Predicting LULC changes and assessing their impact on surface runoff with machine learning and remote sensing data.

Abstract

This study employs an approach to examine the influence of urbanization-induced land use changes on surface runoff. The research leverages the SCS-CN method, integrating remote sensing and machine learning, to analyze land use and cover (LULC) changes over the years 2000 to 2040. Initial land use classification (2000–2020) utilizes the SVM algorithm, while a novel temporal approach is applied to predict LULC for the years 2025, 2030, and 2040. The accuracy of the LULC prediction model is demonstrated to be 85.05% using the Random Forest (RF) algorithm. Notably, built-up areas increase from 11.73% (2000) to 32.96% (2040), whereas cultivated land and grassland decrease from 46.50–26.67%. The study further utilizes the SCS-CN method to model runoff for precipitation return periods of 5, 10, and 20 years, calculating Curve Number (CN) values. The results reveal variations in runoff patterns across different LULC classes and time periods. Higher return periods are associated with expanded runoff areas, with built-up areas contributing to runoff, while forests mitigate it. The study identifies that land factors, such as interception and permeability, exhibit limited influence during intense rainfall events, primarily due to capacity and saturation constraints. These findings have important implications for water resource management and strategies related to flood risk mitigation, benefiting governmental officials, planners, environmental experts, and hydraulic engineers. It's worth noting that a case study in Algeria was selected for its data availability.