Spatio-temporal dynamics of mangrove ecosystems using remote sensing and machine learning in Google Earth Engine; A case study of Tongke-Tongke Indonesia

Abstract

This study aimed to map changes in mangrove ecosystem conditions in Tongke-tongke Village, Sinjai Regency, South Sulawesi, using the Google Earth Engine (GEE) platform. Mangroves are vital coastal ecosystems that are vulnerable to degradation owing to anthropogenic activities. Therefore, efficient and large-scale monitoring of mangrove conditions is essential for sustainable mangrove management. This study analysed changes in mangrove density and health between 2013 and 2023 through NDVI (Normalized Difference Vegetation Index (NDVI) and Mangrove Health Index (MHI) transformations. Accuracy testing was conducted using confusion matrix analysis and ground truthing using hemispherical photography. The NDVI classification results indicated an increase of 6.77 hectares in dense mangrove areas, while MHI analysis revealed 25.07 hectares increase in healthy mangrove areas. The overall classification accuracy reached 88.89%, with a kappa coefficient of 0.83, which was categorized as very good. These findings demonstrate that GEE is an effective tool for long-term and large-scale mangrove monitoring and provides valuable spatial information to support decision-making in coastal zone management.

Access to Document

Other files and links

• Link to publication in Scopus
• Open Access Version Available

Fingerprint