Tropical Forest Degradation Assessment; A Case Study using CLASlite

摘要

Detecting and monitoring forest degradation have implications for forest conservation and management. Mapping and monitoring deforestation have been an operational activity using satellite remote sensing. However, the mapping disturbance and degradation of forests is a major challenge. The present study attempted to quantify the forest disturbances in the Lachhiwala forest range situated in foothills of Himalaya in Uttarakahnd state, India. The study utilizes optical satellite data (Landsat TM, 2010 and Landsat 8, 2015) and spectral unmixing of these datasets gave fractional cover (proportion of vegetation, non-photosynthetic vegetation, and bare soil). An automated tool CLASlite was used to generate fractional cover which adopts Monte Carlo Spectral Unmixing techniques that combine spectral and spatial information was used to enhance the detection and mapping of canopy damage, exposed soil, and dead vegetation. Results show that study area has 86.82% intact forest,1.33% less degraded forest, 6.28% moderately degraded forest and 5.57% high degraded forest. Root mean squared error (RMSE) images were used to assess the performance of the Model. Landsat 8 OLI/TIRS has low RSME (i.e. 0-7%) than Landsat 5TM Landsat 5 TM (i.e. 1-10%).