Comparison of gridded snow water equivalent products with in situ measurements in British Columbia, Canada

Abstract

Regional-scale estimates of snow water equivalent (SWE) are challenging in alpine regions, particularly in areas of high accumulation and dense forest cover, suggesting efforts to improve these estimates may benefit from an evaluation of existing gridded products. Gridded SWE products comprising four reanalysis datasets (ERA-Interim, ERA-Interim/Land, MERRA and MERRA-Land), two land data assimilation system datasets (GLDAS1 and GLDAS2) and two observation-based products (CMC and GlobSnow) have been compared with in-situ measurements over five physiographic regions in British Columbia (BC), Canada. Time series were generated for each survey month (January through June), and median correlation, bias and mean absolute error (MAE) values were found for each product and physiographic region. The best performance in correlation and magnitude of bias and MAE was seen in areas of lowest SWE accumulation and land relief (e.g. the Interior Plains of northern BC), while poorer performance was seen in regions of high accumulation and complex topography (Columbia and Rocky Mountains and Coast Mountains). Overall, the performance ranking order is ERA-Interim/Land (best), GLDAS2, MERRA, CMC, GLDAS1, MERRA-Land, GlobSnow and ERA-Interim.