Sentinel-2 red-edge spectral indexes best suited to discriminate burned from unburned areas in Mediterranean forest ecosystems
Post-fire forest management is been increasingly based on fire damage maps derived from satellite imagery. Though Landsat data have been the most commonly used at medium scale (<100m / pixel), Sentinel 2 satellites provide an opportunity for post-fire damage analysis. MultiSpectral Instrument (MSI) onboard of Sentinel 2 satellites acquires data in red-edge wavelengths, and has higher spatial (10/20m vs. 30m) and temporal (16 vs. 5 days) resolution. Thus, the aim of this study is to check whether Sentinel 2 MSI spectral indexes that include red-edge bands allow a better discrimination between burned and unburned areas than conventional spectral indexes based on red, near infrared and/or short wave infrared. A large forest fire (79.5 km2) occurred in Sierra de Gata (Spain) in August 2015 acted as study area. Official fire perimeter together to Copernicus Emergency Management Service information (ID: EMSR132) provided us a terrain reference. Logistic regression models based on Sentinel 2 MSI spectral indexes (conventional and red-edge based) showed that red-edge spectral indexes outperformed conventional ones in terms of discriminating burned from unburned areas.
Main Authors: | , , , |
---|---|
Other Authors: | |
Format: | actas de congreso biblioteca |
Language: | English |
Published: |
SPIE digital library
2020-09-04
|
Subjects: | Burned area, Forest fires, Logistic regression, Red-edge, Sentinel-2, |
Online Access: | http://hdl.handle.net/10261/344761 https://api.elsevier.com/content/abstract/scopus_id/85093831368 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|