Urbanization-induced warming amplifies population exposure to compound heatwaves but narrows exposure inequality between global North and South cities

The study addresses how urbanization-induced warming alters population exposure to compound heatwaves in cities worldwide and how it affects exposure inequalities between Global North and Global South cities. Motivated by the rapid rise in urban populations and the intensifying frequency, duration, and severity of heatwaves, the authors note limitations in prior global assessments that often: (1) focus on daytime or nighttime heatwaves independently rather than compound events; (2) rely on coarse-resolution climate or reanalysis data that inadequately represent urban processes; and (3) overlook the distinct contribution of urbanization-induced warming to heat exposure and to North–South inequalities. The purpose is to quantify urbanization’s independent warming effect on compound heatwave hazards and population exposure at urban and sub-urban (high vs low urbanization) scales across 1028 cities, and to compare hypothetical scenarios without urbanization warming to real scenarios including it. This is important for accurate risk assessment, equitable adaptation planning, and understanding environmental justice implications of urban heat exposure globally.

The paper situates its contribution within research documenting: (a) growing urban populations and the urban heat island effect that amplifies heat risks; (b) evidence of rapidly increasing urban exposure to extreme heat; (c) the greater health risks of compound heatwaves (concurrent daytime and nighttime extremes) compared to independent events; and (d) common use of heatwave frequency and more comprehensive indices incorporating intensity and duration to capture heat hazards. Prior work often used climate model and reanalysis datasets that poorly resolve urbanization effects, potentially underestimating urban heatwave risk. Station data can reveal urbanization impacts but lack global coverage. Remotely sensed land surface temperature and derived near-surface air temperature products provide high-resolution alternatives used to assess urban warming, exposure inequalities, and mitigation strategies. Gaps identified include quantifying the specific contribution of urbanization-induced warming to urban heat exposure and its effect on Global North–South exposure inequalities, particularly at sub-urban scales.

Study domain and delineations: 1028 urban areas (>100 km²) were selected from the 2018 Global Urban Boundaries (GUB) dataset. Urban and rural extents were balanced using a dynamic buffer D_buffer = (sqrt(2) − 1)*sqrt(S/π), where S is the urban area. Water, wetlands, and ice pixels (MODIS MCD12Q1 IGBP classes) were excluded. To reduce elevation-driven biases, rural pixels >50 m above or below mean urban elevation (GTOPO30 DEM) were removed. Sub-urban classification used GAIA (2018) at 30 m to compute impervious fraction per 1 km temperature pixel: >0.5 classified as high urbanization; 0–0.5 as low urbanization. Data: Daily maximum and minimum near-surface air temperature at 1 km resolution (Earth Engine community dataset; 2003–2020) derived from satellite LST and station data; MODIS MCD12Q1 land cover; GTOPO30 DEM; GAIA impervious area; Population density from GPWv4 (Revision 11). Heatwave definitions: Extended summer seasons (NH: May–Sept; SH: Nov–Mar) from 2003–2019. City-level relative thresholds derived from rural temperatures: hot day (night) if daily max (min) exceeds the 90th percentile of that calendar day computed over 2003–2019 using a ±7-day window. Three heatwave types: independent daytime (≥3 consecutive days with hot days and no hot nights), independent nighttime (≥3 consecutive days with hot nights and no hot days), and compound (≥3 consecutive days with both hot days and nights). Sensitivity tests used 95th and 98th percentile thresholds. Missing temperature values were gap-filled using the mean of the two preceding and following days. Heatwave metrics: HWF (total heatwave days/year), HWD (mean duration of events/year), HWM (mean magnitude, event-average excess above threshold), and CH (cumulative excess heat). HWM = (Σ_i Σ_j (T_ij − T0,j)) / N; CH = Σ_i Σ_j (T_ij − T0,j), where N is number of events, D_i event length, T_ij daily heatwave temperature, and T0,j threshold. Heat exposure index (HEI): HEI = HW × Pop. Three hazard formulations: (1) HW index 1 = HWF; (2) HW index 2 = HWF × HWM × HWD; (3) HW index 3 = CH. Population density from GPWv4; features normalized to 0.1–0.9 via min–max scaling. HEI computed for urban, rural, high-urbanization, and low-urbanization areas. Cities were classified into low, moderate, and high exposure via Jenks Natural Breaks (jenkspy in Python). Urbanization-induced warming scenarios and decomposition: Hypothetical (no urbanization warming): HEI_counter = HW_rural × Pop_urban. Real (with urbanization warming): HEI_urban = HW_urban × Pop_urban. Urban–rural HEI difference decomposed into three components: (i) independent impact of urbanization-induced warming = HW_urban−rural × Pop_rural; (ii) independent impact of population growth = HW_rural × Pop_urban−rural; (iii) interaction = HW_urban−rural × Pop_urban−rural. Relative contributions computed as ratios of each component to total HEI urban−rural difference (RCHW, RCPop, RCint). Inequality analysis contrasted Global North vs Global South cities (following prior classifications) using two-sided Mann–Whitney U tests (scipy). Statistical testing and sensitivity: North–South differences evaluated with Mann–Whitney U tests; robustness tested across heat thresholds (90/95/98%) and heat hazard metrics (HWF, HWF×HWM×HWD, CH), and across urbanization levels (high vs low).

- Urbanization-induced amplification of compound heatwaves: 94.5% of 1028 cities experienced amplified compound heatwaves in urban vs rural areas (2003–2019). Amplification was stronger in high-urbanization areas (5.89 × 10^3) than low-urbanization areas (2.46 × 10^1), and stronger in Global North than Global South cities (6.97 × 10^3 vs 2.78 × 10^3). Results were consistent across thresholds (90/95/98%) and heat indices. - Global exposure levels and trends: 53.9% of cities were low exposure, 40.2% moderate, 6.0% high. High-urbanization areas had more highly exposed cities (8.9%) than low-urbanization areas (1.4%). From 2005 to 2020, average global urban heat exposure risk increased by 11.8%, with larger increases in high-urbanization areas (13.0%) than low-urbanization areas (8.7%). - North–South inequality: Under hypothetical scenarios (no urbanization warming), Global South cities had significantly higher exposure than Global North cities. Including urbanization-induced warming (real scenario) narrowed this inequality due to stronger urban warming in the Global North. Example: the mean North–South exposure difference (North minus South) in urban areas changed from −2.42 × 10^−2 (hypothetical) to −1.98 × 10^−2 (real). The narrowing effect was most evident when using HWF as the heat index; nevertheless, the overall North–South disparity increased over time. - Contribution decomposition: Urbanization-induced warming accounted for 24.2% of the urban–rural HEI difference globally, with a larger contribution in low-urbanization areas (34.2%) than high-urbanization areas (25.8%). By region, the contribution of urbanization-induced warming was higher in the Global North (31.2%) than in the Global South (18.0%). Population contributions to urban–rural HEI differences were more pronounced in the Global South, reflecting larger urban populations. - Exposure distribution by region and urbanization level: Of the 46.2% of cities categorized as high or moderate exposure, 16.5% were in the Global North and 29.7% in the Global South. In high-urbanization areas, of the 53.8% high/moderate exposure cities, 20.1% were in the Global North and 33.7% in the Global South; in low-urbanization areas, of the 20.0% high/moderate exposure cities, 6.6% were in the Global North and 13.4% in the Global South.

The study demonstrates that compound heatwaves, which impose concurrent daytime and nighttime thermal stress, are more strongly intensified by urbanization than independent daytime or nighttime events, elevating urban heat exposure risks. By explicitly incorporating urbanization-induced warming using high-resolution satellite-derived near-surface air temperature, the authors show that standard assessments that omit urbanization effects overestimate the inequality in exposure between Global South and Global North cities. Stronger urban warming in the Global North narrows, but does not eliminate, the exposure gap; in fact, the overall disparity has continued to widen over time due to other drivers, notably larger urban populations in the Global South. The decomposition reveals differing drivers: urban warming contributes more to exposure differences in the Global North, while population growth exerts stronger influence in the Global South. These insights highlight the need for urbanization-aware exposure assessments and tailored mitigation strategies that consider regional urbanization patterns, climate, and demographics, and prioritize environmental justice in both Global North and South contexts.

This work quantifies the distinct role of urbanization-induced warming in amplifying compound heatwaves and urban population exposure across 1028 global cities, and shows that including this effect narrows the apparent North–South exposure inequality because urban warming is stronger in the Global North. Key contributions include: (1) high-resolution, satellite-based assessment of compound heatwaves and exposure; (2) explicit comparison of hypothetical (no urbanization warming) and real (with urbanization warming) scenarios; and (3) decomposition of exposure differences into warming, population, and interaction components. The findings underscore the necessity of integrating urbanization-induced warming into present and future heat exposure and vulnerability assessments. Future research should: extend analyses over longer climate baselines; incorporate comprehensive heat stress metrics that include humidity and wind; integrate population vulnerability (e.g., age structure) into exposure indices; and evaluate the synergistic effects of heatwaves and urban heat islands on exposure. Considering projected urban expansion, demographic aging, and rapid population growth in developing regions, accounting for urbanization will be essential for equitable, effective heat risk mitigation and adaptation planning.

- Temperature-only metric: Exposure assessment used near-surface air temperature derived from satellite LST and station data; it did not incorporate humidity or wind, which are critical for heat stress. Lower urban humidity may reduce heat stress despite higher temperatures. - Vulnerability not included: HEI used population density without explicit vulnerability factors (e.g., age, health, socioeconomic status), potentially biasing risk insights. - Data and representativeness: Near-surface air temperature reconstruction relies partly on unevenly distributed weather stations, which may introduce regional biases, though regional averaging (Global North vs South) and cross-validation were used to mitigate uncertainty. - Time window: Analysis period (2003–2019) is shorter than conventional climatological baselines (≥30 years), warranting caution when interpreting trends. - Scope of urban effects: The study focused on the amplifying role of urbanization-induced warming on compound heatwaves; the full synergistic effects of concurrent heatwaves and urban heat islands on exposure were not directly quantified.