Mapping Urban Sprawl and Land Use Dynamics in Central China Through Random Forest-Based Remote Sensing: A Comparative Analysis of Luoyang and Zhengzhou

This study presents a comparative spatiotemporal analysis of land use and land cover (LULC) changes from 1990 to 2024 in two fast-growing Chinese cities, Luoyang and Zhengzhou. Based on multi-temporal satellite imagery from Landsat-7, Random Forest classification was conducted on the Google Earth Engine (GEE) platform. Classification accuracies between 85 and 91% were achieved through high-resolution visual interpretation and confusion matrix validation. Our findings show opposite trends in the growth of the cities. Driven by aggressive infrastructure and economic zoning, Zhengzhou saw a spectacular 223% rise in built-up area, mostly at the cost of cropland, which fell by more than 22%. As Luoyang also grew, it showed more gradual change, with a 12% decline in cropland and a 222% increase in built-up land, indicating a more restrained urban planning process. The two cities also showed significant growth of surface water by more than 100%, which means that hydrological infrastructure or urban planning priorities have changed. The forest cover was quite intact, and there were slight improvements in Zhengzhou, implying increased concern about greening urban areas. Meanwhile, land degradation or stagnant building sites emerge as a result of the emergence of barren land, especially in Zhengzhou. These results highlight the variation in urban development within the same area, which is shaped by geography, policy, and planning culture.