Addressing deforestation in Myanmar by improving near real-time satellite monitoring

Monitoring forest disturbance in near real-time (NRT) is crucial for mitigation efforts across the world, and this is especially true in countries like Myanmar. As one of the most forested nations in Southeast Asia and home to high levels of biodiversity, the country is disproportionately affected by unchecked forest disturbance, usually from deforestation. Meanwhile, NRT monitoring via remote sensing has been advancing for decades, with recent methods providing promising results. However, there is a clear lack of success with smaller-scale disturbance events, whereby the time lag between an event and high confidence that the event occurred is >2 weeks. This is not ideal for forest managers, particularly in larger protected areas, to intervene before the disturbance event has finished.

This study addresses a new method of progressing NRT monitoring by using a multi-source approach. Specifically, we expand on recent efforts of combining remote sensing data by aggregating Sentinel (1, 2 [top of atmosphere (TOA)]), Landsat (8, [surface reflectance [SR]), and MODIS (MOD09GA) data for the normalized difference vegetation index (NDVI). We apply our method to Chatthin Wildlife Sanctuary in north-central Myanmar due to its biodiversity and our partnership with the park staff. Initial results yield a workflow for robust estimation of disturbance probability across the landscape. The next steps will focus on closing the loop between data users and producers by directly incorporating field observations from local forest managers. To our knowledge, this study represents the first instance of a large multi-source method applied to NRT monitoring.