upper is encouraging. in light of the continuing research. Decrease of salts
layers of soil, which decreases with depth. The conservative chloride anion in depth, most precipitates in upper layers - nol enough leaching to
Syferfontein moves through the different layers at the same degree, without any dilute the sans. Will need high rainfall events {more than evaporation
Optimum Porous cups precipitation or adsorption. factor to leach the sans in soil water. Is lateral movement > vertical.
Controlled irrigation (less than 20mm at a time) will result in the salts
not moving down the soil profile in large quantities. In the case of
Kleinkopje 1 (Major) irrigation with colliery water, the quality of the water is such Ihat it will
Kleinkopje 4 The tensiometers at the virgin soil as well as the spoils idicate a decrease in not leach the salts in the soil, but will result in the accumulation of
soil moisture content in depth during the "drier" winter months when the volume more salts. Effective irrigation will result in salts accumulating in the
Syferfontein of water apUied are more controlled. The distinct clay layers in the soil results in upper soil layers with less effect on the groundwater. Leaching will
Optimum Tensiometers panding above these layers after large rainfall events. only result after heavy rainfall events or after over irrigation.
The water follows the path of least resistance. This means that most
Kleinkopje 1 (Major) 1. At all the pivots flow zones were determined in the saprolite zone. of the water (from the venical as well as the lateral component) flows
Kleinkopje 4 2. Darcy velocities were calculated for all zones in the profile. varying greatly in away along these flow zones until it daylights or where it reaches a
the different zones. The flow zones indicated conductivity values as high as in fracture, then moving downwards into the fractured rock aquifer. This
Syferfontein excess of 100m/d in the main flow zones. is the reason why so little increase in salinity in the groundwater
Optimum Point Dilution Tests 3. The clay layers have very small conductivity values. occurs.
Kleinkopje 1 (Major)
Kleinkopje 4
Syferfontein The upper soil zones have very small concuctivity values, less than 1mId and This implicates that very liille water movement occurs in these upper
Optimum Falling head tests as small as E-2m/d. zones. Most of the flow is along preferred pathways.
Kleinkopje 1 (Major) These values indicate that the deeper rocks are not as weathered as
Kleinkopje 4 is the case with the shallower saprolite. Recharge into these aquiferNormal transmissivity values for Karoo aquifers were obtained in alt the will be a lot less than the lateral movement of water, implicating that
Syferfontein boreholes. These values ranges from 0.2 to 0.8m2/d, with most of them in the very little pollutants will move down into these deeper pans of the
Optimum Pumping tests order of 0.2m2/d. aquifers.
243
1. The EC and sulphate variations over time reveal a very interesting
phenomenon, namely that, where clean water irrigation is employed, the values
decrease over time. The reason for this lies in the fact that the spoils have been
oxidised prior to the experiment. Since they represent the results of the in-situ
spoils at Optimum, it can be expected that oxidation has occurred in these 1. The laboratory columns are useful as they provide a bette
spoils over time and the secondary products from oxidation reactions and the understanding of the real rates at which the processes occur anc
subsequent neutralisation would be contained in the spoils. 2. The celooiseo thus provide a closer "real-time" analogue of the field situation,
cells become undersaturated as secondary products are flushed out, but where although the volumes of water added reflect a far longer time 0
leachate is recirculated or mine water used, the values remain close to irrigation than is usually the case in the field. The good comparlso
saturation. 3. As can be expected in these columns, the carbonates are not with the field situation is encouraging in this regard. 2. I
stressed in the reaction period and calcite and dolomite are at or above does indicate that the initial salt loads in such a system would b
Laboratory Leaching columns -spoils expected equilibrium. very high. with stabilisation occurring in time.
1. This means that not all sulphate will be leached by excessive
irrigation or rain. It must be noted, however. that desorption can
occur and that the 25% sorbed sulphate will not be irreversibly fixed
1. There is a definite decrease in sulphate in both columns. The change in in the soil profile. 2. The acidic conditions in the soit also enhance
these values is not very significant, but may be due to some adsorption. sulphate adsorption. 3. There is no direct correlation between the
Leaching columns - 2. An average sulphate adsorption of 24.9% of the total sulphate extracted. The clay content of the soil and the adsorption values. but the increased
Laboratory adsorption rest (water extracted part) is precipitation. values are definitely a function of the hydraulic properties of the soil.
Without taking evapotranspiration into account, it will take nearly 20
years to leach the salts. Check against John's figures for these
values to calculate a time. Not all adsorped sulphates willi leach -
Leaching columns - salt Most of the soils are leached in the first few years, but more than 6000mm will some will never be retained (40% of 25% total sulphates will
Laboratory leaching be needed to flush the soil. Calculated saturated K value is 0.4 mId remain in soil)
In the case of irrigation with colliery water, the quality of the water is
such that it witt not leach the salts in the soil, but will result in the
accumulation of more salts. The only way to leach the salts is with
heavy rainfall events.
This means that the maximum quantity of gypsum leached from the
soil, is bound by the gypsum saturation of the water moving through
the soil. Calculating the sulphate (1800-2000mglL) and calcium (600-
700mglL) concentrations in the leachate indicates that. if 20 tons
Calculations with PHREEQC indicate that atl the leachate are saturated or over gypsum I ha precipitated in the upper layer of soil, at least 40 - 50
saturated with respect to gypsum. Moist soil results in the leachate being under heavy rainfatt events (75mm) will be needed to leach the precipitated
saturated with respect to gypsum. The increase in Ca and S04 can only occur salts from the soil. Soil type not very important -solubitlity in water
up to a point. where the aqueous solubility of these ions becomes limited by the more important. as wett as movement through soil -eg. clay will leach
solubility of gypsum. The sulphate concentrations in the leachate were all in this at same tempo if clay dry out, resulting in cracks. Tilling of soil witl
Laboratory Leaching drums range. When the soil is wet, soil water not saturated with respect to gypsum also playa role in leaching, as wett as soil water content..
1. Geochemical composition of spoils is the most important
characteristic. 2.
Irrigation with water with elevated alkalinity will assist in preventing
acidification. 3. Iriigation
with gypsiferous water will result in the drainage quality being very
similar to that expected without irrigation in the long term.
4. Irrigation without water which is high in non-gypsiferous
constituents will result in enhanched solubility of minerals and
1. The sulphate response is of great interest and shows that. when the recycled therefor poorer drainage Quality.
water contains significant ions other than Ca and S04, the sulphate can 5. In systems where water will be recirculated, there will be a steady
Optimum - lab Humidity cells increase to very high values over time. decline in quality.
244
9 OPERATION AND MONITORING GUIDELINES
One of the outcomes from this study is to define the conditions under which mine water
irrigation can be implemented and the associated operational and monitoring guidelines that
should be followed. The section that follows explains in detail the rationale used and the rules
that have been established for mine water irrigation. These have been based on the findings
from this study, the fundamental considerations of mine water irrigation, the regulatory
environment and, as far as possible, the practical implementation of mine water irrigation as
part of optimal mine water management.
9.1 Basic Data Input
It is important that the criteria for irrigation operation and monitoring take the fundamentals,
the results from this research and the site specific conditions into account.
The rationale used to provide guidance on site selection, monitoring and operations is
summarized by Figure 9-1.
Can Mine water irrigation be
done sus
Data•tainably ?Input
~
Critical Flaws Yes.. STOP
No ~
Spoils or Virgin Soils
I Soils I ~
Data Input Data Input
~ ~
Indication of Indication of
suitability suitability
<,
Monitoring and Operational -:
Guidelines
Figure 9-1: Rationale for defining site-specific criteria.
This figure indicates that there is basic data that needs to be provided before any kind of
decision on the sustainability of mine water irrigation can be made. Several so-called critical
flaws have also been identified by this research. The presence of these flaws will lead to an
immediate recommendation that irrigation with mine water should not occur. If none of these
245
flaws are present, the decision can be made whether mine water irrigation is to be done on
virgin soils (green fields) or on rehabilitated mine spoils (brown fields). Criteria for each type
of irrigation have been identified and included in this decision making process.
Figure 9-2 shows the type of information considered when defining the suitability of a
particular site and the type of irrigation planned (duration, water quality, crop etc).
Irrigation Water
Quality
Soil properties
Period of
Irriaation
Mining
---c Impact of Mine
Spoils
Water' Irriga,tion Properties
on Groundwa1er'
Critical Flaws
Environmental
Aquifer Considerations
Properties ISetting
Figure 9-2: Data input required to determine Impact and define criteria.
9.2 Selection criteria
The approach is to take information of different types and provide a screening tool.
• This approach is conservative and tiered.
• The screening combines a yes/no approach with the integration of (simplified) key
fundamentals.
• An output of the assessment based on all the criteria is then evaluated.
These data inputs will be discussed in the following sections. The most important question
before irrigation can be introduced is if the land is suitable for irrigation.
246
9.2.1 Identifying land suitable for irrigation (Ie Roux, 2006)
9.2.1.1 Background
Soil is often the only medium for dumping saline water. Using waste water for crop
production under irrigation seems like a perfect solution as it can be financially beneficial as
well. However, the soil is not always the perfect sink and the environment must be protected.
Protection of the environment is often seen as a good reason to pollute the soil with degraded
water, assuming that the soil is a perfect sink andlor the soil is not part of the environment.
The extent to which the soil serves as a good sink varies. Soil can effectively remove
environmental hazards without polluting the sailor the surrounding environment but usually
either the sailor the groundwater is polluted.
Irrigation of soil with gypsum-rich water is an example where the soil may serve the goal well.
If the composition of the irrigation water is favourable for precipitation of gypsum in the soil,
the gypsum may be removed from the water effectively. After precipitation the impact of the
salt on crops are drastically reduced. In the dissolved condition the severity of divalent
salinity is very low compared to sodium chloride solutions.
9.2.1.2 Irrigation requirements of soils
The suitability of soils for irrigation primarily depends on four factors.
o Firstly the final infiltration rate of the soil must be high enough to prevent run off. The
final infiltration rate is mainly determined by the soil texture, soil surface structure and
slope.
o Secondly irrigation soil must have good internal drainage. The water must move
through the root zone freely to avoid water logging. The factors controlling internal
drainage are complicated but fortunately the soil morphology, as accommodated in
the South African soil classification system, is a good indicator of internal drainage.
o Thirdly irrigation soil must have good external drainage as excess water, needed for
leaching, must be able to move out of the system and either joins the groundwater or
surface water in rivers and wetlands. The factors controlling the external drainage are
even more complicated than internal drainage. Soil morphology, specific redox
morphology, also serves as a good indicator of external drainage and therefore the
soil types of the South African soil classification system are classified according to the
degree of drainage. Limited drainage is aggravated by the position in the landscape
or terrain morphological unit (TMU) and an example is illustrated in Figure 9-3.
247
o Fourthly the water ho/ding capacity (WHC) of a soil must be large enough to hold
water for one irrigation cycle, usually one weak. This limitation can be reduced by
systems that can irrigate more often. Mechanical irrigation like drip and micro
irrigation must be avoided as they are probably unsuitable for use with saline water as
the leaching factor can not be upheld with them.
The principles applied in developing these soil suitability ratings were modified to fit the
requirements of the area and quality of the irrigation water. Several factors contribute to the
decision to make the drainage requirements stricter. The rainfall of the area is high,
increasing the risk of water logging. The criteria for soil depth are therefore more strict than
usual. Saline water combined with water logging increases the risk of salinisation and puts
more emphasis on soil depth.
The WHC is mainly determined by the depth and texture of the soil. The texture of the well-
drained soils of this area is in the optimum range and hold 120 mm of plant-available water
per meter of soil. Soils shallower than 500 mm cannot hold enough water to supply the peak
water use in late summer with ET reaching 10 mm per day in a mature crop stand. This
limitation can be addressed to some extent with more frequent irrigation cycles if the system
permits. Shallow soils are also sensitive to water logging and accelerated salinisation.
The degree of management required for successful production of cash crops under irrigation
increases significantly with increased clay content. In clay soils the final infiltration rate is
lower, distribution of water is uneven and workability is poor. The increased aggregation
caused by gypsum in saline waters may improve this parameter to the extent that the
negative effect of high clay content is negligible. Except for differentiating between soil types
in terms of drainage, soil depth and the terrain morphological unit (TMU) are also taken into
account (Figure 9-3). Drainage can be improved by artificial drainage but at a cost roughly
equal to the cost of an irrigation system. Fertility is not considered as a factor in the suitability
of soils for irrigation as fertilisation and/or liming and/or adding gypsum can usually address
it. Salinity is not considered either as the salinity and impact on crop production must be
monitored continuously.
248
Bo.19
1
3 3 1454
15701"'l ~ - - - - - - - - - - - - - - - - - - - - - - - - - -~'- .-._.-.,----
Figure 9-3: Terrain morphological units for soil Ba t9 as an example
9.2.1.3 Soils of the area
The distribution of the soils of South Africa is mapped as land types (Figure 9-4). The soil
data in a land type is presented as the typical pattern of soil distribution in the landscape -
from the crest, midslope and footslope to the valley bottom. The distribution of soils in the
twenty-two land types that cover the area is presented in the land type inventories. The
same soils occur in most of the landscapes of all twenty-two land types. The distribution of
the soils in the land types is rearranged for ease of evaluation.
-25_8
-26.1
-26.4
-26.7
Figure 9-4: Soil types in study area
249
SoHsuitability rating
The soil types present in these land types are divided in three groups according to natural
drainage namely the well drained, moderately drained and poorly drained.
o The group of weI/-drained soils are soils of the Hutton (Hu), Clovelly (Cv), Oakleaf
(Oa), Arcadia (Ar), Valsrivier (Va), Swartland (Sw), Bonheim (Bo) and Dundee (Du)
forms. Exceptions to this rule is that all series (subdivisions of the forms in the 1997
edition) of the Oakleaf, Valsrivier, Swartland, Bonheim and Dundee forms with signs
of wetness and all families (subdivisions of the forms in the 1991 edition) of the
Swartland, Bonheim and Dundee forms of the South African soil classification system
should be considered as moderately drained. Soils of the Dundee form have layered
subsoil deposits and therefore its suitability varies over short distances. Its suitability
should be evaluated by intensive soil survey. The criteria used in the land type survey
make it impossible to evaluate the depth of soils of the Swartland and Valsrivier
forms. Although few of these soils occur in the land types they are not taken into
account for their contribution to land suitable for irrigation making the figures slightly
conservative.
o The moderately drained soils are all the series (1977) and families (1991) of the
Bainsvlei (Bv), Avalon (Av) and Glencoe (Gc) forms of the South African soil
classification system.
o Poorly drained soils are all series (1977) and families (1991) of soils of the Westleigh
(We), Longlands (Lo), Wasbank (Wa), Estcourt (Es), Kroonstad (Kd), Katspruit (Ka),
Rensburg (Rg) and Willowbrook (Wo) forms of the South African soil classification
system.
The soils are classified in four suitability classes namely very suitable (S 1), suitable (S2),
barely suitable (S3) and not suitable (N). The suitability classes are defined as:
_;;;. S1 - Very suitable soils are the soils which are easy to manage under normal irrigation
practices and do not restrict crop production for optimal yields. They have a low risk
of water logging and salinisation but soluble salts that do not precipitate under these
conditions in the soil will leach through the soil and contaminate the ground water.
_;;;.S2 - Suitable soils are soils that require improved irrigation management systems.
These include mechanical systems of irrigation, artificial drainage and/or a high level
of irrigation expertise for example advanced irrigation scheduling.
_;;;.S3 - Soils barely suitable should be avoided if possible. Multiple irrigation problems
may occur. The risk of reduced crop yields, water logging and salinisation is a short-
term hazard.
_;;;. N - Soils not suitable for irrigation are soils and land where the cash crops generally
produced in that area cannot be produced economically even with the best irrigation
practice and most profitable crop generally grown.
250
Table 9-1: Soli sUItability rating gUIde for weI/-drained sous.
Soil factor Soil rating guide Limitation
S1 S2 S3 N
Depth (mm) >1200 1200-800 500-800 <800 WHC
Texture (clay %) <35 35-55 >55 Management
TMU 1+3 4 5 Drainage
Table 9-2: Soli sUItability rating gUIde for moderately drained soils.
Soil factor Soil rating guide Limitation
S2 S3 N
Depth (mm) >1200 >800 <800 WHC
Texture (clay %) <55 >55 Management
TMU 3 1+4 5 Drainage
The following soils and land types are not suitable for irrigation for several reasons namely
rock, Mispah (Ms), Glenrosa (Gs), Mayo (My), Cartref (Cf) and stream beds.
9.2.2 Procedures for selecting land for irrigation
1. Select a land type with a high percentage of land suitable for irrigation.
2. Select a terrain morphological unit in the landscape containing the highest percentage
of land suitable for irrigation. Take into consideration that the percentage may be
high and the area covered quite low.
3. Take slope into account. Different irrigation methods accommodates slope differently.
4. Do a detail soil survey.
5. Select the best land for irrigation.
6. Select crops resistant to salinisation.
In Table 9-3 the percentage of suitable land in the Mpumalanga coalfields for the different
land types as well as the actual size in hectares are portrayed for the different suitability
ratings.
251
Table 9-3: Distribution of soils suitable for irrigation in different land types of the area.
Distribution of soils suitable for irrigation in different land types of the area.
Percentage (%) Area (ha) Total
TMU S1 S2 S3 S1 S2 S3 (%) (ha)
Ab9 1 65 X 5 8779 X 675
3 55 5 15 13619 1238 3714
4 X 40 30 X 1801 1350
5 X X X X X X 69 311761
Ba19 1 30 10 20 8084 2694 5388
3 25 5 10 6736 1347 8083
4 10 10 25 1010 1010 2526
5 20 X X 674 X X 55 36880J
Ba3 1 5 X X 464 X X
3 30 X X 8349 X X
4 40 X X 2783 X X
5 20 X X 464 X X 26 120621
Ba37 1 35 10 10 14368 4105 4105
3 30 10 5 6717 2239 1120
4 5 20 20 373 1493 1493
5 X X X X X X 98 73239J
BaS 1 X 60 X X 9320 X
3 X 65 X X 30290 X
4 X 45 X X 5242 X
5 X X X X X X 58 448521
Bb3 1 40 X 35 17873 X 15639
3 25 X 35 17555 X 24576
4 X 10 10 X 638 638
5 X X X X X X 60 76919J
Bb4 1 15 X 50 10562 X 35209
3 10 X 70 14083 X 98583
4 X X 10 X X 4108
252
5 X X X X X X 64 1484621
Bb11 1 45 X 15 6374 X 2125
3 40 X 15 4249 X 1593
4 X 25 10 X 2479 992
5 X X X X X X 50 178121
Bb12 1 45 X 15 10444 X 3481
3 40 X 15 6962 X 2611
4 X 25 10 X 4062 1625
5 X X X X X X 50 291851
Bb13 1 50 15 X 8063 2419 X
3 40 35 X 7256 6349 X
4 X 30 X X 1209 X
5 X X X X X X 50 25296 1
Bb14 1 X X X
3 20 X X X X X
4 X X X 1988 X X
5 X X X X X X
Bb16 1 10 X 30 X X X 6 1988
1
3 X X 25 1840 X 5520
4 X X X X X 4140
5 X X X X X X
Dc3 1 X X X X X X 31 11500 1
3 X X X X X X
4 X X 10 X X X
5 X X 33 X X 98
Ea5 1 X 20 X X X 98 3 196 1
3 30 10 X X 2790 X
4 X X 90 8372 2790 X
5 X X 40 X X 10045
Ea8 1 X X X X X 1116 30 16741 1
3 X X X X X X
253
4 X X 10 X X X
5 X X 30 X X 114
Ea20 1 X X X X X 85 4 199
1
3 X X X X X X
4 X 10 X X X X
5 X X X X 1746 X
X X X 1 1746
1
Land types Ib15, Ib24, Ib35 and Fa8 do not have land suitable for .Ir.rigation.
9.2.3 Site specific and more general data.
The first criteria required when evaluating each site considered for irrigation is if it is virgin
soil or spoils.
9.2.3.1 Virgin soils:
The baseline information that will decide if land will be feasible and suitable for irrigation is
the following:
1. Name of the colliery and the name of the irrigation site.
2. Irrigation area size (ha)
3. Water quality
PAlk 1 MAlk 1 Cl 1 5041
4. Landtype
5. Geology and structures: If a dyke (barrier or preferred pathway) is present, a warning
should occur. The impact on receptors should be investigated.
6. Water level at the site
7. Depth of soil
8. Annual rainfall
9. Irrigation volumes
10. Crop type
11. Distance to closest community - is the site up gradient or down gradient
12. Community dependence on groundwater
13. Is pivot site in area of land instability?
14. Is there an underlying dolomitic aquifer?
15. Is area undermined? Type of mining: Bord and pillar or high extraction
16. Aquifer classification must be provided
254
17. Environmental data: - River systems and base flow
- Distance to closest receptor
- WOO: user specified or SANS 241 :2005 - (TOS, EC, S04)
- Sensitive site: Will a wetland, nature reserve, Ramsar site or
springs be influenced
18. Current Monitoring systems: - Number of boreholes
- Distance from site
- How many down gradient
- Monitoring frequency
- Current groundwater quality
9.2. 3. 1. 1 Critical Flaw list:
A set of critical flaws for mine water irrigation has been established. These flaws are
situation-related to the location, water quality and receiving environment. If any of these
criteria occur on site, mine water irrigation should only be allowed after the Department of
Water Affairs and Forestry (DWAF) water use permitting system has been followed.
The critical flaws have been defined as follows:
1. No irrigation is allowed with potable water (defined as TOS < 900 or EC< 150)
for basic human needs (BHN).
2. Aquifer classification: If the irrigation is planned on a shallow sole source
aquifer, it is critical flaw.
3. Area of irrigation: If the area is greater than 150 ha, DWAF water use
permitting guidelines rather than the criteria used in this document must be
followed since no current research has been done on such a scale. The
success of current irrigation is also due to the relatively small scale of
application.
4. Water quality: The irrigation water must be suitable for crop production. If the
irrigation water pH < 5, or if (Ca+S04+HC03) in meq is < 60% of the total ionic
composition (unless EC < 200 mS/m), or if the irrigated water is less than one
and a half times that of the background EC, then irrigation cannot be allowed.
Also if the SAR > 15, then this is a critical flaw and irrigation should not be
allowed.
5. Land type: If the assessment for soil suitability is N, this is considered to be a
critical flaw. If there is a clay layer with a clay % more than 10% greater than
the average from that soil type and this occurs less than1 m below the root
zone, this is a critical flaw.
6. Water level: If the groundwater level < 1.5 m, critical flaw.
7. Depth of soil: Less than 0.5 m soils, critical flaw.
8. Community dependence on groundwater: If community dependence on
groundwater> 80%, it is a critical flaw unless community> 2km away or up
gradient of the proposed irrigation or irrigation is planned for a very short period
of time.
255
9. Land stability: If pivot site is in an area of land instability, sinkholes or if there is
an underlying dolomitic aquifer, this would constitute a critical flaw.
10. Environmental: If there is a wetland, nature reserve, Ramsar site or a site
similar down gradient within 2 km, this is critical flaw.
Critical flaws should highlight potential situations where mine water irrigation is likely to cause
long-term problems. However, with detailed determination of the on-site conditions, and
potential monitoring and mitigation, irrigation in such areas may still be possible.
9.2.3.2 Opencast spoils
At opencast sites, there are two main issues to consider. One important factor is the
rehabilitated soils on top, and the other the spoils underneath.
Rehabilitated soils:
• Depth of soil (value in mm)
• Type of soil: The same criteria as for virgin soil in area (this can be carried over from
virgin soil input), or site-specific soil properties (% clay, loam and sand) must be
considered.
Spoils:
• Is it free draining
• Potential for acidity (Assume that it is Potentially Acid Generating (PAG) unless other
data exists). Based on available data, either Potentially Acid Generating, Non-Acid
Generating or Uncertain spoils will be used. Site specific ABA data can be included.
• Proportion of spoils unit to be irrigated (either percentage, or enter total spoils area
draining to decant point and area of mine water irrigation)
• Water use (will decant water be re-used for irrigation)
• Is any decant released to the environment, currently or in the future?
• What % of spoils is flooded at the decant elevation (% value)
• To what degree can gasses enter the spoils
9.2.3.2 1 Rules/Calculations
• Criteria for ABA:
If NNP = NP-AP <0 the sample has the potential to generate acid.
If NNP = NP-AP >0 the sample has the potential to neutralise acid produced.
More specifically, any sample with NNP<20 is potentially acid-generating (PAG), > 40 NAG
and in between is uncertain.
1. Samples with less than 0.3% sulphide-S are regarded as having insufficient
oxidisable sulphide-S to sustain acid generation.
256
2. Neutralising Potential Ratio (NPR) of >4: 1 are considered to have enough
neutralising capability (NAG)
3. NPR ratios below 1: 1 with sulphide-S above 0.3%, are potentially acid generating.
• If applicant indicates decant water will never be released, any irrigation is allowed but
a warning is given that irrigation will accelerate acidification and water quality
deterioration rates.
• For all assessments where spoils are PAG:
o Irrigation on < 10% of spoils is allowable,
o 10-30% is allowed with a warning that irrigation is only a short term measure
and
o Anything> 30% disallowed.
o An exception to this occurs if the % spoils flooded at decant elevation is >80%.
o For marginal spoils this value is >70%
• If the spoils are uncertain then one must consider the likely leachate quality as
obtained from PHREEQC:
o If significant alkalinity exists and the TOS of the leachate is <7000mg/1 or
leached Ca:Na ratio ( in equivalents) >2 then a marginal spoils can be
considered NAG.
o In all other cases a marginal spoils becomes PAG.
• If user indicates water is recirculated then the same rules as above apply except that:
o Irrigation is limited to a 5 year period if irrigation <10.01 % of spoils and
o Irrigation is limited to two years if irrigated area between 10 and 30% of spoils
area contributes to decant. This is only a short term method to decrease water
volume at colliery.
9.3 Monitoring Guidelines
If the potential site fulfills all the criteria set out in the previous sections, appropriate
monitoring must be put in place.
Since the suggested monitoring must meet DWAF requirements, the most appropriate
manner in which to deal with such irrigation is to apply the DWAF Minimum Requirements for
Monitoring at Waste Disposal facilities. Table 9-4 provides an overview of suggested
minimum required monitoring at different types of waste facility.
257
Table 9-4: Minimum monitoring requirements at various types of waste management
facilities (DWAF, 2005).
A t a near surface monitoring Wthinwaste or unsat. zone GrrunÓN atB" rronitoring
g I!! 1]Monitol'ing 2~ * Jil Q) 0 2~ 3'": l(l I:> .!; aRequirements 3: 2 * ~ ~~
Kl ~
3: 1:1 ~.ID Q)
CT 15 I!! c E
ai c !5 l(l
E i J: 3: '
E
3": ~~ u~
1: ~ .'s" 'Q) Q)
c -c ~ .~ l
"i s: '"
~ u
I'S"
c l2 ~c .2 0 6 c 'u"
B ~0 il§ ''"" 0 c ~ ~t ~~ a. u u
'" M 0 Kl .~
1 .s Qi[ .2 E ro
!l E ~2 -e
~ i r'o" sC:l ~ ~* .!li'5 8 ~ *
~ e.!Jl
~ .s]jj s: ]i c 0 2lc:
.s t ~ :~Jl~ '" i u'" E .~ Qi s ~ s: ~Waste environment '" e'"li 0'" ~ '" 0) '0" e ,s 0 ~ ~Q)'" f- bi ~ iIi 6: o'" .'5" f- f}; l e e 00 i5 o dl <!) u, ~
Mines - Reactive environment d d m
Slirres (Slurry) m m y y m yes yes m 3m y y m m
Ore disca"ds d m m y y m yes yes m 3m y y m m
Rock Discards (opereast) m m y y yes yes m 3m y y m m
Rock discards (other) m y y yes yes y y y y y y
Mine watEr" (irrpoundment) m m yes yes 3m 3m y y m m
Mine water (dscha'ged) d y w
Mines -Inert environment d d no yes y y y y y m
SUrres (slurry) m no yes y y y y y m
Rock discards m no yes y y y y y y
Ore disca-ds m no yes y y y y y y
Mine water (dscha'ged) d d
Coal fired power stations d d yes yes 3m 3m y y m m
Coal stockpiling m m yes yes 3m 3m y y m m
Ash disposal (slurry) m yes yes 3m 3m y y m m
Ash disposal (dry) m y yes yes 3m 3m y y m m
Dirty w aer systems m yes yes 3m 3m y y m m
Water discharged d y
General waste
Large (>500 Vd) d d m y y y m m yes yes 3m 3m y y m m
Medium (26 - 500 Vd) d d m y y m yes yes y 3m y y m m
Small (1 - 25 Vd) m yes yes y 3m y y m
Informal «1 Vd) m no yes y 3m y y m
Sewage
Unlined rraturation ponds d yes yes y 3m y y m
Sludge d
Hazardous waste d d d m m m m y m m m m yes yes m m y y m y
waste irrigation d d m m m m y yes yes m m y y m m
Agrcutture (feed lots) d d yes yes m m y y y
Agrcutture (difflse sources) no yes y y y Y
Septic tanks and pit latrines no yes y y y y
Underground storage tan~ yes yes m m m m m
Urban development d m no yes y y y y
Industries Refer to specificwaste atxve, such as gena-ai, hazadocs. irrigation, irrpoundment
Radioactive waste As speclied by the eNS in collaboration w ith the D'M&F
Explanation of codes: d::: daily rronitoring; w =weekly rronitoring; m = rronthly m:>nitoring; 3m = 3-rronthly rronitoring; y :::yearly rronitoring
DWAF (2005) states that monitoring networks at waste management facilities must allow
monitoring of the system on a representative basis, and that the key to successful monitoring
is the linking of point information into larger systems, referred to as monitoring networks. As
such it is recommended that these DWAQF guidance documents should provide the
minimum monitoring requirements at mine water irrigation sites and that the monitoring
"should extend beyond pollution plumes to allow for the delineation of plumes and
investigations into the pollution migration rate. "
258
DWAF guidance for different types of environment is provided below:
Environment No. Holes Distance From Waste Monitoring Frequency
IMines - Keactlve tnvlronment
Slimes (Slurry) 1-3 50-250 m downstream Samples from boreholes every G months. Sample
Ore discards 2-5 50-500 m downstream and above monthly from streams above and below mine. If
Rock discards (open cast) 1/250 ha into water accumulations pollution from mine occurs. install recorders in
Rock discards (other) 1-3 50-200 m downstream streams above and below mine. Measure daily
Mine water (impoundment) 2-6 50-'1000 m downstream flow, EC and pH. Sample farmers' boreholes '1-5 km
Farmers' boreholes within '1-5 km from mine worl<ings radius, initially and when problems are expected.
Mines - Inert Environment
Slimes (Slurry) 0-1 Monthly samples from streams above and below mine
Ore discards 0-'1 If pollution from mine occurs, install recorders in
Rock discards (open cast) 0-'1 streams above and below mine, Measure daily flow,
Rock discards (other) 0-1 EC and pH..Samples farmers' boreholes 1-2 km
Mine water (impoundment) 0-1 radius initially and when problems are expected.
Farmers' boreholes within 1-2 km from mine worl<ings
Coal Fired Power Stations Samples from boreholes every 6 months ..Monthly
Coal stockpiling 2-3 50-500 rn downstream from streams above and below power station. Ii
Ash disposal (wet) 2-3 50-500 m downstream pollution occurs in streams, install recorders in
Ash disposal (dry) 2-3 50-500 m downstream streams above and below power station, Measure
Dirty water systems 2-3 50-500 m downstream d'aily flow, EG and' pH. Sample farmers' boreholes
Private boreholes within '1-5 km from power station 1-5 km radius. initially and when problems arise.
General Waste
Large (>500 tid) 3-6 20-200 m surrounding Samples from boreholes every 6 months or as
Medium (150 - 500 tid) 2-3 20-200 m 1 up- and 1 downstrear specified in permit. Sample water-supply boreholes
Small (25 - 149 tid) 2 20-200 m 1 up- and 1 downstream 1-5 km radius initially and when problems are
Communal «25 tId) 1 20 m downstream expected. Sample surface water as specified in
Private boreholes within '1km from waste permit. Sample monthly for leachate. if any.
Sewage
Unlined maturation ponds 1 20-50 m downstream Samples from boreholes every 6 months. Samples
Sludge 'I 20-50 m downstream monthly from streams above and below sewage work
lHazardous waste 5-10 10-200 m surrounding Site-specific constituents at frequencies
recommended by impact study
Iwaste Irrigation 3-6 50-500 m 1 up- and '1downstrean Samples from boreholes every 6 months. Monthly
samples from streams above and below.
IAgriculture - feed lots 2-3 50-200 m '1up- and 1 downstrearr Samples from boreholes every 5 months. M'onthly
samales from stream above and below.
IAgriculture - diffuse sources 0 Samples from existing water-supply boreholes
when problems are expected.
ISeptic tanks and pit latrines 0 Samples from existing water-supply boreholes
when problems are expected.
IUrban development 0 Monthly EG in streams above and below development
Based on this guidance it is recommended that mine water irrigation be dealt with in the
same manner as "Waste Irrigation" or General Waste, where the monitoring suggested
should be similar to that of a small to medium waste site. The DWAF guidelines are the
absolute minimum requirement that should be adhered to. To properly understand the
interactions on the site, a more extensive monitoing program is recommended
259
Based on the observations from the research and the DWAF Minimum Monitoring
Requirements, it is suggested that for each pivot where mine water irrigation is undertaken,
the following groundwater monitoring is recommended:
o Before installation of monitoring boreholes a detailed conceptual model of the
expected site geohydrology should be constructed so that appropriate monitoring is
put in place.
o As the system is installed, detailed measurements and observations of the geology,
aquifer characteristics (using techniques such as slug or pump testing etc) and should
be done, and the conceptual model verified. This data will also allow a more accurate
determination of the expected impact of the irrigation activities.
o At each pivot site at least one upgradient and two down-gradient boreholes should be
installed, together with a borehole pair near the centre of the pivot itself.
o The boreholes on the outside of the irrigation must be constructed in a manner which
is consistent with DWAF's Minimum Requirements document.
o The borehole pair within the pivot should be installed as follows:
o 1 Shallow borehole that is drilled down to the top of the hard rock underlying
the irrigation area, to below the weathered zone. This borehole should have a
very short length of solid casing followed by slotted casing or a borehole
screen to the bottom of the borehole. If the material is not competent, it is
recommended that a slotted piezometer be installed with a gravel pack to
ensure that the shallow groundwater can be accurately characterized.
o In close proximity a deeper borehole should be drilled into the Karoo
formation. This borehole should be cased off with solid casing to below the
weathered zone.
o At these boreholes proper sanitary seals and/or other preventative measures
should be put in place to prevent the irrigation water from flowing directly into
the boreholes.
o The following parameters should be measured in each borehole:
o Water levels on a monthly basis
o Groundwater sampling for the macro-constituents at least 6 mombty; and
before and after each crop is planted and harvested. The analytical
constituents should include pH, EC, Ca, Mg, Na, K, S04, HC03, Cl, N03 and
P04.
o Data should be compiled into a database and handled as prescribed in DWAF's
Minimum Requirements (2005).
260
9.4 GIMI- A user-friendly decision support application for mine water irrigation
In an attempt to standardise decision-making regarding mine water irrigation, the criteria,
data, rules and fundamentals discussed in the preceding sections have been combined in a
user-friendly tool, called GIMI (Groundwater Impacts from Minewater Irrigation). GIMI is a
screening tool to help users make informed decisions about mine water irrigation. GIMI
consist of a GIS system which contains all the base maps as shape files and an assessment
interface where scenarios are built using a CAD interface. (An "Install" for GIMI is on the
attached DVD).
The aims of this tool can be summarised as follows:
• User-friendly tool to integrate considerations/understanding
• Should take most important lessons into account
• Simplifies complex interactions
• Gives guidance on:
o Applicability
o Monitoring
o Data requirements
The approach followed was to:
• Take information of different types into account
261
• Provide a screening tool
• Employ a conservative methodology to ensure protection of the water resources
• Combine yes/no approach with integration of (simplified) key fundamentals
• Series of input screens with guidance
• Output of assessment
9.4.1 Overview of GIMI operation
The working of the tool is very much in line with that discussed above. Initially users will
provide information on the area, type of irrigation, water quality and similar values. Examples
of the input screens are provided below:
GIMI - Groundwater nnpact from Minewater Irrigation
Eile 'bew l;ssessment Utilities tielp
If ~elect POint ·10 Q.r.phing LlI Generale ileport
H ill Main Layers ...
'_ ~o Towns
DL Name
IProlecledRiversColleries- LandType
- ~.., Name
a AB9
a BAl9
a BA3
BA4
BA5
o BBll
o BB12
BB13
o 8814
• RR17
Figure 9-5: GIMI input screen with c/ickable soils map to link to soil properties.
The clickable soil map will link the detailed soil information directly to the area of irrigation.
262
GIMI - Groundwater Impact from MInewater Irngiltlon
t:telp
~8ve Assessment '. Q.raphing !J1Generate !leport
00"" Topernep [mamsi]0" Waler Level [mam Rainlall [mm/a] 6810. Concenlralion [mgII EC [mS/m] 1Qualily ~AQUJter Cia ~lf1catl None- 0. Virgin ~ Dolomitic AQulter NoPivol
0[1) ~ Sensuwe S1te... NoneSpoils
IjlGeolow 'w'e-atheredRod
Figure 9-6: Example of input screen from GIM/.
9.4.2 Additiona/ features in GIM/
In terms of irrigation water quality, GIMI has a few features in which will greatly assist users in
determining the eventual impacts. The water quality input can either be done manually or can
be linked to a WISH database or Excel sheet. This water quality will be scrutinized for relative
ionic contribution (see "critical flaws"), SAR indices and EC/TDS ratios for comparison to
aquifer water. Additionally a link has been built to the PHREEQC geochemical model. This
will enable correct determination of the concentration and equilibration of the water as
volumes are reduced. This ensures that waters remain in ionic balancer and equilibrium with
gypsum, calcite and dolomite. The mass of precipitated minerals can also be determined in
this manner. The Chemical model used in GIMI is PHREEQC version 1.5 PHREEQC is a
computer program written in the C programming language that is designed to perform a wide
variety of low-temperature aqueous geochemical calculations. PHREEQC is based on an ion-
association aqueous model and has capabilities for (1) speciation and saturation-index
calculations; (2) batch-reaction and one-dimensional (1D) transport calculations involving
reversible reactions, which include aqueous, mineral, gas, solid-solution, surface-
complexation, and ion-exchange equilibria, and irreversible reactions, which include specified
mole transfers of reactants, kinetically controlled reactions, mixing of solutions, and
temperature changes; and (3) inverse modeling, which finds sets of mineral and gas mole
263
transfers that account for differences in composition between waters, within specified
compositional uncertainty limits.
Results of the chemical model are displayed for each Pivot object in the object editor.
The impact on receptors will be determined by using the concentration and load as
determined by the PHREEQC and the water balance equations, and the Domenico (1987)
approach. The time span of irrigation, the distance to the receptor and the water quality
objective (where none- is present the SA drinking water standards for ECITDS are used as
default) at this receptor will be used to assess applicability.
,'" Gral,hing Tool Iël[§~
Primary Secondary
o None o None
@pH O pH
10 1,300
o 1,200Ca S Cl C~
1,100
8 c
OMQ 1,000 o Mg
7
o 900Na o Na
6 800c
(Jl
0
OK 700 ...
as·' '3lO
600 .:.
OCI OCI
500
O S04 ......... :4003·
......... :300
o Gypsum o Gypsum......... ·200
o ......... :100Calcite o Calcite
- :0
o Dolomite o Dolomite
--
[ij 3D [i] Marks vi OK
Figure 9-7 Water Quality and Environmental Data in GIMI.
The Domenico analytical model is based on the advection-dispersion partial-differential
equation for organic contaminant transport processes in groundwater as described below
(Domenico and Robbins 1985):
Where,
C - contaminant concentration in groundwater (mg/L),
t - time (day),
264
v - groundwater seepage velocity (ftIday),
x, y, z - coordinates to the three dimensions (ft),
Ox, Dy, Dz - dispersion coefficients for the x, y, z dimensions (ft2/day),respectively.
To solve equation (1) analytically, under conditions of the steady-state source and finite
continuous source dimension with one-dimensional groundwater velocity, three-dimensional
dispersion, and a first order degradation rate constant, the analytical solution can be
expressed as (Domenico 1987):
Where:
Cx - contaminant concentration in a downgradient well along the plume centerline
at a distance x(mg/L),
Co - contaminant concentration in the source well (mg/L),
x- centerline distance between the downgradient well and source well (ft),
(x, (y, and (z - longitudinal, transverse, and vertical dispersivity (ft), respectively,
Dx=(x·i , Dy=(y·i , Dz = (z·i ,
L - degradation rate constant (1/day),
L=0.693/t1/2 (where t1/2is the degradation half-life of the compound).
i-groundwater velocity (ft/day),
Y - source width (ft),
Z - source depth (ft),
erf - error function,
exp - exponential function.
The Domenico Analytical Model assumes:
(1) The finite source dimension,
(2) The steady state source,
(3) Homogeneous aquifer properties,
(4) One dimensional groundwater flow,
(5) First order degradation rate,
265
(6) Contaminant concentration estimated at the centerline of the plume,
(7) Molecular diffusion based on concentration gradient is neglected,
(8) No retardation (e.g., sorption) in transport process.
Understanding model assumptions is crucial to simulate transport process for a specific
contaminant in groundwater.
9.4.3 Guidelines for operating GIMI
The GIS interface has its own help file and can be accessed through the button III shown in
the screen shot in Figure 9-5.Under Content in the Help file there is a menu with:
• Introduction
• Frequently asked questions
• Utilities
• Models
• Quick start guide
The following section provides a step-by-step guideline to operate GIMI:
Step 1 - Select the point of interest using the" button.
Step 2 - Zoom to required extent of the assessment by placing the GIS in zoom mode
through the use of the ::;p+ button. Hold down the left mouse button and drag a rectangle to
the bottom right that will represent the zoom area.
Step 3 - Click on Create Assessment using the [J button. The number of available
topography points in the selected area will be displayed. Click OK to continue.
The default model grid will be displayed on the assessment area, ready to start building the
assessment:
9.4.3.1 Assessment Tooibar
The assessment tooibar is where the navigation and placement features are accessed from.
The list below provides a summary of the tooibar buttons:
Ál Zoom to full extent of the area
~o Set the zoom factor to 100%
£.1 Selection mode to select object and move them around in the CAD interface
P Zoom mode (dragging a rectangle from top left to bottom right)
266
~ Zoom in 2x from the selected point
J7x Zoom out 2x from the selected point
~ Pan mode
,f/ Place a ruler object on the mine area
A Place a text object on the mine area
ct Rotate selected object
~ Delete selected object
~ Move selected object to the front
~ Move selected object to the back
ll? Use reference (Set reference by holding down Ctrl & Alt and click point)
9.4.3.2 Assessment Objects
The objects are selected from the object menu. The menu is invoked through a right click on
the object tree as shown below:
Irrigation •
Receptors •
Quality •
Crops •
The connecting rules for objects are enforced through the menu system. Thus objects in the
menu are disabled for a particular tree node if the objects are not allowed to connect to the
node in question.
A summary of all object types:
~ Mine or area object which from the parent of all objects in the object tree
Em Model object that represents the transport model
Contour map object (topography, water levels and concentration)
IIQuality object representing the mine water quality
• Virgin soils
!mm Spoils object
267
Pivot object, which is placed on the virgin soils or spoils
Crop object, which is connected to pivot objects
!~A~cid based accounting object connected to the spoils
• Borehole object, which is placed on the area (mine) and acts as a receptor
~ Population object, which is placed on the area (mine) and acts as a receptor
Y Receptor object, which is placed on the area (mine)
9.4.3.3 Object Properties
Each object consists of properties. Some properties is user supplied and some are calculated
from the system. The screen below is the legend describing the various property types:
~
Property legend
t".i"l Locked property updated through calculation of lower levels
ï." Property aquired from a higher level in the assessment tree
t"."~ Property updated through calculation of lower levels
z=~ Calculated property value from object properties
~ Locked property that cannot be altered b}' user
!fil Property aquired from model interface
El Value obtained from TMU
~ User input required via selection
User input required
9.4.3.4 Assessment Interface
The assessment interface consists of 3 main components:
1. The object tree view on left hand side of the screen
2. The modeling area in the center of the screen
3. The object property editor on the right hand side of the screen
268
r~ GIMI • GI'oundwllter Impl'IC11rOm Minewa1er JrrlgBtlon élêlgjj
BI. :!jew assessment U,tllities t::!eJp
&B.esetMl!lp ., ~eled Polrrt ID Run 8S'$essment ~(lve AS$e$$menl ,_ qri!lphlng ~ Oenererte Report
~ Assessment
Jil e JiiJ P ~ ~ f'I~ A ê lOt ) jl!
8 Ga ~ KriellBB41
0111l Model c? Propert)! IVtllue ISpoilsAre<!llha] 100 -IL! "Ii!! Qu,lity J ~ UArea[ha] 1~. Virgin "I Spoil. U,.g.I%1 ·1,000[1] Spoil, ,._,...", ~ SoilDepth [mm] 1
( Cl., Depth [mm] 2
'i i'- Cl" 1"1 20Root Zone [mm) 1
Á weter level [mbgl2Flooded Spoil, 1"11} HNP IkgA CaC03] 1),\ !.!AP Ikg/t C.C031 1 r--
~'--} Chemistr_\) jV.lue
I~ 0,N
File: INone --~I -~--2903553
r---- ...J I .a
Object Tree
The object tree represents the assessment and all objects used is visible in the tree view. By
clicking on each tree node the object properties will update on the right hand side of the
screen in the property editor. The objects with a check mark next to them can be switched on
and off simply by clicking on the check mark. All nodes in the tree can be renamed and this is
important if multiple receptors, pivots and crops are used to distinguish between them in the
report. Click on the node text to highlight it and click on it again to put it in edit mode.
Modeling Area
The modeling area is a CAD interface with a grid placed on top of it which represents the
transport model used in GIMI. The assessment tooibar is used to navigate and place objects
in the assessment interface. Visual objects are displayed on the assessment area as the user
places them. Note not all objects in the object tree is visual objects and may only exist in the
tree.
Property Editor
The property editor is the place where the objects gets populated with information. Some
properties are user input and some are calculated. Each property is color coded and has an
icon next to it depicting the property type.
269
9.4.3.5 Analysing an Assessment
Once the assessment has been built through the available objects and the user is confident
that all data requirements have been addressed the assessment can be run through the use
of the (g) button. Once the assessment has run the user will notice the various contour maps
that are generated and that various object properties have been updated after the model run.
If the user changes any of the object properties the assessment must be re-run to update
relevant object properties.
After an assessment is run the user has the option of saving the assessment for future use
through the use of the button.
Graphing of the pivot parameters can be done through the graphing tool' as shown below:
A report of the assessment highlightingcritical flaws and warnings can be generated through
the use of the ~ button (Figure 9-8). The report can be saved in the custom report format to
be opened through the report generator or can be saved to PDF format for distribution.
Report generated by GIMlon 2006·11-03
Generol
Please reter lo meunreport (avaiable ~...,der hep) tor de tal qvt:lel"es
Mil'\mvm MC<"\iloril'\g Reqviremel'\~:
MOl'\Ihly weterleve readlrgs
6 Menlhly waterqvaity readng
Mlrimvm Borehole Reqviremel'\~:
2 Boreholes 11'\ the pivot at dfferel'\t depths
2 Boreholes dowl'\ gradel'\t trom Ihe pivot
1 Borehole vp gradiel'\t tom the pivot
Critical Flaws & Warnings
Hole that j~r Cl ,y cnficall1aw that exsts, 1'\0 nngallol'\ '$ recccmercled
BOREHOLES- Wall'\hg: No boreholes bVl'\d
VIRGIN SOIL - Critical Aaw: EC" 150, the irrigatiol'\ water ispotcbie
VIRGIN SOIL- Critical Aaw: Soil sutabilityratng = N
VIRGIN SOIL- Critical Aaw: Soildepth" 1500mm
SPOILS-ll'\formatien: Potel'\llaly Acid Gel'\eralhg
SPOILS- Critical Aaw: Tolal inigalgiol'\ creo ~ 12Oha, follow DWAFwater vse permittil'\g gvidelil'\es
SPOILS· Clifical Aaw: Soiloeoth e 1500mm
SPOILS - Wamirg: ~atien will accelerale acidiicatien al'\d wote- qvallty deleriorafiol'\ rales
SPOILS - Wamil'\g: Irrigatien islimited lo a 5 year period
Figure 9-8: Report of the assessment.
270
10 CONCLUSIONS AND RECOMMENDATIONS
This thesis has investigated the impact of irrigation with coal mine water on groundwater
resources. Various field and laboratory experiments were done to determine the impact and
feasibility of irrigation in the short to medium term. Based on the findings from this study, the
results were used to set up guidelines and to develop the GIMI tool. This tool should assist in
the practical implementation of mine water irrigation as part of optimal mine water
management.
10.1 Availability of irrigation water:
Opencast collieries: Approximately 75 000 ha of opencast will eventually be rehabilitated in
the Mpumalanga region. If it is assumed that the recharge to the rehabilitated opencasts will
be in the order of 20% of MAP, and that the annual irrigation water for 1 ha is 5 000 m3, then
in excess of 20000 ha could be irrigated for a single crop.
Underground collieries: Approximately 19500 m3 of water will be in storage when filled for
every hectare mined. Calculated according the average rainfall, shallow mines with a
recharge of 5-10% of MAP will annually have an amount of 350 - 700 m3 /ha available
annually for irrigation. This means that every 7-14 ha will provide enough water to irrigate one
hectare. For deep underground bord-and-pillar areas with a recharge of 4%, 18 hectares will
provide enough water for one hectare of irrigation.
However, not all of this water will be available for irrigation. Three desalination plants are
envisaged for the Mpumalanga coal fields to provide water to the local authorities. Less than
20% will thus be available for other water uses, of which irrigation is only one.
10.2 Water levels and water quality
1. The water levels at all the pivots have varied very little over time.
2. The irrigation water has, as yet, shown a minimal impact on the groundwater. Some
exceptions occur, but none of the boreholes outside the pivot areas show any meaningful
change in water quality from leaching from the irrigated area. This implies a very slow
movement of the constituents in the irrigation water, and that these are attenuated by
different mechanisms between the surface and the aquifers, often in the clay layers. The
271
amount leaching through the soil is small enough to be easily diluted by the regional
groundwater flow. It thus appears that the soil type and morphology plays a very important
role in the vertical movement of the irrigation water constituents. However, in the sandy
aquifer at New Vaal the analyses show that there is a steady and consistent rise in the
electrical conductivity in the groundwater underlying the pivot. This indicates that the
irrigation water is slowly leaching through the more clayey sand layers and is influencing the
groundwater quality directly below the irrigation area.
10.3 Soil and Soil Water Analysis
The soil analysis indicated that most of the salts are captured in the upper one or two meters
of the soil profile. The salts move along the profile in the soil water. The data indicates that
the clay layers, which playa major role in the vertical flux of the water, therefore also have an
influence on the salt distribution through the soil profiles.
According to the soli waterquality the water is saturated with gypsum in the first meter or two,
indicating gypsum precipitation. The calcium decreases much more than the other cations,
indicating the gypsum precipitation.
XRF/XRD analysis indicated that gypsum precipitation occur since Ca values are far higher in
the shallow portions, while Mg values are more evenly distributed. This suggests that the
distribution is chemically controlled rather than by surface interactions since these two ions
should have similar sorptive capacities. This is supported by the Na values, which are evenly
distributed with depth. Research has revealed that bassanite (CaS04 O.5H20, or calcium
sulphate hemihydrate) is present in most of the samples.
Sulphate adsorption: Leaching column experiments indicated sulphate precipitation as well
as adsorption. This means that not all sulphate will be leached by excessive irrigation or rain.
It must be noted, however, that desorption can occur and that the sorbed sulphate will not be
irreversibly fixed in the soil profile.
Soil moisture: Even at sites with shallow water levels there is a visible decrease in soil
moisture during winter when effective irrigation is being practiced. This will result in salts
retained in the upper layers of the soil, as indicated by the soil analyses and the soil water
analyses. The clay layers are important in the downward movement of the soil water. This will
also have an effect on the salts being attenuated in the soil layers above the saprolite zone.
The profile above 1m is less moist than at 1m. Therefore the conditions above one meter will
be much more favourable for gypsum precipitation.
272
Salt balance: One of the critical issues in successful irrigation with the coalmine waters is
how much of the salts are retained in the soil profile over the medium term. Geochemical
modeling has indicated that the soil water in the top one meter is saturated with gypsum, and
precipitation should occur. In the deeper moist layers the water is still undersaturated with
regard to gypsum. The salt balance calculations revealed that 80% of the sulphate applied
during irrigation since 1999 (2001 at other sites) has been retained in the soil profile (also as
gypsum), most of it in the upper two meters. On average, between 75% and 80% of the
sulphate retained are in the pore water (more than 60% of the total sulphate applied) and the
rest precipitates or is sorbed. Hydraulic, adsorption and attenuation factors prevent the salts
in irrigation mine water from being directly and rapidly mobilised down the deeper soil profile
and into the aquifer. New Vaal is the exception where a large amount of the salts is retained
in the clayey soils deeper down the profile. Most of the salts in the clayey soil of Syferfontein
are retained in the upper 0.6 m. At Kleinkopje 4 the salts have migrated a little deeper, with
most remaining at 1-2 m in the moist area above the clayey layer.
Leaching: The conditions in the soil will be drier under normal cultivating conditions (after
irrigation has ceased), and recharge will maximum be 5-6%. It will thus take a very long time
to leach the sulphate from the soil (probably more than a hundred years).
10.4 Hydraulic behaviour in the unsaturated zone
o The chloride values at all the different irrigation sites showed a decrease in chloride
values with depth, implicating bimodal flow.
o Twenty five percent of the total volume of water applied through irrigation reaches the
water level.
o There is a correlation between clay content of the soil and the diffusive component of
the bimodal flow. Due to the moist conditions at irrigation sites, swelling of the clay
results in a very low vertical conductivity value of 10-4 mId.
o Due to lateral flow in the weathered saprolite zone, dilution occurs (lateral flow
contributes approximately 85% towards the dilution process). Water will move along
the weathered zone, and part of the salinity will move deeper down into the fractured
aquifer, some will daylight or seep into adjacent strata. This means that the release of
salinity into the groundwater is a very slow process, which makes irrigation a viable
option over the short to medium term.
273
10.4.1 Kleinkopje 1
In the short to medium term there is a build-up of salts in the soil profile, with most of it (83%)
retained in the soil water. A portion of these salts moves through to the groundwater, which
will limit long-term sustainability of irrigation from an environmental perspective.
Note: The salinity concentrations at this site are a conservative estimate, because the test
site is situated close to a borehole that drains into the mine, thus leaching some of the salts
that build-up in the upper profile, down the mine during high rainfall or irrigation events.
10.4.2 Kleinkopje 4
In the short to medium term there is a build-up of salts in the soil profile, with 77% of it in the
soil water. A portion of these salts moves through to the groundwater, which impact
negatively on a long-term irrigation project. However, the build-up is slow due to dilution
because of the lateral movement of water, which in the short to medium term will not impact
negatively on the aquifer.
However, gypsum improves the soil structure, especially in clayey soils, and replaces salts
like sodium (See Appendix C for an explanation). This make the irrigation option with
gypsiferous type water even more viable.
10.4.3 Syferfontein
In the short to medium term there is a build-up of salts in the soil profile especially in the
upper 1m, with little movement down the profile. A portion of these salts does move through
to the groundwater (as can be seen from the water quality in the well-constructed borehole),
which impact negatively on a long-term irrigation project. Although the swelling clay
(smectite) of well developed Arcadias at Syferfontein can restrict water movement to micro
pore flow, it shrinks when dry. This can cause large cracks in the upper soil profile (Soil
Classification Working Group, 1991) which allows for large volumes of water to enter the soil
at the start of a rainfall event. This is a reason for the higher than expected percentage water
that moves through this very clayey soil profile.
Due to the high sodium content of the water at this very clayey site, the feasibility of irrigation
is questionable. The disruption of the soil structure, together with clay dispersion, greatly
reduces the soil permeability since the larger pores are blocked (See Appendix C).
10.4.4 New Vaal
In the short to medium term there is a slow but definite build-up of salts in the groundwater.
Most of the salts are retained in the shallow clayey layer and the deeper clayey soils because
of the movement of the water through the profile. The salts that move through to the
274
groundwater will impact negatively on a long-term irrigation project. However, the build-up is
slow due to dilution.
10.5 Detailed site evaluation for opencast (spoils) sites
o Models indicate the great importance of the reactive nature of the spoils on which
irrigation will occur. The potential for acid generation should be a very important
consideration of where to apply mine water irrigation, while it is also clear that in high
sulphide areas, the irrigation will enhance reactions and therefore lead to higher salt
loads.
o The re-use of irrigation in certain systems has both advantages and disadvantages.
The advantages of a "closed" or semi-closed loop means that the salt is not released
into the environment. It has also been shown that secondary minerals solubility
constraints will limit the possible upper values. It is therefore clear that the nature of
the recirculating water is important. Where mine water used for irrigation is largely
gypsyferous, the recirculation scenario is feasible under the correct conditions (e.g.
impermeable floor rocks). Where this water contains many other elements such as
sodium chloride, the solubility limits become far higher due to the increase in the ionic
strength of the percolating water. Under these conditions, concentrations will continue
rising to very high levels. This continuous rise is confirmed not only by theoretical
considerations and geochemical models, but also by observations from laboratory
tests. The implication is that the nature of the spoils, and the irrigation water quality,
must be considered in terms of what the likely soluble concentrations of different
elements will be.
o Where mine water of relatively consistent quality is used, the changes in longer-term
quality will be far less dramatic. The system reaches an equilibrium state and can
continue in this state for long periods of time.
Recommendation:
The complexity of the interactions makes it necessary to consider several factors
simultaneously in terms of the feasibility and effects of mine water irrigation on spoils. One of
the important factors seems to be the time of year (during winter when it does not rain) and
the management of the irrigation schedule (not to over-irrigate and wet the spoils too much).
Field tests have indicated that during winter the spoils dry out in depth in the profile, which
will result in an increase of acid generation. Overall the modeling, laboratory tests and field
tests indicate that mine water irrigation on spoils will lead to a further deterioration of spoils
water quality, when compared to rainfall infiltration into the spoils. Where mine water of
relatively consistent quality is used, the changes in longer-term quality will be far less
275
dramatic. It is suggested that irrigation with water containing enough buffering potential be
used for irrigation on spoils.
10.6 Overall Recommendation
Irrigation with mine water must be implemented such that the environmental impacts are
minimized Based on the slow expected salintfy bultd-uo observed at the different sites, it is
recommended that irrigation be done on an alternating basis (i.e. alternating between two
pivots over time), if site criteria selection has been adhered too (See Chapter 9), and that
mine water irrigation should not be done for periods exceeding 10 years in any particular
area. A non-negotiable prerequisite is that appropriate monitoring must be put in place at
such a site. These boreholes must be constructed in such a way that they monitor a/l the
different flow zones and aquifers at these sites.
It is imperitave that the guidelines outlined in Chapter 9, being adhered too. The guidelines
provided should lead to sustainable mine water irrigation.
It is unknown how long this accumulation of salts in the upper layers can continue before
leaching into the underlying aquifers, but in the short to medium term, the evidence from
groundwater monitoring shows that irrigation with mine water does not hold significant threats
for the regional groundwater quality.
To determine the impact of the viability of irrigation with coal mine water in the long term, it is
suggested that monitoring continues at the various test sites. Monitoring should be done as
described in Chapter 9.3.
Due to the relatively short period of research, there are still uncertainties regarding the long-
term sustainability of irrigation, which may be answered if monitoring at these sites continues.
This wil address the following issues:
o The amount of gypsum that can precipitate in the soil, and for how long this will this
sustainable.
o The effect of the gypsum build-up on the structure of the soil?
o How much of the applied salts can precipitate into the upper layers of the profile.
276
o How much of the salts will leach out, and at what rate. If irrigation ceases at some of
the sites (as is the case at New Vaal), this should be investigated. This will also
indicate how the salts will affect the groundwaterover the long term.
This study has shown that it is possible to irrigate with coal mine water in such a manner
that the impacts on groundwater resources is minimal. Therefore it is imperitave that the
regulatory agencies and water managers consider the potential socia-economic benefits
when evaluating coal mine water irrigation.
277
11 REFERENCES
Abdulgawad, G., Bashour, I. and Elmadhon, Y. (1977) Sink holes formation in some salt-
affected soils and their origin. Libyan Journal of Agriculture Vol.6, pp.159-164.
Abrol, LP. and Yaday, Jai Singth Pal and Massoud, F.1. ( 1988) Salt Effected soils and Their
Management. Food and Agriculture Organization of the United Nations, Rome
Adams, F. (1971) Ionic concentrations and activities in soil solutions. Proceedings of the Soil
Science Society of America Vol. 35, pp. 420-426.
Agboola, AA and Corey, R.B. (1973) The relationship between soil pH, organic matter,
available phosphorus, exchangeable potassium, calcium magnesium and nine elements in
the maize tissue. Soli Science Vol.115, pp.367-375.
AI, T.A, Blowes, O.W., Jambor, J.L. (1994a) A geochemical study of the main tailings
impoundment at the Falconbridge, Kidd Creek Division metallurgical site, Timmins, Ontario.
In: Jambor, J.L., Blowes, O.W. (Eds.), Short Course Handbook on Environmental
Geochemistry of Sulfide Mine-Wastes. Mineralogical Association of Canada, 22, 333-364.
AI, T.A, Blowes, O.W., Jambor, J.L. and Scott, J.D. (1994 b) The geochemistry of mine-
waste pore water affected by the combined disposal of natrojarosite and base-metal sulphide
tailings at Kidd Creek, Timmins, Ontario. Can. Geotech. J. Vol. 31, pp. 502-512
Alimaganbetov, S. and Akhanov, Z.U. (1970) Accumulation of gypsum and carbonates and
their effect on soil fertility. Vest. Akad Nauk kazakh SSRVol. 3, pp. 36-41.
Ajwa, H.A. and Tabatabai, M.A. (1995) Metal -induced sulphate adsorption by soils: I. Effect
of pH and ionic strength. Soli Science. Vol. 159(1), pp. 32-42
Allison, G.B. (1988) A review of some of the physical, chemical and isotopic techniques
available for estimating groundwater recharge. In: I. Simmers (ed.) Estimation of Natural
Groundwater recharge. NATO ASI Series C222, D. Reidel Publishing Company, Dordrecht,
pp.49-72.
Allison, G.B. and Hughes, M.W. (1978) The use of environmental chloride and tritium to
estimate total recharge to an unconfined aquifer. Aust. J Soli Res., Vol. 16, pp.181-195.
Allison, G.B., Gee, G.W. and Tyler, S.W. (1994) Vadose-zone techniques for estimating
groundwater recharge in arid and semi-regions. Soil Sci. Soc. Am. J, Vol. 58, pp. 6-14.
278
American Society for Testing and Materials (1996) ASTM Designation: D 5744 - 96 -
Standard Test Method for Accelerated Weathering of Solid Materials Using a Modified
Humidity Cell, ASTM, West Conshohocken, PA, p.13.
Annandale, J.G., Jovanovic, N.Z. and Claassens, A.S. (2002) The influence of irrigation with
gypsiferous mine water on soil properties and drainage water. WRC Report No. 858/1/02.
ISBN No: 1-86845-952-7.
Annandale, J.G., van der Westhuizen, A.M., de Jager, P.C., Rethman, N.F.G., Jovanovic
N.l., Madisane, P.C., Beletse, Y., Tesfamariam, E., Smith, H.J.C., Benadé, N., Hodgson, F.,
Usher, B., Vermeulen, D. and Lorentz, S. (2004) Predicting the environmental impact and
sustainability of irrigation with gypsiferous mine water. Report to the WATER RESEARCH
COMMISSION by Department of Plant Production and Soil Science, Faculty of Biological
and Agricultural Sciences, University of Pretoria, Pretoria, 0001, South Africa. WRC Report
No: 1149.
Annandale, J.G., Beletse, Y., A.M., de Jager, P.C., Jovanovic N.Z., Benadé, N., Lorentz, S.,
Hodgson, F.D.I., Usher, B., Vermeulen, D. and Aken, M.E. (2006) Predicting the
environmental impact and sustainability of irrigation with gypsiferous mine water. Report to
the Water Research Commission, Pretoria, 0001, South Africa. WRC Report No: 1149/02/06.
Appelo, C. A. J. and Postma, D. (1993) Geochemistry, Groundwater and Pollution. A. A
Balkema Publishers, Rotterdam, Netherlands.
Ayars, J.E., Sappe, R.W., Cone, D., and Wichelns, D. (1997) Managing salt load in irrigation
district drainage water, In: Findikakis, A.N. and Stauffer, F. (eds.) Groundwater: An
endangered resource: International Association for Hydraulic Research, 27th Congress, San
Francisco, California, August 10-15, 1997, Proceedings, p. 277-282. Agricultural Drainage
Water Quality: Food and Agricultural Organization of the United Nations, Water Report No.
13, Rome, pp. 29-40.
Azzie B. A. M. (2002) Coal mine waters in South Africa: Their geochemistry, quality and
classification. PhD. Thesis. University of Cape Town.
Bajwa, M.S. and Josan, A.S. (1989) Effects of Alternating Sodic and Non-Sodic Irrigation on
the Build-Up of Sodium in the Soil and on the Crop Yields in Northern India. Experimental
Agriculture Vol. 25, 199-205.
Baldys, Stanley, III, Ham, L.K., and Fossum, K.D. (1995) Summary statistics and trend
analysis of water-quality data at sites in the Gila River Basin, New Mexico and Arizona: Us.
Geological Survey Water-Resources Investigations Report 95-4083, p. 86
279
Banks, D., Younger, P.L. and Dumpleton, S. (1996) The historical use of mine-drainage and
pyrite-oxidation waters in central and eastern England, United Kingdom. Hydrogeology
Journal, Vol. 4, No.4, pp. 55 - 68.
Baker, R. S., and D. Hillel (1990) Laboratory tests of a theory of fingering during infiltration
into layered soil, Soil Sci. Soc. Am. J., Vol. 54, pp.20-30.
Barla, G.B. and Boshkov, S. (1978) Investigations of differential strata movements and water
table fluctuations during longwall operations at Somerset Mine No.60. Columbia Univ., New
York,Department of Energy, NTIS FE-9041-1, p.149.
Barnard,R.O.,Rethman, N.F.G., Annandale J.G., Mentz, W.H., Jovanovic, N.Z. (1998) The
screening of Crop,Pasture and Wetland species for tolerance of polluted water originating
from coal mines WRC report No.582/1/98
Basinski, J.J. (1967) The Russian approach to classification and its recent development. In:
Selected Papers in SoH Formation and Classification. (Ed: J.V. Drew). Soil Science Society of
America Special Publication No.1, pp. 402-414.
Bazuhair, A.S. and Wood, W.W. (1996) Chloride mass-balance method for estimating
groundwater recharge in arid areas: examples from western Saudi Arabia. Journal of
Hydrology, Vol. 186, pp.153-159.
Bennie, A.T.P., Strydom, M.G. and Very, H.S. (1998) Gebruik van rekenaarmodelle vir
landboukudige waterbestuur opekotoopvlak. WRC Report No. TT 102/98, Water Research
Commission, South Africa
Bethke, C. M. (1996) Geochemical Reaction Modelling - Concepts and Applications. Oxford
University Press, New York.
Beven, K., and P. Germann (1982) Macropores and water flow in soils, Water Resour. Res.,
Vol. 18, pp.1311-1325.
Black, C.A. (1965) "Methods of Soil Analysis," Part I, Am. Soc Agron., Madison, Wisconsin.
Blowes, O.W., Reardon, J.E., Jambor, J.L. and Cherry, J.A. (1991) The formation and
potential importance of cemented layers in inactive sulfide mine tailings. Geochim.
Cosmochim. Acta, Vol. 55, pp. 965-978.
Bohn, H.L and Bohn, R.K. (1996) Solid Activity Coefficients of Soil Components.
GEODERMA, Vo1.38, pp. 3-18.
Botha, J.F. (1996) Principles of Groundwater Motion. Unpublished lecture notes, Institute for
Groundwater Studies, University of the Free State, P.O.Box338, Bloemfontein, SA 9300.
280
Bouwer, Herman. (1990) Agricultural chemicals and groundwater quality: Journal of Soil and
Water Conservation, Vol. 45, no. 2, pp. 184-189.
Bouwer, Herman. (1997) Arizona's long-term water outlook: From NIMTO to AMTO: Arizona
Hydrological Society Newsletter, v. 14, April and May issues.
Bouwer, H. and Rice, RC. (1976) A slug test for determining hydraulic conductivity of
unconfined aquifers with completely or penetrating wells. Water Resources Res. Vol. 12,
pp423-428.
Bower, C.A., Reitemeier, RF. and Fireman, M. (1952) Exchangeable cation analysis of
saline and alkali soils. Soil Science Vol. 73, pp.251-261.
Boyadgiev, T.G. (1974) Contribution to the knowledge of gypsiferous soils.
AGON/SF/SYR/67/522. FAO, Rome.
Bredenkamp O.B. (1978) Quantitative Estimation of Groundwater Recharge with Special
Reference to the use of Natural Radio-active Isotopes and Hydrological Simulation.
Unpublished. Ph.D. thesis, UOFS, Bloemfontein, p.367.
Broughton, L. M., Chambers, R W. and Roberston, A. M. (1992) Mine rock guidelines:
Design and control of drainage water quality. Report No. 93301.
Burne, S., Wheater, H.S., Butler, A.P., Johnston, P.M., Wadley, P., Shaw, G. and Bell, J.N.B.
(1994) Radionuclide transport above a near-surface water table: I. An automated Iysimeter
facility for near-surface contaminant transport studies. J. Environ. Qual. 23 6, pp. 1318-1329
Butler, A.P., Chen, J., Aguero, A. and Edlund, O. (1999) Performance assessment of models
for water flow and radionuclide transport in vegetated soils using Iysimeter data. J. Environ.
Radioact. 422, pp. 271-288
Cairncross, B., Hart, RJ. and Willis, J.P. (1990) Geochemistry and sedimentology of coal
seams from the Permian Witbank Coalfield, South Africa; a means of identification.
International Journal of Coal Geology, Vol. 16, pp.309-325.
Campbell, G.S. (1985) Soil physics with basic. Transport models for soil-plant systems.
Elsevier Science Publishers, p.149.
Campbell, R (2001) The chemical response of deep, leached and weathered soils of the
Mpumlanga Highveld, Souh Africa, to irrigation with saline mine water. Unpublished MSc
thesis, Faculty of Science, University of Cape Town, South Africa.
281
Carenas, P.B. and Marfil, R. (1979) Petrografia y geochimica de yesos actuales
continentales de la región manchega. Estudios Geot, Vol. 35, pp.77-91.
Carlson, R.M., Buchanan, J.R., Kapustka, T.E. and Uriu, K. (1974) Displacement of fertilizer
potassium in soil columns with gypsum. Proceedings of the American Society for Hortkultarel
Science, Vol. 99, pp. 221-222.
Caruccio, F.T., Geidel, G. and Williams, R. (1984) Induced alkaline recharge zones to
mitigate acidic seeps. In: Proceedings of the 1984 Symposium of Surface Mining, Hydrology,
Sedimentology and Reclamation, pp. 43-47. Lexington, Kentucky: University of Kentucky
Carson, C.D., Fanning, D.S. and Dixon, J.B. (1982) Alfisols and ultisols with acid sulphate
weathering features in Texas. In: Kittrick, J.A., Fanning, D.S. and Hossner, L.R. (Eds) Acid
Sulphate Weathering. Soil Science Society of America Spec. Pub. No. 10, Madison,
Wisconsin, pp127-146.
Castroviejo, S. and Porta, J. (1975) Apport a I'écologie de la végétation des zones salées des
rives de la Giguela. Colloques Phytosociologiques, Vol. IV, pp.115-139. Les vases salées.
Lille, France.
Childs, E.C. (1967) Soil moisture theory. In: Advances in hydroscience, Vo14. Ed: V.T.Chow,
pp. 73-117. New York: Academic press.
Cooper, H.H., Bredehoeft, J.D. and Papadopulos, l.S. (1967) Response of a finite-diameter
well to an instantaneous charge of water. Water Resources Res., Vol. 3, pp 263- 269.
Cordy, G.E., and Bouwer, Herman, (1999) Where do the salts go? - The potential efects and
management of salt accumulation in south-central Arizona: Us. Geological Survey Fact
Sheet, pp. 170-98, 4 sheets.
Corwin, D.L. (2000) Evaluation of a simple Iysimeter-design modification to minimize sidewall
flow. Journal of Contaminant Hydrology 42, pp. 35-49.
Corwin, D.L., Waggoner, B.L. and Rhoades, J.D. (1992) Simulating the movement of a
reactive solute through a soil Iysimeter column using a functional transport model. J. Environ.
Sci. Health A 277, pp. 1875-1913
Corwin, D.L. and LeMert, R.D. (1994) Construction and evaluation of an inexpensive
weighing Iysimeter for studying contaminant transport. J. Contam. Hydro! 15, pp. 107-123
Davis, J.G and Burgoa, B. (1995) Interactive mechanisms of anion adsorption with calcium
leaching exchange. Soil Science. Vol. 160, No.1, pp. 256-264
282
Davison, M. and House, W.A. (1988) neutralizing Strategies for Acid Waters: Sodium and
Calcium products generate different acid neutralizing capacities. Water Research WA TRAG,
Vol. 22 No.5, pp. 577-583.
De Jager, C. (2005) Personal communication University of Pretoria
De Marsily, C. (1986) Quantitative Hydrogeology. Groundwater hydrology for Engineers.
ACADEMIC PRESS, INC. ISBN 0-12-208916-2
Denagyer-De Smet, S. (1966) Note sur les relations chimiques plante - sol chez Buxus
sempervirens et Pinus sylvestris L. dans les Alpes maritimes. Bull. Soc. Roy. Bot .Belg. Vol.
99, pp. 221-226.
De Villiers, J.M. (1992) Classification of minesoils. 17th Congress, Soil Sci. Soc. S Afr., 28-30
January 1992, Stellenbosch.
DiCarlo, O.A., Bauters, T. W. J., Darnauit, C. J. G., Steenhuis, T.S. and Parlange J. (1999)
Lateral expansion of preferential flow paths in sands. WATER RESOURCES RESEARCH,
Vol. 35, No.2, pp. 427-434, February 1999.
Dieleman, P.J. and De Ridder, N.A. (1963) Studies of salt and water movement in the Bol
Guini Polder, Chad Republic. Journal of Hydrology, Vol.1, pp. 311-343.
Diment, G.A. and Watson, K.K. (1985) Stability analysis of water movement in unsaturated
porous materials. 3: Experimental studies. Water Resour. Res. 21, pp 979-984.
Domenico, P.A. (1987) An Analytical Model for Multidimensional Transport of a Decaying
Contaminant Species. Journal of Hydrology, Vol. 91, pp. 49-58.
Domenico, P.A., and Robbins, G.A.. (1985) A new method of contaminant plume analysis.
Ground Water; Vo1.23, No.4, pp. 476-485.
Domvile,S.J., Li,M.G., Sollner,L.D. and Nesbitt, W. (1994) Weathering behavior of mine
tailings and waste rock:A surface investigation. In: Proceedings of the International Land
Reclamation and Mine Drainage Conference. April 24-29 Pittsburgh, PA. Vo1.1, pp.167-176.
Doorenbos, J. and Pruitt. W.O. (1976) Crop water requirements. Irrigation and Drainage
Paper No. 24. FAO, Rome.
Drever, J.I. (1997) The geochemistry of Natural Waters: Surface and Groundwater
Environments. yd edition. Prentice-Hall, Upper Saddle Rive, New Jersey, USA.
283
Drost, W. and Neumaier, F (1974) Application of single borehole methods in groundwater
research. In Proceedings, Symposium on Isotope Techniques in Groundwater Hydrology,
1974, Vienna, pp. 241-254.
Drost W, Klots D., Koch A., Moser H., Neumaier F. and Rauert W. (1968) Point Dilution
Methods of Investigating Ground Water Flow in Means of Radioisotopes. Water Resources
Research, Vol. 4, NO.1.
Dubrovsky, J.M., Cherry, J.A., Reardon, J.E. and Vivyurka, A.J. (1985) Geochemical
evolution of inactive pyritic tailings in the Eliot Lake uranium district: 1. The Groundwater
zone. Can. Geotech. J Vol. 22, pp. 110-128.
Du Plessis, H.M. (1983) Using lime treated acid mine water for irrigation. Water Sci. Techno/.
Vol. 15, pp. 145-154.
Durkin, T.V. and Herrmann, J.G. (1996) Introduction: Focusing on the Problem of Mining
Wastes. EPA Seminar Publication/625/R-95/007 on Managing Environmental Problems at
Inactive and Abandoned Metals Mine Sites.
DWAF (2005) Minimum Requirements for Water Monitoring at Waste Management Facilities.
Eary, L.E., Runnells, D.D. and Esposito, K.J. (2003) Geochemical controls on ground water
composition at the Cripple Creek Mining District, Cripple Creek, Colorado. Applied
Geochemistry, Vol. 18 pp. 1-24. Elsevier Science Ltd.
Edmunds, W.M. and Walton, N.R.G. (1980) A geochemical and isotopic approach to
recharge evaluation in semi-arid zones- past and present. In: Arid-Zone Hydrology:
Investigations with Isotope Techniques, IAEA, Vienna, pp. 47-68.
Edmunds, W.M., Darling, W.G. and Kinniburgh, O.G. (1988) Solute profile techniques for
recharge in semi-arid and arid terrain. In: I. Simmers (ed.) Estimation of Natural Groundwater
Recharge, NATO ASI Series C222, Reidel, Dordrecht, pp.139-158.
Energy Information Administration (2000) National Energy Information Centre, 1000
Independance Avenue, SW Washington, DC, USA.
Website: http://www.eia.doe.gov
Erasmus, B.J., Hodgson, F.D.I., Kirstein, F.E., Roper, C.B., Smit, J.S., Steyn, P.P.A. and
Whittaker, R.R.L.G. (1981) Geological, hydrological and ecological factors affecting
increased underground extraction of coal. SAIMM Vacation School, Johannesburg.
Erie, L.J., French, O.F., Bucks, O.A., and Harris, K. (1982) Consumptive use of water by
major crops in the southwestern United States: U.S. Department of Agriculture, Agricultural
Research Service, Conservation Research Report Number 29,40 p.
284
Eriksson, E. and Khunakasem, V. (1969) Chloride concentrations in groundwater, recharge
rate and rate of deposition of chloride in the Israel coastal plane. Journal of Hydrology, Vol. 7,
pp.178-197.
Essington, M.E. (2004) Soil and water chemistry: An integrative approach. CRC Press, Boca
Raton, Florida.
Fetter, C.W. (2001) Applied Hydrogeology. 4th Edition, Prentice-Hall, Inc, Upper Saddle
River, New Yersey. ISBN 0-13-088239-9. pp 228-231.
Freeze, RA. and Cherry, J.A. (1979) Groundwater, Prentice-Hall, Englewood Cliffs, NJ
07632 ISBN 0-13-365312-9
Funke, J.W. (1983) The environmental impact of coal mining and combustion. SA Water,
Science and Technology, Vol. 15, pp. 115-144.
Gebrekristros, R (2006) Personal Communication - Point dilution tests
Gehrels, H. (1998) Groundwater level fluctuations. Published Ph.D. thesis, Free University of
Amsterdam.
Glass, RJ., Parlange, J.-Y. and Steenhuis T.S. (1987) Water infiltration in layered soils
where a fine textured layer overlays a coarse sand, in Infiltration Development and
Application, edited by Y.-U. Fok, pp. 66-81, Am. Soc. of Civ. Eng., Reston, Va.
Glass, RJ., Steenhuis T.S. and Parlange, J.-Y. (1989) Mechanism for finger persistence in
homogeneous, unsaturated, porous media: Theory and verification, Soil Sci., Vol. 148, pp.
60-70.
Gomez-Miguel, V. Peres Arias, J., Guerrero, F. and Roquero, C. (1984) The soils and water
table properties of the Polder area "Cestdlo de Dona Blanca'; Puerto de Santa Maria, Cadiz.
Escuela Técnica Superior de Ingenieros Agrónomos, Universidad Politécnica de Madrid.
Green, G.C. (Ed.) (1985) Estimated irrigation requirements of crops in South Africa Part 2,
Highveld region. Directorate of Agricultural information, Plbag X144, Pretoria. ISBN 0 621
0802708 .
Grobbelaar, R (2001) Long term impact of intermine flow from collieries in the Mpumalanga
area. Unpublished MSc thesis, Dept of Geohydrology, University of the Free State,
Bloemfontein, South Africa.
Grobbelaar, R, Cruywagen, L-M., de Necker, E. and Hodgson,F.D.1. (2000) Investigation into
Geohydrological and Hydrochemical aspects relating to Mine Closure for Transvaal
Navigational Colliery. Confidential Report to Ingwe SA.
285
Grobbelaar, R., Usher,B., Cruywagen,L-M., de Necker, E. and Hodgson, F.D.1. (2002) Long
term impact of Intermine flow from Collieries in the Mpumalanga area. WRC Report No
1056/1/04.
Gupta, R.D. and Salaran, G.R. (1971) Effect of gypsum on the microbial population in saline
and alkaline soils of Jammu and Kashmir. Indian Journal of Mycology and Plant Pathology,
Vol.1, pp.137-138.
Hadley, R. and Snow, D. (1974) Water Resources and Problems relating to Mining. American
Water Resource Association, MN.
Hall, D.G.M., Reeve, M.J., Thomasson, A.J. and Wright, V.F. (1977) Water retention, porosity
and density of field soils. Tech. Monograph NO.9, Soil survey of England and Whales.
Havenga, A. (2002) Quantification and management of intermine flow in the Western Witbank
Coalfield: Implication for mine water volumes and quality. Unpublished MSc thesis, Dept of
Geohydrology, University of the Free State, Bloemfontein, South Africa.
Hawkins, J.W. (1998) Hydrolic characteristics of surface-mine spoil. In: Coal Mine Drainage
Prediction and Pollution Prevention in Pennsylvania. Ed: K.B.C. Brady, M.W. Smith and
J.Schueck, Vol. 3, pp.1-3.11 Harrisburg, Pennsylvania: The Pennsylvania Department of
Environmental Protection.
Hawkins, J.W. (2004) Predictability of surface mine spoil hydraulic properties in the
Appalachin Plateau. GROUND WA TER. Vol. 42, NO.1. Jan - Feb 2004, pp.119-125.
Hem, J.D. (1985) Study and interpretation of the chemical characteristics of natural water:
U.S. Geological Survey Water-Supply Paper 2254, pp. 213-217.
Hensley, M. (2004-2006) Personal communication. Dept of Soil Sciences, University of the
Free State, Bloemfontein, South Africa.
Hill, D. E. and Parlange, J.-Y. (1972) Wetting front instability in layered soils, Soil Sci. Soc.
Am. Proc., Vol. 36, pp. 697-702.
Hillel, D. (1982) Introduction to soil physics. ACADEMIC PRESS. INC., 111 Fifth Avenue,
New York. ISBN 0-12-348520-7
Hodgson, F.D.1. Wagner, H. and Shipman, B.J. (1995) "Guidelines for environmental
protection- pollution problems and hydrological disturbances resulting from increased
underground extraction of coal. Chamber of Mines of SA Guideline.
Hodgson, FDI. (1999) Intermine flow between Tavistock and Rietspruit Collieries.
Unpublished report.
286
Hodgson, F.D.1. (2001) "Underground storage of water from pit 5 and 6, at Kriel Colliery."
Unpublished report to Kriel Colliery.
Hodgson, F.D.I and Krantz, R.M. (1998) Groundwater quality deterioration in the Olifants
River Catchment above the Loskop Dam with Specialised Investigations in the Witbank Dam
Sub-Catchment. WRC Report No 291/1/98.
Hodgson, F.D.I, Wagner, H. and Shipman, B.J. (1985) GUIdelines for Environmental
protection. Pollution problems and Hydrological disturbances resulting from increased
underground extraction of coal. UOFS Press, Bloemfontein, South Africa.
Hodgson F.D.I., Usher,B.H., Scott, R., Zeelie,S., Cruywagen, L-M. and de Necker, E. (2001)
Prediction techniques and preventative measures relating to the post -operational impact of
the underground mines on the quality and quantity of groundwater resources. WRC Report
No 699/1/01.
Hoffman, G.J. (2004) Waster quality criteria for Irrigation. University of Nebraska-Lincoln.
http://ianrpubs.unl.edu/irrigation/ec
Hough, J.J.H. (2002) Evaluation of Management Options for Intermine Flow and associated
Impacts in the Central Witbank Coalfield. Unpublished MSc thesis, Dept of Geohydrology,
University of the Free State, Bloemfontein, South Africa.
Hutson, J.L. (1983) Estimation of hydrological properties of South African soils. Unpublished
PhD thesis, University of Natal, Pietermaritzburg, South Africa.
Hvorslev, M.J. (1951) Time lag and soil permeability in groundwater observations. US Army
Corps of Engineers, Bulletin 36, Waterways Experiment Station, Vicksburg, MS.
ISO 11048 (1995) Soil quality -Determination of water-soluble and acid-soluble sulphate.
International Standard, First edition, May 2005. Ref. No ISO 11048: 1995(E)
James, A. R. (1996). The Prediction of Pollution Loads from Coarse Sulphide-Containing
Waste Materials. Report to the Water Research commission by Steffen, Robertson and
Kirsten. WRC Report No. 559/1/97. Water Research Commission, Pretoria.
Jaynes, D.B., Rogowski, A.S. and Pionke, H.B. (1984) Acid Mine Drainage from reclaimed
coal strip mines. I. Model description, Water Research Resources, Vol. 20, p. 233.
JCPDS International Centre for Diffraction Data (1980a) Mineral Powder Diffraction File Data
Book. JCPDS International Centre for Diffraction Data, USA.
JCPDS International Centre for Diffraction Data (1980b) Mineral Powder Diffraction File
Search Manual. JCPDS International Centre for Diffraction Data, USA. pp. 58, 260.
287
Johnson, C.D., Hurle, D.H. Hudson, D.R. and Height, M.L. (1983) Water movement through
preferred athways in lateritic profiles of the Darling Plateau, Western Australia. Technical
Paper NO.1, pp.1-34.
Jovanovic, N.Z., Barnard, R.O., Rethman, N.F.G. and Annandale, J.G (1998) Crops can be
irrigated with lime-treated acid mine drainage. Water SA Vol. 24, NO.2 pp. 103-107.
Jovanovic, N.Z., Annandale, J.G., Claassens, A.S., Lorentz, S.A., Tanner, P.D., Aken, M.E.
and Hodgson, F.D.I. (2002) Commercial production of crops irrigated with gypsiferous mine
water. Water SA, Vol. 28, NO.4, October 2002. ISBN 0378-4738
Kelly, E.J., Bell, 0.1. and Johnstone, T.L. (1998) Lysimeter studies to investigate the leaching
of Am-241 from low level radioactive waste. J. Alloys Compd. 271, pp. 227-230
Kertell, C.R., Wagner, J.F. and Novotny, S.J. (1988) Treatment of Coal ASH/mine Refuse
Leachate: A case history .Proceedings of the 4;!1d Industrial Waste Conference. Purdue
University, West Lafayette, Indiana, May 12-14 1987, Lewis Publishers Inc., Chelsea,
Michigan, pp. 919-926.
Kirchner, J., Van Tonder, G.J. and Lukas, E. (1991) Exploitation Potential of Karoo Aquifers.
Report to the WRC, Report 170/1/91, Water Research Commission, Pretoria, p. 283.
Klein, M., Miller, M., Dust, M. and Gorlitz, G. (1997) Validation of the pesticide leaching
model Pelmo using Iysimeter studies performed for registration. Chemosphere 35 11, pp.
2563-2587
Kovda, V.A. (1954) La géochimie des déserts de I'URSS. Communication au 5me Congres
International de la Science du Sol. L'Académie des Sciences, Moscou.
Krantz, R. (1993) An investigation into water quality deterioration in opencast collieries of the
Eastern Transvaal, with special reference to the Olifants River catchment. (Confidential
report)
Kruseman, G.P and de Ridder, N.A (1994) Analysis and Evaluation of Pumping Test Data,
Second Edition (Completely revised), pp237. ISBN 9070754207
. Kuo, E.Y. and Ritchie, A.I.M. (1999) The impact of convection on the overall oxidation rates in
sulfidic waste rock dumps. In: Proceedings of Conference: Sudbury '99-Mining and the
Environment Paper AD2, Vol. 1, pp. 9-18.
Lamontagne S., Dighton J., Ullman W. (2002) Estimation of groundwater velocity in riparian
zones using point dilution tests. CSIRO Land and Water. Technical Report 14/02.
288
Lawes JB, Gilbert JH. Warington R (1882). On the Amount and Compositi on of the Rain and
Drainage Water Collected at Ro thamsted.Williams, Clowes and Sons, London. UK
Latilla+Olivier (1998) Report on stability of pillars for mine closure of Ermelo Mine Service.
Confidential Report to Ingwe S.A.
Lawrence, R.W. and Day, S. (1997) Chemical prediction techniques for ARD. Short course
#2. Fourth International Conference on Acid Rock Drainage. May 31-June6, Vancouver, BC.
Lax, A., Diaz, E. and Castillo, V. and Albaladejo, J. (1994) Reclamation of Physical and
Chemical Properties of a Salinized Soil by Organic Amendment. And Soil RehablYitation, Vol.
8, pp. 9-17.
LeGrand, H.E. and Rosen, L. (2000) Systematic Makings of Early Stage Hydrogeological
Conceptual Models. Ground Water, Vol. 38, NO.6, pp. 887 - 893.
Leij, F.J. and Van Genuchten, M.Th. (1999) Solute Trnsport, Section A, chapter 6. In:
Sumner, M.E. (Ed.) Handbook of Soil Science. CRC Press. Boca Raton, USA.
Le Roux, P.A.H (2004-2006) Dept. of Soil Science, University of the Free State,
Bloemfontein, South Africa. Personal communication and site visit.
Levebvre, R., Hockley, D., Smolensky, J. and Lamontagne, A. (2001) Multiphase transfer
processes in a waste rock piles producing acid mine drainage: 2. Applications of numerical
simulation. J Contaminant. Hydrology, Vol. 52, pp. 165-186.
Lichtner P. C. (1996) Continuum Formulation of Multicomponent-Multiphase Reactive
Transport in Reviews in Mineralogy, Vol. 34, Reactive Transport in Porous Media. Lichtner,
P. C., Steefel, C. I. and Oelkers, E. H. (Editors) Mineralogical Society of America.
Lindsay, W.L. (1979) Chemical equilibria in soils. John Wiley and Sons, New York.
Liu, Y., Steenhuis, T.S. and Parlange, J.-Y. (1994) Formation and persistence of fingered
flow fields in coarse grained soils under different moisture contents, J. Hydrol., Vo1.159, pp.
187-195.
Lorentz, S., Goba, P. and Pretorius, J. (2001) Hydrological process research: Experiments
and measurements of soil hydraulic properties. WRC Report No 744/1/01
Loxton, R.F. (1961) A modified chart for the determination of basic soil textural classes in
terms of the International classification for soil separates. S. Afr. JAgric Sci., Vol. 4, pp 507 -
512.
289
Lu, T. X., Biggar J. W., and Nielsen, D. R. (1994) Water movement in glass bead porous
media, 2, Experiments of infiltration and finger flow, Water Resour. Res., Vol. 30, pp. 3283-
3290.
MacVicar, C.N., de Villiers, J.M., Loxton, R.F., Verster, E., Lambrechts, J.J.N., Merryweather,
F.R., le Roux, J., van Rooyen, T.H. and Harmse, H.J. von M. (1977) Soil classification. A
binomial system for South Africa. Science bulletin, 390, Dept. of Agric. Tech. Services,
Pretoria, South Africa.
Maini, T. and Hocking, G. (1977) An examination of the feasibility of hydrologic isolation of a
high level waste repository in crystalline rocks. Invited paper, Geologic Disposal of High
Radioactive Waste Session, Ann. Meet, Geol Soc. Am, Seattle, Washington.
Martin, P.J. (1994) Modelling of the North Waterloo Multi-Aquifer System. MSc. Thesis,
University of Waterloo. Waterloo, Ontario.
Matar, A.E. and Jabbour, G. (1982) Effect of soil gypsum on leaching of potassium, calcium
and magnesium from soil, growth and mineral content of alfalfa and corn. Techreen
University. Review for Scientific Research, Vol. 5, pp. 107-116. Lattakia, Syria.
McBride, M.B. (1994) Environmental chemistry of Soils. Oxford University Press. New York
and Oxford
McClaren, R.G. (1995) GRID BUILDER 4.0, Users Guide, A generator for 20, triangular,
finite-element meshes and mesh properties. Waterloo Centre for Groundwater Research,
University of Waterloo.
Midgley, D.C., Pitman, WV., and Middleton, B.J. (1994) Surface Water Resources of SA
1990. WRC Report No 298/1.4/94 Second Edition.
Mills, C. (1998b). "The Role of Micro-organisms in Acid Rock Drainage"
http://www.enviromine.com/ard/Midroorganisms/roleof.htm
Molson, J.W., Martin, P.J. and Frind, E.O. (1995) WATFLOW/3D: A Three Dimensional
Groundwater Flow Model Users Guide, Version 1.0. Waterloo Centre for Groundwater
Research, University of Waterloo, p. 31.
Moret, L. (1979) Précis de géologie. Masson, Paris.
Morin, K.A. (1983) Prediction Of Subsurface Contaminant Transport In Acidic Seepage From
Uranium Tailings Impoundments. Ph.D. Thesis, Department of Earth Sciences, University of
Waterloo, Ontario.
290
Morin, K.A. and Hutt, N.M. (1997) Environmental Geochemistry of Minesite Drainage:
Practical Theory and Case studies, Minesite Drainage Assessment Group, MDAG
Publishing, Vancouver, Canada.
Morin, K.A. and Hutt, N.M. (2001) Prediction of minesite-drainage chemistry through closure.
Journal of Geochemical Exploration, Vol. 73, pp. 123-130.
Motram, R., Hutson, J.L. and Goodman, P.S. (1981) Water retension of some Natal soils as
related to soil texture and organic matter content. Crop Production, Vol. 10, pp. 47-50.
Mott, C.J.B. (1981) Anion and Ligand Exchange, chapter 5. In: Greenland, D.J. and Hayes,
M.H.B. (Eds.) The chemistry of Soil Processes. Wiley-Interscience. Chichester.
Naney, J.W., Ahuja, L.R. and Barnes, B.B (1983) Variability and interrelation of soil water and
some related soil properties in a small watershed. In: Advances in Infiltration. Proceedings of
the National Conference on Advances in Infiltration, Chicago, Illinois, pp. 92-101.
Newbrough, P. and Gammons, C. H. (2002). An experimental study of water-rock interaction
and acid rock drainage in the Butte mining district, Montana. Environmental Geology (2002)
Vo141, pp. 705-719, Springer-Verlag.
Nordstrom, D.K., Plummer, L.N., Langmuir, D., Busenberg, E., Jones, B.F. and Parkhurst,
D.L. (1990) Revised Chemical Equilibrium Data for Major Water-Mineral Reactions and their
Limitations. In: Chemical Modelling of Aqueous Systems II, Melchior, D.C. and Bassett, (Eds)
ACS Ser. 416, American Chemical Society, Washington, DC. pp. 398 - 413.
Nordstrom, D. K. and Alpers, C. N. (1999a) Negative pH, efflorescent mineralogy, and
consequences for environmental restoration at the Iron Mountain Superfund site, California
Proc. Nat/. Acad Sci. USA. Vol. 96, pp. 3455-3462.
Ologunde, O.O. and Sorensen, R.C. (1982) Influence of concentrations of K and Mg in
nutrient solutions on sorghum. Agronomy Journal, Vol. 74, pp. 41-46.
Oster, J.D. and Frenkel, H. (1980) The chemistry of the reclamation of sodic soils with
gypsum and lime. Soil Set:Soc.Am.J Vol. 44, pp. 41-45.
Parkhurst, D.L. and Appelo C.A.J. (1999) User's guide to PHREEQC (version 2)-A computer
program for Speciation, Batch-Reaction, One-dimensional Transport, and Inverse
Geochemical Calculations. Water-Resources Investigations, Report 99-4259, U. S.
Geological Survey, U. S. Department of the Interior.
Pavan, M.A., Bingham, F.T. and Pratt, P.F. (1984) Redistribution of exchangeable calcium,
magnesium and aluminium following time or gypsum applications to a Brazilian oxisol.
Journal of the SoH Science Society of America, Vol.48, pp. 33-38.
291
Petersen, G.W., Cunningham, R.L. and Matelski, R.P. (1968) Moisture characteristics of
Pennsynania soils. Soil Sci. Soc. Amer. Proc. Vol. 32, pp. 271-275
Powell, A. R. and Parr, S.W. 1919. A study of the forms in which sulphur occurs in coal. Univ.
ItI. Bul., Vol. 16, No. 34, Eng. Exp. Sta. Bul. 111
Price, W.A. (1997) DRAFT Guidelines and Recommended Methods for the Prediction of
Metal Leaching and Acid Rock Drainage at Minestfes in British Columbia, British Columbia
Ministry of Employment and Investment, Energy and Minerals Division, Smithers, BC, (April),
p. 143.
Prueqer, J.H., Hatfield, J.L., Aase, J.K. and Pikul, J.L. (1997) Bowen-ratio comparisons with
Iysimeter evapotranspiration. Agron. J. 89 5, pp. 730-736.
Pulles, W., Howie, D., Otto, D. and Easton, J. (1995) A Manual on Mine Water Treatment and
Management in South Africa. Water Research Commission Report No. TT 80/96, Pretoria,
South Africa.
Quirk, J.P. (1994) Interparticle forces: A basis for the interpretation of soil physical behaviour.
Advances in Agronomy, Vo1.53, pp. 121-183.
Rahn, P.H. (1976) Potential of coal strip-mine spoils as aquifers in the Powder River Basin.
Billings, Montana: Old West Regional Commission, project No. 10470025
Reddy, M.R., Raju, N.J., Reddy, Y.v. and Reddy, T.V.K. (2000) Water resources
development and management in the Cuddapah district, India. Environmental Geology, Vol.
39, pp 342 - 252.
Reeve, R.C. (1953) A method of determining the stability of soil structure based upon air and
water permeability tests. Soli Sci. Amer. Proc. Vol. 17, pp. 324-329.
Rivers, E.D. and Ship, R.F. (1978) Soil water retention as related to particle sie in selected
sands and loamy soils. Soli Sci. Vol. 126, pp. 94-100.
Riemann K. (2002) Aquifer parameter estimation in fractured-rock aquifers using a
combination of hydraulic and tracer tests. PhD thesis, submitted to the Institute for
Groundwater Studies, University of the Free State, South Africa.
Rietra, R.P.J.J., Hiemstra, T. and Van Riemsdijk, W.H. (1999) Sulphate adsorption on
goethite. Journal of Col/oid and Interface Science, Vol 184.
Richards, L.A. (1931) Capillary cionduction of liquids through porous mediums. Physics
1:318-333
292
Ritchie, A.I.M. (1994) The Waste-rock Environment, in Environmental Geochemistry of
Sulphide Mine-wastes, Mineralogical Association of Canada Short Course Handbook (Eds. J.
L. Jambor and D. W. BIowes,), Vol. 22, pp. 133-161.
Ritsema, C. J. and Dekker, L. W. (1994) How water moves in a water repellent sandy soil, 2,
Dynamics of fingered flow, Water Resour. Res., Vol. 30, pp. 2519-2531.
Rouppet, B. (2003) Gypsum: The Essential Amendment.
http://westernfarmpress.com/mag/farming essential amendment highl
Rust, l.G. (1975) "Tectonic and sedimentary framework of Gondwana Basin in Southern
Africa." Papers from the third Gondwana Symposium Canberra, Australia.
SA Drinking water Standards 241 (2001). Edition. 5. ISBN 0-626-12908-7
Saffigna, P.G., Kenney, D.R. and Tanner, C.B. (1977) Lysimeter and field measurements of
chloride and bromide leaching in an uncultivated loamy sand. Soil Sci. Soc. Am. J. 41, pp.
478-482.
Schoeman, J.L., June 2001: Synopsis of Agricultural Research on Rehabilitated Coal Mined
Land (1994-2001) Conducted in Terms of the Kraai van Niekerk Report, Coaltech 2020
Schoeman, J.L, Matlawa, S.M. and Howard, M.D. (2002) Quantification of the water balance
of selected rehabilitated mine soils under rainfed pastures in Mpumalanga. WRC Report No
798/1/02, ISBN No 1 86845 909 8.
Schwartz, F.W., Crowe, A.S., Hendry, M.J. and Chorley, O.W. (1987) A Case Study to
Assess the Potential for Saline Soil Development Due to Irrigation. Journal of Hydrology, Vol.
91, pp. 1-27.
Segalen, P. and Brion, J.C. (1981) Pédochimie:quatriemepartie, T 1.ORSTOM, Paris.
Seif, J.M. (2002) The Science of AMD and Passive Treatment. http://www.dep.state.pa.us
Sekhara, R., Chandra, S. and Patil, S.v. (1976) Effect of calcium and sulphur and certain
minor nutrient elements on the growth, yield and quality of groundnut. Oleagineux, Vol. 35,
pp.507-510.
Selaoio, E.T. (1998). Tracer studies and groundwater recharge assessment in the Eastern
fringe of the Botswana Kalahari. Published Ph.D. thesis, Free University of Amsterdam.
GRES project publication.
293
Shafer, W.M., Smith, S., Luckay, C. and Smith, T. (1994) Monitoring gaseous and liquid flux
in sulfide waste rock. In: Proceedings of Pittsburg Conference. U.S. Bureau Mines Special
Publications, SP 06A-94, pp. 410-418.
Shannon, M., Cervinka, V. and Daniel, D.A. (1997) Drainage water reuse, In: Madramootoo,
C.A., Johnston, W.R. and Willardson, L.S. (eds.) Management of Agricultural Drainage Water
Quality: Food and Agricultural Organization of the United Nations, Water Report No. 13,
Rome, pp. 29-40.
Sharma, M.L. and Hughes, M.W. (1985) Groundwater recharge estimation using chloride,
deuterium and oxygen - 18 profiles in the deep coastal sands OfWestern Australia. Journal of
Hydrology, Vol. 81, pp.93-109.
Shternina, E.B. (1960) solubility of gypsum in aqueous solutions of salts. InterN. Geol .. Rev.,
1, pp. 605-616.
Singh, B.R. and Taneja, S.N. (1977) Effects of gypsum on mineral nitrogen status in alkaline
soils. Plant and Soil, Vol. 48, pp. 315-321.
Smith, L., Lopez, D.L., Beckie, R., Morin, K., Dawson, R. and Price, W. (1995) Hydrogeology
of waste rock dumps, Report for Dep. of Natural Resources Canada, Ontario, Canada.
Smith, D.A.M. and Whittaker, R.R.L.G. The Springs-Witbank Coalfield. In: Anhaeusser, C.R.
(Ed.) (1986) Geology Mineral Deposits of SA.
Snyman, J.P (1998) Coal. In" The mineral resources of South Africa': Eds: Wilson, M.G.C
and Anhaeusser, C.R. pp. 136-161.
Soil Classification Working Group (1991) Soil Classification. A Taxonomic System for South
Africa. Mem. Nat. Agric. Resources S. Afr. No. 15, Department of Agricultural Development,
Pretoria.
Sommerfeldt, T.G. and Chang, C. (1980) Water and Salt Movement in a Saline-Sodic Soils in
Southern Alberta. Canadian Journal of Soil Sciences, Vol. 60, pp. 53-60.
Sracek, O. (1997) Hydrogeochemical and isotopic investigation of acid drainage from waste
rock at Mine Doyon, Quebec, Canada, PhD thesis. Université Laval, Québec, Canada.
Sracek, 0., Choquette, M., Gélinas, P., Lefebvre R. and Nicholson, RV. (2004) Geochemical
characterization of acid mine drainage from a waste rock pile, Mine Doyon, Québec, Canada
Journal of Contaminant Hydrology, Vol. 69, pp.45-71
294
Steenhuis, T.S., Geohring, L.D., Richards, B.K., Walter, M.T. and Peranginangin, N. (2002)
Preferential flow, extension and educational module. Dept. of Biological and Environmental
Engineering, Cornell University.
Steffen, Robertson and Kirsten (1989) Draft Acid Rock Drainage. Technical Volume 1-
Technical Guide
Steffen, Robertson and Kirsten (1989) Draft Acid Rock Drainage Technical Volume 1-
Technical Guide
Steffen, K. and Savina, M. (1996) The Effects of Secondary (Irrigation Induced) Soil
Salinization: Will Secondary Soil Salinization Make Irrigated Agriculture in Arid and Sub-Arid
Areas Non-Sustainable on a Global Scale?
Stevenson, P. (1997). Heavy Metals: What are they? And why are they dangerous? http:
//www.avon.net.au/globe/features/heavymet.htm
Streetly, M. (1998) Dewatering and environmental monitoring of the extractive industry.
Quarterly Journal of Engineering Geology, Vol. 31, pp125 - 127.
Stockwell, J., Beckie, R and Smith, L. (2001) Hydrogeology of an unsaturated waste rock
pile, Key Lake, Saskatchewan. Proc.IAH Conf., Calgary, pp.1582-1587.
Suarez, D.L. (1989) Impact of agricultural practices on groundwater salinity. Agriculture,
Ecosystems and Environment. Vol. 26, pp. 215-227
Swenson, H.A. and Baldwin, H.L. (1965) A primer on water quality: U.S. Geological Survey,
Washington, D.C., p. 27.
Szabolcs, I. (1986) Agronomical and Ecological Impact of Irrigation on Soil and Water
Salinity. Advances in Soil Sciences, Vol. 4, pp. 189-215.
Tabatabai, M.A. (1982) Sulphur, chapter 28. In: Page, A.L., Miller, RH. and Keeney, D.R
(Eds.). Methods of Soil Analysis, Part 2: Chemical and Microbiological Properties, ?d Edition.
American Society of Agronomy, Inc. and Soil Science Society of America, Inc. Madison,
Wisconsin, USA.
Tanner, P.D. (1993) Listing of Arncoals trials and investigations pertinent to the rehabilitation
of coal strip-mines. Report to the Working Group on research needs for the rehabilitation of
high extraction coal mining areas.
Taylor, G.F. (2000) Pollution control and minesite rehabilitation in surface coal mining.
295
Terman, G.L., Allen, S.E. and Bradford, B.N. (1975) Nutrient dilution-antagonism effects in
corn and snap beans in relation to rate and source of applied potassium. Proceedings of the
Soil Science Society of America, Vol. 39, pp. 680-685.
Thellier, C., Holtclaw, K.M., Rhoades, J.D. and Sposito, G. (1990) Chemical effects of saline
irrigation water on a San Joaquin Valley soil: I Column studies. " Field soil samples. J.
Environ, Qual. Vo1.19. pp. 50-56.
Thornton-Dibb S, and Lorentz, S. (2002) Manual for the operation of Tensiometer nests,
using the HOBO Logger System. School of Bioresources Engineering andd Environmental
Hydrology (SBEEH), University of Natal, Pietermaritzburg
Tom, A.A. and Blowes, O.W. (1999) The hydrogeology of a tailings impoundment formed by
central discharge of thickened tailings: implications for tailings management. Journal of
Contaminant Hydrology, Vol. 38, (1999) pp. 489-505.
Tuovinen, O.H., Niemela, S.1. and Gyllenberg, H.G. (1971) Tolerance of Thiobacillus
ferrooxidans to some metals. Antonie van Leeuwenhoek, Vol. 37, pp. 489-496.
US EPA (1994) EPA and Hardrock mining: A Source Book for Industry in the Northwest and
Alaska
Usher, B.H. (2003) The evaluation and development of hydrogeochemical prediction
techniques for long-term water chemistry in South African coalmines. Unpublished PhD
thesis, Dept of Geohydrology, University of the Free State, Bloemfontein, South Africa.
Usher, B.H., Cruywagen, L-M., de Necker, E. and Hodgson, F.D.1. (2003) On-site and
Laboratory Investigations of Spoil in Opencast Collieries and the Development of Acid-Base
Accounting Procedures. WRC Report No 1055/1/03
Usher, B.A., Pretorius, J.A. and van Tonder, G.J. (2006) Management of a Karoo fractured-
rock aquifer system - Kalkveld Water User Association Water SA Vo1.32, N01, pp. 9-19.
Van Genuchten, M.Th. (1980) A closed-form equation for predicting the hydraulic conductivity
of unsaturated soils. Soil Sci Soc. Am. J. Vol. 44.
Van Schilfgaarde, J. (1994) Irrigation- a Blessing or a Curse. Agricultural Water Management
25, pp. 203-219.
Van den Ende, J. (1991) Supersaturation of soil solutions with respect to gypsum. Plant and
Soil. Vo1.133, pp. 65-74
296
Van der Merwe, AJ. (1993) The effects of underground high extraction coal mining on
agriculture. Report to the Working Group on research needs for the rehabilitation of high
extraction coal-mining areas.
Van der Westhuizen, M. (2005) Institute for Mining and Metallurgy Conference, Carnival City,
Boksburg S.A 12 July 2005
Van Tonder, G.J. (2003- 2006) Personal communication
Van Tonder, G., Bardenhagen, I., Rieman, K., van Bosch, J., Dzanga, P. and Xu, Y. (2002)
Manual on pumping test analysis in fractured-rock aquifers. WRC Report No 1116/1/02.
Van Tonder, G., Riemann, K. and Dennis, I. (2002b) Interpretation of Single-well tracer tests
using fractional-flow dimensions. Part 1: theory and mathematical models. Hydrogeology
Journal, Official Journal of the International Association of Hydrogeologists. Vol. 10, NO.3.
Van Tonder, G.J. and Vermeulen, P.O. (2005) The applicability of slug tests in fractured-rock
formations. Water SA, Vol. 31, No.2, pp157, April 2005. ISSN 0378-4738
Van Tonder G, Vermeulen 0, Cogho Vand Kleynhans J (2003) Prediction of the decant rate
and sulphate concentration from rehabilitated open cast coalmines in South Africa. 6TH
ICARD, Cairns, Queensland, 12-18 Ju1y2003.
Velasco, F., Polo, A, Ladero, M. and Almendros, G. (1980) La humificación en diversos
ecosistemas forestales representativos de la provincia de Toledo. Anales del Jardin Botánico
de Madrid, Vol. 37, pp. 129-141.
Vermeulen, P.O. (2003) Investigation of water decant from underground collieries in
Mpumalanga. Unpublished MSc thesis, Dept of Geohydrology, University of the Free State,
Bloemfontein, South Africa.
Vermeulen, P.O. and Usher, B.H. (2006) Sulphate generation in South African underground
and opencast collieries. Environmental Geology, Vol. 49, N04, Feb. 2006, pp. 552-569.
Vieillefon, J. (1976) Inventaire critique de sols gypseux de Tunisie. ORSTOM, Division des
Sols, Tunisie, ES 98 ORES.
Viljoen, J.N.J (1992) Soil development in rehabilitated coal mine spoil: preliminary
observations. 17th Congress, Soil Sci. Soc. S. Afr., pp. 28-30 January 1992, Stellenbosch
Viljoen, J.N.J (1993) Summary of research and consultant reports: Trans-Natal Coal
Corporation Limited. Report to the Working Group on research needs for the rehabilitation of
high extraction coal mining areas.
297
Vivier, J.J.P., van Tonder, G.J. and Botha, J.F. (1995) The use of slug tests to predict
borehole yields: correlation between the recession time of slug tests and borehole yields. In
Conference Proceedings: Groundwater'9S' Groundwater Recharge and Rural Water Supply,
Midrand, South Africa
Wagner, BNH (1980) Pillar design in coal mines. J.s. Afr. Inst. Min. MetaII., Vol. 80, pp.37-45.
Wainerdi, R.E. and Uken, E.A. (eds) (1971) Modern methods of geochemical analysis.
Plenum Press, New York, p. 397.
Websites:
http://apec-egcfe.fossil.energy
http://www .miningwatch. org/emcbc/pu blications
http://www .acad. ca rieton. edu/curricu lar/G EOL/ciasses/geo258/studentswork/Steffen.
http://www.gcrio.org/geo/grndwt.html- GCRIO
http://wwwrcamnl.wr.usgs.qov/uzf/unsatflow.htmIUSGS Unsaturated zone flow project
http://www.weathersa.co.za/SA Weather Bureau
www.bee.comell.edu/swlab/SoiIWaterWeb/Research Lawes, J.B. Hilbert, J.H. and
Warington, R. (1882)
http://wwwrcamnl.wr.usgs.qov/uzf/unsatflow.htmIUSGS (2001) Unsaturated -Zone Flow:
Definitions and Details
http://www.wikipedia.org/wiki/Darcy's_law
http://www.soilmoisture.com/whatsws.htmISkaling and Wagner (1961) What is a Lysimeter
anyway? Soil Report Newsletter of Soilmoisture Equipment Corp (2002).
Williams, E.G., Rose, A.W., Parizek, R.R and Waters, S.A.(1982) Factors controlling the
generation of acid mine drainage. Pennsylvania State University Report to the U.S. Bureau of
Mines, p. 265.
WISA (2002) Conference on Mine Closure Report. Randfontein Estates Sports Club (23-24
October 2002).
Wood, W.W, Rainwater, K.A. and Thompson, O.B. (1997) Quantifying Macropore Recharge:
Examples from a semi-arid region. Ground Water; Vol. 35 No.6, pp.1 097-11 06.
298
Younger, P.L. (1993a) Possible environmental impact of the closure of two collieries in
County Durham. Journal of the Institution of Water and Environmental Management, Vol. 7,
pp. 521 - 531.
Younger, P.L. (1998b) Coalfield abandonement: geochemical processes and hydrochemical
products. In: Nicholson, K., Energy and the Environment. Geochemistry of Fossit, Nuclear
and Renewable Resources. Society for Environmental Geochemistry and Health. McGregor
Science, Aberdeen, pp. 1-29.
299
Summary
Key words: Virgin soils, spoils, gypsiferous water, irrigation, precipitation, diffusive and
bimodal flow, leaching, monitoring.
Huge volumes of mine water, impacted on by acid mine drainage, are presently produced as
a result of mining activities. These waters are mostly neutralised either as a result of seepage
through neutralising geological strata, or artificially by the addition of lime. Mine drainage is,
therefore, often saturated with gypsum. Gypsiferous mine water can thus be regarded either
as one of the greatest problems associated with mining, or as a potential asset. Large
amounts of waste water could be made available to the farming community, and utilised for
the irrigation of highly productive soils in the coalfields. Concentrating the gypsiferous soil
solution through evapotranspiration, thereby precipitating gypsum in the soil profile, will
reduce environmental pollution, as these salts are removed from the water system. However,
the regulatory authorities have raised some concerns regarding the possible impact of larger-
scale irrigation on water resources. This research attempts to provide meaningful answers.
Five different irrigation sites were selected for research. Monitoring indicates that very little
salinity applied on surface during irrigation reaches the water table. Soil analysis indicates
that most of the salts are captured in the upper one or two meters of the soil profile. The data
indicate that the clay layers, which playa major role in the vertical flux of the water, also have
an influence on the salt distribution through the soil profiles. The porous cups data suggest
that the majority of the salts is currently contained within the uppermost portions of the soil
profile, showing a steady decrease in sulphate concentrations with depth. Geochemical
modelling indicates that soil water in the uppermost meter is saturated with gypsum, and
precipitation occurs.
Virgin solts:
The chloride values at all the different irrigation sites showed a decrease value with depth,
indicating bimodal flow. The saline water moving through the soil profile contributes
approximately 15% or less to the mixing process of the water. Lateral flow in the highly
conductive F zones on the contact of the weathered zone with the solid rock, results in
salinity dilution. Twenty-five percent of the total irrigation water volume reaches the water
level. There is a direct correlation between the clay content of the soil and the diffusive
300
component of the bimodal flow. Due to the moist conditions at irrigation sites, swelling of the
clay results in the very low vertical conductivity value of 10-4 mid.
The average estimated moisture content with depth is:
Depth 1000mm 2000mm 3000mm 4000mm
Kleinkopje 1 0.32 0.373 0.34 0.34
Kleinkopje 4 0.358 0.339 0.34 0.28
Syferfontein 0.384 0.35 0.274 0.262
The salt balance revealed that 80% of the sulphate applied during irrigation is retained in the
soil profile, most of which occurs in the upper two meters. 80% of this retained sulphate
occurs in the soil water. After irrigation ceases, it can take more than 100 years to leach the
sulphate from the soil profile, due to bimodal flow and low recharge.
Conclusion: In the short to medium term, irrigation is feasible on virgin soils, as most of the
salinity is captured, and little is released in the groundwater.
Spoils:
The tests and models indicate the great importance of the reactive nature of the spoils on
which irrigation will occur. Where mine water of relatively consistent quality is used, the
changes in longer-term quality will be far less dramatic.
Based on the slow salinity build-up at the different sites, it is recommended that irrigation be
alternated (i.e. between two pivots over time) if site criteria selection has been adhered to,
and that mine-water irrigation should not be done for periods exceeding 10 years in any
particular area. A non-negotiable prerequisite is that appropriate monitoring must be in place.
301
Opsomming
Groot volumes water, geimpakteer deur suurvorming, word huidiglik geproduseer as gevolg
van mynbou aktiwiteite. Hierdie water word meestal geneutraliseer as gevolg van vloei deur
geologiese strata wat neutralisering aanhelp, of deur die toevoeging van kalk. Water
afkomstig van mynbou is dus dikwels versadig met gips. Hierdie gipsryke water kan beskou
word as een van mynbou se grootste probleme, of as 'n potensiële bate. Groot hoeveelhede
afval water kan tot die beskikking van die landbou gemeenskap gestel word vir gebruik in
besproeiing op die hoë-potensiaal gronde van die steenkoolveld. Deur die gipsryke water te
konsentreer deur evapotranspirasie, kan gips in die grond gepresipiteer word. Dit sal die
omgewingsbesoedeling verminder deurdat saute verwyder word uit die waterstelsel. Die
betrokke owerheid is egter besorg oor die impak wat grootskaalse besproeiing kan hê op die
grondwater reserves. Hiedie navorsingsprojek poog om sulke vrae te beantwoord.
Vyf verskillende besproeiings persele is geselekteer vir die projek. Monitering van die persele
oor tyd toon dat baie min van die saute wat deur die water op die grond toegedien word, die
water tafel bereik. Grond analises toon dat meeste van die saute in die boonste twee meter
van die grond profiel vasgevang is. Die klei lae speel 'n groot rol in die vertikale beweging
van die water. Analises toon ook dat dit 'n invloed het op die sout verspreiding deur die
grond profiel. Poreuse koppie data het hierdie bevinding bevestig, en aangetoon dat daar 'n
afname van sulfaat konsentrasie met diepte is. Geochemiese modelering toon dat die
grondvog in die boonste meter van die profiel versadig is met gips, en dat presipitasie
plaasvind.
Onversteurde grond:
Die chloried konsentrasie by al die verskillende persele wys 'n afname in konsentrasie met
diepte. Dit dui op voorkeur vloei. Die soutryke water wat afbeweeg in die profiel dra ongeveer
15% of minder by tot die vermenging van die water om die laer chloriedwaardes te verkry.
Die hoë geleidingsones op die kontak tussen die verweerde sane en die soliede rots
veroorsaak die laterale vloei van water wat bydra tot die verdunning van die
soutkonsentrasie. Vyf-en-twintig persent van die toegediende besproeiingswater bereik
die water tafel. Daar is 'n direkte verband tussen die klei inhoud van die grond en die
diffusie komponent van die bimodale vloei. As gevolg van die natter toestande op
302
,
besproeiingsperseie, veroorsaak die swelling van die klei dat die vertikale
konduktiwiteitswaarde baie laag is, nl. 10-4 mld.
, Die gemiddelde vog inhoud met diepte is:
Depth 1000mm 2000mm 3000mm 4000mm
Kleinkopje 1 0.32 0.373 0.34 0.34
Kleinkopje 4 0.358 0.339 0.34 0.28
Syferfontein 0.384 0.35 0.274 0.262
Die soutbalans wys dat 80% van die sulfaat wat op die perseel toegedien is, in die grond
vasgevang is. Die meeste hiervan is in die eerste twee meter van die profiel. Nadat
besproeiing gestaak is, sal dit meer as 100 jaar neem om die sulfaat uit die grondprofiel te
loog, omdat bimodale vloei en lae reën aanvulling loging sal vertraag.
Gevolgtrekking: In die kort tot medium termyn, is besproeiing met die gipsryke water
haalbaar. Die meeste van die sout is vasgevang, en baie min loog deur na die water tafel.
Versteurde grond'
Die proewe en modelle wys dat die reaktiewe aard van die versteurde grond waarop
besproeiing plaasvind, baie belangrik is. Die verandering in langtermyn water kwaliteit sal
, baie minder wees as water van redelik konstante kwaliteit gebruik word vir besproeiing.
Gebasser op die stadige opbou van soute in die grondprofiel, word wisseling van besproeiing
tussen twee persele voorgestel as alle reëls vir die kies van geskikte persele nagekom is.
Besproeiing moet nie langer as tien jaar per perseel geskied nie. 'n Ononderhandelbare
voorwaarde vir besproeiing is dat geskikte monitering in plek moet wees.
uv .. UFS
TEIN
303