Fire Severity and Vegetation Recovery Determination Using GEE and Sentinel-2: The Case of Peschici Fire

We propose an accurate and rapid methodology for the extraction of spatio-temporal fire features using Sentinel 2 products and the Google Earth Engine (GEE) platform. All Sentinel 2 images available in the GEE platform were clipped using the fire area mask and then the NBR, NDVI. dNBR and RdNBR indices were derived. The differential values of NBR, NDVI, dNBR and RdNBR were obtained by calculating the difference of the index values between two temporally adjacent images. The use of all available images in GEE restricted the time of occurrence of the images 5 days, excluding cloud-covered images and shortening the processing time of each satellite image. The results obtained showed that the proposed methodology allows for the rapid and accurate identification and classification of burnt areas, and also allows for the efficient and accurate extraction of the spatio-temporal characteristics of post-fire vegetation recovery. The results obtained can be used to implement targeted post-fire vegetation restoration practices.