Spatiotemporal modeling for enhancing winter wheat yield and water productivity in dryland farming with supplemental irrigation under variable rainfall conditions

Abstract

Water scarcity induced by climate change poses a significant challenge to sustainable crop production in dryland regions. This study employed the water-driven AquaCrop model to simulate wheat yield in the southwestern Iran divided into dryland and irrigated regions. Using long-term meteorological data, wheat grain yield (GY) and crop water productivity (WPc) were simulated under three water level scenarios including dryland, supplemental irrigation (SI), and fully irrigation. In this study we introduced a newly defined modified rainfall shape index (MRSI), which takes into account both rainfall amount and temporal distribution, especially early- or late-season events- that strongly influence dryland wheat growth. Analysis showed a very significant correlation of MRSI with GY (r = 0.41), particularly in the moderate-rainfall regions (e.g., 300–500 mm). Simulations revealed that applying 50 mm supplemental irrigation before flowering notably increased irrigation water productivity (SIWP) to 0.6–2.4 kg m^-³ . Despite the model's insights, reliance on monthly rainfall data led to over- or under-estimations of GY due to the AquaCrop interpolation method. Furthermore, a regression model showed transpiration-based and evapotranspiration water productivity of 1.86 and 1.64 kg m^-³ , suggesting to apply agronomic practices in dryland regions to reduce soil evaporation and select high-transpiration-efficiency wheat genotypes. In the irrigated regions, deficit irrigation maintained yields while improving WPc. Inverse modeling of light extinction coefficient (k) ranged from 0.43 to 0.68, emphasizing the importance of canopy structure in optimizing water use efficiency. Overall, the study highlights the important role of rainfall pattern and irrigation management in sustaining wheat grain yield in semi-arid regions.