Identifying Snowfall Elevation Gradients by Assimilating Satellite-Based Snow Depth Observations

Abstract

Precipitation in mountain regions is highly variable and poorly measured, posing important challenges to water resource management. Traditional methods to estimate precipitation include in-situ gauges, doppler weather radars, satellite radars and radiometers, numerical modeling and reanalysis products. Each of these methods is unable to capture complex orographic precipitation. Here, we propose a novel approach to characterize orographic snowfall over mountain regions. We use a particle batch smoother to leverage satellite information from Sentinel-1 derived snow depth observations and to correct various gridded precipitation products. This novel approach is tested using a simple snow model for an alpine basin located in Trentino Alto Adige, Italy. We quantify the precipitation biases across the basin and found that the assimilation method (i) corrects for snowfall biases and uncertainties, (ii) leads to cumulative snowfall elevation gradients that are consistent across precipitation products, and (iii) results in overall improved basin-wide snow variables (snow depth and snow cover area) and basin streamflow.

Keywords: mountain hydrology, snowfall correction, data assimilation