PNNL

Abstract

Agricultural irrigation is steadily expanding in the traditionally non-irrigated US Corn Belt, a globally important region for maize and soybean production. While irrigation boosts crop productivity and stabilizes yields amid weather variability, its significant consumptive water use has implications for regional hydrology. Understanding the spatial dynamics of irrigation spread and its effects on crop yields is crucial for balancing agricultural productivity, farm economics, and water resources under global change. However, the region lacks official data on irrigated outcomes. Fortunately, satellite data can now track both annual irrigation activity and crop health at the field level. Here, we first analyze annual maps of sub-field level (30 m) irrigation across the US Corn Belt from 1997 to 2017 to identify patterns in irrigation expansion. We then leverage annual 30 m crop yield maps to quantify yield impacts using a machine learning causal inference approach. We find proximity to existing irrigation strongly predicts new irrigation locations, and that irrigation has been preferentially installed on lower productivity, fast-draining soils. On average, irrigation has increased maize and soybean yields by 12.6% and 7.9%, respectively, with greater benefits on lower quality soils and in dry weather conditions. Irrigation benefits have increased over our study period, likely due to shifts in cultivars and management. Combined with anticipated increases in yield-limiting weather, these trends suggest that incentives to irrigate are likely to grow. We conclude that continued irrigation expansion is probable in this region and that further study is needed to inform practices and policies promoting sustainable growth.