A data-driven, cloud-based approach for forest aboveground biomass mapping using GEDI and other earth observation data: an ecoregion-specific Investigation across the state of Alabama, USA

Forest aboveground biomass (AGB) is a crucial indicator for carbon monitoring. This study presents a data-driven, cloud-based workflow for AGB density (AGBD) mapping using Global Ecosystem Dynamics Investigation (GEDI) data combined with Earth observation (EO) datasets. Focusing on the state of Alabama, the specific objectives were to: (1) evaluate and compare the effectiveness of ecoregion-specific models versus a statewide model for estimating AGBD, and (2) generate a 30-m AGBD map based on the optimal model and assess agreement with forest inventory and global-scale information. GEDI data preparation was performed locally, while the analysis was conducted within GEE using the Random Forest algorithm. The ecoregion-specific models achieved higher accuracy (R2: 0.20-0.71; RMSE: 66.67-48.99 Mg/ha) than the statewide model (R2: 0.41; RMSE: 62.09 Mg/ha). These findings underscore the importance of ecoregion-specific modeling and combining EO data in cloud computing for advancing AGBD mapping.