Spatiotemporal Patterns and Random Forest Prediction of Tropospheric NO₂ in Ankara, Türkiye

This dataset contains high-resolution tropospheric NO2 column density predictions for Ankara, Turkey, for the year 2025. The data was generated using a multi-sensor data fusion approach, integrating Sentinel-5P TROPOMI atmospheric data with Sentinel-2 multispectral surface reflectance. A Random Forest (RF) regression model was employed to downscale the coarse Sentinel-5P data (5.5 km) to a high-spatial resolution (100m) grid. The methodology relies solely on satellite-derived spectral signatures, providing a scalable and cost-effective solution for urban air quality monitoring without the need for ground-based stations or auxiliary meteorological data.