Assessment of Human Thermal Stress Dynamics and Regional Drivers Across the Arabian Peninsula During 1990–2022 From Observations

Human thermal stress is intensifying in the Arabian Peninsula (AP); however, a lack of consensus in observational evidence continues to limit a quantitative understanding of trends and the relative influence of the meteorological drivers underlying regional variability. In this study, we used 33 years (1990–2022) of in situ meteorological observations from 75 stations across the AP to assess the variability and trend of human thermal stress, expressed through the Universal Thermal Climate Index (UTCI), together with an information flow analysis quantifying the directional information transfer to UTCI from temperature, mean radiant temperature (MRT), relative humidity, and wind speed. Because the station archive reports daily summaries, all quantities are evaluated on a daily‐mean basis; the analysis therefore characterises the climatology of daily‐mean thermal conditions and does not resolve peak sub‐daily exposure or operational heat hazard. Results reveal systematic human thermal stress intensification across all seasons, with summer (June–August) exhibiting the strongest significant regional‐mean warming (0.64°C decade −1 ), followed by spring (0.46°C decade −1 ) and autumn (0.36°C decade −1 ). The number of days on which daily‐mean UTCI (UTCIdm) exceeds the strong heat stress threshold of 32°C increased by approximately 5 days decade −1 , with the intensification concentrated in the shoulder seasons (spring and autumn). Information flow analysis indicates contrasting patterns among the four station groups: temperature carries the largest resolved flow to UTCI in 11 of the 16 group–season combinations, a modest positive wind flow is resolved in 13, and the MRT contributes less than temperature everywhere and changes sign between groups. The findings reveal a region‐specific mitigation strategy due to heterogeneity in the forcing across regions. This approach offers a template for observation‐based diagnosis of thermal stress in data‐sparse arid regions.