Since World War II, the rapid expansion of irrigation throughout the Great Plains has threatened the sustainability of the Ogallala Aquifer as evidenced by a dramatic decline in the water table by more than 40 m. Center pivot irrigation practices have accounted for much of the withdrawals from the Ogallala, especially in the western part of the Great Plains domain. Irrigation has been shown to change the surface energy and water budgets by altering the partitioning of latent and sensible heating. An increase in latent heating with irrigation contributes to a cooler and more humid surface, which has competing impacts on convection. In this study, the Weather Research and Forecasting (WRF) model was modified to simulate the effects of irrigation at sub-grid scales. Nine April-October simulations were completed for different hydrologic conditions over the Great Plains. May through September precipitation increased on average over the Great Plains by 4.97 mm (0.91%), with localized increases of up to 20%, especially in Nebraska where irrigation is most heavily applied. The largest increases occurred during wet years (6.14 mm; 0.98%) and the smallest increases occurred during drought years (2.85 mm; 0.63%). Over irrigated areas, precipitation increased by 7.86 mm (1.61%). Large precipitation increases occurred over irrigated areas during normal and pluvial years, while decreases occurred during drought years. These findings suggest the presence of a soil moisture threshold whereby irrigation results in the suppression of convection over irrigated areas when antecedent soil moisture is extremely low and the enhancement of convection when antecedent soil moisture is relatively high.

Data from the WRF simulations were assimilated into a back-trajectory analysis to identify where evapotranspired moisture from irrigated fields predominantly falls out as precipitation (i.e. irrigation-induced precipitation). On average, only 15.8% of evapotranspired moisture from irrigated fields fell out as precipitation over the Great Plains, resulting in 5.11 mm of May-September irrigation-induced precipitation and 6.11 mm of recycled precipitation. Decreases in non-recycled precipitation suggest that irrigation reduces precipitation of advected moisture. The heaviest irrigation-induced precipitation occurred in Nebraska and was coincident with simulated and observed precipitation increases, suggesting that observed precipitation increases are related to evapotranspiration of irrigated water. Water losses with an increase in evapotranspiration were much larger than irrigation-induced precipitation and recycled precipitation increases, confirming that irrigation results in a net export of water over the Great Plains.