Evaluación de las potencialidades y limitaciones de los satélites Sentinel-2 y Landsat-8 para el estudio del grado de severidad de incendios forestales en la zona andina

Abstract

Forest fires are catastrophic events of great impact because they affect the vegetation and the forest landscape. They also cause biodiversity loss, land degradation and ecological imbalance. Post-fire assessment using severity indices is essential to find the effects of fire on vegetation and implement mitigation strategies. This study evaluates the potential and limitations of images from Sentinel-2 and Landsat-8 satellites for the study of the degree of severity of forest fires in the Andean zone of Ecuador, from 2017 to 2020 using the Google Earth Engine analysis platform. Two algorithms were used, the first was used to calculate and compare spectral indices of forest fire severity (dNBR and RdNBR), and the second to identify burned areas. The results were then compared with fire data provided by MAE and the "Amazonía Sin Fuego" program. The RdNBR index underperformed compared to the dNBR index. Therefore, it was determined that the dNBR index calculated for Landsat-8 was the most suitable for the identification of burned areas in the study area. With respect to the detection of burned areas, it was observed that both satellites obtained specificity values greater than 90%. Finally, a workflow for using GEE in the study is presented.