ABSTRACT
The tropical forest contributes around 5% to 15% of atmospheric carbon emissions, which are mostly anthropogenic. But there are large uncertainties in the quantification of these emissions from its sources. The remote-sensing data offers a practical opportunity to monitor and assess different forest disturbances. Western Himalayan forest is often affected by fire events, mostly during (pre-monsoon) dry and warm periods. In this study, we present a way to monitor the forest degradation condition using spectral mixture analysis (SMA) and surface reflectance of Landsat-8 data from 2014 to 2019. The Normalized Degradation Fraction Index (NDFI) has been performed by using spectral end member fractions of green vegetation (GV), non-photosynthetic vegetation (NPV), soil, and shade in the Google Earth Engine (GEE) cloud platform. The NDFI shows considerable spatial correspondences with clusters of fire spots during the pre-monsoon period. Around 3% to 9% of the forest burned area transformed to partially to highly degraded forest. The overall trend of degradation fraction (NDFI) over total forest cover shows a significant negative trend over a considerable area. Thus, Landsat-8-based SMA and NDFI demonstrate a potential way to identify forest degradation mediated by forest fires, although remote sensing-based approaches are limited in their capacity to accurately detect forest disturbances. Furthermore, field-based studies are needed to monitor the potentialities of the NDFI approach in forest degradation identification.
