A macroscale evapotranspiration benchmark based on spatial and temporal enhancements to the Budyko framework

The Budyko framework estimates macroscale evapotranspiration (ET) through hydrological partitioning, which determines what proportion of precipitation becomes ET or passes overland as runoff. This method has been widely used for modelling catchment ET and calibrating satellite-based ET algorithms at multi-annual timesteps, as it effectively accounts for how the relationship between water and energy balance changes with climatic aridity. However, substantial deviations from the predictive Budyko curve are observed as the spatiotemporal resolution increases, suggesting that additional environmental controls beside aridity also significantly influence hydrological partitioning over space and time. In this study, we assessed the impact of 14 environmental indicators on hydrological partitioning from the Budyko framework across 45 catchments over central-western Europe. The combination of catchment slope and two climatic factors − cumulative moisture surplus and rainfall erosivity − effectively explained spatial differences in hydrological partitioning (r = 0.83, rRMSE = 11.66 %). Interannual variations in the enhanced vegetation index and the fraction of precipitation falling as snow captured the temporal scatters (r = 0.37, rRMSE = 14.18 %). Based on these findings, we introduced a two-step modification involving a spatial adjustment of curve parameter (w) and a temporal correction of curve scatter (α) into the Budyko model, which led to improved ET prediction (Δr = 0.30, ΔRMSE = -21.62 mm yr⁻¹) compared to the original framework. The revised Budyko model can be further applied to map ET beyond catchment scale, which facilitates water balance studies at regional scales. The Budyko-predicted ET, independent from existing energy-balance ET (ET_EB) products, can also serve as a macroscale water-balance ET benchmark for satellite-based products in data-scarce regions.