Advertisement

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Diurnal asymmetry in heat stress intensification across Bangladesh, 1985–2024: Accelerated nighttime warming and emerging urban risk

Abstract

Bangladesh’s rapidly growing cities are becoming hotter, but how heat stress is changing over the day–night cycle has remained unclear. Using 40 years (1985–2024) of hourly Universal Thermal Climate Index (UTCI) data from ERA5-HEAT, we examined long-term changes in physiologically relevant heat stress across Bangladesh and its major cities. Results show a clear day–night imbalance in warming: nighttime heat stress (UTCIₘᵢₙ) is rising faster than daytime extremes. National trends indicate increases of +0.03 °C per decade for UTCIₘᵢₙ, +0.02 °C for daily mean UTCI, and +0.01 °C for UTCIₘₐₓ, with the strongest warming occurring in the early morning hours. These national spatially averaged trends reflect the mean across all 194 grid cells; individual grid cells show local trends of +0.1 to +0.4°C per decade, and the cumulative nighttime warming over the full 40-year period reaches approximately 1.0–1.7°C across most of the country. This signals a steady loss of nighttime cooling that people rely on for physical recovery. The most pronounced nighttime warming occurs in western and southern Bangladesh. Major cities—including Dhaka, Rajshahi, Khulna, Chattogram, and Sylhet—show additional intensification consistent with, but not directly attributed to, urban heat-island dynamics at the spatial scale of this analysis. The number of very strong heat stress days (UTCI > 38 °C) has increased by 4–15 days per decade, and cities such as Rajshahi and Dhaka now experience more than 150 such days annually. Together, these findings indicate a transition from occasional heat extremes to persistent, 24-hour heat stress, increasing risks to health, labor productivity, and urban resilience. By identifying when heat stress is rising fastest and where it is concentrated, this study provides evidence to support city-specific heat-action plans, early-warning systems, and climate-responsive urban design in rapidly warming regions.

Citation: Kamruzzaman M, Abdullah AYM, Islam AKMS, Talukder A, Wahid S, Shahid S (2026) Diurnal asymmetry in heat stress intensification across Bangladesh, 1985–2024: Accelerated nighttime warming and emerging urban risk. PLOS Clim 5(8): e0000848. https://doi.org/10.1371/journal.pclm.0000848

Editor: Lakshmi Kumar T V, Jawaharlal Nehru University, INDIA

Received: February 4, 2026; Accepted: June 12, 2026; Published: August 5, 2026

Copyright: © 2026 Kamruzzaman et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All data underlying the findings of this study are fully and openly available without restriction. The Universal Thermal Climate Index (UTCI) data were obtained from the ERA5-HEAT reanalysis dataset provided by the Copernicus Climate Change Service (C3S) Climate Data Store. These data are publicly accessible at: https://cds.climate.copernicus.eu/datasets/derived-utci-historical All processed datasets generated during the analysis (including daily minimum, mean, and maximum UTCI metrics, climatological aggregates, and city-level time series) can be reproduced directly from the ERA5-HEAT source data using the methods described in the manuscript. No third-party proprietary datasets were used. There are no legal or ethical restrictions on data sharing.

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Thermal comfort represents the subjective perception of how individuals experience temperature, humidity, and airflow, and it is a key dimension of urban liveability and sustainability. However, as global temperatures rise, urban populations increasingly face conditions that exceed not only comfort thresholds but also the limits of physiological tolerance. While thermal discomfort reflects a decline in perceived well-being, heat stress occurs when environmental conditions impair the body’s ability to dissipate heat, triggering measurable physiological strain. This distinction is critical: a population may feel uncomfortable at moderate temperatures, but health risks escalate sharply when conditions surpass biophysical thresholds for heat dissipation, particularly in humid, densely built environments [1–4]. Extreme heat is now associated with rising mortality, reduced labour productivity, accelerated environmental degradation, biodiversity loss, and broader economic instability [5–11]. In rapidly urbanizing regions, land-use change and limited heat-resilient infrastructure further intensify exposure, underscoring the importance of monitoring human heat stress—not comfort alone—to assess vulnerability and inform targeted urban adaptation strategies [12].

Although more than sixty indices have been developed to quantify and assess heat stress [13], the widely used indices such as the wet-bulb globe temperature, effective temperature, and discomfort index often oversimplify the thermal environment settings by neglecting factors that influence thermal comfort experienced by individuals, such as radiant heat and evaporative cooling [14–18]. The indices also fail to account for individual variation in heat perception due to acclimatization, hydration, clothing, or cultural practices. The Universal Thermal Climate Index (UTCI) overcomes these shortcomings by incorporating meteorological inputs (air temperature, humidity, wind speed, and mean radiant temperature) with a dynamic physiological model of thermoregulation [19–21]. By simulating the human heat budget, incorporating conduction, convection, radiation, and latent heat loss, the UTCI expresses physiological strain as an equivalent air temperature under standardized conditions. Its validity across diverse climates has led to its widespread adoption in North America, Europe, Asia, and the Middle East, but its application to the South Asian context remains limited [22–31].

In this regard, the urgency of studying heat stress in South Asia stems from rapid urbanization, high population density, and limited heat-adaptive capacity. Recent evidence shows UTCI warming of 0.25–0.75 °C per decade during the pre-monsoon and monsoon seasons in India and Pakistan, which coincides with peak outdoor labour periods in these countries and exacerbates health risks [32,33]. Another densely populated country in South Asia, Bangladesh, is also highly vulnerable to heat stress due to its rising urban population in major cities like Dhaka, Chattogram, Rajshahi, Sylhet, and Khulna (25% in the 1980s to nearly 40% by 2020), exacerbating urban heat island effects caused by vegetation loss and unplanned urban expansions [34,35]. Recent projections from the International Labour Organization (ILO) suggest a 4.8% reduction in working hours by 2030 due to heat stress, and that 37.5% of the population will be at high risk from heat exposure from outdoor agricultural activities [36]. Other studies report that 21 billion working hours could be lost annually due to a 1 °C temperature rise, disproportionately affecting outdoor labourers such as construction workers and rickshaw pullers [36–38]. Health burdens are likely to grow sharply in the country, as evidenced by the 138% increase in elderly mortality (comparing between 2000–2004 vs 2018–2022), which overlaps with 8–9 heatwave day exposures per year during the 2014–2023 period [39–41].

Despite growing interest in thermal risk in South Asia, Bangladesh still lacks a mechanistic, long-term understanding of human heat stress that reflects both atmospheric conditions and urban morphological change. First, most heat-stress studies in Bangladesh rely on satellite-derived land surface temperatures (LST), which do not capture physiological discomfort [42–45]. Second, most studies focused primarily on Dhaka and used narrow timeframes that were often insufficient to capture long-term spatiotemporal trends and changes in alignment with urban and population growth [45,46]. As a result, secondary cities remain understudied, leaving critical gaps in understanding urban heat exposure and resilience in the country [42,43]. While recent studies have characterised daytime UTCI warming in South Asia [Kyaw et al., 2023; Ullah et al., 2022], none have systematically quantified diurnal asymmetry in UTCI trends — specifically the differential rates of change between nighttime minimum (UTCImin) and daytime maximum (UTCImax) thermal stress — across four decades at national scale in Bangladesh. As used in this study, ‘diurnal asymmetry’ refers to the condition in which UTCImin trends consistently exceed UTCImax trends across the majority of the study domain, implying progressive contraction of the nocturnal physiological recovery window. When this window shrinks, the cumulative 24-hour heat burden intensifies even if daytime peaks plateau or decline in some regions. This study fills that gap through three specific novel contributions: (1) the first 40-year national-scale diurnal UTCI climatology for Bangladesh using ERA5-HEAT, providing both spatial and hourly resolution; (2) the first quantification of nighttime recovery loss as a distinct and measurable dimension of heat-stress risk, defined as the reduction in daily hours below the UTCI comfort threshold (26°C); and (3) city-specific evidence for five major cities documenting the escalating transition from Strong to Very Strong heat-stress categories and its differential impact across day and night.

Materials and methods

Study area

The study area is Bangladesh, situated between latitudes 20.5°–26.5°N and longitudes 88.0°–92.7°E, characterized by a diverse agroecological and climatic landscape under a tropical monsoon climate. The country has hot, humid summers and mild winters, with temperatures ranging from 5°C in winter to over 40°C in summer [47]. Due to its high population density, rapid urbanization, and low elevation, Bangladesh is highly susceptible to climate-related health risks, notably heat stress [2].

This study analyzes thermal stress in five major cities in Bangladesh: Dhaka, Chattogram, Rajshahi, Khulna, and Sylhet. These cities were selected because they span diverse geographic and climatic zones across Bangladesh, from coastal lowlands to inland plains, and because of their demographic and socioeconomic significance. This diversity enables a comprehensive evaluation of heat-stress dynamics, as each city presents unique risk factors. For instance, Dhaka, the capital city, suffers from a strong UHI effect due to its high population density and dense infrastructure. The coastal megacity of Chattogram faces heightened thermal risks due to high humidity and urban congestion, whereas Rajshahi, located in the arid northwest, is characterized by consistently high daytime temperatures. Finally, Sylhet and Khulna represent varied topographies and urban development patterns in the hilly northeast and the coastal southwest, respectively (Fig 1).

thumbnail
Download:
Fig 1. Elevation map of Bangladesh depicting the key city corporations—Dhaka, Chattogram, Sylhet, Rajshahi, and Khulna—highlighted by their administrative limits.

https://doi.org/10.1371/journal.pclm.0000848.g001

Data

This study employed the UTCI from the ERA5-HEAT dataset, a global gridded reanalysis product developed by the European Centre for Medium-Range Weather Forecasts (ECMWF) and distributed by the Copernicus Climate Change Service (C3S). ERA5-HEAT derives hourly UTCI values by integrating meteorological reanalysis (ERA5) with a polynomial approximation of the Fiala multi-node thermoregulation model [28,48]. ERA5 itself assimilates satellite, ground-based, balloon, and buoy observations using advanced four-dimensional variational (4D-Var) data assimilation techniques, producing a consistent and high-quality global climate record.

ERA5-HEAT provides UTCI estimates at a spatial resolution of 0.25° × 0.25°, covering the globe between 90°N–60°S and 180°W–180°E from 1940 to the present. For this study, hourly data for Bangladesh were extracted from January 1985 to December 2024. A total of 194 grid points within the country’s boundaries were retained, allowing high-resolution spatiotemporal analysis of thermal stress across diverse urban morphologies. The dataset was accessed via the Copernicus Climate Data Store (https://cds.climate.copernicus.eu/datasets/derived-utci-historical).

To validate ERA5-HEAT against independent observations, ERA5-derived monthly Tmax, Tmin, and relative humidity (RH) were compared against Bangladesh Meteorological Department (BMD) station records for Dhaka, Chattogram, Rajshahi, Khulna, and Sylhet over 1985–2020 (S3 Fig). ERA5 reproduced the monthly seasonal cycle of all three variables reasonably well across the five major cities. Correlation coefficients were consistently high for Tmax (r = 0.941–0.984), Tmin (r = 0.989–0.994), and relative humidity (r = 0.896–0.944), indicating strong temporal agreement between ERA5 and station observations. ERA5 exhibited a systematic cold bias for Tmax across all cities, ranging from −1.35°C in Khulna to −3.85°C in Sylhet. Tmin was better represented, with smaller biases in most cities, although ERA5 underestimated Tmin in Sylhet by 2.19°C. Relative humidity was reproduced well in terms of seasonality, though city-level bias varied: ERA5 slightly overestimated RH in Dhaka, Chattogram, and Sylhet, and underestimated it in Rajshahi and Khulna. These systematic biases are largely time-invariant and do not substantially affect long-term trend detection over the 40-year study period. Overall, ERA5-HEAT is suitable for representing monthly climate variability and long-term trend analysis in Bangladesh, though Tmax and relative-humidity biases should be considered in city-level interpretation.

Universal thermal climate index (UTCI)

The UTCI is a physiologically based index designed to capture the dynamic interaction between meteorological conditions and human thermoregulation [19,20]. It integrates air temperature, wind speed, relative humidity, and mean radiant temperature with physiological factors such as metabolic rate and clothing insulation. The UTCI is expressed as an equivalent air temperature that elicits the same physiological response as a standardized reference environment (50% relative humidity, 0.5 m/s wind).

ERA5-HEAT computes UTCI using a sixth-order polynomial regression function approximating the complete Fiala model, with a root mean square error of 1.1 °C [20,28,49]. The UTCI is expressed as an equivalent air temperature that accounts for the combined physiological effects of thermal radiation, wind speed, and humidity, as defined by Equation (1):

(1)

where ΔToffset represents the offset from air temperature computed through a sixth-order polynomial regression of mean radiant temperature (Tmrt), wind speed (va), and relative humidity (RH). The polynomial approximation has a root mean square error of 1.1°C relative to the complete Fiala multi-node thermoregulation model (Di Napoli et al., 2021). ERA5-HEAT provides UTCI values calculated using this validated polynomial implementation.

Where:

Globally, UTCI values are categorized into ten thermal stress classes, ranging from extreme cold stress to extreme heat stress [19]. However, Bangladesh’s subtropical climate exhibits a UTCI range of approximately 0–46°C. To assess the occurrence of values outside this range, the distributions of daily UTCImax and UTCImin were calculated across 194 grid cells during 1985–2024, comprising 2,816,686 grid-day observations. Daily UTCImax exceeded 46°C in 764 observations (0.027%), primarily in northwestern Bangladesh during April–June, whereas daily UTCImin fell below 0°C in 284 observations (0.010%), mainly during January–March. The complete distributions and corresponding tail probabilities are shown in S4 Fig.. For this reason, the analysis adopts a regionally appropriate subset of five categories (Table 1), spanning Slight Cold Stress to Very Strong Heat Stress. This approach aligns with existing practices for tailoring indices to regional climates in which cold stress is negligible [28,50].

thumbnail
Download:
Table 1. UTCI assessment scale applied in this study.

https://doi.org/10.1371/journal.pclm.0000848.t001

Processing of daily thermal metrics

To characterize diurnal heat exposure, hourly UTCI values were aggregated into daily minimum (UTCIₘᵢₙ), daily mean (UTCImean), and daily maximum (UTCIₘₐₓ). UTCIₘᵢₙ represents nighttime cooling capacity, UTCIₘₐₓ indicates peak daytime heat stress, and UTCImean reflects the 24-hour thermal environment. Hourly data were converted from Coordinated Universal Time (UTC) to Bangladesh Standard Time (UTC + 6) prior to diurnal analysis.

Climatological periods

Daily UTCI values were averaged into two 20-year climatological periods: 1985–2004 and 2005–2024. This split reflects observed post-2000 acceleration in regional warming, rapid urban expansion, and intensifying urban heat-island effects across South Asia. Treating these as separate climatological windows enables detection of structural shifts in baseline thermal conditions.

Spatial and seasonal characterization

For each climatological period, spatial climatologies of UTCIₘᵢₙ, UTCImean, and UTCIₘₐₓ were generated at annual and seasonal scales (pre-monsoon, monsoon, post-monsoon, and winter). This approach identifies spatial patterns of heat stress, including thermal hotspots and contrasting conditions across the inland, northern, and coastal regions.

Trend detection

Long-term monotonic trends in annual UTCI metrics were quantified using Sen’s slope estimator, which provides a robust measure of trend magnitude. Statistical significance was assessed using the Modified Mann–Kendall test, which accounts for serial correlation in climatological time series [51]. Trends were mapped at the grid-cell level to capture spatial heterogeneity.

City-level analysis

Five major urban centres—Dhaka, Chattogram, Rajshahi, Khulna, and Sylhet—were selected for detailed analysis. Annual and seasonal UTCI metrics were extracted from grid cells corresponding to each city’s administrative boundary UTCI metrics were extracted from ERA5-HEAT grid cells whose centroids fell within each city’s administrative boundary polygon. Grid cell counts per city: Dhaka (4 cells), Chattogram (2 cells), Rajshahi (2 cells), Khulna (2 cells), and Sylhet (2 cells). City-level values represent the unweighted spatial mean across included grid cells. To assess sensitivity to boundary definition, the analysis was repeated using a ±0.25° buffer zone beyond each administrative boundary; resulting trend magnitudes differed by less than 5% from primary estimates, confirming robustness to grid-cell selection uncertainty. All city boundary shapefiles and grid-cell assignment tables are provided in the supplementary data repository to enable full reproducibility.

Results

Spatiotemporal patterns of heat stress in Bangladesh

Climatological changes between early and recent periods.

Fig 2 compares climatological UTCI conditions between two multi-decadal periods (1985–2004 and 2005–2024) and reveals substantial changes in both nighttime and daytime thermal environments across Bangladesh. The most pronounced increases are observed in daily minimum UTCI (UTCIₘᵢₙ). Relative to the earlier period, UTCIₘᵢₙ increased by approximately 1.0–1.5 °C across most regions, with local increases approaching 2 °C in parts of central and southern Bangladesh.

thumbnail
Download:
Fig 2. Climatological UTCI patterns in Bangladesh for two separate periods—(A) 1985–2004 and (B) 2005–2024.

Panels (a–c) display the spatial distributions of daily minimum (UTCIₘᵢₙ), mean (UTCImean), and maximum (UTCIₘₐₓ) UTCI for each period, whereas panel (d) presents the monthly climatological means aggregated across all grid cells, with error bars indicating interannual variability.

https://doi.org/10.1371/journal.pclm.0000848.g002

Daily maximum UTCI (UTCIₘₐₓ) also increased between the two periods, accompanied by a clear spatial expansion of high-temperature zones. Areas experiencing UTCIₘₐₓ values above 35 °C, previously concentrated mainly along the southern coastal belt, now extend inland across central, southwestern, and western regions. In these areas, daytime UTCIₘₐₓ increased by approximately 1–2 °C relative to the earlier climatology.

Seasonal comparisons indicate that warming is strongest during the pre-monsoon season (March–May). During this period, UTCIₘₐₓ increased by approximately 1.0–1.5 °C, with May emerging as the month exhibiting the highest climatological UTCI values, exceeding 36–37 °C in several regions. Winter conditions have also shifted, with January–February UTCIₘᵢₙ increasing by ~2 °C in many parts of the country. Collectively, these changes indicate a systematic warming of minimum, mean, and maximum UTCI conditions across seasons.

Decadal evolution of thermal conditions.

Fig 3 illustrates the decadal evolution of UTCI across Bangladesh from 1985 to 2024. During the 1985–1994 decade, nighttime UTCIₘᵢₙ values generally ranged from 18.5–22 °C, with the lowest values occurring in northern and northwestern districts. By 2015–2024, most regions exhibited increases of 1.0–1.5 °C, and large areas of southern and southwestern Bangladesh consistently recorded nighttime UTCIₘᵢₙ values above 22.5 °C.

thumbnail
Download:
Fig 3. Decadal spatial distributions of UTCI values across Bangladesh from 1985 to 2024.

Each row corresponds to a decade: 1985–1994, 1995–2004, 2005–2014, and 2015–2024. Columns represent (from left to right) daily minimum (UTCIₘᵢₙ), mean (UTCImean), and maximum (UTCIₘₐₓ) UTCI values.

https://doi.org/10.1371/journal.pclm.0000848.g003

Daytime conditions followed a similar spatial progression. In the earliest decade, UTCIₘₐₓ values across most regions ranged between 31–34.5 °C. In contrast, during the most recent decade, extensive areas of central, western, and southern Bangladesh regularly exceeded 35.5 °C, with localized increases of up to 1–2 °C relative to earlier decades.

Daily mean UTCI (UTCImean) also increased steadily over time. Whereas earlier decades exhibited mean values between 24.8–27.5 °C, recent decades show increases of 0.8–1.2 °C, with mean UTCImean exceeding 28 °C in southern districts. These decadal changes indicate a progressive shift toward higher baseline thermal conditions across both nighttime and daytime periods.

Long-term heat stress trends and frequency shifts

Annual and seasonal warming trends.

Fig 4 presents long-term trends in daily minimum (UTCIₘᵢₙ), maximum (UTCIₘₐₓ), and mean (UTCImean) UTCI across Bangladesh for the period 1985–2024. Spatial distributions of Sen’s slope estimates (°C decade−¹) are shown in Fig 4a–4c, and Fig 4d displays the corresponding national mean time series.

thumbnail
Download:
Fig 4. Spatial and temporal trends of UTCI indices over Bangladesh during 1985–2024, showing (a) Sen’s slope estimates (°C per decade) for daily minimum (UTCIₘᵢₙ), (b) maximum (UTCIₘₐₓ), and (c) mean (UTCImean) UTCI values.

Filled black circles indicate statistically significant trends (p < 0.05) based on the Modified Mann–Kendall test. Panel (d) presents annual national mean UTCI time series for UTCIₘᵢₙ (green), UTCIₘₐₓ (red), and UTCImean (blue), with linear trends annotated by Sen’s slope values.

https://doi.org/10.1371/journal.pclm.0000848.g004

Trends in UTCIₘᵢₙ show widespread positive values across the country (Fig 4a). Grid-cell Sen’s slope estimates range from +0.1 to +0.4 °C decade−¹, with higher values concentrated in southern and western regions. More than 80% of grid cells exhibit statistically significant positive trends (p < 0.05). The spatially averaged national trend in UTCIₘᵢₙ is +0.03 °C decade−¹.

Trends in UTCIₘₐₓ exhibit pronounced spatial variability (Fig 4b). Southern and southeastern regions show weak positive trends of approximately +0.1 °C decade−¹, whereas northern and northwestern regions display near-neutral to slightly negative trends ranging from −0.1 to 0.0 °C decade−¹. Approximately 50% of grid cells show statistically significant trends. The national mean UTCIₘₐₓ trend is +0.01 °C decade−¹ (p < 0.05).

Daily mean UTCI (UTCImean) exhibits positive trends across much of Bangladesh (Fig 4c). Sen’s slope estimates range from +0.05 to +0.25 °C decade−¹, with higher values observed in central and southern regions. The national mean UTCImean trend is +0.02 °C decade−¹ (p < 0.05).

Seasonal trend patterns are shown in Fig 5. UTCIₘᵢₙ displays positive trends in all seasons, with magnitudes ranging from +0.05 to +0.3 °C decade−¹. The largest seasonal trends occur during the post-monsoon period, particularly in northwestern and southern regions, where values approach +0.3 °C decade−¹. Seasonal UTCImean trends follow similar spatial patterns but with lower magnitudes. Annual Sen’s slope values range from −0.05 to +0.15 °C decade−¹, with a national mean of +0.02 °C decade−¹ (p < 0.05).

thumbnail
Download:
Fig 5. Seasonal Sen’s slope maps showing trends in UTCI values over Bangladesh from 1985 to 2024.

Each row represents the annual average for the pre-monsoon (March–May), monsoon (June–September), and post-monsoon (October–November) periods. Columns correspond to daily minimum (UTCIₘᵢₙ), mean (UTCImean), and maximum (UTCIₘₐₓ) UTCI trends. Black dots indicate regions with statistically significant trends (p < 0.05), and colour gradients represent the rate of change in °C per decade.

https://doi.org/10.1371/journal.pclm.0000848.g005

Seasonal UTCIₘₐₓ trends are generally weak, with neutral to mildly negative values inland (−0.1 to 0.0 °C decade−¹) and small positive trends (≤ +0.05 °C decade−¹) in coastal and metropolitan regions. The contrast between stronger UTCIₘᵢₙ trends and weaker or mixed UTCIₘₐₓ trends indicates a long-term convergence of nighttime and daytime UTCI trends.

Shifts in the frequency of hazardous heat days.

Fig 6 summarizes long-term changes in the annual frequency of UTCI-based thermal stress categories across Bangladesh for the period 1985–2024. Analysis of daily mean UTCI (UTCImean) indicates a widespread reduction in the number of days classified as No Thermal Stress, with decreases of approximately 5–15 days per decade, most pronounced in northern and western regions. Concurrently, the frequency of days classified as Strong Heat Stress increased by approximately 5–15 days per decade, with the largest increases observed in southern, coastal, and deltaic regions.Changes in the upper range of thermal stress are most evident for daily maximum UTCI (UTCIₘₐₓ). The frequency of Very Strong Heat Stress days increased substantially over the study period. In northwestern and central regions, including Rajshahi and Dhaka, increases range from 8 to 15 days per decade. Spatial patterns of these changes are shown in S1 Fig.Overall, the results indicate a redistribution of UTCI stress categories over time, characterized by declining frequencies of thermally neutral conditions and increasing frequencies of strong and very strong heat stress across large portions of Bangladesh.

thumbnail
Download:
Fig 6. Trends in annual frequency of UTCImean and UTCIₘₐₓ stress categories (1985–2024), in days per decade, assessed using the Modified Mann–Kendall test (p < 0.05).

Filled circles indicate statistically significant grid cells. Positive values (warm colours) indicate increasing frequency; negative values (cool colours) indicate decreasing frequency.

https://doi.org/10.1371/journal.pclm.0000848.g006

Urban heat stress dynamics and diurnal intensification

Intensification of the diurnal heat cycle.

Fig 7 compares climatological hourly mean UTCI values across Bangladesh between two periods, 1985–2004 and 2005–2024. Across nearly all hours of the day, UTCI values are higher during the later period. The largest differences occur during early morning hours (05:00–07:00 LT), with national mean UTCI increases of approximately 0.6–0.8 °C. Afternoon UTCI values between 13:00 and 15:00 LT increased by approximately 0.6–0.7 °C. During most hours, mean UTCI values in 2005–2024 exceed those of 1985–2004 by 0.5–1.0 °C.

thumbnail
Download:
Fig 7. Climatological hourly mean UTCI across Bangladesh for two periods (1985–2004 vs 2005–2024).

https://doi.org/10.1371/journal.pclm.0000848.g007

Fig 8 presents analogous diurnal UTCI profiles for five major cities: Dhaka, Rajshahi, Chattogram, Sylhet, and Khulna. All cities exhibit higher UTCI values across the diurnal cycle in the later period. Morning-hour increases range from 0.5–1.0 °C, while afternoon increases range from 0.5–0.8 °C. Rajshahi records the highest midday UTCI values, whereas Dhaka shows comparatively larger increases during evening hours. S2 Fig shows consistent long-term increases in daily maximum UTCI across all five cities.

thumbnail
Download:
Fig 8. Hourly mean Universal Thermal Climate Index (UTCI) for five major cities in Bangladesh (Dhaka, Rajshahi, Chattogram, Sylhet, and Khulna), comparing two climatological periods: 1985–2004 and 2005–2024.

https://doi.org/10.1371/journal.pclm.0000848.g008

City-specific heat stress profiles and trends.

Fig 9 shows the distribution of daily maximum UTCI values by thermal stress category for five cities over the period 1985–2024. Median UTCI values within the Very Strong Heat Stress category are highest in Rajshahi (≈40.7 °C) and Sylhet (≈40.4 °C). Extreme values exceeding 45 °C are observed in Rajshahi. Coastal cities, including Chattogram and Khulna, exhibit narrower distributions and lower upper extremes relative to inland cities. Limited overlap among stress-category distributions indicates distinct UTCI ranges associated with each category.

thumbnail
Download:
Fig 9. Daily maximum UTCI by stress category for five cities (1985–2024).

https://doi.org/10.1371/journal.pclm.0000848.g009

Fig 10 presents annual frequencies of UTCI thermal stress categories for each city. Days classified as Strong or Very Strong Heat Stress dominate the annual distribution across all five cities. Very Strong Heat Stress days frequently exceed 150 days per year in Rajshahi. Days classified as No Thermal Stress occur infrequently, typically fewer than 20 days per year in all cities. Temporal trends in these categories are shown in Fig 11.

thumbnail
Download:
Fig 10. Annual frequency of Universal Thermal Climate Index (UTCI) thermal stress categories: No Thermal Stress, Moderate Thermal Stress, Strong Thermal Stress, and Very Strong Thermal Stress for five major cities in Bangladesh (1985 – 2024).

https://doi.org/10.1371/journal.pclm.0000848.g010

thumbnail
Download:
Fig 11. Trends in the number of days per decade for UTCI thermal stress categories: No Thermal Stress, Moderate Thermal Stress, Strong Thermal Stress, and Very Strong Thermal Stress across five major cities in Bangladesh.

https://doi.org/10.1371/journal.pclm.0000848.g011

Across all cities, the frequency of Very Strong Heat Stress days increased over the study period, accompanied by declines in Strong and Moderate Heat Stress categories. Chattogram exhibits the largest increase (+12.7 days decade−¹), offset by reductions in Strong (−10.2 days decade−¹) and Moderate (−4.7 days decade−¹) categories. Dhaka shows an increase of +8.2 Very Strong Heat Stress days decade−¹, alongside a decline of −9.2 Strong Heat Stress days decade−¹. Khulna, Rajshahi, and Sylhet exhibit increases of +6.6, +4.5, and +5.0 days decade−¹, respectively.

Table 2 quantifies pronounced changes in both daytime and nighttime thermal exposure across all five major cities between 1985 and 2024. During daytime hours, the proportion of time classified as Very Strong Heat Stress increased substantially in every city. Rajshahi experienced the highest daytime exposure, rising from 15.7% to 26.4% of daytime hours (+10.7 percentage points), while Dhaka showed an increase from 8.2% to 20.7% (+12.5 percentage points). Comparable increases were observed in Khulna (+11.3 percentage points), Sylhet (+10.0 percentage points), and Chattogram (+11.2 percentage points). These results indicate that extreme daytime heat stress has become markedly more frequent across both inland and coastal urban environments.

thumbnail
Download:
Table 2. Intensification of daytime and nighttime heat stress (UTCI%) in major Bangladeshi cities, 1985–2024.

https://doi.org/10.1371/journal.pclm.0000848.t002

Nighttime conditions exhibit an opposing but equally consequential trend. The share of nighttime hours classified as No Thermal Stress declined sharply in all cities, indicating a progressive loss of thermally neutral recovery periods. The largest reductions occurred in Chattogram (−25.5 percentage points) and Dhaka (−21.1 percentage points), followed by Khulna (−18.9 percentage points), Sylhet (−13.6 percentage points), and Rajshahi (−9.7 percentage points). As a result, cities that previously experienced predominantly comfortable nighttime conditions now spend a substantially greater fraction of the night under thermally stressful environments.

Fig 12 presents the diurnal heatmaps of UTCI-derived thermal stress categories across months for five major cities in Bangladesh. Thermally Comfortable conditions dominate the annual cycle in all cities, accounting for 70–77% of total hours. Chattogram exhibits the highest proportion of comfortable conditions (77%), followed by Dhaka and Rangpur (73%), Sylhet (74%), and Rajshahi (70%).

thumbnail
Download:
Fig 12. Diurnal heatmap of UTCI-derived thermal stress categories across months for five major cities of Bangladesh.

Each row represents the hour of the day (0–23), and each column represents months (January to December). Color codes indicate UTCI-based thermal stress levels. The percentage of total time classified as ‘Comfortable’ is shown for each city.

https://doi.org/10.1371/journal.pclm.0000848.g012

Across all cities, daytime heat stress intensifies seasonally, with Warm, Hot, and Very Hot categories emerging during the pre-monsoon months (March–May). The expansion of heat stress occurs primarily between late morning and late afternoon hours. Rajshahi shows the longest and most persistent daytime exposure to Very Hot conditions during April and May, spanning several consecutive daytime hours. Dhaka and Sylhet also exhibit repeated occurrences of Hot and Very Hot categories during the same period, though with shorter durations. Rangpur displays extended Warm and Hot daytime conditions during the pre-monsoon season, with limited persistence of Very Hot stress.

Chattogram displays a comparatively distinct diurnal pattern, with daytime heat stress present during the pre-monsoon and monsoon seasons but shorter durations of Very Hot conditions relative to inland cities. In all cities, nighttime hours are predominantly classified as Comfortable during winter months, while Warm nighttime conditions become more frequent from April through September, particularly during early morning hours.

Synthesis of Results: Taken together, the findings reported in this section reveal three structural shifts in Bangladesh’s thermal environment over 1985–2024. At the national scale (Figs 4–6), diurnal asymmetry is established: UTCImin trends are positive and widespread while UTCImax trends are spatially heterogeneous, narrowing the day–night thermal contrast across the country. At the city scale (Figs 9–11 and Table 2), this asymmetry is amplified: Very Strong Heat Stress daytime hours have increased by 10–12.5 percentage points in Dhaka and Chattogram, while thermally comfortable nighttime hours have simultaneously declined by 13–25 percentage points. At the hourly scale (Figs 7–8), the early morning window (05:00–07:00 LT) now shows some of the largest absolute UTCI increases of any time period. Together, these three scales of evidence converge on a single conclusion: Bangladesh is transitioning from episodic heat episodes to a persistent, round-the-clock heat stress regime. The Discussion below contextualises this finding within regional literature and draws out implications for public health, labour productivity, and heat adaptation policy.

Discussion

The results demonstrate that heat stress in Bangladesh is intensifying asymmetrically across the diurnal cycle: UTCIₘᵢₙ is rising at three times the national-mean rate of UTCIₘₐₓ (+0.03 vs. +0.01°C decade−¹), with UTCIₘₐₓ showing no statistically significant trend at the national scale (95% CI crosses zero). This finding has three analytically significant implications. First, it identifies erosion of the nocturnal recovery window as a distinct and previously unquantified dimension of heat risk in Bangladesh — entirely absent from existing assessments based on land surface temperature or daytime maximum temperature [Imran et al., 2021; Uddin et al., 2022]. Second, the data show that southern and western regions experience the sharpest diurnal convergence between UTCIₘᵢₙ and UTCIₘₐₓ, indicating that the day–night thermal contrast is narrowing most rapidly in areas already among the most vulnerable to compound heat exposure. Third, city-level evidence (Table 2) reveals that Dhaka and Chattogram simultaneously face the largest increases in daytime Very Strong Heat Stress hours and the largest reductions in thermally neutral nighttime hours, creating a convergence of exposure intensification and recovery loss that compounds cumulative physiological burden.

The study shows that heat stress has increased significantly between 1985 and 2024, and it is not occurring evenly across the country. The southern regions, which are densely populated and urbanized, have been disproportionately affected. There is warming trends in both nighttime temperatures (UTCIₘᵢₙ) and daily averages (UTCImean), with the national average increasing by approximately 0.03°C per decade and 0.02°C per decade, respectively. In some parts of the country, the warming has been even more dramatic—up to 0.4°C per decade. This fits with what is happening across South Asia more broadly. Kyaw et al. [32] found that UTCI has been rising by 0.25–0.75°C per decade during the pre-monsoon and monsoon seasons, exactly when people are doing the most outdoor labor.

One of the most interesting findings of the study is that nighttime and daytime warming are not happening at the same rate. Nighttime temperatures (UTCIₘᵢₙ) are going up pretty much everywhere, but daytime highs (UTCIₘₐₓ) show a mixed picture—there is slight warming along the coast but actually some cooling in the north. This means the difference between day and night temperatures is getting smaller, which is significant for public health. Early-morning UTCI has increased by 1.7°C over the past 40 years, indicating that people are losing the crucial cool nighttime hours their bodies need to recover from daytime heat. Without nighttime relief and with the entire 24-hour heat cycle intensifying, people experience cumulative heat strain, increasing their risk of heat-related illnesses. The study, Khan et al. [52] recently identified increasing early-morning heat stress as one of the main drivers of rising thermal risk in South Asia, which supports our findings.

The shift toward a hotter climate really shows up in the numbers of dangerously hot days, especially in cities. The results revealed that the number of days with “Very Strong Heat Stress” (UTCI above 38°C) has increased by up to 15 days per decade in places such as Rajshahi and Dhaka. In Rajshahi, more than a quarter of all daytime hours in 2024 fell into this potentially deadly category—that is up from just 15.7% in 1985. What is particularly concerning is that days that were previously classified as “Strong Heat Stress” are now moving into the “Very Strong” category. This shift is most evident in Chattogram, where it increased about 12.7 “Very Strong” heat-stress days per decade while reduced almost the same number of “Strong” and “Moderate” heat-stress days.

This intensification in urban areas makes sense in terms of the urban heat island (UHI) effect that Dewan et al. [35] explored in detail. Heat from human activities and the concrete and asphalt in cities raise temperatures above those in surrounding rural areas. The elevated UTCI values observed in densely populated cities such as Dhaka and Khulna align with findings from other rapidly growing urban areas worldwide. For instance, Krüger et al. [50] linked increases in UTCI in Brazil to the expansion of paved surfaces. The spatial patterns in the study, with the highest UTCI values concentrated in major cities, show that urban populations are hit with a double whammy—they face both regional climate warming and the localized heating from the urban environment itself.

The increasing frequency and intensity of heat stress have real, serious consequences for public health and economic productivity. UTCI is closely linked to physiological strain—such as changes in core and skin temperatures—which makes it a strong indicator of health risk, as noted by Zare et al. [13]. Cities such as Dhaka and Rajshahi are now experiencing prolonged periods of intense heat, which places enormous strain on people’s bodies. This is especially critical for the millions of people working outdoors in informal jobs—construction workers, transportation workers, agricultural laborers—who sustain the urban economy. For these workers, spending long hours exposed to UTCI levels above 32°C during peak working hours (9 AM to 5 PM), as our diurnal heatmaps show during the pre-monsoon season, dramatically increases their risk of heat exhaustion, heat stroke, and physical inability to work effectively.

The trends observed in this study align with regional analyses indicating a sharp rise in high-heat-stress days across South Asian cities and with projections of significant losses in labor productivity. The ILO [36] has projected substantial reductions in working hours in Bangladesh because of heat stress. The Grantham Institute policy brief [53] emphasized that informal workers in Dhaka often lack access to air conditioning, proper insulation, or effective heat-warning systems, leaving them incredibly vulnerable to extreme heat. Our findings, which show UTCI levels consistently rising across major metropolitan areas nationwide, underscore the urgent need for interventions such as shaded rest areas, improved dissemination of early warnings, and adjustments to work hours to protect these at-risk populations.

Given the number of people at risk, the growing heat-stress crisis in Bangladesh’s urban areas demands urgent, evidence-based policy action. National and local governments should begin incorporating UTCI-based monitoring into their climate adaptation planning. This could help identify high-risk populations and ensure resources for public health interventions are allocated where they are needed most. Ahmedabad’s Heat Action Plan demonstrates how early warning systems, microclimate monitoring, and targeted cooling interventions can reduce heat-related mortality. Dhaka South City Corporation’s Climate Action Plan 2024 outlines similar priorities—including reflective roofing, urban greening, and social protection for vulnerable groups—but rapid and decisive implementation is now essential. Local governments should also prioritize investing in nature-based solutions, more shade infrastructure, and community cooling initiatives in the most heat-exposed urban areas that have been identified. These strategies directly support Sustainable Development Goals for Health (SDG 3), Sustainable Cities (SDG 11), and Climate Action (SDG 13), and can help us build urban areas that are more climate-friendly, sustainable, and resilient.

While this study provides important evidence on heat stress and exposure across Bangladesh, it has some limitations. ERA5-HEAT introduces multiple sources of uncertainty that must be acknowledged: (1) The polynomial approximation used to compute UTCI carries an inherent RMSE of 1.1°C relative to the full Fiala thermoregulation model [Di Napoli et al., 2018]. (2) ERA5 reanalysis may underestimate temperature extremes in data-sparse regions; however, Bangladesh’s relatively dense observational network reduces this bias. (3) Comparative studies of ERA5-derived UTCI against observed station UTCI in South Asia report seasonal mean biases of ±1–2°C [Kyaw et al., 2023; Ullah et al., 2022]; such biases are largely systematic and time-invariant, and would not substantially affect linear trend detection over the 40-year study period. (4) ERA5-HEAT does not explicitly resolve the urban canopy layer, meaning the dataset may underestimate actual UTCI in densely built city cores. The dataset has a spatial resolution of 0.25° (approximately 27 km), which means it cannot reliably capture microclimatic differences—such as the cooling effects of green spaces, how building layouts affect temperature, or how urban density matters—at the neighbourhood scale. Despite these constraints, ERA5-HEAT remains the most comprehensive, temporally consistent, and publicly available gridded UTCI product for Bangladesh; the 40-year record provides sufficient statistical power to detect robust warming signals well above the ERA5 uncertainty threshold. If higher-resolution modeling and data frameworks become available, these gaps can be addressed. For example, coupled traffic-emission-dispersion modeling has shown it is possible to capture fine-scale urban atmospheric dynamics [54], and land-use studies demonstrate that local surface cover can substantially change neighborhood thermal conditions [55]. Future work should aim for sub-kilometer-resolution thermal mapping or urban-scale modeling to better understand exposure gradients within cities. Second, our findings would be more meaningful if it is possible to incorporate socioeconomic vulnerability indicators (e.g., health data or productivity metrics), which are crucial for identifying which subpopulations are most at risk of heat exposure. However, such a detailed socioeconomic vulnerability assessment is beyond the scope of this initial study, which focuses primarily on quantifying and characterizing heat stress in Bangladesh. Future research could integrate socioeconomic vulnerability indicators with UTCI trends to develop composite heat risk indices, thereby generating more contextually relevant and equitable policy recommendations.

Despite these limitations, this study presents strong evidence that heat stress in Bangladesh is increasing, as measured by UTCI. The trends of rising UTCI, longer periods of thermal stress, and the concentration of urban hotspots point to an intensifying public health crisis. By putting these findings in the context of regional and global research, this study provides a critical evidence base to inform targeted urban planning, public health interventions, and climate adaptation strategies. These findings highlight the need for evidence-based adaptation strategies to address escalating urban heat stress.

Conclusion

This study presents four decades of evidence indicating that heat stress in Bangladesh is rising, directly impacting daily living and public health. UTCI trends from 1985 to 2024 indicate a significant alteration in the 24-hour thermal cycle, characterized by nighttime warming surpassing daytime variations. The nighttime UTCI has risen by approximately 1.7°C, diminishing chances for physiological recovery and exacerbating persistent heat stress. Cities like Rajshahi and Dhaka now endure an increase of 8–15 days per decade of “Very Strong Heat Stress,” with inland regions—particularly the northwest and southwest—warming at a rate surpassing that of coastal areas. The urban heat island phenomenon exacerbates temperatures in Dhaka and Khulna, endangering outdoor laborers and economically disadvantaged groups the most. These findings emphasize the necessity for prompt, evidence-informed heat adaption. Incorporating UTCI-based monitoring into national climate strategies would facilitate the identification of at-risk populations and enable focused interventions. This analysis is constrained by the spatial resolution of ERA5-HEAT data; yet, the overarching conclusion is unequivocal: heat stress is increasing unevenly yet significantly. The immediate priority is to convert this evidence into prompt policy measures to safeguard health, livelihoods, and urban resilience.

Supporting information

S1 Fig. Spatial visualization framework of decadal UTCI heat stress classification for Bangladesh (1985–2024).

The figure systematically arranges UTCI-based categories—No Thermal Stress, Moderate Heat Stress, Strong Heat Stress, and Very Strong Heat Stress—across minimum (nighttime), mean (average daytime), and maximum (peak daytime) thermal conditions. Each column represents a stress category, while each row corresponds to successive decades.

https://doi.org/10.1371/journal.pclm.0000848.s001

(TIF)

S2 Fig. Framework for long-term temporal representation of daily maximum Universal Thermal Climate Index (UTCI) across five major Bangladeshi cities (Dhaka, Rajshahi, Chattogram, Sylhet, and Khulna) over the period 1985–2024.

The figure applies a 30-day moving average to depict intra-annual variability and multidecadal continuity in daily maximum UTCI values. Distinct color-coded traces represent individual city series, providing a standardized visualization structure for interpreting seasonal cycles and long-term fluctuations in urban thermal environments.

https://doi.org/10.1371/journal.pclm.0000848.s002

(TIF)

S3 Fig. Monthly comparison of ERA5 reanalysis against Bangladesh Meteorological Department (BMD) station observations for the five major cities of Bangladesh (Dhaka, Chattogram, Rajshahi, Khulna, and Sylhet) over the period 1985–2020.

Each panel corresponds to one city and shows the 12-month mean climatology of three variables: daily maximum temperature (Tmax, °C), daily minimum temperature (Tmin, °C), and relative humidity (RH, %). BMD station observations are shown as solid lines with filled markers; ERA5 reanalysis values are shown as dashed lines with open markers. The left y-axis corresponds to temperature (°C) and the right y-axis to relative humidity (%). BMD = Bangladesh Meteorological Department; Tmax = daily maximum temperature; Tmin = daily minimum temperature; RH = relative humidity.

https://doi.org/10.1371/journal.pclm.0000848.s003

(TIF)

S4 Fig. Daily maximum and minimum UTCI distributions and tail probabilities across Bangladesh.

Panels (a) and (b) show the probability-density distributions of daily maximum UTCI (UTCImax) and daily minimum UTCI (UTCImin), respectively, with kernel density estimate curves. Vertical dashed lines and background shading indicate the applicable UTCI stress-category boundaries. Panel (c) shows the right-tail exceedance probability, P(UTCImax > threshold), on a logarithmic scale, while panel (d) shows the left-tail cumulative probability, P(UTCImin < threshold), on a logarithmic scale. Of 2,816,686 grid-day observations from 194 ERA5-HEAT grid cells during 1985–2024, 764 observations (0.027%) exceeded 46°C and 284 observations (0.010%) were below 0°C. These rare values were identified in northwestern Bangladesh during April–June and January–March, respectively. KDE, kernel density estimate; UTCI, Universal Thermal Climate Index.

https://doi.org/10.1371/journal.pclm.0000848.s004

(TIF)

References

  1. 1. Lukić M, Filipović D, Pecelj M, Crnogorac L, Lukić B, Divjak L, et al. Assessment of outdoor thermal comfort in Serbia’s urban environments during different seasons. Atmosphere. 2021;12(8):1084.
  2. 2. Shahid S, Wang X-J, Harun SB, Shamsudin SB, Ismail T, Minhans A. Climate variability and changes in the major cities of Bangladesh: observations, possible impacts and adaptation. Reg Environ Change. 2015;16(2):459–71.
  3. 3. Ekra AT, Hamed MM, Ali Z, Muhammad MKIB, Khan MMH, Kamruzzaman M, et al. Changes in human heat discomfort and its drivers in Bangladesh. Urban Climate. 2024;55:101884.
  4. 4. Islam HMT, Kamruzzaman M, Shahid S, Mainuddin M, Alam E, Islam ARM d T. Spatiotemporal changes in temperature projections over Bangladesh using multi-model ensemble data. Front Environ Sci. 2023;10.
  5. 5. García FC, Bestion E, Warfield R, Yvon-Durocher G. Changes in temperature alter the relationship between biodiversity and ecosystem functioning. Proc Natl Acad Sci U S A. 2018;115(43):10989–94. pmid:30297403
  6. 6. García-León D, Casanueva A, Standardi G, Burgstall A, Flouris AD, Nybo L. Current and projected regional economic impacts of heatwaves in Europe. Nat Commun. 2021;12(1):5807. pmid:34608159
  7. 7. Hamed MM, Nashwan MS, Shahid S. Projected changes in thermal bioclimatic indicators over the Middle East and North Africa under Paris climate agreement. Stoch Environ Res Risk Assess. 2022;37(2):577–94.
  8. 8. Islam HMT, Mainuddin M, Affan A, Ahmed S, Rahman MA, Sadeque A, et al. Thermal bioclimatic transformations in the coastal regions of Ganges delta: insights from CMIP6 multi-model ensemble. Sci Rep. 2025;15(1):20569. pmid:40593949
  9. 9. Kamruzzaman M, Wahid S, Shahid S, Alam E, Mainuddin M, Islam HMT. Predicted changes in future precipitation and air temperature across Bangladesh using CMIP6 GCMs. Heliyon. 2023;9.
  10. 10. Kamruzzaman M, Shariot-Ullah M, Islam R, Amin MGM, Islam HMT, Ahmed S, et al. Projections of future bioclimatic indicators using bias-corrected CMIP6 models: a case study in a tropical monsoon region. Environ Sci Pollut Res Int. 2024;31(56):64596–627. pmid:39541022
  11. 11. Szewczyk W, Mongelli I, Ciscar J-C. Heat stress, labour productivity and adaptation in Europe—a regional and occupational analysis. Environ Res Lett. 2021;16(10):105002.
  12. 12. Duanmu L, Sun X, Jin Q, Zhai Z. Relationship between human thermal comfort and indoor thermal environment parameters in various climatic regions of China. Procedia Eng 2017.
  13. 13. Zare S, Hasheminejad N, Shirvan HE, Hemmatjo R, Sarebanzadeh K, Ahmadi S. Comparing Universal Thermal Climate Index (UTCI) with selected thermal indices/environmental parameters during 12 months of the year. Weather Clim Extrem. 2018;19.
  14. 14. Steadman RG. The assessment of sultriness. Part I: a temperature-humidity index based on human physiology and clothing science. J Appl Meteor. 1979;18(7):861–73.
  15. 15. Kenny GP, Yardley J, Brown C, Sigal RJ, Jay O. Heat stress in older individuals and patients with common chronic diseases. CMAJ. 2010;182(10):1053–60. pmid:19703915
  16. 16. Budd GM. Wet-bulb globe temperature (WBGT)-its history and its limitations. J Sci Med Sport. 2008.
  17. 17. Spangler KR, Liang S, Wellenius GA. Wet-bulb globe temperature, universal thermal climate index, and other heat metrics for US counties, 2000–2020. Sci Data. 2022;9.
  18. 18. Song W, Zhong F, Calautit JK, Li J. Exploring the role of skin temperature in thermal sensation and thermal comfort: a comprehensive review. Energy Built Environ. 2025;6(4):762–81.
  19. 19. Błazejczyk K, Jendritzky G, Bröde P, Fiala D, Havenith G, Epstein Y. An introduction to the Universal thermal climate index (UTCI). Geogr Pol. 2013;86.
  20. 20. Bröde P, Krüger EL, Rossi FA, Fiala D. Predicting urban outdoor thermal comfort by the Universal Thermal Climate Index UTCI--a case study in Southern Brazil. Int J Biometeorol. 2012;56(3):471–80. pmid:21604151
  21. 21. Fiala D, Havenith G, Bröde P, Kampmann B, Jendritzky G. UTCI-Fiala multi-node model of human heat transfer and temperature regulation. Int J Biometeorol. 2012;56(3):429–41. pmid:21503622
  22. 22. Mölders N. Outdoor universal thermal comfort index climatology for Alaska. ACS. 2019;09(04):558–82.
  23. 23. Park S, Tuller SE, Jo M. Application of Universal Thermal Climate Index (UTCI) for microclimatic analysis in urban thermal environments. Landscape Urban Plan. 2014;125:146–55.
  24. 24. Wu F, Yang X, Shen Z. Regional and seasonal variations of outdoor thermal comfort in China from 1966 to 2016. Sci Total Environ. 2019;665.
  25. 25. Zhang L, Wei D, Hou Y, Du J, Liu Z, Zhang G, et al. Outdoor thermal comfort of urban park—a case study. Sustainability. 2020;12(5):1961.
  26. 26. Caminade C, McIntyre KM, Jones AE. Impact of recent and future climate change on vector-borne diseases. Ann N Y Acad Sci. 2019;1436(1):157–73. pmid:30120891
  27. 27. Basarin B, Lukić T, Matzarakis A. Quantification and assessment of heat and cold waves in Novi Sad, Northern Serbia. Int J Biometeorol. 2016;60(1):139–50. pmid:26062816
  28. 28. Di Napoli C, Pappenberger F, Cloke HL. Assessing heat-related health risk in Europe via the Universal Thermal Climate Index (UTCI). Int J Biometeorol. 2018;62(7):1155–65. pmid:29546489
  29. 29. Dunjić J. Outdoor thermal comfort research in urban areas of Central and Southeast Europe: a review. Geogr Pannon. 2019;23(4):359–73.
  30. 30. Pantavou K, Theoharatos G, Santamouris M, Asimakopoulos D. Outdoor thermal sensation of pedestrians in a Mediterranean climate and a comparison with UTCI. Build Environ. 2013;66.
  31. 31. Roshan G, Yousefi R, Błażejczyk K. Assessment of the climatic potential for tourism in Iran through biometeorology clustering. Int J Biometeorol. 2018;62(4):525–42. pmid:29058081
  32. 32. Kyaw AK, Hamed MM, Kamruzzaman M, Shahid S. Spatiotemporal changes in population exposure to heat stress in South Asia. Sustain Cities Soc. 2023;93:104544.
  33. 33. Kyaw AK, Hamed MM, Shahid S. Spatiotemporal changes in Universal Thermal Climate Index over South Asia. Atmos Res. 2023;292.
  34. 34. Arifeen HM, Phoungthong K, Mostafaeipour A, Yuangyai N, Yuangyai C, Techato K, et al. Determine the land-use land-cover changes, urban expansion and their driving factors for sustainable development in Gazipur Bangladesh. Atmosphere. 2021;12(10):1353.
  35. 35. Dewan A, Kiselev G, Botje D, Mahmud GI, Bhuian MdH, Hassan QK. Surface urban heat island intensity in five major cities of Bangladesh: Patterns, drivers and trends. Sustain Cities Soc. 2021;71:102926.
  36. 36. de Jouvenel H. OIT (Organisation internationale du travail), Working on a Warmer Planet: The Effect of Heat Stress on Productivity and Decent Work, Genève: OIT, juillet 2019, 98 p. https://www.ilo.org/wcmsp5/groups/public/---dgreports/---dcomm/---publ/documents/publication/wcms_711919.pdf. Consulté le 27 février 2020. Futuribles. 2020;N° 436. doi:10.3917/futur.436.0133g
  37. 37. Parsons LA, Shindell D, Tigchelaar M, Zhang Y, Spector JT. Increased labor losses and decreased adaptation potential in a warmer world. Nat Commun. 2021;12(1):7286. pmid:34907184
  38. 38. Mamin FA, Dey G, Das SK. Health and safety issues among construction workers in Bangladesh. Int J Occup Saf Health. 2019;9(1):13–8.
  39. 39. Burkart K, Breitner S, Schneider A, Khan MMH, Krämer A, Endlicher W. An analysis of heat effects in different subpopulations of Bangladesh. Int J Biometeorol. 2014;58(2):227–37. pmid:23689928
  40. 40. Nissan H, Burkart K, de Perez EC, Van Aalst M, Mason S. Defining and predicting heat waves in Bangladesh. J Appl Meteorol Climatol. 2017;56.
  41. 41. Romanello M, Napoli C d i, Green C, Kennard H, Lampard P, Scamman D. The 2023 report of the Lancet Countdown on health and climate change: the imperative for a health-centred response in a world facing irreversible harms. Lancet. 2023.
  42. 42. Tabassum A, Hong S-H, Park K, Baik J-J. Simulating urban heat islands and local winds in the Dhaka metropolitan area, Bangladesh. Urban Climate. 2025;59:102284.
  43. 43. Tabassum A, Park K, Seo JM, Han J-Y, Baik J-J. Characteristics of the urban heat island in Dhaka, Bangladesh, and its interaction with heat waves. Asia-Pac J Atmos Sci. 2024;60(4):479–93.
  44. 44. Imran HM, Hossain A, Islam AKMS, Rahman A, Bhuiyan MAE, Paul S. Impact of land cover changes on land surface temperature and human thermal comfort in Dhaka city of Bangladesh. Earth Systems and Environment. 2021;5.
  45. 45. Uddin ASMS, Khan N, Islam ARMT, Kamruzzaman M, Shahid S. Changes in urbanization and urban heat island effect in Dhaka city. Theor Appl Climatol. 2022;147.
  46. 46. Kamal ASMM, Fahim AKF, Shahid S. Changes in wet bulb globe temperature and risk to heat-related hazards in Bangladesh. Sci Rep. 2024;14(1):10417. pmid:38710893
  47. 47. Mortuza MdR, Selmi S, Khudri MdM, Ankur AK, Rahman MdM. Evaluation of temporal and spatial trends in relative humidity and dew point temperature in Bangladesh. Arab J Geosci. 2013;7(12):5037–50.
  48. 48. Hersbach H, Bell B, Berrisford P, Hirahara S, Horányi A, Muñoz-Sabater J. The ERA5 global reanalysis. Q J R Meteorol Soc. 2020;146.
  49. 49. Ullah S, You Q, Wang G, Ullah W, Sachindra DA, Yan Y, et al. Characteristics of human thermal stress in South Asia during 1981–2019. Environ Res Lett. 2022;17(10):104018.
  50. 50. Krüger E, Gobo JPA, Tejas GT, da Silva de Souza RM, Neto JBF, Pereira G, et al. The impact of urbanization on heat stress in Brazil: a multi-city study. Urban Climate. 2024;53:101827.
  51. 51. Hamed KH, Ramachandra Rao A. A modified Mann-Kendall trend test for autocorrelated data. J Hydrol (Amst). 1998;204:182–96.
  52. 52. Khan N, Kamruzzaman M, Shahid S. Diurnal pattern of heat stress over South Asia: a wet bulb globe temperature‐based analysis from 1984 to 2023. Int J Climatol. 2025;45(6).
  53. 53. Letsch L, Dasgupta S, Robinson EJ. Policy brief Adapting to the impacts of extreme heat on Bangladesh’s labour force. 2023.
    • 54. Landolsi J, Rehimi F, Kalboussi A. Urban traffic and induced air quality modeling and simulation: Methodology and illustrative example. Urban Clim. 2017;21:154–72.
    • 55. Khanh DN, Varquez ACG, Kanda M. Impact of urbanization on exposure to extreme warming in megacities. Heliyon. 2023;9(4):e15511. pmid:37128331