Wall-to-wall Mapping of Boreal Forest Heights (year 2019-2020) based on quality controlled ICESat-2 and Sentinel data in Fairbanks, Alaska

This dataset provides the ICESat-2 observations (between 2019 and 2020) with quality enhanced by a multivariate quality control approach, known as the Cook's distance, and its based boreal forest mapping in combination with Sentinel-derived explanatory variables for a study site in Fairbanks, Alaska. The regression model used for canopy height modeling is the random forest. The refined ICESat-2 and its based canopy height modeling achieve high Pearson's correlation coefficients (~0.8, ~0.85) to airborne lidar measured counterparts. The refined ICESat-2 data can be used for modeling validation and the mapped forest height products can be used for eco-related studies, such as biomass estimation and forest height monitoring.