Neighborhood-scale thermal variability of urban heat islands: distinct roles of horizontal and vertical urban morphology
SUSTAINABLE CITIES AND SOCIETY, cilt.150, sa.2026, ss.1-22, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 150 Sayı: 2026
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.scs.2026.107936
- Dergi Adı: SUSTAINABLE CITIES AND SOCIETY
- Derginin Tarandığı İndeksler: Scopus, Science Citation Index Expanded (SCI-EXPANDED), Compendex, Geobase, INSPEC
- Sayfa Sayıları: ss.1-22
- Çanakkale Onsekiz Mart Üniversitesi Adresli: Evet
Özet
Urban Heat Island (UHI) dynamics are strongly influenced by urban land cover composition and configuration, yet the relative roles of horizontal surface sealing and vertical urban structure in shaping the neighborhood-scale daytime thermal variability remain insufficiently understood. This study examines the influence of twodimensional (2D) and three-dimensional (3D) urban morphology, together with anthropogenic indicators, on UHI intensity in Çanakkale, Türkiye. Landsat-derived Land Surface Temperature (LST) data were integrated with detailed urban morphology metrics to evaluate thermal patterns across neighborhoods. Initial bivariate analyses identified strong positive correlations between spring LST and morphological indicators, notably Gross Population Density (GPD, rs = 0.74) and Floor Area Ratio (FAR, r = 0.79). To overcome severe multicollinearity (VIF > 30) and quantify independent contributions, a dual-stage multivariate machine learning framework integrating Ridge Regression and a Random Forest (RF) algorithm with Leave-One-Out Cross-Validation was implemented. Feature importance analysis revealed a complex thermal dynamic: population density (GPD, RF importance: 0.312) and total vertical volume (FAR, 0.289) act as the primary independent drivers of general thermal load. Conversely, localized extreme heat anomalies (Maximum UHI) are uniquely driven by a combination of demographic concentration (GPD, 0.245) and horizontal surface coverage (Building Coverage Ratio (BCR), 0.198). These findings demonstrate that urban morphology controls distinct thermal processes: general surface warming is driven by volumetric density, whereas localized extreme anomalies are dictated by horizontal surface sealing and localized demographic crowding. The results underscore the necessity of multidimensional, plan-based multivariate modeling for climate-responsive urban planning.