Aboveground biomass and carbon stock of evergreen forests in six socio-economic regions of Vietnam: an approach combining multispectral optical and radar remote sensing.

This research assesses the spatial distribution of aboveground biomass (AGB) and aboveground carbon (AGC) in evergreen forests across six socio-economic regions of Vietnam. The classification outcomes of evergreen forests utilizing the Support Vector Machine (SVM) model achieved an overall accuracy (OA) of 86.59% and CI (F1_forest) of 0.845-0.955, suggesting that evergreen forests represent over 44% of the total forest area in Vietnam. The amalgamation of diverse data sources, encompassing optical indices, radar, and topographical information, considerably enhanced the precision of AGB estimations, with the Genetic Algorithm-Adaptive Neuro Fuzzy Inference System (GA-ANFIS) model reaching an R2 value exceeding 0.82. The findings indicate a cumulative AGB figure of 1,377,020.22 tons/ha, which is 2.2 times higher than the total AGC figure of 605,026.24 tons/ha. Regions characterized by lower AGB and AGC values are primarily concentrated in the Northern Midland and Mountain areas (NMR). Conversely, areas with elevated AGB and AGC values are predominantly located in the North Central and Central Coast (NCR), Central Highlands (CHR), and specific provinces within the Southeast region (SER). While the GA-ANFIS model exhibited commendable performance, a gap persists between forecasted and actual figures, with an overall discrepancy of 16,749.81 ha for AGB and 8156.6 tons/ha for AGC, likely attributable to challenges in field data acquisition, particularly in pristine and mature forest ecosystems. This study serves as an instrumental resource for extensive forest resource management and advocates for targeted restoration actions in zones exhibiting diminished AGB and AGC levels.