Multi-scale assessment of heat stress indices in the Lyon metropolis using geo-environmental datasets and clustering analysis
Abstract
Heat stress represents a major impact of extreme heat events, affecting human health through heat-related illnesses, reduced productivity, and increased mortality. In Europe, the rising frequency and intensity of heatwaves are driven by both natural climate variability and anthropogenic climate change, with rapid urbanization further amplifying local temperature extremes through urban heat islands (UHIs). The Lyon metropolitan region in France has experienced several severe heatwaves since the 2000s, leading to significant socio-environmental and health impacts. This study investigates the spatiotemporal variability of heat stress in Lyon during the summer months (June–August) over 2000–2023 (with station-based trend analysis extended to 1950–2023 and MODIS-based surface analysis covering 2003–2023) by integrating ground-based observations, ERA5 reanalysis, and satellite-derived land surface temperature (LST) datasets. Three complementary indices are analyzed: (1) Steadman’s Heat Index (HI) from in-situ temperature and humidity, (2) the Universal Thermal Climate Index (UTCI) from ERA5, and (3) Surface Urban Heat Index (SUHI) intensity derived from MODIS LST using single-channel algorithms with NDVI-based emissivity corrections. Multi-temporal satellite imagery was fused to enhance spatial coverage and consistency. In addition, K-means clustering of temperature–humidity conditions was applied to identify dominant summer thermal regimes. Comparative analyses reveal distinct spatial and temporal patterns across indices. Summer maximum temperatures increased at 0.054 °C yr −1 in station observations and 0.052 °C yr −1 in ERA5 over 1950–2023, a cumulative warming of approximately 4.0 °C across the 74-year record, and all major heatwaves except 2003 occurred after 2015. Standardized annual HI and UTCI agree closely ( 𝑟 = 0 . 9 7 ) and rank the same summers highest (2003, 2019, 2022, 2023), whereas SUHI varies inversely with atmospheric heat ( 𝑟 = − 0 . 7 7 against HI), representing a complementary surface dimension of urban heat. Mean daytime SUHI is 3.98 °C and is robust to the rural reference: an elevation-matched reference selected by land-cover and vegetation criteria yields 3.99 °C. A distributed lag non-linear model fitted to daily age- and sex-stratified mortality gives a cumulative relative risk over lags 0–10 of 1.75 (95% CI 1.56–1.97) for the 99th versus the 50th percentile of observed 𝑇 m a x , 1.79 (1.59–2.03) for UTCI and 1.14 (1.03–1.25) for SUHI. HI and UTCI capture atmospheric and physiological heat stress, while SUHI highlights surface-driven amplification in dense urban areas. Extreme heat events and higher potential heat-health risk are concentrated in peak summer months, with urban zones consistently exhibiting higher thermal stress than surrounding rural areas. Summer age- and sex-standardized mortality peaked during the 2003 heatwave and corresponds with HI- and UTCI-identified extreme years, although the aggregated seasonal indicator precludes causal attribution. These findings demonstrate the value of multi-source, multi-scale heat indicators for assessing urban heat exposure and its health implications. The results emphasize the importance of high-resolution, multidimensional assessments to guide climate adaptation, urban planning, and heat-mitigation strategies in European cities such as Lyon.
Related
- Where should the null live? Surrogate tests for clustering of threshold-derived air-quality indices, with application to PM2.5 exceedances in Chile
- Comprehensive Global Assessment of 24 Gridded Precipitation Datasets Across 18 428 Catchments Using Hydrological Modeling
- Assessment of anthropogenic impacts on river water quality in Palopo City Indonesia using physicochemical parameters and pollution indices
- Spatio-temporal assessment of mining-induced environmental changes using remote sensing: a case study of the Talcher Coalfield, Angul District, India
- Energy, environmental, and economic assessment of sustainable fuels from wet biomass via hydrothermal, sorbent-enhanced gasification, and Fischer-Tropsch route
BibTeX
@article{geesupalli2026multiscale,
title = {{Multi-scale assessment of heat stress indices in the Lyon metropolis using geo-environmental datasets and clustering analysis}},
author = {Purna Durga Geesupalli and Catherine Pothier and Aman Arora},
journal = {Remote Sensing Applications: Society and Environment},
year = {2026},
month = oct,
doi = {10.1016/j.rsase.2026.102246},
url = {https://hal.science/hal-05781815v1},
note = {CC BY},
}
RIS
TY - JOUR
TI - Multi-scale assessment of heat stress indices in the Lyon metropolis using geo-environmental datasets and clustering analysis
AU - Purna Durga Geesupalli
AU - Catherine Pothier
AU - Aman Arora
PY - 2026
DA - 2026/10/06
JO - Remote Sensing Applications: Society and Environment
DO - 10.1016/j.rsase.2026.102246
UR - https://hal.science/hal-05781815v1
AB - Heat stress represents a major impact of extreme heat events, affecting human health through heat-related illnesses, reduced productivity, and increased mortality. In Europe, the rising frequency and intensity of heatwaves are driven by both natural climate variability and anthropogenic climate change, with rapid urbanization further amplifying local temperature extremes through urban heat islands (UHIs). The Lyon metropolitan region in France has experienced several severe heatwaves since the 2000s, leading to significant socio-environmental and health impacts. This study investigates the spatiotemporal variability of heat stress in Lyon during the summer months (June–August) over 2000–2023 (with station-based trend analysis extended to 1950–2023 and MODIS-based surface analysis covering 2003–2023) by integrating ground-based observations, ERA5 reanalysis, and satellite-derived land surface temperature (LST) datasets. Three complementary indices are analyzed: (1) Steadman’s Heat Index (HI) from in-situ temperature and humidity, (2) the Universal Thermal Climate Index (UTCI) from ERA5, and (3) Surface Urban Heat Index (SUHI) intensity derived from MODIS LST using single-channel algorithms with NDVI-based emissivity corrections. Multi-temporal satellite imagery was fused to enhance spatial coverage and consistency. In addition, K-means clustering of temperature–humidity conditions was applied to identify dominant summer thermal regimes. Comparative analyses reveal distinct spatial and temporal patterns across indices. Summer maximum temperatures increased at 0.054 °C yr −1 in station observations and 0.052 °C yr −1 in ERA5 over 1950–2023, a cumulative warming of approximately 4.0 °C across the 74-year record, and all major heatwaves except 2003 occurred after 2015. Standardized annual HI and UTCI agree closely ( 𝑟 = 0 . 9 7 ) and rank the same summers highest (2003, 2019, 2022, 2023), whereas SUHI varies inversely with atmospheric heat ( 𝑟 = − 0 . 7 7 against HI), representing a complementary surface dimension of urban heat. Mean daytime SUHI is 3.98 °C and is robust to the rural reference: an elevation-matched reference selected by land-cover and vegetation criteria yields 3.99 °C. A distributed lag non-linear model fitted to daily age- and sex-stratified mortality gives a cumulative relative risk over lags 0–10 of 1.75 (95% CI 1.56–1.97) for the 99th versus the 50th percentile of observed 𝑇 m a x , 1.79 (1.59–2.03) for UTCI and 1.14 (1.03–1.25) for SUHI. HI and UTCI capture atmospheric and physiological heat stress, while SUHI highlights surface-driven amplification in dense urban areas. Extreme heat events and higher potential heat-health risk are concentrated in peak summer months, with urban zones consistently exhibiting higher thermal stress than surrounding rural areas. Summer age- and sex-standardized mortality peaked during the 2003 heatwave and corresponds with HI- and UTCI-identified extreme years, although the aggregated seasonal indicator precludes causal attribution. These findings demonstrate the value of multi-source, multi-scale heat indicators for assessing urban heat exposure and its health implications. The results emphasize the importance of high-resolution, multidimensional assessments to guide climate adaptation, urban planning, and heat-mitigation strategies in European cities such as Lyon.
SN - 2352-9385
ER -