Abstract. Accurate estimation of aboveground biomass (AGB) is crucial to understanding the carbon cycle in conservation areas. This study developed a predictive model for AGB density in La Joya - La Barreta Ecological Park, Querétaro, by integrating LiDAR data from the GEDI mission with passive sensor information (Sentinel-1 and Sentinel-2) and topographic variables. This park is a vital space for recreation, environmental education, and carbon credit generation. We used Global Ecosystem Dynamics Investigation (GEDI) AGB density data from April 2024, with values ranging from 2.5 to 368.2 Mg/ha. For modelling, we processed satellite imagery and a digital terrain model, generating a comprehensive set of 178 spectral indices and topographic variables. Through Recursive Feature Elimination (RFE), five optimal covariates were selected: the LS-Factor, Analytical Hillshading, and Channel Network Base Level (topographic variables), along with Sentinel-2's Triangle Water Index (TWI) and Enhanced Modified Bare Soil Index (EMBI) spectral indices. The Quantile Regression Forest (QRF) model predicted AGB density with an RMSE of 1.8 Mg/ha and an R2 of 0.59, indicating a robust predictive capability for local applications. AGB density predictions ranged from 5.2 to 113.4 Mg/ha, with a mean of 32.9 Mg/ha. A total biomass of 8171.9 Mg was estimated for the park, containing 3652.8 Mg of organic carbon. These results provide a cost-effective basis for monitoring and verifying the park's conservation projects, highlighting the importance of topography and spectral indices in biomass distribution.
Rivera et al. (Tue,) studied this question.