Answers are grounded in this paper and cite their sources.
Aim The study quantifies when and where urban parks lower the radiative heat load of their surroundings, how this depends on climate and season, and which park and neighbourhood attributes control it. The response variable is seasonal Surface Upward Longwave Radiation (SULR, W m⁻²) at 30 m. 2. Study area and sample The analysis covers 1,208 OpenStreetMap parks (≥ 1 ha, green in summer, fully covered by thermal data) in the 31 provincial capitals of Iran, spanning six Köppen–Geiger classes from hot desert to humid subtropical. 3. Input data • Thermal and vegetation: seasonal Landsat 8/9 SULR and NDVI composites (spring, summer, fall, winter) at 30 m. • Emissivity correction: ASTER Global Emissivity Dataset v3, ERA5-Land downwelling longwave radiation at 07:00 UTC and Landsat red reflectance. • Urban form and cover: ESA WorldCover (10 m), Meta/WRI canopy height, Global Building Height Map (10 m), global Local Climate Zones. • Terrain and climate: DSM-derived slope and aspect, elevation, KGClim Köppen–Geiger classes and climate normals, ERA5 near-surface climate. • Parks: OpenStreetMap leisure=park polygons. 4. Analyses All layers were co-registered to the 30 m SULR grid. For each park and season, park cooling intensity (PCI) was computed as the SULR difference between a 30–300 m ring and the park, together with sector PCI and ring profiles to 600 m. More than 100 predictors described park geometry, composition, canopy, tree configuration, greenness, terrain and solar illumination, neighbourhood land cover and building volume, and city climate. PCI was analysed with class × season matrices, variance partitioning, gradient boosting with SHAP, linear mixed models, the optimal Geodetector and logistic models of minimum effective park area. Distance profiles were normalised with background models trained on pixels far from parks to isolate the park signal. 5. Outputs • Rasters: final SULR, emissivity and downwelling radiation, solar incidence, cast shadow and hillshade for each city and season (GeoTIFF). • Vectors: parks with seasonal PCI, SULR and cooling class; all park attributes; city summaries; Iran boundaries (Shapefile, GeoPackage). • Tables: descriptive statistics, variance partitioning, SHAP importance, mixed-model coefficients, Geodetector results, minimum effective areas, adjusted profiles and per-park metrics (CSV, XLSX). • Figures: 12 main and 7 supplementary figures (PNG, 300 dpi). 6. Code and documentation Python and Google Earth Engine scripts are organised in five stages: data acquisition, preprocessing, park features and PCI, analysis models, and figures and tables, with a run script.
No takes yet. Share an insight, caveat, or question.
Karamoozian et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: