How different are the Sierra Nevada snowpack estimates from four land surface models with three forcing datasets?

Abstract:

Seasonal snow and its melt are dominant water sources in many regions of the northern hemisphere. Despite the importance of seasonal snow, estimating total snowpack over large areas remain a challenge, particularly in mountainous areas. As land surface modeling (LSM) has evolved rapidly with the advances in high-performance computing, LSM provides a unique opportunity to quantify spatially distributed snow water equivalent (SWE) with high spatiotemporal resolution over the large extent. In this study, twelve Sierra Nevada’s snowpack estimates from four LSMs with three forcing datasets are compared to three “reference” products, the Sierra Nevada Snow Reanalysis (SNSR), Snow Data Assimilation System (SNODAS), and University of Arizona (UA) SWE, over the period 2010 – 2017. Four different LSMs were run using the NASA Land Information System: (1) Noah version 2.7.1, (2) Noah-Multi-Parameterization, version 3.6, (3) Catchment version 2.5, and (4) the Joint UK Land Environment Simulator. Three different forcing datasets were used to drive each of the LSMs: (1) Modern Era Retrospective Analysis for Research and Applications, version 2, (2) Global Data Assimilation System, and (3) the European Centre for Medium-Range Weather Forecasts. The results provide insight into how differences in snow physics in the four LSMs combine with the three meteorological forcing datasets change snowpack and runoff estimates in order to improve water balance modeling and data assimilation.