Improving the Retrieval of Wintertime Snow Depth and LAI through the Synergy of ICESat-2, GEDI, MODIS, and Surface-Based Measurements

Seasonal snowpack (snow depth and snow water equivalent) is important for weather, climate, and hydrometeorology, but it is difficult to estimate reliably from satellite measurements. Our recent study (Hu et al. 2022) developed an innovative method to directly retrieve snow depth from the ICESat-2 lidar backscattering. Our subsequent study (Lu et al. 2022) demonstrated the applicability to the Arctic, subarctic, and midlatitude land surfaces of the Northern Hemisphere. Our most recent study (Mitchell et al. 2024, submitted) comprehensively evaluated this retrieval over the U.S. and found the good overall performance but the struggle of the retrieval in late season and high density snowpacks. In particular in mountainous regions, we identified several terrestrial factors related to canopy presence and complexity of terrain that negatively impacted the retrieval’s performance.

At the same time, we take this opportunity to evaluate the potential of using snowpack and lidar data to potentially improve the MODIS LAI retrieval for two reasons: (a) The retrieval of MODIS LAI is known to have some issues over the winter months. For instance, over high-latitude evergreen forest defined by the IGBP land cover classification, the MODIS LAI can approach zero. (b) it is unclear if this is caused by the misclassification of Land Cover Type or retrieval deficiency. Using canopy metrics from both GEDI (25 m) and ICESat-2 (11 m) in combination with snow depths from the ICESat-2 lidar (7 m) and in-situ observation-based snow depth from the University of Arizona’s 4-km dataset, we found that the near-zero 500 m MODIS LAI for the Land Cover Type (IGBP) of evergreen forest is inconsistent with the canopy metrics from the collocated ICESat-2 and GEDI lidar measurements. In this talk, we will present the ICESat-2 snow depth retrieval evaluation and the MODIS LAI/land cover evaluation in winter along with our preliminary solution on the correction of wintertime LAI or land cover type