|dc.description.abstract||South Africa is classified as a semi-arid country characterized by low and erratic rainfall. This poses major limitations to crop productivity, especially for smallholder farmers who rely on rainfed agriculture. This is worsened by lack of knowledge regarding best management practices that can improve crop yields attained by smallholder farmers. In addition, smallholder farmers lack access to markets and do not participate in the agricultural value chain. The Biofuel Regulatory Framework (DoE, 2014) seeks to include smallholder farmers in the biofuel feedstock value chain. However, a prerequisite to their meaningful participation in the value chain would be to increase their current levels of crop and water productivity. The main aim of this study was to estimate the yield and water use of soybean (Glycine max L.) under rainfed and smallholder farming conditions using the AquaCrop model. Secondary to this, the effect of mulch and fertilizer on soybean water use efficiency was assessed. Lastly, the Soil Water Balance model (SWB) was used to compare simulations made by AquaCrop for the non-mulched, full fertilizer treatment. Thereafter, the water use efficiency of soybean was calculated from crop water use and the final yield. The soybean trial was carried out at Swayimane, KwaZulu-Natal. The model simulations of crop water use and reference crop evapotranspiration were also used to calculate crop coefficients under non-standard conditions. Crop growth and yield parameters were measured to calibrate and evaluate model performance. Soil water content was monitored using Watermark sensors, along with climatic variables. An analysis of variance (ANOVA) was used to detect significant interactions between treatments, while statistical indicators were used to evaluate model performance of AquaCrop and the SWB model. Mulching improved soil water content and reduced soil water evaporation, although the final yield and total water use efficiency was reduced. It is believed the yield reduction in mulched plots was mostly affected by nitrogen immobilization as a result of decaying straw mulch. Increasing soil fertility improved crop yield and water use efficiency in both mulched and non-mulched treatments. The AquaCrop model simulated the final yield and biomass fairly well, except in mulched treatments. The model simulated the highest yield in the mulched, fully fertilized plots, which is contrary to what was observed. This is because the model only accounts for improved soil water content and does not account for the complex interactions between the soil and mulch residue that resulted in nitrogen deficiency. The SWB model simulated fairly similar crop water use and yield to AquaCrop. The water use efficiencies obtained in this study were compared to that derived by Mengistu et al. (2014) for the same cultivar grown in a commercial farming environment at Baynesfield, KwaZulu-Natal.
In comparison to commercial farmers, smallholder farmers tend to produce lower water use efficiencies. The modelled water use efficiency reported for Baynesfield was 1.277 kg m-3, compared to 0.359 kg m-3 obtained in this study for the non-mulched, full fertilizer treatment. According to AquaCrop, the mulched, full fertilizer treatment had a water use efficiency of 0.485 kg m-3. It is believed that the latter water use efficiency could have been achieved had enough nitrogen been available to the crop. In conclusion, implementing best management practices can help narrow the yield gap between smallholder and commercial farmers. It was evident from this study and others that agronomic practices have a significant impact on crop yield and ultimately, water use efficiency.||en_US