Assessment of AquaCrop model in simulating maize canopy cover, soil-water, evapotranspiration, yield, and water productivity for different planting dates and densities under irrigated and rainfed conditions

A well-tested crop model can be an important tool in assessment of different crop management scenarios to improve crop yield and water productivity. In this study, the FAO AquaCrop model was evaluated for predicting maize (Zea mays L.) canopy cover (CC), available soil-water (ASW), grain yield, crop evapotranspiration (ET_c), and water use efficiency (WUE) when grown at three different planting dates and densities under rainfed and subsurface drip-irrigated conditions in Nebraska, USA. The model adequately simulated CC in 2011 with root mean squared error (RMSE) and model efficiency (EF) in the range of 5.3 to 12.7% and 0.42 to 0.94, respectively; while in drier 2012, CC was substantially under-predicted with higher RMSE of 24.4%. ASW was consistently underestimated in both years with negative EF value for most treatments in 2011, while in 2012, this trend was more pronounced in irrigated treatments, especially during mid-season. ET_c estimates were marginal with prediction error of up to 15% in 2011 and 18% in 2012, and overall RMSE of 35 mm for both years. Simulated grain yield was acceptable in 2011 with deviations from measured data up to 8% while in 2012, the deviations were much higher up to 40% for rainfed treatments with overestimations in both years. The errors in ET_c and yield simulations are considered high, especially when considering the model application in water-limiting or/and rainfed conditions, and low-yielding areas in which these errors can be a substantial portion of the total yields produced and in seasonal total ET_c from production fields and can cause challenges for growers and decision-makers. The overestimation of yield and underestimation of ET_c resulted in overestimation of WUE, producing inconsistent estimates in both years. Detailed analyses of model performance and potential reasons for the discrepancies and areas that require improvements in the model are presented.