Introduction

Cities and urban areas generate over 60% of global greenhouse gas (GHG) emissions1, and are critical to tackling climate change2,3. Despite the leadership of many cities and city networks4, their climate actions remain uneven and insufficient to meet global targets5. There is an operationalization gap between the promise and practice of city climate action6. Regional differences in the extent and ambition of city climate commitments5 and in their resilience to economic pressures and crises risk widening this gap, particularly where capacity and resources are limited7. Addressing these disparities is essential for scaling city climate action worldwide, especially in the global South, where rapid urban growth and climate vulnerability are converging2.

Scaling city climate action requires understanding what drives the adoption of climate policies and actions in diverse urban contexts. These factors include geographic, institutional, political, and socio-economic conditions, which vary at all scales, but especially between regions. Understanding the conditions and drivers that allow cities worldwide to adopt climate policies and implement climate actions can enhance their uptake and effectiveness. If we understand what factors are more relevant in specific contexts, practitioners can design regionally appropriate climate strategies and scale climate actions in ways that leverage rather than trip over regional particularities.

Despite the importance of understanding what drives city climate action in diverse contexts, the global picture remains fragmented and incomplete8,9. Research on the factors that drive city climate action is geographically and methodologically uneven10,11. Global studies typically examine a narrow subset of factors related to local leadership, capacity, and willingness to act, as well as city climate networks9,12, although recent studies suggest that, for example, environmental factors and local business support may also be important7. National studies suggest regional particularities such as how administrative practices and political will may have the most impact in Brazilian cities13, while public support, population, density, and education are important in US cities14. However, few studies have comprehensively studied these factors or compared them across regions, especially at a global scale, and disproportionately focus on North America and Europe9,14. Historically, in the context of climate adaptation, large-N quantitative studies, which are valuable because they allow for wider comparability and generalisation, have been rare in the global south, where research has tended to focus on single case studies15,16. To build a more effective, globally-relevant evidence base, we must move beyond one-size-fits-all explanations and account for the contextual complexities and place-based nuances that drive or hinder progress.

Expanding on a comprehensive framework originally developed for U.S. cities by Yeganeh and colleagues14, we address this gap by systematically investigating the factors that are associated with city climate action. We identify the factors that determine whether cities adopt, continue, or increase climate policies and actions (from hereon we collectively refer to these various choices as “climate action”). We analyze how these factors have been studied across regions and research methods, and the extent to which methodological choices interact with geographic context. We also analyze how consistently and strongly these factors are associated with climate action within and across regions. In contrast to earlier studies, which have largely focused on cities in the global North9,12,14,17, our study examines drivers of urban climate action across both developed and developing countries. While North-South research gaps remain relevant, we move beyond this binary to identify patterns across world regions. Given the limited studies from the global South, regional claims should be made cautiously. Still, recognising regional similarities and differences can inform efforts to scale urban climate action.

This paper seeks to move beyond North-South disparities in research to reveal a more regionally nuanced picture about the factors linked to city climate action, although claims about these differences need to be qualified by the paucity of studies in the global South. Rather than transferring insights derived mainly from Northern cases, scaling city climate action globally requires addressing North-South research disparities and tailoring interventions to local contexts. Attending to these similarities and regional differences will help generate more context-sensitive, targeted ways to enable cities to scale climate action.

Results

Geographic distribution and methods of studies

Of the 103 articles in our review that identified factors related to city climate action (see Table S4), most studies were based on cities in Europe and North America, respectively accounting for 36% and 34% of all studies (Fig. 1). We only found 14% of studies focused on Asia, 7% on Latin America and the Caribbean, 7% on Africa, and 3% on Oceania. Across all studies, cities in 108 countries were mentioned; 75% were OECD countries, and 25% were non-OECD countries. We also found that most studies (73%) focused on cities in a single country or region, 18% focused on multiple regions, and 8% had a global focus.

Fig. 1: Number of studies by geographic region and type of research methods.
figure 1

Circle size is scaled to the number of studies on the factors affecting city climate action in each region.

Full size image

Studies used a variety of methods for data collection and analysis. Of the 103 studies, 55% used only qualitative methods, 17% used only quantitative methods, and 27% used both qualitative and quantitative (mixed) methods. Most (58%) qualitative studies used primary data sources and small sample sizes, while most (67%) quantitative studies used secondary data and large sample sizes; mixed methods studies typically (61%) used both primary and secondary data with intermediate sample sizes. Many studies used more than one method of analysis. Qualitative studies often (51%) used document analysis. Quantitative studies primarily (79%) employed a variety of regression analyzes, while some (11%) used basic statistical tests; none used causal inference methods. Mixed methods studies combined quantitative and qualitative methods in diverse ways to generate insights. These methodological differences are important because they varied by geographic region (Fig. 1). Qualitative and mixed methods studies were common across all regions except Oceania. In contrast, quantitative-only studies appear primarily in relation to cities in North America, Europe, and global evaluations, with only one study focused on Asia18.

Factors linked to city climate action

Here we report on the range of factors found to be associated with city climate action (from hereon “factors identified”). From the 103 relevant articles, we identified 551 instances of factors linked to climate action, which were categorized according to the descriptions in the taxonomy (for more information about these categories and factors, see Table S3 in supplementary material). In addition to the 90 factors named in the taxonomy, we identified seven unique factors which we grouped according to the category descriptions. These were social capital, ability to influence energy use and GHG emissions, research facilities, legitimacy of ideas from knowledge producers, COVID-19, vertical and horizontal collaboration, and interaction of factors. The number of factors identified increased over time, with a large increase after 2016, especially in the category government capacity.

Overall, the category most frequently linked to city climate action was government capacity, accounting for 34% of the 551 factor occurrences identified (Fig. 2). Other commonly identified categories of factors included government interests (16%) and intergovernmental factors (14%). Community support (11%) and government structure (9%) were less frequently identified. The categories least often linked to climate action were the built environment (2%), environmental factors (6%), and city structure (7%). Two articles identified interactions among factors, showing, for example, how external funding, risk perceptions, and local leadership combine to enable climate action19.

Fig. 2: Factors associated with city climate action organized by category.
figure 2

Bars indicate the number of times a factor was identified.

Full size image

There were differences within and between categories in terms of factors identified (Fig. 2). In some categories (government capacity, government interests, and community support) most or all factors were identified, while in other categories (environmental factors, city structure, and built environment) more than half of all the factors in each category were identified once or not at all. In the categories for which factors were less frequently identified, there were gaps in the factors identified. For example, in the category city structure, population and income were frequently identified, but factors related to diversity and justice (e.g., social vulnerability, education, inequality, racial diversity) were rarely or never identified. Likewise, environmental factors such as coastal area, mileage, or proximity and severity of weather and natural events were identified more often, while factors such as water quality, precipitation, and drought were rarely or never identified.

At the individual factor level, occurrences varied widely. A small subset of factors (15 of the 90 factors in our taxonomy) accounted for 348 (63%) of the instances in which factors were identified, whereas over half of the factors were identified 0–2 times across all studies (Fig. 2). Of these, staff capacity, stakeholder involvement, and policy and institutional frameworks were the most frequently identified factors. Fiscal capacity, city climate network membership, technical capacity, and attitudes and motivations of politicians were also frequently identified.

Factors and geography

There was clear regional variation in how often different categories of factors as well as individual factors were linked to city climate action. All categories of factors were identified in Europe, North America, Asia and Oceania. In contrast, in Latin America and the Caribbean, no factors in the category built environment were identified, while in Africa, no factors in the categories city structure and built environment were identified. Global studies lacked attention to government structure and community support.

In terms of individual factors by region, we found 226 instances of individual factors (38%) identified in Europe, followed by 32% in North America, 10% in Asia, 8% in Latin America and the Caribbean, 6% in Africa, 2% in Oceania, and 3% in global studies.

The factors that were most frequently identified were similar in the global North and South, but there were important differences between specific regions (Fig. 3). The five factors most frequently identified in the global North (North America and Europe) were policy and institutional frameworks, staff capacity, stakeholder involvement, city climate network membership, and fiscal capacity. The five factors most frequently identified in the global South (Latin America, Africa, Asia, and Oceania) were fiscal capacity, staff capacity, stakeholder involvement, technical capacity, and (tied for fifth) administrative structure and attitudes or motivations of politicians and government staff.

Fig. 3: The five factors associated with city climate action that are most frequently identified in each geographic region and in global studies.
figure 3

Circles sizes reflect the number of times a factor was identified. Factors shown are those with the greatest absolute count in a region. When more than one factor had the same count in a region, the factor with the higher regional concentration ratio is shown, calculated as Regional concentration ratio = [(count for factor in that region) / (total count of all factors in that region)] / (count for factor overall / count for all factors overall).

Full size image

Overall, the factors that were identified most often in each region were largely similar. North America and Europe shared four of their top five factors, including policy and institutional frameworks, staff capacity, stakeholder involvement, and city climate network membership. Similarly, North America and Asia shared four top factors (fiscal capacity rather than city networks membership). Africa and Asia also shared four top factors, including technical capacity, staff capacity, fiscal capacity, and policy and institutional frameworks. Thus, the top factors identified in the global North (North America and Europe) and the global South (Latin America, Africa, Asia, and Oceania) were almost identical, with the exception that city climate network membership was a top factor in the global North but not in the global South.

However, some regions had a distinct focus (Fig. 3). Two of the top factors identified in Latin America (administrative structure and form of government) were not part of the top ten factors identified overall and were not focused on in the global North. Technical capacity was a top factor in Asia and Africa, but nowhere else. City climate network membership was a top factor in North America and Europe, but not a focus in Latin America and Africa. Finally, three of the five top factors identified in Oceania (extreme event frequency or type, city economic structure, and population) were not top five factors in any other region. However, since our review identified few studies in the global South, these geographic differences need to be qualified because the factors that were most frequently identified in some regions were only identified 1–3 times.

Consistency and direction of association of factors

Of the 551 instances where factors were identified, the factor’s direction of association on climate action was positive 336 times (61%), negative 189 times (34%), and not identified 26 times (5%). Of the ten factors identified most often, direction of association was consistent across regions (positive in all but one region, mixed or negative in the other) for city climate network membership, attitudes or motivations of politicians and government staff (i.e., pro-environmental was positive, anti-environmental was negative), and population. Direction of association was also consistent across regions for intergovernmental relationships and dependence (i.e., supportive relationships were positive, lack thereof was negative); income; (intergovernmental) level of involvement; coastal area, mileage, or proximity; issue salience to government staff; and issue salience to the public. Consistent is defined as all but one region having the same overall direction of association, or all regions having the same overall direction of association when direction was identified in only 3 regions. For all other factors, the direction of association was inconsistent across regions. The direction of association was not identified for two factors.

Factors and their strength of association

Here we report on how strongly factors were found to be associated with city climate action. While some studies identified how strongly a factor was linked to climate action, over two-thirds did not (Table S5). Strength of association, within and across regions, was calculated as the total number of instances strength was reported as very strong and strong minus weak and very weak. Overall, the factors that were most frequently found to be linked to city climate action (e.g., the 10 most frequently identified) also had the strongest associations. Fiscal capacity, city climate network membership, and attitudes or motivations of politicians and governments had the strongest relationships, with studies identifying their influence as very strong or strong more often than moderate, weak, or very weak. Other factors that were frequently identified and whose strength of association was very strong or strong were stakeholder involvement and policy and institutional frameworks. None of the 30 factors that were most frequently identified had weak associations with city climate action. Of the 30 factors that were rarely identified (1–2 times), 8 had strong associations, 7 had weak associations, and the strength of association was not identified for 15. The strength of association was not specified for 22 of the 77 factors found to influence city climate action.

Beyond this general association between factors more often found to influence city climate action and those that had a stronger influence, we found several exceptions. One exception is that, of the factors most often identified, population had strong associations with city climate action an equal number of times as it had moderate, weak, or very weak associations. Legal and regulatory structure, human resources, and country action plan or GHG target also had mixed strength of association, with equal numbers of studies identifying their influence as strong and as weak. Another exception is that some factors that were identified less often (3–6 times) had strong associations, including coastal area, mileage, or proximity; electoral features; issue salience to the public; (community) attitude toward climate change; mayoral leadership; transportation infrastructure; and education.

Factors and their strength of association by region

Some of the factors that were strongly associated with city climate action were similar across regions (Fig. 4). Fiscal capacity, stakeholder involvement, and technical capacity had strong associations in all or almost all regions. City climate network membership and public support had strong relationships in North America and Europe, but there was less or no data about their influence in other regions. Policy priorities and administrative structure had strong associations in Europe, Africa, and Latin America, with data gaps in other regions.

Fig. 4: The five factors most strongly associated with city climate action in each geographic region and in global studies.
figure 4

Circles sizes reflect the number of times a factor was identified. Factors shown are those with the greatest overall strength of association in a region, calculated as Strength of association = [very strong + strong] - [weak + very weak]. All factors shown have strong association overall except the three factors in Oceania, for which strength of association is weak.

Full size image

We also found important regional differences and gaps (Fig. 4). In North America, issue salience to the public and education had strong associations, while policy priorities and population had strong associations in Europe. Mayoral leadership was important in Asia, form of government was important in Latin America, and legal and regulatory structure and administrative structure were both important in Africa and Latin America.

In other regions, we found major gaps that limit conclusions about which factors are most important. In Oceania, (1) population; (2) income, poverty, or wealth; and (3) extreme event frequency or type were the only factors for which strength of association was identified, and it was weak for all three factors. In global studies, city climate network membership and city economic structure were the only factors for which strength of association was identified and it was strong for both factors.

Comparing frequency factors were identified and their strength of association by region

Comparing how often a factor was identified with how strongly it was related to city climate action, the factors that had strong regional associations overlapped with the factors that were identified most frequently most of the time (57%). While this result suggests that factors that are more often identified to be related to city climate action do have greater influence, we found exceptions and/or gaps in all regions that limit this claim. In Asia and Latin America, four of the five most frequently identified factors were also factors that had the strongest associations. However, the strength of association for policy and institutional frameworks was not identified in any studies that focused on Asia, and strength was not identified for attitudes or motivations of politicians and government staff in Latin American studies. In North America, Europe, and Africa, the most frequently identified factors also typically had strong associations, but strength was mixed, weak, or not reported for 1–3 of the most frequently identified factors in each region. In Europe, all the most frequently identified factors had strong associations except staff capacity, which had a mixed influence. In North America, three of the most frequently identified factors had strong associations, but policy and institutional frameworks had mixed associations and strength of association was not identified for staff capacity. In Africa, three of the factors that were most often identified had strong associations, but staff capacity had weak influence, and strength of association was not identified for policy and institutional frameworks. In Oceania and in global studies, we had no data about strength of association for most of the factors that were most often found to have an influence.

While these comparisons generally confirm regional differences about factors that are most important, we found no clear differences between the global North and South. We found that city climate network membership was identified to be related to city climate action more often in the global North, and technical capacity was identified more often in the global South, but we did not find similar North-South differences in terms of strength of association. In the global South, the strength of association of city climate network membership was identified once as high, but it was not identified to have an influence in Oceania or Latin America and only once in Africa where strength was not specified. In the global North, the strength of association of technical capacity was identified three times, each time as high. Thus, while some factors have been found to be related to city climate action more often in the global North or South, these factors did not have stronger associations in the global North or South.

Discussion

Our analysis reveals two key insights. First, the literature is geographically and methodologically skewed in ways that may hamper efforts to scale city climate action. Few studies examined multiple regions or adopted a global perspective, and there were stark differences between the global North and South. Consistent with previous research9, we showed that over two-thirds of all reviewed studies (n = 72, 70%) focused on North America or Europe, while comparatively few studies (n = 31, 30%) focused on Latin America and the Caribbean, Africa, Asia, or Oceania. Overall, the literature is strongly skewed towards cities in the global North, whereas much less is known about cities in the global South.

Since we know more about the factors associated with city climate action in the global North, there is a greater opportunity to use those insights to scale climate action in the global North. However, if those insights are used to inform city climate action in the global South, there is a risk of not considering contextual nuances in the global South. North-based knowledge may embed socio-political, economic, infrastructural, or ecological assumptions that either favor or are unique to the global North, and that may not translate to other contexts. To mitigate these risks, research needs to attend to regional differences and contextual nuances and the ways they influence city climate action.

These North-South research differences are compounded by gaps in the methods used. A balance of qualitative, quantitative, and mixed methods was used in studies in North America and Europe, as well as global studies. In contrast, quantitative, large-N studies were largely absent from studies in Latin America and the Caribbean, Africa, and Oceania. This deficit of studies with strong and complementary methods in the global South limits comparability and ability to make inferences about scaling.

There is a pressing need for more quantitative, large‑N studies that can generate comparative insights about how climate action can be scaled within and across regions. While qualitative studies are essential to capture contextual nuance, the paucity of large-N studies of cities in the global South limits comparability and scaling within and across regions. A recent study by O’Garra et al.7 offers a strong example of how such analyzes can be conducted effectively. Drawing on a global dataset of 793 cities, half of which are in the global South, the authors used multilevel regression with extensive robustness checks to identify the key factors that influenced the durability of city climate commitments during COVID-19. This example illustrates how to compare the efficacy and applicability of factors both within and between regions. Such quantitative studies would complement the contextually rich qualitative insights about city climate action typical of studies in the global South.

Research on factors associated with city climate action has increased over time, yet it remains disproportionately focused on government capacity (e.g., staff capacity, policy and institutional frameworks, and stakeholder involvement). In contrast, researchers have paid much less attention to environmental factors (e.g., coastal area, temperature change, and precipitation) and to the built environment (e.g., transportation infrastructure, manufacturing dependence). This imbalance may impede the scaling of climate action because factors in these different categories are related, with potential complementarities. Factors related to government capacity help us understand cities’ overall ability to act on climate change, while the built environment and environmental factors are related to how vulnerable cities are to climate change and how much they may need to adapt, issues especially relevant to scaling climate action in the global South2,20.

Our second key finding is that some factors are globally important, while others have distinct regional importance. We found that some factors are more consistently linked to city climate action across regions than others. Direction of association was consistent across regions for nine factors: city climate network membership; attitudes or motivations of politicians and government staff; population; intergovernmental relationships and dependence; income; (intergovernmental) level of involvement; coastal area, mileage, or proximity; issue salience to government staff; and issue salience to the public. These consistencies can be leveraged to scale city climate action across regions, while attempts to scale city climate actions based on factors with less consistent associations across regions would require attention to regional differences.

An overlapping but different set of factors was most strongly associated with city climate action, some globally and others regionally. We found that fiscal capacity, stakeholder involvement, and technical capacity had strong associations across all or almost all regions. City climate network membership, public support, policy priorities, and administrative structure also had strong associations in many regions. Other factors were strongly associated with city climate action in specific regions. For example, education was important in North America, mayoral leadership was important in Asia, and form of government was important in Latin America. Only city climate network membership was found to be strongly and consistently linked to city climate action across regions. Building on existing research9,12,14, our results provide a global understanding of the factors that matter in different regional contexts.

These results can be used to connect and align diverse contexts with diverse city climate actions. To scale climate actions, practitioners might leverage regional similarities while attending to regional differences. For example, policy-makers could replicate climate policies among African, Asian, or Latin American cities that leverage stakeholder involvement and fiscal capacity, but modify them to account for regional differences (e.g., the importance of form of government in Latin America).

Comparing how often factors were identified with their strength of association, we found that researchers identified different factors in different regions, but those regional foci only partially reflected the factors most strongly associated with city climate action. Overall, we found that factors that are most often identified to be associated with climate action are also those that are most important, but with several exceptions. Importantly, while we found North-South differences in how often different factors were identified, these factors had the same strengths of association in the global North and South. This result reveals not only major gaps in knowledge in the global South but also that differences in how often factors are identified do not always mean those factors are more important. One surprising result was that there was greater alignment between the factors most often identified and the factors that were most important in Asia and Latin America than in North America. We also identified gaps where specific factors were frequently identified to be associated with city climate action in a region, but none of those studies identified the factor’s strength of association. For example, in North America, staff capacity was frequently identified, but data on its strength was absent, while education was rarely identified but had a strong association with climate action. Similarly, in Asia, policy and institutional frameworks were frequently identified but data on their strength was lacking, while mayoral leadership was rarely identified but had a strong association.

These exceptions and data gaps qualify and limit what we can conclude. Since strength of association was not specified in most (68%) of the instances in which factors were identified in relation to city climate action, we have an incomplete picture of what factors matter the most. Our findings that some factors are strongly associated with city climate action in some regions do not necessarily mean that those factors are not important in other regions; instead, there was often no data about strength of association in the other regions. Moreover, given data limitations, we cannot conclude that other factors for which strength was not reported do not also have strong relationships. These data gaps limit our ability to draw more robust conclusions about what factors are important, especially in Oceania and in global studies.

Moreover, while we identified factors linked to city climate action, other factors not identified may also be associated with climate action. Since our results are based on factors reported to be associated with city climate action, we were unable to identify and distinguish between factors studied and found not to influence city climate action versus factors not studied that, if studied, would have been found to have an influence. To disentangle what is studied from what is identified as having an influence, researchers would need to collect data on null findings about factors that were studied but found not to influence city climate action.

While our results tell us what is important in different contexts, they do not tell us how or why those factors are important. Our analysis suggests several possible reasons for this regional variability. These reasons include: (1) local/regional context, (2) data availability and quality, (3) how factors influence climate action, and (4) researcher biases.

First, some factors are more relevant in certain contexts. For example, extreme weather events are clearly important for climate adaptation in Oceania. Other factors better reflect the characteristics, politics, and priorities of some regions than others. For instance, Lehmann and colleagues21 show that, while the overall categories of barriers to climate adaptation are similar across regions, the focus on government structure in Latin America is linked to, for example, a lack of coordinating organizations. Our analysis reveals which factors are important globally and in certain regions, but not why. Building on this analysis, if we understand how and why some factors are more relevant in specific contexts, practitioners can design regionally appropriate climate strategies and more effectively scale climate actions. These differences need to be qualified and further explored because, as we have shown, knowledge about the factors that influence city climate action is more complete in the global North, whereas there are important gaps in the global South. Do these commonalities reflect genuine, shared challenges and approaches that could inform scaling, or are they merely a product of uneven research and a lack of studies that compare factors across these regions?

Second, differences in factors studied and those linked to city climate action may be related to differences in the type, quality, and availability of data. For example, quantitative data on population and income are readily available, whereas data on diversity is less so. Our results show both a lack of quantitative analyzes and that researchers often collect their own data in the global South, highlighting limited access to quantitative data on which to base analyzes. This gap is, in many instances, compounded by limited capacity to generate, manage, and standardize city climate data21. A study of Chilean cities overcame these constraints by using targeted data inquiries (correspondence, phone calls) to supplement government data and construct a panel database22. Researchers and city practitioners can improve city climate data through collaboration, capacity building on data management, and data-driven governance initiatives such as the Global Climate Action platform10.

Third, the relationship between factors and city climate action is multifaceted and complex. Disentangling the primary drivers of city climate action from secondary and intervening variables remains a challenge. For instance, Fraser and colleagues23 showed that disasters, risk, and exposure are correlated, but the causal relationships between these factors are complex and may depend on other intervening factors such as city capacities24,25. In a U.S. study, Dilling et al. 19 demonstrate how researchers can use mixed methods to explore these complex relationships. Dilling et al. combined qualitative and quantitative methods (hypothesis testing using regressions and in-depth case studies) to show how multiple factors interact and act in combination to enable cities to respond to weather- and climate-related risk. Researchers could build on these studies to explore how factors interact to enable or impede city climate action, and whether these interactions differ in different contexts. For example, how do city climate networks interact with different forms of government and administrative structures? Are these interactions different in Europe versus Latin America, and what does this tell us about the potential for scaling within and across regions?

Finally, researcher assumptions and biases about what matters in different contexts may influence what is studied and how. For example, while our results indicate that the factors linked to climate action and their strength of association are related, an outstanding question is why. Do researchers tend to confirm that the factors they study have a strong influence, whereas studies reporting low strength of association either tend not to be submitted or tend to be rejected? Our analysis identified the population as an exception. Is the population less subject to biases and why (e.g., data availability and quality)? And how can researchers account for or avoid these biases? For example, are there quantitative methods with more objective ways of measuring strength of association? To leverage these insights and better support scaling of city climate action, researchers should be explicit about why certain factors are studied and why they matter.

Leveraging these insights and answering these questions may enable researchers and practitioners to develop targeted city climate actions and to effectively scale them.

Methods

We conducted a systematic review to identify the factors that influence or help to explain the adoption, continuation, increase, or change of city/urban/municipal climate mitigation and adaptation policies, commitments, or actions (collectively referred to as “city climate action”). In accordance with the Preferred Reporting Items for Systematic Review Recommendations (PRISMA)26, the following identification, screening and inclusion procedures were used (see Fig. S1 in Supplementary Material):

Identification

To identify literature, we conducted a systematic search in Web of Science using the following search terms and criteria:

  • Indexes = SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH, ESCI

  • Timespan = 2014–2021

  • ((AB = (cities OR urban OR mayor* OR “local government” OR municipa*) OR TI = (cities OR urban OR mayor* OR “local government” OR municipa*)) AND (AB=climat* OR TI=climat*) AND (AB = (action OR polic* OR commitmen* OR adap* OR mitigation OR resilien* OR “low carbon” OR risk OR initiativ* OR smart OR sustaina* OR plan OR green) OR TI = (action OR polic* OR commitmen* OR adap* OR mitigation OR resilien* OR “low carbon” OR risk OR initiativ* OR smart OR sustaina* OR plan OR green)) AND (AB = (adopt* OR uptake OR driver OR predictor) OR TI = (adopt* OR uptake OR driver OR predictor)))

  • AND LANGUAGE: (English)

  • AND DOCUMENT TYPES: (Article OR Book OR Book Chapter OR Data Paper OR Editorial Material OR Proceedings Paper)

Our search identified 1792 journal articles. A manual search using identical terms in Google Scholar identified 14 additional articles that were added for a total of 1806. These results were imported into Mendeley for screening.

Screening

Six coders screened the articles based on their relevance to the research focus. Coders reviewed titles and abstracts and excluded articles that clearly did not fit the subject of “cities” and “climate action/measures”; and generally factors influencing climate action in urban contexts (from hereon “the search criteria”). Coders retained articles with titles and abstracts that might fit the search criteria. Screening produced 125 articles that were clearly relevant and 147 articles that were potentially relevant, for a total of 272.

Inclusion

The coders assessed the eligibility of the 272 full-text articles for inclusion based on the criteria that articles explicitly or implicitly identified factors affecting city climate action. The 272 articles were divided among six coders. Each coder read the full text of the article to identify whether factors affecting city climate action were explicit (relevant) or implicit (marginally relevant), or neither (not relevant). Of the 272 articles, 125 (46%) were relevant, 60 (22%) were marginally relevant, and 87 (32%) were not relevant. Coders used a Google Form to record the relevance of each article and then, if relevant or marginally relevant, extracted data on the article and the factors it identified. When assessment differed between initial title and abstract screening and assessment of full paper for eligibility regarding relevance, a third coder validated the assessment. Assessment of eligibility and validation produced 103 articles that identified factors associated with city climate action, all of which had been identified as clearly relevant in title and abstract screening. Marginally relevant articles were checked to ensure no relevant articles were missed.

Data extraction and coding

Data was extracted and coded in two steps. In Step 1, the 272 full-text articles were divided among six coders to extract descriptive statistics and details about factors affecting city climate action (see Supplementary Material Table S1). In Step 2, using the data extracted in Step 1, factors were identified, categorized, and specific details about those factors were extracted (Table S2). From the 103 relevant articles, a total of 551 factors were identified. Factors were categorized according to the modified Yeganeh14 taxonomy (Table S3), and categorizations were validated by three independent coders. If a factor was not clearly named in the taxonomy (seven instances), it was categorized according to the category description and included as a unique factor before being validated by three coders.

Data synthesis and analysis

Data were synthesized and analyzed at both the article and factor levels according to the taxonomy of factors as well as research themes and questions predetermined by the research team. First, data on articles and factors were synthesized according to the taxonomy of factors, first by the categories and then the factors identified. Second, the data was synthesized and analyzed according to several research themes and questions predetermined by the research team, including: geographic focus (world region or global and OECD versus non-OECD); the focus of study (climate adaptation and/or mitigation) and emergent themes (resilience, nature-based solutions, maladaptation); and study design (methods of data collection and analysis, including qualitative and/or quantitative approach, specific methods of analysis, and types of relationships identified). The country was used to identify the world region(s) associated with each article using a modified version of the United Nations M49 standard geoscheme27. The direction and strength of association of each factor were also identified and compared to factors identified (Table S3). Third, the data was analyzed by exploring potential relationships between different themes, including: between factor categories/factors and geography; between methods and geographic distribution, including the regional distribution of qualitative versus quantitative and specific methods used; between factors, methods and geography; and between how frequently factors were identified and their strength of association.