Combining UAV-SfM, SAR, MSI and field surveys for estimation of above ground biomass in mangrove forest of Chonburi, Thailand

Abstract

Mangrove biomass is a key indicator for quantifying carbon cycling in blue-carbon ecosystems, yet conventional approaches face significant challenges. To improve large-scale mangrove biomass assessment and provide a baseline for targeted conservation, present study proposes a Single-tree–Plot–Community–Region (AGBT/F~U~S) upscaling method that integrates UAV-SfM, SAR, MSI, and field surveys, and applies it to Chonburi, Thailand. In 2023, total mangrove aboveground biomass in Chonburi Province was 145.24 kt, with a mean AGB density of 101.61 Mg/ha, slightly below the global mangrove average. Long-term records reveal an initial decline followed by post-2015 recovery to about 85% of the 1996 level. Relative to the conventional plot–satellite model, the AGBT/F~U~S framework substantially improves estimation performance and reduces prediction error (ΔR²≈0.47; ΔRMSE ≈ 66.03 Mg/ha), and remains robust under limited training data, with accuracy gains saturating once plot numbers exceed a moderate threshold. These results demonstrate that multi-scale upscaling provides a transferable pathway for mangrove biomass mapping in data-scarce regions and offers a practical baseline for blue-carbon accounting and targeted restoration planning.