Urban Heat Island Dynamics and Blue-Green Infrastructure Planning Using High-Resolution Geospatial Analytics

Abstract

This paper reviews recent literature applying high-resolution geospatial analytics, satellite remote sensing, Geographic Information Systems (GIS), and machine learning, to characterize urban heat island (UHI) dynamics and inform blue-green infrastructure (BGI) planning. Drawing on a 306-reference systematic review of BGI cooling mechanisms, a GIS-based cartographic modeling study from Olsztyn, Poland achieving one-meter vector-data precision, an ECOSTRESS-based tropical-megacity study of five parks in Kuala Lumpur, an XGBoost machine-learning study of cooling intensity in Hue City, Vietnam achieving a coefficient of determination of 0.97, and a Google Earth Engine-based land-surface-temperature study of blue-green spaces in Bhubaneswar, India, this review reports specific, quantified findings rather than general claims about BGI’s cooling benefit. Particular attention is given to the named local climate zone (LCZ) classification framework used to contextualize land-surface-temperature (LST) data, the specific geospatial data sources (Landsat, ECOSTRESS, Google Earth Engine) each reviewed study employed, and the named machine-learning architectures, specifically XGBoost with SHAP interpretability, used to quantify nonlinear relationships between landscape configuration and cooling performance. Comparative tables map named geospatial platforms and sensors to the specific UHI or BGI research question each addressed, cross-reference BGI types (bioswales, infiltration trenches, green bus stops, urban parks, water bodies) against their documented cooling mechanisms, and set reported quantitative outcomes, coefficients of determination, precision levels, temperature reductions, against the specific study and city that produced them. The review concludes that high-resolution geospatial analytics has produced increasingly precise, city-specific BGI planning tools, but that translating these tools into standardized, cross-city planning frameworks remains constrained by local climate zone heterogeneity and inconsistent adoption of interpretable machine-learning methods.