Gridded drought response assessment of winter wheat in Oklahoma using big data and AquaCrop-OS

Winter wheat is the most dominant crop in Oklahoma and critically important to the economy of agricultural industry in this state and the region. However, weather anomalies such as droughts, which are frequent in Oklahoma, pose serious threats to winter wheat yield. This study was conducted to assess the effects of droughts on the simulated yield of dryland winter wheat (Y_sd) in Oklahoma employing a gridded approach with the AquaCrop Open Source (AquaCrop-OS). Long-term (30-year) simulations showed that the average Y_sd ranged from 0.9 to 7.1 Mg ha^-1 across the 3281 grids of the Oklahoma winter wheat belt. At the climate division level, Y_sd ranged from 2.9 Mg ha^-1 in Panhandle to 5.5 Mg ha^-1 in Central divisions. Fluctuation in Y_sd generally agreed with the actual dryland yield measured at several research fields across the study area. However, the model tended to underestimate Y_sd during certain drought events at some sites. The coefficient of variation (CV) for Y_sd ranged from 0.1 to 1.6. Panhandle climate division had the largest CV, mainly due to inconsistency in growing season precipitation and temperature related stresses. In addition, grids with smaller total available water in the root zone, such as those with coarser textures near streams, had lower yields and larger CV. The strongest correlations between Y_sd and drought indices (SPEI, Z-index, PDSI) existed in the months of March to May, when the crop is highly sensitive to water stress. Largest drought sensitivity of winter wheat, based on Y_sd-drought index regression slopes, was estimated in West Central division for short to long-term droughts.