Forest cover dynamics (1998 to 2019) and prediction of deforestation probability using binary logistic regression (BLR) model of Silabati watershed, India

Abstract

Forest is a renewable resource and an imperative part of the total environment. But in the last few decades, forest areas are being highly converted and depleted due to excessive anthropogenic stress particularly for modern societal development. The magnitude of deforestation has been exaggerated particularly in tropical and sub-tropical developing countries of the world. The main objective of the study is to identify the probable areas of deforestation along with the analysis of forest cover dynamics within Silabati watershed of India applying Binary Logistic Regression Model. The forest cover map (1998 and 2019) has been extracted with the help of satellite imageries which have been obtained from USGS. The remote sensing techniques coupled with multivariate statistical analysis (Binary Logistic Regression Model) have been used to prepare probable deforestation areas within the watershed. The statistical analysis also signifies the correlation among dependent and independent variables. The model shows that the middle and lower middle part of the watershed is highly vulnerable (approximate 24%) whereas the upper basin area is less vulnerable (approximate 48%). The current study reflects that high population density, road construction and initiation of cash crops are responsible principal factors behind probable deforestation. The model is validated by the receiver operating characteristic (ROC) curve and the Area Under Curve (AUC) method and it reflects around 80% overall accuracy of this model. This model is very rational and effective for policy makers and conservators before the execution of any sustainable forest conservation plan.

Topics & Concepts