Can a model reveal what people feel? Coupling of thermal walks with high-fidelity urban modelling

Abstract

Urban heat is worsening liveability in cities worldwide, yet effective mitigation requires spatially precise data to target interventions at the scales where people actually experience thermal discomfort. Traditional urban climate research methods focus on measuring or modelling the thermal environment, while current research increasingly prioritises human thermal experience. We present the first systematic and quasi-real-time integration of high-resolution PALM simulations (1-meter/1-minute) with an extensive thermal walk campaign: 3,666 thermal sensation vote locations collected in Prague-Holešovice (Czech Republic) on a hot summer day. This approach bridges subjective thermal perception and objective microclimate modelling to identify heat stress patterns (thermal hotspots) at the microscale. Our analyses identified modelled mean radiant temperature (MRT) as the dominant predictor of reported thermal sensation, exhibiting the highest spatiotemporal variability. In contrast, air temperature showed minimal spatial variation (<0.3 °C within streets) and far lower predictive power. Distinct spatial patterns emerged: the northern sides of east-west-oriented streets without trees showed persistent heat stress, whereas mature street trees reduced heat stress by more than 50%, providing cooling comparable to that from building shade. Open impervious areas exhibited heat stress similar to the worst streets, while accessible green spaces just dozens of metres away offered significant thermal relief. This integrated approach identifies microscale heat adaptation priorities with unprecedented spatial precision, enabling argeted interventions where people need them most.