Integrating Machine Learning with Sentinel-2 Imagery for Mangrove Health Index Assessment

Abstract

Mangrove ecosystems are increasingly threatened by climate change and land-use conversion, highlighting the need for accurate and scalable health monitoring tools. This study addresses this challenge by integrating machine learning (ML) techniques with multispectral data from Sentinel-2 imagery to develop a Mangrove Health Index (MHI). The methodology includes preprocessing satellite imagery, extracting spectral characteristics of mangrove vegetation, and validating results with ground-truth data from field surveys. Three ML models—support vector regression (SVR), random forest regression (RFR), and k-nearest neighbor (k-NN)—are evaluated. Among them, the RFR model demonstrates the highest accuracy, achieving a root mean square error (RMSE) of 6.00, a mean absolute error (MAE) of 4.73, and a coefficient of determination (R2 ) of 0.82. These findings support the use of ML methods combined with medium-resolution satellite imagery to enable scalable, precise assessments of mangrove health, offering valuable tools for environmental monitoring and conservation planning.