Comparative Analysis of the Spatio-Temporal Distribution of Thermal Stress Derived from ERA5-Land Reanalysis and Observational Data in Iran

Thermal comfort plays a critical role in human well-being, productivity, and urban as well as environmental planning. This study evaluates the performance of ERA5-Land reanalysis data in detecting thermal stress compared with synoptic station observations across Iran from 1991 to 2020.

Daily records of temperature, humidity, and wind speed from both sources were used to calculate the Effective Temperature and Baker indices. Spatial and temporal patterns were analyzed seasonally, with additional comparisons for cold–dry and warm–humid climatic zones.

ERA5-Land demonstrated strong agreement with observations in reproducing the spatio-temporal distribution of thermal indices, particularly in spring and summer (correlation coefficient = 0.80, p < 0.005). In winter, both datasets indicated widespread cold stress: observations reported 79% of the country under mild to moderate cold stress, compared with about 95% in ERA5-Land. In the cold–dry zone, observed “very cold” days declined from around 100 to fewer than 10, while “no stress” days rose from 50 to more than 150. ERA5-Land, however, maintained nearly constant values (80–100 “very cold” days). During summer, observations recorded over 62% and ERA5-Land about 43% of Iran under mild to moderate heat stress. In the warm–humid zone, observed “very hot” days rose from fewer than 50 in 2000 to over 350, whereas ERA5-Land reflected only about 150 days. Overall, ERA5-Land reasonably captures broad thermal stress patterns but tends to underestimate extreme local climatic changes.

ERA5-Land reasonably represents general thermal stress patterns but underestimates extreme local climatic changes.



Extended Abstract

Introduction

understanding thermal conditions is essential for promoting human health, productivity, and overall well-being, as well as for guiding urban and environmental planning. These conditions, driven by temperature, humidity, wind, and solar radiation, can generate both benefits, such as supporting economic and social activities, and adverse outcomes, including heat- and cold-related illnesses and mortality. To assess thermal comfort and stress, numerous bioclimatic indices have been developed. Thier accuracy, however, depends heavily on the quality and resolution of input data.

In recent years, reanalysis datasets such as ERA5-Land have been widely applied in climate research due to their extensive spatial and temporal coverage. These datasets, produced by combining numerical modeling with meteorological observations, provide a consistent depiction of atmospheric conditions. Nevertheless, their reliability at local scales and across diverse climatic zones requires validation. By contrast, observational data from synoptic stations—despite limitations such as incomplete spatial coverage and occasional missing values, generally provide more accurate local-scale measurements. They thus serve as an essential benchmark for evaluating reanalysis datasets.

Previous studies in Iran have largely relied on observational data, demonstrating the strong influence of thermal stress on energy demand, building design, and human comfort. Findings also indicate a significant rise in heat stress in recent decades, particularly in lowland areas. Nevertheless, no comprehensive study has directly compared thermal indices derived from reanalysis wirh those based on station observations in Iran.

The present study addresses this gap by evaluating the performance ofERA5-Land data in calculating thermal indices and comparing results with observational records across Iran’s diverse climates. This comparison is particularly relevant for applications in public health, energy management, and environmental planning.

Methodology

Two datasets were employed: ERA5-Land reanalysis data and synoptic meteorological observations from of the Iran Meteorological Organization (IRIMO) covering 1991–2020. ERA5-Land, developed by the European Centre for Medium-Range Weather Forecasts (ECMWF), provides hourly data with a of 0.1° spatial resolution. Observational data consisted of monthly mean air temperature, wind speed, and relative humidity from 130 synoptic stations with complete records. Two thermal comfort indices Net Effective Temperature (NET) and the Baker Index (CPI)—were selected, as both are computable from reanalysis variables. These indices were calculated on an hourly basis for both datasets using Python. ERA5-Land values were extracted for grid points nearest to the station coordinates and for specific times of day (00, 03, 06, 09, 12, 15, 18, and 21 UTC). Seasonal maps of the indices were generated using Inverse Distance Weighting (IDW) interpolation to illustrate spatial patterns. In addition, mean values of the indices were compared across two representative climatic regions: a cold–arid zone in northwestern Iran and a hot–humid zone along the southern coasts. The reliability of ERA5-Land was further assessed by computing correlation coefficients, coefficients of determination (R²), and statistical significance levels (p-values).



Results and discussion

Both observational and ERA5-Land datasets displayed consistent spatial and seasonal patterns, although ERA5-Land generally produced slightly cooler values. During winter, most of Iran was affected by cold stress, with near-comfort conditions confined to southern coastal areas. In summer, central, southern, and eastern regions experienced widespread heat stress. ERA5-Land performed better in warmer seasons (spring and summer), showing stronger correlations with station-based data, while accuracy declined during autumn and especially winter.

Spatial analyses indicated that ERA5-Land underestimates topographic effects in mountainous areas, often reporting lower-than-observed temperatures. Climatic comparisons revealed that observational data more effectively captured warming trends, particularly the increase in extremely hot days in the hot–humid south after 2000. In contrast, ERA5-Land underestimated this rise.

Statistical evaluations confirmed these findings: correlations were highest in spring and summer (r ≈ 0.8), but weaker in autumn and winter (r ≈ 0.5–0.7). Overall, ERA5-Land provides a reliable source for assessing thermal comfort during warmer periods but requires calibration and integration with station observations for colder seasons and regions with complex topography.



Conclusion

This study demonstrates that ERA5-Land data can reliably reproduce large-scale patterns of thermal comfort across Iran during 1990–2020, with the strongest agreement in spring and summer. However, ERA5-Land tends to overestimate cold stress in winter and underestimate heat stress in summer. Discrepancies were most pronounced in mountainous areas and the hot–humid southern coastal region. Long-term analyses further revealed that observational data more accurately capture local warming signals, such as the marked increase in extremely hot days after 2013, which ERA5-Land failed to reflect.

In summary, ERA5-Land is a valuable resource for identifying general thermal comfort distributions and large-scale climate conditions but has limitations in representing localized variations and long-term warming trends. Integrating reanalysis datasets with ground-based observations is therefore crucial for applications in urban planning, energy management, and climate change adaptation.