Comparative analysis of machine learning algorithms in Google Earth Engine for urban land use dynamics in rapidly urbanizing South Asian cities

Abstract Land use/land cover (LULC) dynamics play a crucial role in understanding the complex interactions between ecosystems and climate. This study demonstrates the effective integration of Google Earth Engine (GEE) and machine learning (ML) algorithms for monitoring LULC changes in two rapidly urbanizing cities in Bangladesh. By combining Landsat imagery with classification and regression trees, random forest (RF), and support vector machine algorithms within the GEE platform, we analyzed LULC changes from 2001 to 2021. Our analysis revealed significant urban expansion in both cities, with built-up areas showing the highest increase, while natural land covers experienced notable declines. The RF classifier consistently demonstrated superior performance, with the overall accuracy exceeding 93%. The GEE-based approach significantly reduced the processing time compared to traditional methods, while the integration of multiple ML algorithms enhanced the classification accuracy. This research advances environmental monitoring by showcasing the effectiveness of cloud-based geospatial analysis for rapid and accurate LULC change detection. The methodology presented herein offers valuable insights for urban planners and policymakers, particularly in rapidly urbanizing regions, contributing to Sustainable Development Goals 11 (Sustainable Cities and Communities) and 15 (Life on Land).