Using LiDAR Remote Sensing to Evaluate and Improve the Retrieval of Snow Depth, Leaf Area Index, and Land Cover Types for Hydrometeorological Studies

Abstract

Recent NASA satellite missions such as the Ice Cloud and Land Elevation Satellite version 2 (ICESat-2) and Global Ecosystem Dynamics Investigation (GEDI) have been deployed to survey the Earth’s surface with mission goals of monitoring changes in glacier ice, sea ice, and vegetation for ICESat-2 and retrieval of 3-D structure of mid-latitude and tropical canopies globally for GEDI. Both instruments have provided the community with unprecedented high-resolution active remote sensing measurements of variables relating to processes in the water cycle. These advancements in spaceborne lidar technology motivate the works performed in this dissertation that demonstrate the importance of using these instruments for the retrieval of hydrometeorological variables and provide motivation for future spaceborne lidar missions. Mitchell et al. (2025a) evaluated snow depths retrieved from ICESat-2 multiple lidar scattering measurements, a new and novel technique developed by Y. Hu et al. (2022) and Lu et al. (2022). Snow depths from ICESat-2 are compared to the in-situ measurement – derived University of Arizona (UA) product for two distinct regions of the contiguous U.S. (CONUS): the Mountain West (complex terrain) and the Great Lakes (homogeneous terrain). Biases between the snow products are co-located with several terrestrial datasets (i.e., Moderate Resolution Imaging Spectroradiometer (MODIS), GEDI, ICESat-2, and USGS LANDFIRE) and then evaluated in terms of the time of snow season (December – April) and snow density to understand the performance of the retrieval. The retrieval performance performed well overall, but results showed the performance decreased with increasingly complex terrain and in the presence of tall canopies. Additionally, the retrieval’s performance decreased later into the snow season and with higher snow densities. The findings provided insights into future corrections that can be made to the retrieval in future studies. Mitchell et al. (2025b) co-located GEDI spaceborne lidar canopy measurements and snow depths from UA with MODIS LAI and land cover (LC) products over the CONUS for a two-year period (2019-2021) to address the questions of if the underestimation of the MODIS LAI data for evergreen forest are due to deficiencies related to the misclassification of the input LC data or the LAI retrieval itself? Comparisons between GEDI plant area index (PAI) and MODIS LAI highlighted the MODIS retrieval deficiencies in evergreen forests, where the median GEDI PAI and MODIS LAI winter/summer ratios are 0.87 and 0.29 respectively. The sensitivity of LAI to snow cover is highest in evergreen forests where LC analyses also demonstrate the highest potential for misclassified pixels according to the International Geosphere Biosphere-Programme LC classification using GEDI canopy metrics. Corrections to wintertime LAI using the winter/summer PAI ratios are applied to tall forest LC types and showed the greatest improvements over evergreen needleleaf forest. Finally, a decision tree approach leveraging several GEDI canopy metrics showed potential to reclassify the MODIS-misclassified LC pixels and demonstrate the advantage of leveraging active spaceborne lidar measurements to improve passive remote sensing data. Following Mitchell et al. (2025b), corrections are made to MODIS LAI prescribed to the Community Land Model version 5 (CLM5.0) for evergreen trees in the third study, to investigate the impact of using LAI datasets improved by spaceborne lidar measurements on land modeling. The findings show promising results in the improvement of the representation of LAI for evergreen throughout the year. In boreal evergreen forest, changes to the LAI substantially impacted snowpack, evaporation, and runoff by shifting the seasonal cycles by a month. For tropical evergreen forest, the largest changes were seen in wet season partitioning of evaporation, but overall changes were relatively small . These studies highlight the need for continued improvement of the retrieval of hydrometeorological properties from spaceborne lidar and the importance of continuing future spaceborne lidar missions with new advancements in lidar technology.