Estimating aboveground biomass and biodiversity of a protected area in moist tropical forest ecosystem using remote sensing application

• Multispectral and SAR data used to map AGB in tropical forest sanctuary. • High tree density and low basal area suggest recovering secondary forest. • Biodiversity indices improved AGB model estimates. • Sentinel-2 outperformed LISS-IV in AGB prediction despite lower spatial resolution. • Microcos paniculata dominance indicates past disturbances in the protected area. Moist tropical forest ecosystems serve as substantial repositories for high carbon stocks and are also rich in biodiversity. Protected areas in these ecosystems increase the population of native floral and faunal communities and conserve biodiversity. This study aimed to provide a quantitative baseline mapping of aboveground biomass stock and tree diversity in the Rowa Wildlife Sanctuary (RWLS) of Tripura, northeast India using multispectral and SAR remote sensing (Sentinel-1, Sentinel-2, LISS-IV) and ensemble machine learning models. We established 126 plots (10 m × 10 m) for field sampling of trees, herbs, shrubs, deadwood, and leaf litter. Local allometric models were used to calculate tree-level AGB in the field. We also estimated different indices pertaining to biodiversity such as the Importance Value Index, Family Importance Value, Pielou’s Evenness index, Margalef’s richness index, Simpson’s Dominance index and Shannon-Wiener Diversity Index. Our results showed 68 tree species in the area, indicating moderately high species richness. The tree density of 1412 trees ha −1 was significantly higher than usual, while the mean basal area of 20.11 m 2 /ha was low. These findings suggest a forest dominated by smaller diameter trees, potentially indicating a recovering secondary forest subject to frequent disturbances. Remote sensing analysis demonstrated that Sentinel-2 outperformed LISS-IV in predicting AGB, despite LISS-IV’s higher spatial resolution, highlighting the importance of spectral resolution in estimating biomass. Incorporating biodiversity indices (Shannon-Wiener, Simpson’s, and Margalef’s) also significantly improved AGB estimates (R 2 = 0.59). The mean AGB of RWLS was estimated as 122.60 ± 42.75 Mg ha −1 . The present study also observed a dominance of the species Microcos paniculata, reflected through high Simpson index (D = 0.18) and low evenness (0.6), which suggests its colonization after past disturbances in the area. The outcomes of the study underscore the need for targeted conservation efforts in RWLS, considering its challenges as a small protected area under anthropogenic pressure. We recommend implementing silvicultural treatments, measures to prevent disturbances, and establishing a regime of repeated monitoring using advanced remote sensing technologies.