Mangrove forests play a vital role in mitigating global climate change. However, despite growing awareness and conservation efforts, deforestation and degradation continue to threaten these ecosystems. A key challenge is the lack of an efficient, accessible monitoring system to assess deforestation and restoration. This study aims to develop indicators for deforestation and restoration monitoring based on Sentinel-2 time series using the Continuous Change Detection and Classification (CCDC) algorithm. The research was conducted at two sites in Lampung, Indonesia, i.e., a large natural mangrove patch and a restoration project area. The primary dataset consists of Sentinel-2 time series from 2019 to 2024. Based on this data and workflows implemented in Google Earth Engine (GEE), the study successfully developed mangrove monitoring maps at 10 m resolution: a deforestation monitoring map with eight classes (overall accuracy 86%) and a restoration monitoring map with nine classes (overall accuracy 90%), serving as indicators to track status, disturbance, and recovery. Further sensitivity test via systematic parameter perturbation (±10% and ±20%) demonstrated high model stability, with accuracies ranging from 78% to 90% across all scenarios. Results show that most natural mangroves are in healthy stable condition, while restored mangroves are showing sustainable growth. However, some small areas were found to have experienced disturbances, requiring timely intervention to prevent further loss. By providing the proposed indicators, our study will not only significantly strengthen local management capacity by eliminating technical barriers but also provide a monitoring tool to mitigate mangrove deforestation and ensure the long-term sustainability of restored mangroves. • Maps track mangrove degradation and recovery for timely management practices. • Deforestation and restoration maps reached 86% and 90% accuracy, respectively. • 85% of existing mangroves (3,455 ha) remain healthy, but 173 ha are disturbed. • Restoration monitoring shows 74% success (133 ha), while 14 ha face disturbance.
Fauzi et al. (Fri,) studied this question.