ASSESSING TEMPORAL DYNAMICS, GEOMETRY, AND COMPOSITION OF KATWA MUNICIPALITY LANDFILL SITE USING HIGH-RESOLUTION MULTISPECTRAL DATA AND MACHINE LEARNING-BASED SPECTRAL ANALYSIS

Abstract

This study attempts to assess and characterize landfill site under Katwa Municipality in Purba Burdwan district, West Bengal, India. The landfill site has grown rapidly in recent years, in an unmanageable manner. The majority of the municipal solid waste is dumped here. Areal and geometric changes over time (from 2017 to 2025) have been assessed and characterized by digitizing the site for the present day and at two other points in the past (2017 and 2021) using Google Earth and its History tool. Existing geometric properties have also been characterized using different indices, including the Shape Index (SI), Compactness (C), Fractal Dimension, and Elongation Ratio (ER). The composition of the waste has been characterized using remotely sensed data by comparing two advanced techniques: Linear Spectral Unmixing and Random Forest-refined Spectral Unmixing. Field visits were conducted, and the materials of composition were identified as follows 1. Plastic, 2. Raw and decomposed organic material, humus, and barren soil and 3. Thin and sparse vegetation cover. The landfill site has grown in size, possibly because more waste is being made. The evaluation of site geometry reveals a deterioration in compactness and regularity, indicating potential deficiencies in planning and waste management oversight. It is clear from what we see that there is too much plastic, either in the form of fully plastic parts or materials that are covered in plastic. The combination of RF and LSU worked better for finding plastic and plants. But LSU by itself might work better for the soil. The results provide important decision-making aid to municipal authorities in improving their landfill practices and waste management activities through remediating waste.