Satellite-based mapping of evapotranspiration, water stress, and crop productivity in mixed irrigated–rainfed agricultural watersheds of central minnesota

Accurate characterization of actual crop evapotranspiration under non-standard conditions (ETₐ), water stress, and crop productivity is essential for agricultural water management in irrigated–rainfed systems. This study applied a Google Earth Engine–based Surface Energy Balance Algorithm for Land (SEBAL) framework to estimate ETₐ and a water stress index (WSI) across two central Minnesota watersheds during 2021–2024. A moisture-sensitive vegetation index (MVI) approach generated 30 m spatially disaggregated corn and soybean yield estimates, which were then combined with ETₐ to calculate crop water productivity ( WP c ). SEBAL-derived ETₐ showed strong consistency with the methodologically related OpenET geeSEBAL product (R = 0.87–0.97), and moderate agreement with independent field water-balance ETₐ (R = 0.74–0.75; NSE = 0.53–0.54). Clearwater–Elk exhibited a stronger irrigation signal, whereas Sauk showed higher growing-season water stress, likely reflecting differences in irrigation, soil water storage, crop development, and rainfall timing. Yield disaggregation showed strong agreement with observed field-scale patterns for corn (R = 0.91; NSE = 0.81), whereas soybean showed high correlation but more modest predictive efficiency (R = 0.80; NSE = 0.42). Higher yield was generally associated with higher ETₐ and lower WSI, while irrigated fields showed higher WP c , particularly for corn. However, higher WP c in Sauk mainly reflected lower seasonal ETₐ rather than higher yields. Yield–ETₐ thresholds indicated limited yield increases beyond certain ETₐ levels, with irrigated fields approaching the upper range of observed yields at lower seasonal ETₐ than rainfed fields. Overall, this integrated framework supports irrigation performance assessment and watershed-scale agricultural water planning.