Tracking mangrove ecosystem dynamics: A remote sensing approach for species classification and conservation assessment

Abstract

Biodiversity among mangrove species is crucial for ecosystem health and resilience but is increasingly threatened by human activities and inadequate restoration efforts. To successfully evaluate conservation efforts, we must understand temporal dynamics. This study assessed changes in mangrove species diversity and the effectiveness of rehabilitation using a multi-sensor remote sensing approach integrated with field-based ecological data. We analyzed 36 years (1987–2022) of Landsat data to map long-term changes in mangrove extent and stand types, alongside low-tide Sentinel-1 SAR and Sentinel-2 MSI composites to reduce tidal influence on spectral and backscatter signals. A systematic sampling design incorporating 257 field plots and hierarchical clustering was used to identify six distinct species communities. Tidal inundation had a significant impact on backscatter and reflectance, highlighting the importance of low-tide imagery for accurate species discrimination. The Simplified Automatic Regrowth Monitoring Algorithm (SARMA) revealed that mangrove extent increased by 1518 ha, from 7030 ha in 1987–8548 ha in 2022 representing a 21.59 % expansion relative to the 1987 baseline. Within this increase, rehabilitated and regenerated areas added 837 and 681 ha, respectively. However, the rehabilitated stands were largely monoculture and structurally less complex, signaling the need for more diverse and ecologically informed restoration strategies. Species mapping using random forest classification achieved an accuracy of 83.57 %. By integrating long-term remote sensing with ecological clustering, this study provides a scalable tool for monitoring mangrove diversity and demonstrates how rehabilitation outcomes can guide evidence-based conservation and sustainable management policies.