Assimilating Subsampled Airborne Lidar: How Much Lidar is Enough?

Airborne lidar snow depth retrievals are vital for water resource management in basins with limited snow observations. However, airborne lidar remains impractical to collect frequently over large domains due to the high economic cost. In this study, we investigated the extent to which lidar coverage improved modeled snow evolution using a distributed model and assimilation scheme. Full-coverage Airborne Snow Observatory snow depth data in Tuolumne, California and the Olympic Mountains of Northwest Washington State were used as a baseline in which to test the improvement in modeled snow water resources when optimizing flight frequency, timing, and spatial coverage. Collections over multiple seasons in Tuolumne were also used to investigate the impact when assimilating observed snow patterns. Our results indicate that errors in distributed models make snow depth difficult to determine at fine spatial resolutions. However, patterns from lidar in previous seasons are informative enough to train modeled accumulation in following years, therefore reducing the need for repeated, full lidar collections.