Seasonal snowpack is an important runoff source in many mountain watersheds, making accurate snowpack water storage data essential to streamflow prediction. Traditional ground-based observations of snow water equivalent (SWE), however, are insufficient in number and distribution to characterize the water supply from snowpack at larger spatial scales, from watersheds to global regions. Satellite, airborne, and drone-based remote sensing provide additional observations of snowpack variables at different spatial, temporal resolutions and accuracies to supplement ground-based observations. Airborne lidar provides spatially extensive and detailed but infrequent snapshots of snow depths across a basin. There is high utility in developing techniques that can integrate these infrequent snapshots with in situ SWE time series and snowpack models to improve SWE estimation.
Here, we describe a data assimilation approach to combine spatially-extensive snow observations from lidar with snowpack models to better quantify SWE in mountain watersheds. Using lidar flights over different baby直播app basins, we compare snow depth from lidar with outputs from model ensembles (SNOW 17) to define weights based on the accuracy of each run relative to the Lidar snow depths for each model parameter set. We then extend through time yielding a lidar-informed SWE and depth data cube. Finally, the lidar informed weights and evaluate strategies for combining these disparate weights to generate long-term gridded daily datasets of SWE and snow depth.