<![CDATA[Mangrove forests are vital coastal ecosystems that protect shorelines and help maintain coastal environmental balance. Mangrove forests undergo changes every year due to degradation or restoration, so monitoring needs to be carried out. This study analyzes the spatiotemporal dynamics of mangrove density in Demak Regency from 2020 to 2025 using Sentinel-2 imagery and the CART (Classification and Regression Tree) algorithm processed through the Google Earth Engine (GEE) platform. Six spectral indices were used as classification inputs, namely NDVI, NDWI, MVI, MI, AMMI, and CMRI. AMMI is the best index in identifying mangroves as a whole, both in narrow and large areas. Based on classification, the mangrove cover area has been restored or significantly increased from 1644.011 ha in 2020 to 2530.522 ha in 2025. The largest increase occurred in the high canopy density class of 960.157 ha. Meanwhile, the medium and low canopy density classes showed a decrease in area. Accuracy assesment showed an overall accuracy value of 99.92% for 2020 and 99.91% for 2025, with kappa accuracy above 97% in both years. These results show that the classification method with the support of cloud computing can be relied on in spatiotemporal monitoring of mangrove changes efficiently and accurately. Keywords: Mangrove, Sentinel-2, Machine Learning, Classification and Regression Tree, Cloud Computing]]>
Copyrights © 2025
