Use of GLOBE Observer Citizen Science Data to Validate Continental-Scale Canopy Height Maps Derived from ICESat-2 and GEDI

Mei-Kuei Lu^1*Sorin Popescu¹Lonesome Malambo²Brian Campbell³

• ¹Texas A and M University, College Station, United States
• ²Texas A&M University, College Station, United States
• ³NASA Goddard Space Flight Center, Greenbelt, United States

Citizen science plays a crucial role in enhancing the spatial and temporal resolution of environmental observations. The GLOBE (Global Learning and Observations to Benefit the Environment) Observer (GO) app, a mobile extension of the GLOBE Program, empowers the public to collect environmental data that supports both scientific research and educational outreach. In this study, we leverage citizen science data from the GLOBE Observer program to evaluate and validate three canopy height models (CHM): the ICESat-2 (Ice, Cloud, and land Elevation Satellite-2) product, a GEDI (Global Ecosystem Dynamics Investigation)-Landsat derived global map (GL), and a GEDI-Sentinel-2 fusion map (GS-2). GLOBE tree height measurements, filtered for forested areas using LANDFIRE's (Landscape Fire and Resource Management Planning Tools) Existing Vegetation Cover data, were analyzed across multiple ecozones within the contiguous United States, and manual comparisons were performed with airborne lidar data at selected sites. Our findings indicate that GLOBE data, while providing extensive temporal and spatial coverage, exhibit low general agreement with airborne lidar reference heights (R² = 0.14), primarily due to geolocation inaccuracies and measurement inconsistencies inherent in citizen-collected data. Validations with spaceborne lidar derived canopy height maps (ICESat-2, GL, GS-2) showed generally low correlations (R² = 0.08-0.17), though filtering for higher location accuracy (0-25 m) improved agreement (up to R² = 0.22). Notably, filtering for higher location accuracy (0-25 m) enhances the correlations, though challenges persist. These results underscore both the potential and the limitations of using citizen science data for the validation of spaceborne lidar-derived canopy … p. 3 height maps, and they highlight the need for enhanced data collection protocols to improve geolocation accuracy for future ecological monitoring efforts.