HAPpy Hour Seminar : Filling the Gaps: Predicting Snow Water Equivalent with Machine Learning and ASO Data

Abstract: The Airborne Snow Observatory (ASO) has mapped high-resolution distributed snow depth and snow-water equivalent (SWE) measurements for many basins in the Western United States using airborne lidar. However, these lidar flights are typically limited to 1–5 times per water year per basin, leaving temporal gaps in the SWE record. In this study, we apply a simple machine learning approach that combines ASO’s spatially rich but temporally sparse SWE data with in-situ SWE measurements, which are temporally continuous but spatially limited. The goal is to generate continuous predictions of basin-aggregate SWE across elevation bands. We also examine how the algorithm’s performance depends on the number of ASO flights, the number and quality of in-situ data, and how these methods can inform the placement of future snow measurement stations.


View recordings of previous meetings here

View the schedule (and sign up to give a talk!) here