Response of soil carbon fractions to land use systems under arid to semi-arid climates in South Africa
Unprecedented pressure on food-producing ecosystems as a result of increasing population has resulted in substantial losses of soil carbon (C). Carbon loss as a basic and major precursor of soil degradation, is more prevalent in the semi-arid to arid environments, which cover more than half of South Africa’s total land surface. Low inputs due to disposal of crop residues by burning in cultivated soils or by overgrazing in rangelands together with climatic conditions interact to influence the quantity and quality of soil C and food production. Management systems with the underlying goal to maintain high levels of soil C and reverse soil degradation in the dryland ecosystems are available and have been tested under different agro-ecological settings locally and abroad. However, there is still limited information on the relationship between soil C quality and quantity, especially in the semi-arid and arid regions. This study was therefore carried out to characterize soil C fractions and evaluate their response to different land use systems under semi-arid to arid climates in South Africa. Soil samples used in this study were selected from previous collections under three different studies by other researchers. Study one: Applied treatments included two methods of straw management (unburned and burned), three methods of tillage (no-tillage, mouldboard ploughing and stubble mulch) and three methods of weed control (chemical and mechanical) in a long-term wheat trial in the Eastern Free State near Bethlehem. Selected samples were collected in different treatment combinations at the 0-50 mm soil layer. Study two: Selected soil samples were collected in the primary grasslands, croplands and secondary perennial pastures at the 0-200 mm soil layer in three agro-ecosystems: Harrismith, Tweespruit and Kroonstad. Study three: Three farms with two rangeland conditions (poor and good) were selected in two ecosystems: Thaba Nchu (grassland) and Kuruman (savanna), and samples were taken at the 0-50 mm soil layer in poor and good rangeland conditions. All soil samples were analyzed for soil organic (SOC) and inorganic (SIC) carbon, permanganate oxidizable carbon (POXC), cold (CWEC) and hot (HWEC) water extractable carbon, extractable humic substances (CEX), humic acids (CHA) and fulvic acids (CFA). Cross polarization magic angle spinning (CPMAS) 13C nuclear magnetic resonance spectroscopy (NMR) was used for structural characterization of SOC. Humification (HI) and polymerization (PI) indices as well as alkyl C/O-alkyl C ratio were calculated as indicators of the extent of SOC decomposition. Results demonstrated that adoption of proper crop residue management (unburned straw in this case) in no-tillage treatments can reverse the current trend of soil degradation in arable landscapes as revealed by an increase in different soil C fractions and SOC quality, compared to treatment combinations that involved mouldboard ploughing. Although indicators of the extent of decomposition were generally similar under the applied treatment combinations, accumulated SOC in no-tilled treatments was less humified, and thus suggested high lability and susceptibility to losses if the soil can be brought under intensive cultivation again. Our results further indicated that conversion of primary grasslands into arable cropping had a negative impact on the quantity and quality of soil C. However, the magnitude of loss for most C fractions generally followed the order: Harrismith > Tweespruit > Kroonstad, which was unexpected because the increase in mean annual rainfall (MAR) and clay content occurred in the opposite direction. O-alkyl C, a measure of lability of SOC, remained almost the same in Harrismith despite prolonged soil cultivation. This could be explained by higher mean annual rainfall and clay content. Reversion of cultivated soils to secondary perennial pastures demonstrated capability to restore historic C fractions and improve SOC quality to comparable levels to primary grasslands. However, there were no clear trends regarding the magnitude of gain across the three agro-ecosystems. Besides soil management and climatic or clay content, losses or accumulations of the measured C fractions were modulated by saturation deficits and vegetation quality. Rangeland conditions in the grassland and savanna ecosystems did not influence soil C fractions, due to rotational grazing, which allowed regeneration of overgrazed areas. However, vegetation regeneration in the savanna was delayed due to hot-dry conditions, hence lower soil cover. Significant changes in the measured C fractions arose when the two ecosystems were compared, and our results indicated that the hot-dry sandy savanna ecosystem was more vulnerable to degradation as revealed by 2-4 times lower C fractions, compared to the cool-moist clayey grassland ecosystem. The CPMAS 13C NMR spectroscopic results highlighted that SOC composition was affected by decomposition in the grassland ecosystem. The four SOC functional groups did not display clear trends in the savanna ecosystem, probably due to heterogeneous (grass-shrub-trees) vegetation composition. Based on the results of this study it is evident that maintenance of a good permanent soil cover can restore lost C fractions and counteract soil degradation processes in the drought prone ecosystems. This means that for arable cropping, adoption of no-tillage with proper residue management could be an option. No-tillage or secondary pasture management can also be used to reclaim degraded cultivated soils. In the rangelands, permanent soil cover can be maintained or improved by rotational grazing depending with not only availability of rangeland resources, but also with prevailing climatic and soil conditions. Where possible reseeding of grass species can be implemented to avoid desertification and erosion losses of C fractions.