A neural network-based Universal Thermal Climate Index for reliable global thermal-stress classification across extreme weather
Abstract
Extreme temperatures are the leading cause of climate-related mortality world-wide. Climate-health research and operational weather forecasting require accurate estimates of human thermal stress. The Universal Thermal Climate Index (UTCI) is among the most sophisticated and widely used feels-like temperature metrics. However, its ubiquitous polynomial approximation does not generalize well to extreme weather conditions. Here, we introduce Neural-UTCI, a neural network that calculates UTCI with substantially higher accuracy across global conditions at a lower computational cost for operational use. Neural-UTCI reduces the polynomial approximation RMSE from 2.78°C to 0.36 °C, an 87% improvement, and lowers thermal stress misclassification rates from 5.3% to 1.7%, with consistent performance across resampling experiments. These differences affect thermal exposure metrics. For example, during the 2003 European heatwave summer in Rome, Italy, the number of very strong heat stress days increases from 15 to 35 days when using Neural-UTCI compared to operational products like ERA5-HEAT. Simultaneously, Neural-UTCI reliably classifies extreme cold stress conditions, allowing continuous global application. By improving UTCI accuracy, Neural-UTCI can strengthen climate-health risk assessments and public weather warning systems, especially as global warming increases the incidence of extreme events.
Related
- No Universal Urban Form Heat Relationship: Contrasting Thermal Controls in African Cities
- Thirty-five years of urban expansion and thermal transformation: Integrating LULC change climate variability and population growth
- Numerical study of a Saharan dust plume impact on radiation, urban climate and thermal comfort in the Paris region during an heatwave event
- Evaluate the thermal performance of green roofs as an urban heat island mitigation strategy under arid climate conditions in Bahrain
- Predicting hourly thermal demand of buildings for digital twin-based energy network modeling using a steady-state model capturing thermal inertia
BibTeX
@misc{pastine2026neural,
title = {{A neural network-based Universal Thermal Climate Index for reliable global thermal-stress classification across extreme weather}},
author = {Bikem Pastine and Milan Klöwer and Tianning Tang and Sarah Wilson Kemsley and Louise Slater},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.35949},
eprint = {2609.35949},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.35949},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - A neural network-based Universal Thermal Climate Index for reliable global thermal-stress classification across extreme weather
AU - Bikem Pastine
AU - Milan Klöwer
AU - Tianning Tang
AU - Sarah Wilson Kemsley
AU - Louise Slater
PY - 2026
DA - 2026/09/28
PB - arXiv
DO - 10.48550/arxiv.2609.35949
UR - https://arxiv.org/abs/2609.35949
AB - Extreme temperatures are the leading cause of climate-related mortality world-wide. Climate-health research and operational weather forecasting require accurate estimates of human thermal stress. The Universal Thermal Climate Index (UTCI) is among the most sophisticated and widely used feels-like temperature metrics. However, its ubiquitous polynomial approximation does not generalize well to extreme weather conditions. Here, we introduce Neural-UTCI, a neural network that calculates UTCI with substantially higher accuracy across global conditions at a lower computational cost for operational use. Neural-UTCI reduces the polynomial approximation RMSE from 2.78°C to 0.36 °C, an 87% improvement, and lowers thermal stress misclassification rates from 5.3% to 1.7%, with consistent performance across resampling experiments. These differences affect thermal exposure metrics. For example, during the 2003 European heatwave summer in Rome, Italy, the number of very strong heat stress days increases from 15 to 35 days when using Neural-UTCI compared to operational products like ERA5-HEAT. Simultaneously, Neural-UTCI reliably classifies extreme cold stress conditions, allowing continuous global application. By improving UTCI accuracy, Neural-UTCI can strengthen climate-health risk assessments and public weather warning systems, especially as global warming increases the incidence of extreme events.
ER -