Quantifying Uncertainties of Seasonal and Regional Evapotranspiration Estimates from Operational Simplified Surface Energy Balance Model in the Contiguous USA

An operational simplified surface energy balance (SSEBop) model can provide site to regional and daily to seasonal evapotranspiration (ET) estimation. Like other ET models, SSEBop model also faces a challenging issue—uncertainty from inevitable input errors, poorly defined parameters, and inadequate model structures. Recent uncertainty assessment at multiple Ameriflux tower sites showed the uncertainties of seasonal ET estimates by the SSEBop model were within a reasonable range (less than 20%). In this study, we further quantified the uncertainties of seasonal SSEBop ET estimates at 1-km pixel resolution for the contiguous U.S.A. (CONUS) from 2001 to 2014. The uncertainty analysis focused on quantifying uncertainties from errors of key parameters (i.e., difference between cold and hot temperature boundaries (dT)) and primary driving forces, including the 8-day composite 1-km Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature (LST) data, Global Land Data Assimilation System daily air temperature data, and USGS/EROS daily short grass reference ET data. The sensitivity analysis indicated that relative errors of ET estimates from errors of input data and parameters were less than 20% in most areas of CONUS, except for western mountain areas with complex topography, southwestern bare desert areas with high albedo and dry climate conditions, and southeastern areas with high emissivity and humid climate conditions. The relative errors of ET estimates were around 30% in those specific areas where the large ET errors stemmed from errors of input data and parameters, such as LST data with errors over 1.5K, and dT with errors over 2.5K. The uncertainty assessment showed that SSEBop model is a reliable method for wide-area ET calculation. Reduction of errors from input variables (e.g., LST from MODIS) and key parameters (e.g., dT) in topographically complex areas and high albedo and emissivity surfaces can significantly improve the ET calculation of the SSEBop model.