Attribution of Snowpack Errors to Simulated Temperature and Precipitation in E3SMv1 Over the Contiguous United States

Abstract

Snow water equivalent (SWE), temperature, and precipitation biases and trends are evaluated in the atmosphere-land simulations of the Energy Exascale Earth System Model version 1 (E3SMv1) in comparison to the Community Earth System Model version 2 (CESM2) and two other models using the ground measurement-based University of Arizona (UA) snow product. SWE, temperature, and precipitation biases are highest in magnitude in the Western contiguous United States (CONUS). SWE errors are attributed to temperature and precipitation through multiple linear regressions of normalized errors, the coefficients of which represent the sensitivities to temperature and precipitation errors. SWE errors are more sensitive to temperature errors throughout the CONUS. Model SWE and temperature trends are generally opposite from UA product trends in the Western CONUS. SWE trend errors are also attributed to temperature and precipitation trend errors using multiple linear regressions of normalized trend errors. SWE trend errors are more sensitive to those of precipitation at higher elevations (>1,500 m) in the Western CONUS in these simulations. Thus, the sensitivity to temperature and precipitation differ for SWE errors and its trend errors. Furthermore, the SWE trend errors are more sensitive to temperature and precipitation in the atmosphere-ocean coupled simulations in which the atmosphere-land is coupled to an active ocean model. These results suggest that both errors in simulated temperature and precipitation contribute to SWE errors. © 2021 The Authors. Journal of Advances in Modeling Earth Systems published by Wiley Periodicals LLC on behalf of American Geophysical Union.