Sub-Annual Scale LandTrendr: Sub-Annual Scale Deforestation Detection Algorithm using Multi-Source Time Series Data

In cloudy and rainy regions, frequent cloud cover limits clear data obtained using a single optical sensor, posing substantial challenge for detecting deforestation events on sub-annual scale. this study, scale detection algorithm, namely, the LandTrendr (SSLT) change was developed synergies from multiple sources. First, combined time series constructed by combining Landsat Sentinel-2 data. Second, sliding window applied to spatially normalize normalized burn ratioand eliminate effects of forest phenological changes sensor differences. Finally, an integrated created fit SSLT trajectory, root mean square error (RMSE) fitted trajectory calculated determine segmentation threshold. Pixels with magnitude greater than RMSE three consecutive times were marked as pixels. Application algorithm subtropical low density observations resulted in spatial temporal accuracies 88% 92.8%, respectively. Conclusively, method provides accurate timely identification