干旱区生态环境知识资源中心

The aim of this study was twofold: in the first part, we adjusted and tested AquaCrop 4.0 version parameters for durum wheat under semi-arid conditions. Grain yield, biomass and the evolution of soil water content (0-90 cm layer) in an irrigated perimeter were simulated. The experiment was conducted in the Tadla region, Morocco, between 2009 and 2012, using 15 fields. The comparison between observed and simulated grain yield and above-ground biomass using a leave-one-out cross-validation (LOOCV) approach gave a normalized root mean square error of 4.1% (0.2 t ha(-1)) and 5.7% (0.8 t ha(-1)), respectively. Similarly, the difference between observed and modelled soil water content has, on average, an RMSE of 8.2%. In the second part, the analysis of irrigation scenarios showed the potential of crop modelling to schedule irrigation water according to a threshold for water deficit. It also displayed logical trends in the relationship between grain yield and the amount, frequency and timing of irrigation water. Scheduling irrigations during the cropping season significantly improved the grain yield and increased water productivity. We concluded that AquaCrop could be a suitable tool for forecasting yield under semi-arid conditions and to improve crop and irrigation management. Copyright (C) 2016 John Wiley & Sons, Ltd.