Detecting Deforestation Using Logistic Analysis and Sentinel-1 Multitemporal Backscatter Data

A novel method using Sentinel-1 C-band backscattering data reveals that deforestation events can be accurately detected and dated in tropical forests, such as Iguazu National Park, with an F1-score of 0.93. This approach models temporal variability in backscatter intensity using a logistic function, identifying the inflection point as the deforestation date. While effective for tropical biomes, accuracy may drop by up to 20 percentage points in temperate forests due to environmental factors like rainfall and snow. The study demonstrates the potential of Sentinel C-band data for large-scale deforestation monitoring, with a slight delay of two months in estimating event dates.

Dascălu, A., Catalão, J., & Navarro, A. (2023). Detecting Deforestation Using Logistic Analysis and Sentinel-1 Multitemporal Backscatter Data. Remote Sensing, 15(2), 290. https://doi.org/10.3390/rs15020290