A remote sensing framework for predicting and managing waste generation in tropical highland agroecosystems: case study of land use transitions in cameron highlands, Malaysia

Abstract

Urbanisation, agriculture, and tourism are rapidly transforming land use and land cover (LULC) in tropical highland agroecosystems. In the Cameron Highlands, Malaysia, these transitions are closely associated with rising volumes of solid waste, yet the spatial dynamics remain poorly understood. This study presents a remote sensing-based framework to predict and analyse solid waste generation using multi-temporal satellite-derived LULC data and georeferenced waste datasets from 2020 to 2024. Annual LULC maps were generated via supervised classification of Sentinel-2 and Landsat-8 imagery using Google Earth Engine (GEE), while overlay analysis in ArcGIS Pro 3.2 linked waste data to specific land use zones. Correlation and regression analyses revealed strong positive relationships between cropland expansion and waste generation (r = 0.89, p < 0.05), and a significant negative correlation with forest cover (r = − 0.95, p < 0.05). High-waste areas were consistently located within 500 m of built-up and agricultural zones, particularly in Brinchang, Tanah Rata, and Kea Farm. A spatio-temporal model achieved 92% accuracy in predicting waste-intensive zones. Waste peaked in 2023, aligning with major land conversion trends, highlighting the influence of land dynamics on waste proliferation. This integrated approach demonstrates the utility of remote sensing and GIS for highland waste prediction and provides evidence for spatial zoning, ecological buffer preservation, and tailored circular economy strategies. The framework offers a replicable tool for sustainable waste governance in other ecologically sensitive or rapidly urbanising highland regions.