Data for: Leveraging probabilistic machine learning to enable large scale assimilation of ICESat-2 snow depth observations

Data of the paper: the input data that are not available elsewhere (SRTM DEM of the downscaling, the model and ICESat-2 cells, the 50 m predictor grid of the Gaussian-process model), the outputs of the four data-assimilation experiments (MuSA, ICESat-2 snow depth, East River basin, Colorado, water years 2023 and 2024; Gaussian-process and Gaspari-Cohn prior correlations: prior and posterior snow depth ensembles of 100 members at the 5400 inducing cells and 400 ICESat-2 representative cells, and mapped to the 50 m grid on the ASO survey dates) and the data files of the interactive website. Each folder of the record is one zip file; see README.md for the structure and the file formats, and the code repository for the processing chain. ASO snow depth is not included.