Grid-level Dataset of 3D Urban Configuration and Land Surface Temperature in Hefei, China

This dataset contains the processed, grid-level geospatial data used in the research article "Geospatial analytics of 3D urban configuration: Decoupling solar exposure and aerodynamic drag on land surface temperature," submitted to GIScience & Remote Sensing. It provides a comprehensive multi-source spatial database designed to investigate the complex thermodynamic interactions within the vertical urban space of Hefei, China.

The data is structured as a cleaned fishnet grid, integrating satellite-derived Land Surface Temperature (LST) with innovative 3D morphological indicators, specifically the Building Effective Solar Area (BESA) and Aerodynamic Frontal Area Density (AFD). Alongside these core metrics, the dataset includes various control variables such as Sky View Factor (SVF), Albedo, Nighttime Light (NTL), population density, and 2D landscape indices. All variables have undergone rigorous data cleaning and multicollinearity checks to ensure high quality and reliability for geospatial modeling.

This dataset was specifically prepared to support an interpretable machine learning framework (XGBoost-SHAP) to decouple the independent and non-linear impacts of solar radiation capture and aerodynamic drag on the urban thermal environment. By making this data publicly available, we aim to comply with the Open and FAIR data policies, allowing researchers to reproduce our findings, validate the identified non-linear thermal thresholds, and further explore the V-shaped reversal effect of 3D aerodynamic drag in climate-resilient urban planning.