Data and code for: "Rethinking the Urban Heat Island: Local Surface Structure and a City-Scale Background Signal"

Data and analysis code accompanying the manuscript "Rethinking the Urban Heat Island: Local Surface Structure and a Weak City-Scale Background Signal" by Hauschopp et al. (Leibniz University Hannover, Institute of Meteorology and Climatology; manuscript in preparation). The study analyses one year (November 2024 to October 2025) of 10-min near-surface air temperature observations from seven sites in Germany — the urban–rural reference station pairs of Berlin and Hannover (DWD) and three low-cost stations of the LUMOS network in villages of the rural Wedemark municipality — together with a 1-m-resolution analysis of the surface structure around each site. CONTENTS data/stations: Station air temperature data (NetCDF, 10-min, UTC): DWD stations Berlin-Alexanderplatz (399, incl. the long-term record 1992–2025), Berlin Brandenburg (427), Hannover-Langenhagen (2014), Hannover-Nordstadt/Weidendamm (15811), the three LUMOS village stations (Mellendorf, Elze, Abbensen), and the sensor-validation pair Herrenhausen (LUMOS vs. aspirated research-grade reference, Dec 2025–Jul 2026). data/derived: Surface-class fractions and mean obstacle height as a function of scan radius (1–3000 m) for all sites; Delta-T statistics including block-bootstrap confidence intervals. data/surface_types: Surface classification rasters (1 m GeoTIFF; classes: high vegetation, low vegetation, built-up, bare ground, water) for the Hannover region (EPSG:25832) and the two Berlin sites (EPSG:25833). data/accuracy: Accuracy assessment of the classification following Olofsson et al. (2014): 600 labelled reference points, stratum weights, and full results (confusion matrices, area-weighted accuracies). code: Python scripts for all figures and statistics (code/figures, run_all.sh), the geoprocessing workflows including download scripts for the open geodata (code/geoprocessing), the accuracy-assessment tooling (code/accuracy), and the sensor-validation processing (code/validation). See README.md in the archive for the full structure, the reproduction chain, and software requirements. LICENSING AND ATTRIBUTION Own data (LUMOS station data, derived products, classification rasters, accuracy reference data) are provided under CC BY 4.0; the code is provided under the MIT License (see README.md). DWD station data are redistributed in processed form under Datenlizenz Deutschland – Namensnennung – 2.0 (source: Deutscher Wetterdienst, Climate Data Center, opendata.dwd.de). The geospatial input data (orthophotos, elevation models, land cover, ALKIS) are open data of LGLN Lower Saxony, Geoportal Berlin, LGB Brandenburg, and BKG and are not included; they can be retrieved with the included download scripts.