Machine Learning Maps eThekwini Landslide Risks as Urban Growth Threatens Safety

Researchers in eThekwini, South Africa, have combined machine learning with satellite data to identify the city's most landslide-prone areas, according to a new study in Scientific Reports. The analysis found that urban development—including road construction, property expansion, and vegetation loss—is significantly increasing landslide susceptibility across parts of the municipality. By pinpointing these high-risk zones, the study offers a data-driven tool for planners and emergency services to prioritize mitigation efforts in a region where informal settlements often occupy steep, unstable slopes. The findings highlight the hidden costs of rapid urbanization, as the removal of natural vegetation and the alteration of drainage patterns can destabilize hillsides. With climate change expected to bring more intense rainfall to KwaZulu-Natal, the need for proactive land-use planning becomes even more urgent. The research underscores how machine learning can turn satellite imagery into actionable risk maps, helping local authorities build safer communities. As eThekwini continues to grow, these insights could guide zoning decisions and infrastructure investments to reduce future landslide disasters.