The increasing frequency, intensity, and duration of heatwaves due to climate change pose significant threats to ecosystems and human societies, hindering sustainable development. Extreme heat exposure poses a serious risk to human health, increasing morbidity and mortality rates. Cities, housing 55% of the world's population and projected to house 68% by 2050, are particularly vulnerable due to the urban heat island effect, which amplifies heat risks for city dwellers. Previous research shows a substantial increase in the global urban population exposed to extreme heat. Accurate assessment of population exposure to extreme heat is crucial for developing effective mitigation strategies and improving urban livability. Evaluating extreme heat exposure requires understanding both the thermal environment and population distribution. The Clinton Risk Triangle framework (Hazard-Exposure-Vulnerability) highlights this. While many studies use daytime or nighttime heatwaves, compound heatwaves (simultaneous extreme daytime and nighttime temperatures) are more hazardous to human health. Various heatwave indicators, such as heatwave frequency (HWF), are employed, but these may not fully capture the complexities of urban heat hazards. Urbanization alters the surface energy balance, leading to higher temperatures in urban areas compared to surrounding rural areas (urban heat island effect). The combined effects of heatwaves and urban heat islands exacerbate thermal threats. While previous studies have examined the impact of urban warming on heat exposure, the specific influence of urbanization-induced warming remains inadequately explored and is often overlooked in assessments. Global climate model data and reanalysis data have limitations due to their coarse spatial resolution, failing to accurately capture urbanization effects. Meteorological station data suffers from limited spatial distribution. Satellite-based surface temperature data provides high spatial and temporal resolution, making it a valuable tool for capturing urbanization effects. This study aims to address the gap in understanding the impact of urbanization-induced warming on heat exposure and the inequality of heat exposure between cities in the global South and North, focusing on compound heatwaves using a satellite-derived near-surface air temperature dataset.

Existing literature extensively documents the escalating global risks associated with heatwaves, highlighting their severe impacts on both ecosystems and human health. Studies show a marked increase in heat-related mortality and morbidity, particularly in urban areas. The urban heat island effect, a well-established phenomenon, significantly contributes to heightened heat exposure in cities. Previous research has assessed global urban population exposure to heatwaves, often using daytime or nighttime heatwaves as indicators. However, the impact of compound heatwaves, characterized by simultaneous extreme daytime and nighttime temperatures, has received comparatively less attention, despite their demonstrably greater health risks. The literature also reflects a variety of methods for quantifying heat exposure, including heatwave frequency, intensity, and duration. Existing studies have noted disparities in heat exposure between global North and South cities, with the global South generally exhibiting higher levels of exposure. However, the influence of urbanization-induced warming on this inequality has been understudied. Many studies utilize climate model data, but these models often lack sufficient resolution to capture the unique thermal characteristics of urban environments. The use of satellite-derived temperature data offers a more spatially precise assessment of urbanization’s influence. Overall, the literature provides a foundation for this study by highlighting the critical need for a comprehensive analysis of heat exposure that incorporates the complexities of urbanization and the unique risks posed by compound heatwaves.

This study utilizes a satellite-derived near-surface air temperature dataset (1 km resolution, 2003-2020) to assess heat exposure in 1028 cities globally. Urban and rural areas were delineated using the 2018 Global Urban Boundaries (GUB) dataset, with dynamic buffer distances calculated to ensure comparable area sizes. Water bodies, wetlands, and ice pixels were excluded. To account for elevation effects, pixels in rural areas significantly deviating from the average urban elevation were removed. The 2018 global artificial impervious area (GAIA) dataset was used to classify urban areas into high and low urbanization zones. Heatwaves (compound, independent daytime, and independent nighttime) were identified for the extended summer season (May-September in the Northern Hemisphere, November-March in the Southern Hemisphere) using a relative temperature threshold (90%, 95%, and 98% percentiles) based on the average maximum/minimum temperatures in the city relative to rural areas. Heatwave frequency (HWF), averaged heatwave magnitude (HWM), average heatwave duration (HWD), and cumulative heat (CH) were calculated to characterize heatwave events. A heat exposure index (HEI) was developed by multiplying heat hazard (using HWF, HWF×HWM×HWD, or CH) and population density from the Global Population Grid dataset (GPWV4). Features were normalized. The Jenks Natural Break method was used to classify cities into high, moderate, and low heat exposure categories. To analyze urbanization-induced warming's effects, a counterfactual scenario (no urbanization-induced warming) was constructed by assuming urban heat hazards equal to rural areas, allowing for a comparison with the actual heat exposure. The urban-rural difference in HEI was decomposed into three components: independent impact of urbanization-induced warming, independent impact of population growth, and their interaction. The Mann-Whitney U-test was used to compare heat exposure differences between the global North and South cities.

The analysis of 1028 cities revealed that more than half (53.9%) had low heat exposure risk, while 6.0% and 40.2% faced high and moderate risks, respectively. High urbanization areas showed significantly higher heat exposure (8.9% high exposure) than low urbanization areas (1.4%). Globally, urban heat exposure risk increased by 11.8% from 2005 to 2020 (13.0% in high urbanization areas vs 8.7% in low urbanization areas). Urbanization amplified compound heatwaves in 94.5% of cities, more strongly in high urbanization areas (amplification effect 5.89 × 10³ vs 2.46 × 10¹ in low urbanization areas). The amplification effect was stronger in global North cities (6.97 × 10³ ) than in global South cities (2.78 × 10³). Under a hypothetical scenario ignoring urbanization-induced warming, global South cities showed higher heat exposure than global North cities. However, considering urbanization-induced warming, this inequality narrowed due to stronger warming in global North cities. The contribution of urbanization-induced warming to heat exposure was 24.2% overall, higher in low urbanization areas (34.2%) than in high urbanization areas (25.8%). This contribution was significantly stronger in global North cities (31.2%) than in global South cities (18.0%). The study also found that compound heatwaves showed stronger increasing trends than independent heatwaves, especially in HWM and CH. The urban HWF trend in compound heatwaves was approximately double the rural trend, highlighting the disproportionate impact of urbanization on compound heatwave events.

The study's findings demonstrate that urbanization significantly amplifies the risk of compound heatwaves, a particularly hazardous form of extreme heat. The stronger amplification in global North cities, despite initially higher heat exposure in the global South, underscores the complex interplay between climate change, urbanization, and population distribution in shaping heat exposure patterns. This highlights the need to consider urbanization-induced warming in assessments. The narrowing of the inequality in heat exposure between the global North and South when accounting for urbanization-induced warming suggests that current approaches which do not account for this factor might be overestimating the disparity. The decomposition of the urban-rural difference in heat exposure into the independent contributions of urbanization-induced warming, population growth and their interaction revealed that urbanization-induced warming plays a significant role, particularly in the global North. The results emphasize the importance of incorporating urbanization-induced warming into assessments of heat-related risks and informing policy interventions.

This study demonstrates the significant role of urbanization-induced warming in amplifying population exposure to compound heatwaves, particularly in high-urbanization areas. While the global South faces higher heat exposure in a scenario without considering urbanization effects, the inclusion of urbanization-induced warming shows a narrowing of the disparity between global North and South cities. The stronger contribution of urbanization-induced warming in global North cities is notable. This study underscores the critical need to incorporate urbanization-induced warming into future heat exposure and vulnerability assessments to ensure more accurate and equitable risk evaluations. Future research could investigate the impact of various adaptation and mitigation strategies in reducing the heat exposure differences between global North and South cities, as well as exploring other environmental factors like humidity in addition to temperature.

This study primarily focuses on near-surface air temperatures derived from satellite data. While this approach offers high spatial resolution and global coverage, it does not fully capture the microclimatic variations within urban areas and the complexities of the urban heat environment. The study uses population density as a proxy for exposure, but this might not fully reflect the vulnerability of different population subgroups to heat stress. The time frame of the study (2003-2019) is relatively short, limiting the analysis of long-term temporal trends. Finally, other environmental factors, such as humidity and wind speed, that significantly influence heat stress were not directly included in the heat index calculation.