# Optimizing spatial equity of urban park cooling services: Integrating landscape metrics with K-means and PSO algorithms in Nanchang, China

## TL;DR

This study improves the fair distribution of urban park cooling services in Nanchang, China, using landscape metrics and optimization algorithms to address heat disparities.

## Contribution

A novel framework combining K-means and PSO algorithms with landscape metrics to optimize urban park cooling equity.

## Key findings

- 71 UPGS showed significant cooling effects, with optimal thresholds at 60 hm2 area and 3 km perimeter.
- Water coverage strongly correlates with lower land surface temperatures (R2 = 0.4284).
- Only 71.2% of residents can access cooling services within a 15-minute walk, with suburban areas lagging.

## Abstract

Urban parks and green spaces (UPGS) provide critical cooling services to mitigate urban heat islands, yet their equitable distribution remains poorly addressed. This study integrated landscape metrics with spatial optimization algorithms to quantify and enhance the cooling equity of UPGS in Nanchang, China—a city experiencing severe heat stress. Using Landsat 8TIRS data (2021), we analyzed 85 UPGS to extract cooling indicators (LST, PCI, PCA, PCG) and correlated them with landscape composition (area, perimeter, impervious/green/water coverage) and pattern indices (PD, LPI, etc.). Network analysis based on road networks and 3,024 settlements evaluated accessibility to cooling ranges. Results showed 71 UPGS exhibited significant cooling effects (P < 0.05), with optimal thresholds at 60 hm2 area and 3 km perimeter. Water coverage was most strongly associated with lower LST (R2 = 0.4284), while complex green patch morphology extended cooling distance. Crucially, only 71.2% of residents could access cooling services within a 15-min walk, revealing severe suburban disparities (e.g., 59.1% coverage outside Second Ring Road vs. > 73% intra-city). To address gaps, we combined K-means clustering (identifying 18 optimal UPGS additions) and Particle Swarm Optimization (locating placements prioritizing suburban demand). This framework bridges micro-scale UPGS design (e.g., maximizing water bodies) and macro-scale algorithmic spatial planning, offering actionable strategies for thermally equitable cities.