Forest cover dynamics in India quantified using Google earth engine
Student name: Mr Ankit Khajuria
Guide: Dr Neeti
Year of completion: 2016
Host Organisation: Jawaharlal Nehru University, New Delhi
Supervisor (Host Organisation): Dr P. K. Joshi
Abstract: Quantification of forest change in India has been inadequate despite the recognized
importance of forest ecosystem services India has a diverse range of climate and
vegetation types with respect to a varied topography. In this study we have classified
6 main types of forest found in India i.e. Dry Deciduous, Moist Deciduous,
Temperate Broadleaved, Temperate Conifer, Tropical Evergreen and Tropical
Semi-evergreen. Random Forest classification technique has been used to classify
forests of India. The classification of vegetation has been carried out based on the
Champion and Seth (1968) scheme of forest types in India. The main purpose of this
study is to consistently quantify forest cover change across Indian region, since the
1980s based on the Landsat data archive. An algorithm has been developed to
simultaneously process data from different Landsat platforms and sensors
(TM,ETM+, OLI) to map annual forest cover of India using open source Google
Earth Engine cloud computing platform. Accuracy assessment was performed and
classification accuracy of almost 89% has been achieved. This spatially explicit
database will be highly useful for the studies related to changes in various forest
types across India.
Keywords: Google Earth Engine, Random Forests, Forest type, Cloud computing
