High-resolution mapping of snow cover in montane meadows and forests using Planet imagery and machine learning

Mountain snowpack provides critical water resources for forest and meadow ecosystems that are experiencing rapid change due to global warming. An accurate characterization of heterogeneity in these requires snow cover observations at high spatial resolutions, yet most existing datasets have a coarse resolution. To advance our observation capabilities meadows forests, we developed machine learning model generate snow-covered area (SCA) maps from PlanetScope imagery about 3-m The achieves median F1 score 0.75 103 cloud-free images across four different sites the Western United States Switzerland. It is more (F1 = 0.82) when areas excluded evaluation. We further tested performance 7,741 mountain two study Sierra Nevada, California. achieved 0.83, with higher accuracy larger simpler geometry than smaller complexly shaped meadows. While mapping SCA regions close or under canopy still challenging, can accurately identify relatively large gaps (i.e., 15m < DCE 27m), 0.87 sites, shows promising very (>10m) edges. Our highlights potential high-resolution satellite forested meadows, implications advancing ecohydrological research world expecting significant changes snow.