A new opportunity to measure snow depth from space: evaluation of retrievals from ICESat-2 using airborne laser-scanning data

The unprecedented precision of the altimetry satellite ICESat-2 and the increasing availability of high-resolution elevation datasets open new opportunities to measure snow depth in the mountains, a critical variable for ecosystems and water resources monitoring. We retrieved snow depth over the upper Tuolumne basin (California, USA) for three years by differencing ICESat-2 ATL06 snow-on elevations and various snow-off elevation sources, including ATL06 and external digital elevation models. The snow presence of each ATL06 segment (i.e. point measurements regularly spaced every 20 m) can be determined from the number of photons returned by the ground surface. Snow depth derived from ATL06 data only (snow-on and snow-off) provided a poor temporal and spatial coverage, limiting its utility. However, using airborne lidar or satellite photogrammetry elevation models as snow-off elevation source yielded an accuracy of ~0.2 m (bias), a precision of ~0.5 m for low slopes and ~1.2 m for steeper areas, compared to eight reference airborne lidar snow depth maps. The snow depth derived from ICESat-2 ATL06 will help address the challenge of measuring the snow depth in unmonitored mountainous areas.