Assessment of Land Use and Land Cover Change Dynamics Using Google Earth Engine
Shashi Bhushan Kumar
Agricultural and Food Engineering Department, IIT Kharagpur, West Bengal, India.
Krishna Mondal
College of Agricultural Engineering and Post Harvest Technology, Central Agricultural University Imphal, Ranipool, Sikkim, India.
Ashok Mishra
Agricultural and Food Engineering Department, IIT Kharagpur, West Bengal, India.
Suyog Babasaheb Khose *
Krishi Vigyan Kendra Narayangaon, Pune, Maharashtra, India.
*Author to whom correspondence should be addressed.
Abstract
This study assessed land use and land cover (LULC) dynamics in the Subarnarekha River Basin, eastern India, using multi-temporal satellite datasets and cloud-based geospatial analysis. The study aimed to map LULC changes during 2019–2021, evaluate their spatial and temporal variability, and compare the performance of Random Forest and Support Vector Machine classifiers within Google Earth Engine. Sentinel-2 multispectral imagery and Sentinel-1 synthetic-aperture radar data were integrated to reduce limitations associated with cloud cover and spectral ambiguity. Seasonal, annual and combined LULC maps were generated for seven classes: trees, shrubland, grassland, cropland, built-up areas, bare/sparse vegetation and permanent water bodies. The overall classification accuracy ranged from 65.16% to 73.87%, while the kappa coefficient ranged from 0.61 to 0.69. Combined multi-temporal datasets produced higher classification accuracy than single-season datasets, indicating the value of temporal integration for reducing classification uncertainty. The results showed a gradual decline in forest cover, accompanied by increases in cropland and built-up areas. Grassland and bare/sparse vegetation showed moderate variability, whereas permanent water bodies remained comparatively stable with seasonal fluctuations. These changes indicate increasing anthropogenic pressure associated with agricultural expansion and urban development. The findings provide baseline spatial information for understanding land-transformation processes and supporting sustainable land and water resource management in the Subarnarekha River Basin.
Keywords: Land use and land cover change, Google Earth Engine, Sentinel-1, Sentinel-2, Random Forest, Support Vector Machine, Subarnarekha River Basin, multi-temporal classification, remote sensing, classification accuracy, watershed management