Assessing and correcting bias in gridded reference evapotranspiration over agricultural lands across the contiguous United States

Gridded reference evapotranspiration (ETo) data are widely used for agricultural water management and remote sensing evapotranspiration (RSET) models, but biases can arise where coarse meteorological inputs fail to capture agricultural microclimates. We investigated biases in the gridMET ETo product across irrigated agricultural areas of the contiguous United States using ETo calculated from 793 agricultural weather stations, then used those stations to develop monthly correction surfaces. Results show that gridMET systematically overestimates ETo by 10–20% at most cropland sites, while pockets of underestimation appear in some arid western regions, primarily due to wind speed bias. Wind speed bias was the dominant driver, with secondary effects from solar radiation, vapor pressure, and maximum air temperature that varied regionally. Compared with independent micrometeorological data from 79 eddy covariance sites, correction reduced monthly ETₒ mean absolute error at 24 of 30 cropland sites. Corrected ETₒ also improved the three ETₒ-dependent OpenET RSET models (eeMETRIC, SIMS, SSEBop), reducing monthly model mean absolute error at 47–67% of cropland sites and reducing mean bias error by up to 9.5 mm/month. Across natural land cover types including forests, wetlands, grasslands, and shrublands, MAE and RMSE also improved for all RSET models. These results show that gridded ETo bias should be addressed in agricultural water management, including irrigation-demand estimation and RSET workflows, especially where spatially complete ETₒ inputs are required.