Simulating yield and water use of a sorghum-cowpea intercrop using APSIM

Growth, yield and crop water use (ET) of a sorghum-cowpea intercrop system were evaluated using APSIM and data from field experiments conducted at Ukulinga Research Farm, South Africa over two seasons (2013/14 and 2014/15). Weather and soil data were observed in situ and input into APSIM. Data from optimum experiments (2013/14) was used for local adaptation of APSIM. Thereafter, the model was tested using data obtained from 2014/15 under various water management strategies [rainfed (RF), deficit (DI) and full irrigation (FI)]. Model simulations were evaluated using observed data for phenology, leaf number, leaf area index (LAI), biomass, yield, ET and water use efficiency (WUE). Model performance was assessed using R2, root mean squared error (RMSE) and its components (RMSEs and RMSEu) and the d-index. The model simulated phenology satisfactorily for sorghum (R2 = 0.98, RMSE = 6.62 days) and cowpea (R2 = 0.86, RMSE = 13.67 days) across different water regimes. The model underestimated LAI (36.98%); this was associated with defoliation of crop canopy due to hail damage. Satisfactory simulations (RMSE RF = 145.38 kg ha−1 < DI = 37.38 kg ha−1 < FI = 35.99 kg ha−1) of sorghum yield (RF = 953.63 kg ha−1 < DI = 1079.32 kg ha−1 < FI = 1082.52 kg ha−1). The model also gave satisfactory (RMSE RF = 3.06 kg mm−1 ha−1 > DI = 1.97 kg mm−1 ha−1 > FI = 1.66 kg mm−1 ha−1) simulations of system WUE across the water regimes [RF = 17.54 kg mm−1 ha−1 >DI = 16.43 kg mm−1 ha−1 > FI = 16.75 kg mm−1 ha−1]. The APSIM model was able to simulate growth, yield and WU of an intercrop system under varying water regimes, However, it is still limited with regards to rainfed conditions since it overestimated biomass (6.25%), yield (14.93%) and WU (7.29%) and under estimated WUEb (−14.86%). APSIM can still be used to determine best management practices for intercropping under water scarce environments.