Simulating long-term effects of sowing date on the yield of dryland and irrigated winter wheat

Dryland winter wheat is the most widely grown crop in Oklahoma, where water availability and temperature variations are the major factors affecting the yield of this crop. These effects were assessed using the AquaCrop model. Model calibration and validation were performed using field measurements collected from 11 site-years. The validated model satisfactorily simulated soil water content, maturity biomass, and yield, with overall normalized root mean square errors of 11%, 5%, and 9%, respectively. The model was then used in long-term (1994-2019) simulations of winter wheat at three locations in Oklahoma's wheat belt under five sowing dates (15 Sept., 25 Sept., 5 Oct., 15 Oct., and 25 Oct.) and three soil water contents (SWC) of 40%, 70%, and 100% of the total available water at sowing. The highest simulated average dryland yields (6.4 to 6.6 Mg ha-1) were obtained for early to mid-October sowing dates under the largest SWC at sowing. The lowest average yields (2.67 to 4.15 Mg ha-1) belonged to earlier sowing (mid-September) under the smallest SWC at sowing. Irrigation had a positive impact on the yield of winter wheat and this impact increased for later sowing. The yield of winter wheat sowed in mid-September increased by 0.28 Mg ha-1 under irrigation when averaged for all years and sites, while late-October average yield was 0.96 Mg ha-1 higher than dryland average yield. Irrigation water productivities also increased with later sowing dates.