# Cooling effect of urban green spaces: LCZ-based assessment comparing four cities at similar latitudes via hotspot and regression models

## TL;DR

This study compares how urban green spaces reduce heat in four cities with similar latitudes, showing that cooling effects depend on local urban structure and green space placement.

## Contribution

The study introduces a context-sensitive framework using LCZ and spatial analysis to assess and compare urban green space cooling effects across cities.

## Key findings

- Green areas significantly lower land surface temperatures, with the strongest cooling within 400 m of parks.
- Cooling magnitudes vary across local climate zones and urban contexts, with up to 4.2°C reduction in compact residential zones.
- Heat clusters are concentrated in dense built-up zones, while cold spots are in natural areas, guiding green infrastructure priorities.

## Abstract

Urban green spaces (UGSs) are widely recognized as a key strategy to mitigate urban heat islands (UHIs) through evapotranspiration and shading. However, their cooling effectiveness depends not only on size but also on spatial configuration and surrounding urban morphology. This study provides a cross-latitude, local climate zone (LCZ)-based comparative analysis of UGS cooling effects in four medium-sized cities located at similar latitudes but with distinct morphological patterns: Salt Lake City (USA), L’Aquila (Italy), Krakow (Poland), and Kastamonu (Türkiye). Using MODIS-derived land surface temperature (LST), LCZ classification, and Getis-Ord (Gi*) hotspot analysis, we quantified spatial clustering of heat. We identified where parks and natural landscapes are most effective in lowering temperatures. Furthermore, Ordinary Least Squares (OLS) regression was used to establish global relationships between green space coverage and LST, followed by Geographically Weighted Regression (GWR) to capture local variations and spatial heterogeneity in this relationship. Results reveal that green areas significantly lower LST, with the strongest cooling observed within 400 m of parks. Rather than treating this buffer distance as a universal planning threshold, the analysis demonstrates that cooling magnitudes and spatial consistency vary systematically across LCZ types and urban contexts. In Salt Lake City, for instance, UGSs reduced temperatures by up to 4.2 °C in compact residential zones (LCZs 2–3). Hot spot analysis further shows that statistically significant heat clusters (95–99% confidence) are concentrated in dense, built-up LCZ types, while cold spots are predominantly associated with natural LCZ classes (A–D). Together, these findings suggest that urban morphology not only influences the presence but also the effectiveness of cooling buffers around green spaces. By integrating LCZ mapping, spatial statistics, and a comparative, cross-latitude design, this study moves beyond distance-based generalisations. It advances urban climate research by demonstrating a transferable yet context-sensitive framework for diagnosing UHI intensity and prioritising green space interventions. The results provide actionable guidance for planners: park siting and design should prioritize heat-prone, high-density zones to maximize cooling benefits and support climate-resilient urban development.

The online version contains supplementary material available at 10.1007/s00484-026-03172-x.

The spatial distribution of the cooling effects of UGS on four cities was determined.

The relationship between UGS and LST was evaluated comparatively using LCZ.

The cooling effects of UGS were compared among LCZs.

Getis-Ord Gi* analysis revealed significant heat/cool clusters, indicating priority zones for green infrastructure.

The strongest cooling was observed within 400 m of parks, particularly in compact, dense LCZs.

The online version contains supplementary material available at 10.1007/s00484-026-03172-x.