Evaluation and Enhancement of Surface Flux Equilibrium Method for Estimating Evaporative Fraction

Abstract

Surface Flux Equilibrium (SFE) has emerged as a promising method for estimating evaporative fraction (EF) in data‐scarce regions by only using observations of temperature and specific humidity. In this study, we perform a comprehensive evaluation of the underlying assumptions of the SFE theory, analyze the effects of assumption violations on biases in SFE‐derived EF, and identify the physical variables which can be used to correct these biases. We conduct the study using ECMWF reanalysis 5 (ERA5) data set over the Indian landmass. Our analysis reveals that the bias in SFE‐based EF relative to ERA5 data set is predominantly controlled by the imbalance between surface moistening through latent heat and surface heating through sensible heat. We find that frequent imbalances between surface moistening and heating result in poor performance in SFE‐based EF estimation relative to ERA5 EF across India in both pre‐monsoon and monsoon seasons. We find that the physical variables which reflect the imbalance in surface fluxes vary considerably across regions and are dependent on the region's ET regime and rainfall characteristics. The surface flux imbalances are captured most effectively by relative humidity and temperature gradient between ground and air in regions with water‐limited ET regime and by net radiation in regions with energy‐limited ET regime. The EF predictions from Random Forest models which use specific humidity, air temperature and the identified region‐specific predictors significantly outperform SFE‐derived EF across all regions and in both seasons, which demonstrates the effectiveness of empirical models with region‐specific predictors for EF estimation.