Decadal aboveground biomass change (2014–2024) across a montane–lowland gradient in southeastern Kenya using airborne LiDAR 

The Taita Hills in southeastern Kenya are a critical biodiversity hotspot within the Eastern Arc Mountains, characterized by a complex mosaic of montane forest fragments, exotic plantations, and agroforestry systems transitioning into semi-arid grasslands and Acacia-Commiphora bushland. This landscape with elevations ranging from approximately 750 m to 2200 m exemplifies competing land-use interests, where a growing population and agricultural expansion have historically driven forest and tree cover loss. Accurate monitoring of these biomass dynamics is essential for quantifying carbon stocks, informing climate mitigation strategies, and guiding contemporary conservation and natural forest regeneration efforts.This study employs an extensive multi-temporal dataset to quantify aboveground biomass (AGB) changes across the Taita Hills and adjacent lowlands. We analyzed data from 38 airborne LiDAR flights conducted between 2014 and 2024, covering 1,600 km², with 650 km² of overlapping coverage for change detection. Field-measured AGB plots (2013–2018) and LiDAR data from 2014/2015 were used to generate a baseline AGB map. A Random Forest model, calibrated on this baseline and LiDAR metrics, was then applied to predict AGB from 2022/2024 acquisitions. These predictions were validated using independent field measurements collected in 2024–2025. Finally, we analyzed annual AGB change rates in relation to high-resolution canopy height model changes, elevation zones, and land cover types to characterize spatial AGB dynamics and identify drivers of gain and loss.Preliminary analysis reveals heterogeneous AGB dynamics across the landscape. The highest positive change rates were observed in young forest plantations, while agroforestry systems exhibited modest gains, indicating successful tree retention and maturation. Notably, native montane forest fragments remained relatively stable, with forest cover losses primarily concentrated within exotic plantations. Conversely, localized AGB reductions were identified in foothill areas and along riverine corridors. The multi-temporal LiDAR approach proved robust for capturing these fine-scale spatial patterns. This ongoing analysis will further refine the magnitude and drivers of decadal carbon stock fluctuations, providing critical evidence for landscape-level conservation and climate mitigation strategies in the region.