Improving Weather Predictions in Rapidly Urbanizing Regions Using the AI-Powered Normalized Difference Urban Index Plus (NDUI+) Dataset

The rapid pace of urbanization worldwide has significantly altered local and global weather-climate interactions. Expanding urban areas modify land-atmosphere interactions, leading to changes in temperature, precipitation patterns, and extreme weather events. To better understand and mitigate urban-induced weather and climate effects, long-term, frequently updated urban datasets are essential. Such datasets enable accurate monitoring of urban expansion and its impact on atmospheric processes, ultimately improving weather forecasting capabilities. The Weather Research and Forecasting (WRF) model is a widely used numerical weather prediction tool, yet its urban representation remains constrained by the limited availability of continuous, high-resolution urban data. The accuracy of weather forecasts, particularly in and around urban areas, is dependent on how well the model represents urban land cover and surface characteristics. In this study, the authors present Normalized Difference Urban Index+ (NDUI+) dataset, a 30-meter, long-term, continuously updated urban dataset designed to enhance urban representation within WRF. This dataset uses AI-calibrated DMSP-VIIRS nighttime light images merged with Landsat NDVI (Normalized Difference Vegetation Index) to generate Normalized Difference Urban Index (NDUI) metric from 1999 to present. The integration of NDUI+ into WRF improves the characterization of urban areas, leading to more precise simulations of weather conditions. Results demonstrate that the enhanced urban representation significantly refines key meteorological variables such as temperature, humidity, and wind speed, yielding more reliable and accurate forecasts. The improved model performance underscores the necessity of incorporating high-resolution, frequently updated urban datasets to advance weather prediction capabilities, especially in rapidly urbanizing regions. By bridging the gap between urban data availability and numerical weather modeling, this study highlights the critical role of urban datasets in improving the accuracy of weather forecasts and understanding micro and meso-level urban-climate interactions.