Accuracy of remote sensing-based aboveground biomass and canopy height products in eastern African forest fragments

Eastern African forest ecosystems are under severe pressure from deforestation and forest degradation, which negatively affects biodiversity and forest carbon stocks. While field-based data for forest monitoring are often scarce, the accuracy of global remote sensing products remains poorly understood in this region. To address this gap, we established field plots (r=15m) in three contrasting forest types: Dry Afromontane Forest (Tara Gedam, Ethiopia, n=30), Montane Cloud Forest (Taita Hills, Kenya, n=59), and Coastal Forest (Kay Kambe, Kenya, n=20). We quantified forest structure and aboveground biomass (AGB) and compared field-derived estimates with global and continental biomass products (ESA BIOMASS 2022; African Biomass 2017) and global canopy height products (Lang et al., 2023; Tolan et al., 2024). To account for uncertainty in visually estimated tree heights, we developed a simulation-based correction procedure informed by species-specific maximum heights. Forest structure and biomass differed markedly among sites. Mean AGB varied between 117.6 Mg ha⁻¹ (Tara Gedam), 203.6 Mg ha⁻¹ (Kaya Kambe) and 411.0 Mg ha⁻¹ (Taita Hills). Both biomass products systematically underestimated field-derived AGB across all forests, with underestimation increasing strongly at higher biomass values. In contrast, both canopy height products showed substantially better agreement with field measurements, with best results for Lang et al. (2023). Our results underline that current global and continental biomass products do not adequately capture fragmented eastern African forests, whereas recent canopy height models show considerable promise for forest monitoring and future biomass estimation efforts. Meanwhile, we recommend a closer collaboration and coordination between field-based and remote sensing studies.