Predicting Mangrove Degradation Under Anthropogenic and Climatic Stressors: A Google Earth Engine - Based Case Study from Pichavaram

Abstract.
Mangrove forests are vital ecosystems that protect coastlines, support biodiversity, and capture carbon, yet they are increasingly threatened by human activities and climate change. This study presents a data driven approach to assess and predict mangrove degradation in the Pichavaram region of Tamil Nadu, India, using Google Earth Engine (GEE) and Random Forest classification. By combining satellite imagery from Landsat and Sentinel-2 with vegetation indices (NDVI and NBR), elevation, slope, salinity, tidal dynamics, and distances to roads, rivers, and settlements, developed a model to classify mangrove areas into stable, degraded, and regenerating zones. The classification model was associated with a very high accuracy (98.14%), strong agreement (Kappa = 0.972), and AUC scores above 0.96 for every class. According to the results, degraded zones were usually close to settlements, roads, and aquaculture, whereas regenerating patches lied close to rivers. A custom Climatic Stress Index integrating vegetation trends, salinity, and tidal variability provided added insight into environmental pressures.This research shows a transferable, cloud-based methodology for real-time mangrove monitoring, which thereby can provide useful tools for conservation planning, restoration prioritisation, and climate adaptation exercises in accordance with the Sustainable Development Goals (SDG 13 and 15). It shows how remote sensing and machine learning can be used to direct ecosystem management in areas where there is a lack of.