Derived Datasets and Code for Seasonal Net Surface Shortwave Radiation and 3D Urban Morphology Analysis in Beijing (2023)

# Derived Datasets and Code for Seasonal Net Surface Shortwave Radiation and 3D Urban Morphology Analysis in Beijing (2023) ## Description This dataset supports the paper: **Investigating the impact of two-dimensional and three-dimensional urban structures on seasonal net surface shortwave radiation** by Yang Liu, Shisong Cao*, Jingyi Chen, Huiping Sun, Ling Li, and Nuan Wang (submitted to *GIScience & Remote Sensing*) The dataset contains derived 30 m resolution grids and tabular data for Beijing within the Fifth Ring Road in 2023. ## Contents - **3D_USPs.gdb**: File Geodatabase containing eight 3D urban structural parameters at 30 m resolution: - MBH (Mean Building Height) - MBV (Mean Building Volume) - SVF (Sky View Factor) - FAR (Floor Area Ratio) - SCD (Spatial Crowdedness Degree) - BSA (Building Surface Area) - BU (Building Uniformity) - BSI (Building Structural Index) - **NSSR_0116.gdb**, **NSSR_0305.gdb**, **NSSR_0828.gdb**, **NSSR_1124.gdb**: Net Surface Shortwave Radiation (NSSR, unit: W/m²) grids for four representative dates in winter, spring, summer, and autumn 2023. - **Landcover_2D.gdb** (or raster file): Two-dimensional land cover classification data for the study area (derived from Dynamic World or similar). - **Albedo.gdb** (or raster file): Surface albedo grids used in the NSSR retrieval process. - **BJ_DSR_CSV.csv** (or .xlsx): Tabular data combining seasonal NSSR values with corresponding 3D urban structural parameters for machine learning analysis. - **MODISpoint.xlsx**: Validation data extracted from MODIS products and ground observations. - **Python scripts** (*.py): Key codes for machine learning models (XGBoost, LightGBM, Random Forest, etc.) and SHAP interpretability analysis. ## Software Requirements - GIS software: ArcGIS or QGIS (to open .gdb files) - Python 3.x with libraries: pandas, numpy, xgboost, lightgbm, scikit-learn, shap, matplotlib, etc. ## Usage Example ```python import pandas as pd df = pd.read_csv('BJ_DSR_CSV.csv') print(df.head())