Evaluation of the CSM-CROPSIM-CERES-Wheat Model as a Tool for Crop Water Management

Abstract

Development and implementation of improved methodologies for crop water management will conserve valuable water resources in agricultural regions that depend on irrigation. To address this problem for conditions in central Arizona, we have evaluated the CSM-CROPSIM-CERES-Wheat model using measured wheat growth and soil water data from plot-level irrigation scheduling experiments conducted during the winters of 2003-2004 and 2004-2005. During each season, wheat plots were managed using two FAO-56-based irrigation scheduling approaches at three planting densities (~75, ~150, and ~300 plant m-2) and at two nitrogen application rates (~80 and ~215 kg ha-1 year-1). For these treatments, the calibrated model simulated wheat yield with relative root mean squared errors (RRMSE) of 7.4% and 1.7% for the 2003-2004 and 2004-2005 seasons, respectively. Time series plots of measured and simulated Zadoks number, leaf number, leaf mass, stem mass, spike mass, and green leaf area index demonstrated favorable wheat development and growth responses to experimental treatments and seasonal weather and management variability. The model was able to quantify average soil water contents in eight soil layers to a depth of 210 cm with RRMSEs ranging from 3.3% to 18.9% for the 2003-2004 season and from 2.7% to 11.3% for the 2004-2005 season. Evapotranspiration was simulated with RRMSEs of 2.4% and 3.2% for the 2003-2004 and 2004-2005 seasons, respectively. Deficiencies were demonstrated in the ability of the model's automatic irrigation routines to reproduce the FAO-56 irrigation schedules devised during field experimentation. With further development, the CSM-CROPSIM-CERES-Wheat model could become a valuable central component for decision tools designed to evaluate alternative water management scenarios and improve water management for irrigated agricultural systems.