Pioneering approach for mangrove forest monitoring: Enhancing accessibility and assessment methodology in developing countries

In the era of climate crisis, precise spatial data is essential for implementing international agreements and initiatives to mitigate climate change. Developing countries have significant challenges in acquiring spatial data due to limited technical expertise, infrastructure, and financial resources. Timor-Leste, a small island nation in Southeast Asia, has mangrove forest with substantial potential for climate change mitigation. However, a lack of accurate spatial data on the distribution of mangrove forests has hindered effective protection and restoration strategies. This study presents a cost-effective method for developing countries to monitor their mangrove forests using earth-observing satellite imagery. We acquired open-source satellite imagery from the Google Earth Engine platform and applied U-Net deep learning algorithm to create mangrove forest distribution maps of mainland in Timor-Leste for the years 2013, 2019, and 2023. To verify our model and results, we conducted field visits in Timor-Leste to explore mangrove habitats and constructed ground truth data to evaluate the accuracy. The overall accuracy of the mangrove forest classification results by year was 0.890, 0.869, and 0.901, respectively, indicating high accuracy. The estimated areas of mangrove forests of mainland in Timor-Leste were 6,378.2 ha in 2013, 4,175.1 ha in 2019, and 3,504.7 ha in 2023, describing an annual decline of 4.51% across the nationwide. This study demonstrates the potential of using advanced remote sensing technologies to monitor changes in mangrove forests in countries with limited scientific expertise and financial infrastructure. It is expected to support conservation and restoration projects, contributing to global efforts to mitigate climate change.