Portal de Periódicos da CAPES

Above-ground biomass estimation models of mangrove forests based on remote sensing and field-surveyed data: Implications for C-PFES implementation in Quang Ninh Province, Vietnam

2021; Elsevier BV; Volume: 48; Linguagem: Inglês

10.1016/j.rsma.2021.101985

2352-4855

Nguyen Hai Hoa, Thị Thu Hien Nguyen,

Land Use and Ecosystem Services

The free charge and open-source of remote sensing imagery, including Landsat-8 and Sentinel-2, offer new opportunities for forest-based AGB mapping and monitoring, especially mangrove forests in the tropics. Modelling relationships between mangrove AGB estimation-based survey and remote sensing data (spectral bands and vegetation indices) have not been evaluated in Quang Ninh Province, Vietnam. In this study, we evaluated the capability of Landsat-8 and Sentinel-2 data for the retrieval and predictive mapping of mangrove AGB in Mong Cai Coast, Quang Ninh Province as a case study. We used 2019 Landsat-8 and Sentinel-2 to develop AGB estimation models through stepwise linear regression approaches in R statistics. We developed models each from spectral bands and vegetation indices derived from Landsat-8 and Sentinel-2 imagery. The results showed that the models based on spectral bands and vegetation indices derived from Sentinel-2 were more accurate in predicting the overall AGB of mangrove forests than those of Landsat-8 data. High correlation values between AGB and Sentinel-2-derived vegetation indices (Model 6.6, r2=0.973; Model 6.7, r2=0.982; Model 6.8, r2=0.988) and Landsat-8-derived vegetation indices (Model 6.1, r2=0.927; Model 6.2, r2=0.927; Model 6.3, r2=0.939); and between AGB and Sentinel-2-derived spectral bands (Model 5.5, r2=972; Model 5.4, r2=0.975); between AGB and Landsat-8 derived spectral bands (Model 5.2, r2=0.913; Model 5.3, r2=0.935; Model 5.1, r2=0.855) were obtained. The developed AGB estimation models have high prediction accuracy, agreements of observed and predicted AGB values of 89.88% for Landsat-8 derived vegetation index (Model 6.2), 96.51% for Sentinel-2 derived vegetation index (Model 6.6). Overall, both Landsat-8 and Sentinel-2 provide satisfactory results in the retrieval and predictive mapping of mangrove AGB. Our study suggests that mangrove conservation under C-PFES schemes should be applied over Quang Ninh Coast based on AGB estimation models developed.