Determining thresholds of vegetation indexes for mangrove cover mapping and change detection in Ha Tinh province

Tóm tắt

Accurate monitoring of mangrove ecosystems is crucial for coastal zone management, yet region-specific methodologies remain underdeveloped for many vulnerable coastlines. This study developed optimized vegetation index thresholds from Sentinel-2 imagery for accurate mapping of mangrove cover changes in Ha Tinh province, Vietnam. Eight vegetation indices were calculated and evaluated using field survey data combined with remote sensing data. Through threshold analysis and accuracy assessment, the Combined Mangrove Recognition Index (CMRI ≥ 0.15) was identified as the most effective index for mangrove detection in the study area. The mangrove classification for 2025 achieved high accuracy (Overall Accuracy = 95%, Kappa = 0.91). Change detection analysis between 2020 and 2025 revealed a net increase in mangrove area by 27.86% (from 687.97 ha to 879.65 ha). This expansion is attributed to active government-led restoration programs and supporting initiatives. However, persistent threats from illegal logging, aquaculture expansion, and climate-induced hazards highlight ongoing challenges. This study proposes a practical and transferable remote sensing approach to mangrove monitoring. It underscores the need to integrate robust spectral thresholding with effective policy enforcement to ensure long-term mangrove conservation in the vulnerable coastal areas of central Vietnam.