Abstract
Urban heat island effects and climate change are worsening outdoor thermal environments, particularly for groups required to wear personal protective equipment (PPE) during outdoor work, thus increasing their heat-stress risk. A summer field study was conducted in a hot-humid region to assess individuals’ thermal and physiological responses under two clothing conditions (PPE and ordinary clothing) and four cooling scenarios, along with concurrent measurements of thermal environmental parameters. Machine learning models were developed and interpreted with Shapley Additive Explanations (SHAP) to quantify the influence of microclimatic, physiological and personal variables on thermal sensation. Among the tested algorithms, CatBoost yielded the highest prediction accuracy, followed by XGBoost. For PPE wearers, mean skin temperature predominates at neutral thermal sensation levels (SHAP = 0.900), whereas BMI and air temperature become the main drivers at higher thermal sensation levels. Under ordinary clothing, BMI remains the most influential factor across all levels. Across cooling scenarios, BMI shows the largest overall impact, with its SHAP value peaking (0.259) at the highest thermal sensation level under the shading-only scenario, indicating higher heat-stress risk for individuals with higher BMI. Under the misting + shading scenario, gender is a key factor at neutral thermal sensation levels, whereas high humidity becomes the primary driver at the highest thermal sensation levels when forced convection is absent. Therefore, effective outdoor cooling strategies should consider individual characteristics and prioritize composite approaches combining shading, fans, and misting to enhance both heat mitigation effectiveness and equity. © 2026 Published by Elsevier B.V.
| Original language | English |
|---|---|
| Article number | 116997 |
| Number of pages | 22 |
| Journal | Energy and Buildings |
| Volume | 354 |
| Online published | 12 Jan 2026 |
| DOIs | |
| Publication status | Published - 1 Mar 2026 |
Funding
This work was supported by the National Natural Science Foundation of China (Project No. 52278097 ). The authors express their gratitude to all the subjects who participated in the survey.
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 11 Sustainable Cities and Communities
-
SDG 13 Climate Action
Research Keywords
- Machine learning
- Outdoor thermal comfort
- Personal protective equipment
- Physiological indicators
- Thermal mitigation strategies
Fingerprint
Dive into the research topics of 'Determinants of outdoor thermal sensation across cooling scenarios in a hot-humid region: A SHAP-based analysis'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver