INTEGRATING REMOTE SENSING AND FIELD DATA FOR BIOMASS ESTIMATION IN COCOA AGROFORESTRY SYSTEMS IN ATLANTIC FOREST, BRAZIL.

Vegetation biomass plays a crucial role in sequestering carbon, helping to mitigate climate change by reducing atmospheric CO₂ levels. This is particularly relevant in agroforestry systems, which provide sustainable methods for cultivating and preserving tree cover. In this study, we modeled above-ground biomass (AGB) using 520 field inventory plots across agroforestry systems and forested areas, along with optical (Sentinel-2), SAR (Sentinel-1), and spaceborne Lidar data (GEDI) from 2019 to 2023 as predictors in a random forest model. The model achieved a strong predictive performance, with an R² of 0.94 and a general RMSE of ±29.21 Mg ha⁻¹ across the study area. However, predictions underestimated forest biomass due to high AGB variability. To improve accuracy, we recommend incorporating additional variables, such as canopy height and climate data on precipitation and temperature.