Megacities are causal pacemakers of extreme heatwaves

Urbanization is a major driver of modern societal development, with urban areas now housing 56% of the world's population and producing a significant portion of global carbon emissions. Global climate change exacerbates the environmental challenges faced by cities, including increasingly intense and frequent heatwaves. Heatwaves have widespread adverse impacts on human health, infrastructure, and ecosystems. While natural factors like high-pressure systems and climate variability contribute to heatwaves, anthropogenic emissions from human activities significantly increase their intensity and frequency. This study uses a network approach to understand the dynamics of heatwave propagation across interconnected U.S. cities, focusing on the causal role of megacities in the generation and spread of extreme heat events. The causal analysis of human activities and their relationship with heatwaves, particularly in urban areas, remains a crucial area of research. This investigation aims to reveal how megacities act as central nodes influencing the spread of extreme heat across urban networks in the U.S.

Previous research has established the link between global climate change and the increased frequency, intensity, and duration of heatwaves. Natural contributors to heatwaves include high-pressure systems, soil moisture memory, and large-scale teleconnections like ENSO and PDO. Anthropogenic emissions, primarily from fossil fuel combustion, have contributed significantly to global warming. Studies have shown increasing trends in U.S. temperature extremes, largely due to anthropogenic forcings, with projections indicating a substantial increase in the global population exposed to hazardous warming by 2100. Existing literature also highlights the interconnectedness of urban environments and the concept of 'teleconnections', where geographically distant cities exhibit similar thermal patterns. This interconnectedness is relevant for understanding the spread of heatwaves across urban areas. While traditional methods like correlation analysis have been used, this study adopts a more advanced causal network analysis approach to uncover the underlying drivers of heatwave propagation, offering more insight than linear correlation.

This study analyzes two major heatwave events in the contiguous U.S. (CONUS) between 1980 and 2020, examining their impact on approximately 300 U.S. urban areas during the summer months (May-September). Extreme heatwaves were identified using daily maximum temperature data from the NOAA Climate Prediction Center (CPC), defining heatwaves as periods of at least 3 consecutive days exceeding the 90th percentile of daily maximum temperatures during warm seasons. Twelve heatwaves were selected for detailed analysis. Hourly air temperature data for CONUS urban areas were obtained from the Historical Comprehensive Hourly Urban Weather Database (CHUVD-H). The data included 390 weather stations within urban areas and 130 stations in nearby areas, with data quality control and gap-filling processes performed. The long-term average temperature for each hour in each urban area was removed to isolate the heatwave impacts. The Convergent Cross Mapping (CCM) method was used for causal inference. CCM, based on Taken's lag-coordinate embedding theory, analyzes time series data to determine causal relationships between variables. This method, suitable for nonlinear dynamic systems with weak or moderate coupling, helps to quantify the directional influence between the temperatures of different cities. Parameters in CCM, such as time delay and embedding dimension, were determined using the correlation integral and dimension methods. Network analysis techniques were then applied. Each of the 520 urban sites was treated as a node in a directed network, with edges representing causal connections based on the CCM results. Network metrics such as indegree (incoming links), outdegree (outgoing links), and PageRank centrality were calculated to characterize the role of cities in the heatwave network. Finally, correlations between network metrics, city population (total and density), and the causality strength were examined. Population data were drawn from the U.S. Census Bureau and WorldPop dataset.

The analysis revealed the spatial patterns of heatwaves across the CONUS, showing that megacities acted as central nodes in the heatwave network. Figures 3-5 illustrate the indegree, outdegree, and PageRank centrality of the 520 urban sites. Megacities like New York City and Chicago showed exceptionally high values for all three metrics, indicating their central role in spreading heatwave effects. These cities served as both heat sinks (receiving heat from many other locations) and heat sources (sending heat to many other areas). The high indegree values in these megacities indicate their vulnerability to heat from other urban areas, while their high outdegree values point to their influence on a large number of other urban sites. Figure 6 demonstrates a significant positive correlation between the causal outdegree (the number of cities influenced by a given city) and population metrics (total and density) for 53 large U.S. cities with populations exceeding 200,000. This finding strongly suggests a link between large populations, high anthropogenic activity, and the spread of extreme heat. The findings show that megacities act as pacemakers of extreme heatwaves, significantly influencing the network structure and the spread of extreme heat across the CONUS. The positive relationship between city population and the strength of the causal influence further supports the hypothesis that anthropogenic activities play a crucial role in generating and spreading extreme heat in urban areas.

The findings of this study directly address the research question by demonstrating the causal role of megacities in the propagation of extreme heatwaves. The identification of megacities as pacemakers offers a novel insight into the dynamics of urban heat vulnerability. The positive correlation between population and causality suggests a significant contribution of anthropogenic factors, such as waste heat and greenhouse gas emissions, to the spread of extreme heat. These results are significant as they highlight the disproportionate impact of megacities on the broader urban heat landscape, underlining the need for targeted mitigation strategies in these areas. The network approach provides valuable information for urban planning, policymaking, and stakeholder engagement, enabling the development of more effective heatwave mitigation strategies. The identification of long-range teleconnections among urban areas offers the potential for improving predictions of heatwaves, analogous to early warning systems based on climate drivers like El Niño events. The study also adds to the broader understanding of inter-municipal and inter-urban dynamics during extreme climatic events.

This study demonstrates that megacities serve as causal pacemakers for extreme heatwaves in the CONUS. The strong positive correlation between population size and causality strength indicates a significant influence of anthropogenic factors. The findings highlight the need for targeted mitigation strategies in megacities and offer insights for improving heatwave prediction and adaptation strategies. Future research could focus on expanding this analysis to other regions and exploring the effectiveness of various mitigation strategies in reducing the causal influence of megacities on heatwave propagation.

The study primarily focuses on summertime heatwaves in the CONUS, limiting its generalizability to other seasons or regions. The accuracy of the findings relies on the quality and availability of data, and the choice of network analysis methods could influence the interpretation of results. Further research is needed to account for other confounding factors that might affect heatwave dynamics.