1 Introduction

Unprecedented increases in global human well-being have and will continue to have destabilizing impacts on important regulating earth system processes [1,2,3,4]. Although food security and access to clean water and shelter are basic human rights, increases in population, affluence, and consumption levels represent fundamental threats to the climate that has provided stability and human advancement during the Holocene [5]. Increasingly, broad coalitions of policymakers, scientists, and other stakeholders have called for explicit limits to future planetary warming levels with most goals to limit future warming to 1.5°—2 °C [6,7,8,9,10]. To achieve these targets, initiatives such as the Paris Agreement call on global nation’s carbon emissions to peak as soon as possible [11,12,13].

Because of the large-scale challenges humanity faces, efforts to increase sustainability have largely focused on macroscale complex global agreements with diverse national stakeholders. To be sure, significant progress has been made over the last few decades on global agreements to limit the effects of human-mediated climate change [14,15,16,17,18]. However, adaptation to climate change has been characterized as local and fragmented [19]. For example, the Paris Agreement is often referred to as the first universal and legally binding treaty for decreasing carbon emissions [20]. However, the structure of the agreement and its dependence on individual national policy to foster climate change mitigation and adaptation has led to questions about its effectiveness [21].

International collaborative initiatives such as the Paris Agreement are undeniably important and provide a unifying framework for the globe to address climate change, although fundamental disagreements remain amongst nations with respect to implementation [22,23,24]. The challenge to structure truly global and binding agreements presents an opportunity for parallel sustainability initiatives focused on a smaller scale [25, 26]. Furthermore, and to a large degree, global sustainability depends on sustainability initiatives implemented on sub-global or regional scales [27, 28]. We argue that through collaboration, bottom-up local and regional sustainability efforts can scale across areas and contribute to a nation’s pathway toward broader sustainability (e.g. [29].). To this end, the purpose of this paper is to assess variation in sustainability metrics across an urban–rural continuum across the entire United States and to highlight a few case studies illustrating impactful regional collaborations between different county types.

Beyond inching toward national and global sustainability goals, collaboration across regions—urban and rural areas—represents one way to address persistent disparities between urban and rural populations [30,31,32]. In the United States, popular media conversations surrounding the urban–rural divide often focus on politics, masking real impacts of urbanization—rural to urban area migration for economic opportunity—and the subsequent economic change that have driven disparities between urban and rural areas [33,34,35,36].

Rural areas in the United States often lack access to the internet, which can result in mortality rate disparities [37, 38] and limit economic opportunity [39]. However, rural areas also provide essential ecosystem services such as food production [40,41,42], biodiversity protection [43, 44], and carbon sequestration [45], outdoor recreation [46], and relief from the COVID-19 pandemic [47]. Similarly, urban areas have higher: air pollution as measured by PM 2.5 and NOx emissions, levels of food insecurity and inequality, homicide rates, and homelessness [48]. People in urban areas tend to be more educated and more highly compensated, have better access to the internet, and more equal gender representation amongst elected officials [48,49,50,51]. Thus, regional sustainability depends on both rural and urban areas and increasing collaboration between them has the potential to harness the benefits of both urban and rural areas.

Assessing the extent to which local and regional initiatives promote progress toward higher scale sustainability goals (for instance, across rural and urban areas in a region) requires a framework that marries humanity’s social needs within the earth’s biophysical and finite constraints. One framework that is most often applied globally and nationally that could be applied regionally is the doughnut model (DM). This framework is focused on balancing social needs of humanity—“social foundations”—with the planetary boundary “ecological ceiling” framework introduced by Rockstöm et al [52]. The inner ring of the DM contains 12 metrics that represent basic human needs and include measures for access to food, housing, political representation, jobs, social equity, gender equality, educational attainment, life expectancy, water quality, waste management, peace/justice, and network connections [53]. The outer ring of the DM contains planetary boundaries including climate change, ocean acidification, chemical pollution, nitrogen and phosphorus loading, freshwater withdrawals, land conversion, biodiversity loss, air pollution, and ozone layer depletion [52]. The doughnut model provides a way to conceptualize how to sustainably meet the needs of humanity while not overshooting planetary boundaries. Thus, the DM can be used to provide an assessment of regional sustainability by holistically and comprehensively quantifying social needs of humanity within the bounded reality of our planet. This approach is easily adaptable based on data availability and can be used to quantify variation in a suite of metrics across different regions.

Despite the importance of sub-national level sustainability initiatives toward promoting global sustainability, there are few studies that focus on quantifying comprehensive regional social-ecological sustainability. Instead, regional sustainability studies tend to focus on a few different metrics instead of a suite of social-ecological metrics. Previous studies have investigated the importance of collaboration and equity in regional housing planning in California [54], implemented proxies using input–output resource matrices to quantify self-sufficiency in regions of Germany [55], and assessed regional sustainability based on land and water availability [56]. Previous studies have also provided theoretical frameworks for assessing regional sustainability including: using Metropolitan Statistical Areas as system boundaries [57], providing a series of ecological and social parameters to assess urban sustainability [58], reviewed best practices for assessing regional sustainability [59], using human carrying capacity for assessing regional sustainability [60], conceptual approaches for downscaling the planetary boundary framework for national assessment [61], and coupling various frameworks for sustainable development together [62].

Although there is an increasing focus on quantifying global planetary thresholds for sustainability across a variety of metrics, there are few studies that quantify sustainability efforts on smaller scales and collaboration between areas to scale sustainability efforts. This work builds on a previous study that used the DM to quantify sustainability metrics across county types in the Upper Midwestern United States [48]. The main objective of this work is to quantify variation in DM sustainability metrics across the United States. In this study we use the DM to quantify regional sustainability and collaborative efforts across 32 metro areas and 180 counties in the United States. Across these metro areas and counties, we compile data from various county, state, and national governmental agencies and databases to quantify sustainability across regions using 8 ecological metrics and 12 social metrics. Our second objective is to highlight how variations in these metrics provide an avenue for collaboration to increase regional sustainability.

Finally, our third objective is to systematically scan the 180 county websites to see to what extent different county types (urban, peri-urban, rural) explicitly acknowledge collaboration with other counties and county types on two different sustainability initiatives: (1) to enhance resilience to climate change and (2) to improve human health outcomes. Finally, we present two example case studies highlighting cross county collaboration to increase regional sustainability. By highlighting a few case studies, we aim to illustrate the impact of different county types collaborating to enhance regional sustainability.

2 Methods

2.1 Study area, study approach, and data collection

The U.S. is the 3rd largest country in the world by land area (or fourth depending on criteria) and it consists of ~ 9.5 × 106 km2 of land. Across this vast land area, there are complex networks of the built environment with cities, roads, and bridges all designed and interacting within their geographic, ecological, and climatic context. To get a better sense of how geographic and climatic factors influence regional sustainability and opportunities for collaboration toward sustainability goals across an urban-peri-urban–rural gradient, our analysis focused on urban regions categorized by both U.S. Census Regions and Census Divisions. We selected large metropolitan regions within Census Regions and Census Divisions and categorized counties within these metropolitan regions as urban, peri-urban, or rural depending on population density. We defined a county as urban with population density greater than 300 people km−2; counties as peri-urban with between 100 and 299 people km−2; and counties as rural with population densities of less than 100 people km−2. Based on these criteria, we created a database of 180 counties around 32 metropolitan areas in the United States (Fig. 1).

Fig.1
figure 1

Map of 180 counties included in study across 32 metropolitan areas in the United States. Orange pins designate urban counties, yellow pins designate peri-urban, and green pins designate rural counties

Full size image

2.2 Data collection

We assessed regional sustainability across the urban–rural gradient in this network of metropolitan areas by collating a variety of ecological and social parameters similar to other national analyses[3, 53]. We modified ecological ceiling and social foundation metrics in a similar manner as described by Chapman et al [48]. Briefly, we compiled ecological ceiling and social foundation metrics from various national and state level databases including from the U.S. Geological Survey, U.S. Environmental Protection Agency, Centers for Disease Control and Prevention, and the United States Census Bureau. Below we describe specific methodologies for each of the 8 ecological and 12 social parameter data, analysis, and visualization.

2.3 Ecological ceiling metrics

Climate change, Freshwater withdrawals, Nitrogen and phosphorus cycles, Biodiversity loss, Air quality (PM 2.5, Ozone, AQI).

To assess variation in climate change along our study system, we used county-level carbon dioxide data from the 2017 National Emissions Inventory Data from the U.S. Environmental Protection Agency (EPA).

We collected freshwater withdrawal data from the U.S. Geological Survey (USGS) on the USGS NWIS web database. Within the database, we summed multiple freshwater withdrawal usages for each county including public supply total self-supplied withdrawals, total domestic total self-supplied withdrawals plus deliveries, industrial total self-supplied withdrawals, total thermoelectric power total self-supplied withdrawals, total livestock total self-supplied withdrawals, fresh irrigation, golf courses consumptive use for golf courses, fresh irrigation, and crop consumptive use for crops fresh. We report freshwater withdrawals per capita for each county in m3 y−1 per person.

We estimated nitrogen and phosphorus fertilizer use by county type through the U.S. Geological Survey’s county level estimates of nitrogen and phosphorus from commercial sales from 1987 to 2012. We summed total fertilizer use for nitrogen and phosphorus from both non-farm and farm use for each county and reported values from the last year of the dataset (2012) in kg per year.

We used the percentage of non-conserved land in each county to assess biodiversity loss across our study regions. County-level data were compiled from the 2018 US Geological Survey’s PAD-US (Protected Areas Database of the US).

We assessed variation in air quality across county type through several different manners: (1) by using PM 2.5 data obtained from the U.S. EPA Air Quality Systems database for cities and counties from 2019; (2) we quantified ozone concentration across county types by compiling 8-h O3 data from the U.S. EPA; and (3) using annual Air Quality Index data from the U.S. EPA from 2019–2021.

2.4 Social foundation metrics

Food, Health, Education, Jobs, Peace/justice, Political voice, Social equity, Gender equality, Housing, Networks, Energy/ waste management, and Water quality.

We assessed food insecurity rates by compiling county-level data from Feeding America, a hunger relief organization in the United States. Feeding America’s Map the Meal project estimates county-level food insecurity by creating a model from state-level data based on unemployment rate, poverty rate, demographic information, disability rate, and percentage of homeowners in a given county [63].

We obtained average life expectancy data by county from the U.S. CDC’s Center for Health Statistics. To calculate ‘deficit years,’ we found the difference in life expectancy between each county compared to the average life expectancy of a U.S. citizen. Put another way, if our calculated variable for ‘deficit years’ were negative, the interpretation is that the life expectancy of an average individual in that county was greater than the average life expectancy of a U.S. citizen.

We measured variation in education across county types by quantifying the percentage of individuals 25 years and older within each county that have a bachelor's degree or higher. We obtained data between 2016 and 2020 from the U.S. Census Bureau and we quantified the percentage of individuals who do not have a bachelor’s degree or higher by subtracting the U.S. Census Bureau data by 100.

The jobs metric sought to quantify differences in employment opportunities across county types. To assess variation in the jobs metric across county type, we compiled data from the U.S. Bureau of Economic Analysis. This metric represented the difference between the annual income of the county relative to the national average income in US dollars in 2019. We reported this metric as a US dollar deficit; counties that had average annual incomes lower than the national average income had a positive value and counties that had higher incomes relative to the national average were negative.

We calculated peace/justice by compiling death rates by homicide across county types. We collected peace data from the National Center for Health Statistics of the Center for Disease Control and Prevention (CDC). Data were collected by filtering group results by state and county and we reported the rate of homicide per 100,000 individuals across county types.

We represented political voice as voter non-turnout. We collected county-level voter turnout data from the 2020 U.S. election cycle from the U.S. Census Bureau. We subtracted voter turnout by 100 to calculate the percentage of voter non-turnout.

We measured social equity using Gini index, a unitless statistical measure of how evenly distributed income is across a region where 0 is completely equal distribution and 1 is perfectly unequal distribution. We compiled county-level data from the American Community Survey of the U.S. Census Bureau from 2013 to 2017 depending on the county.

We calculated the female representative deficit, our measure of gender equality, from state house (or assembly) and state senate elected representatives as of 2020. First, we quantified the total proportion of female representatives from each state legislature (house/ assembly plus senate members). Then, we quantified the deficit of each county to 50% (equal gender representation) by subtracting the proportion of female representatives from 50%. The interpretation of a negative female representative deficit was that the state legislature was more than 50% female.

For housing, we quantified the number of homeless individuals per county using a few different approaches. If a county reported data as the number of homeless individuals after a point in time count (defined as a count of the number of sheltered or unsheltered homeless individuals on a single night in January), we used the reported number of homeless individuals and divided that by the county population to find the number of homeless individuals per county. However, if a particular county were a part of a larger Continuum of Care (CoC)— a program through the U.S. Housing and Urban Development to address homelessness using a community-wide approach—we used a different approach. If the county were a part of a CoC that reported the number of homeless individuals as a total, we used a continuum map to determine which counties made up the particular CoC. Once we determined which counties were in each CoC, we developed a simple equation to calculate the mean number of homeless individuals per county. The mean number of homeless individuals in a particular county (MHC) were equal to the number of homeless individuals in a CoC (HIC) divided by the total population of the CoC (TP) and then multiplied by the mean population of the counties in a particular CoC (MPC) [Eq. 1].

$${\text{MHC}} = {\text{ HIC}}/{\text{TP }}*{\text{ MPC}}$$
(1)

We measured the extent to which communities are connected in networks by using the percentage of households with a broadband internet subscription or connection. County-level network data were collected from the U.S. Census Bureau and we reported the deficit percentage of households within the county that do not have a broadband internet subscription or connection by subtracting those with broadband by 100.

We used municipal waste production as a proxy for how counties manage waste disposal. We compiled the quantity of waste generated by counties through county websites and reported the kg of waste generated per capita from county population data.

We used the number of impaired water bodies in a county as a proxy for water quality. We compiled data from state level agencies tasked with reporting and maintaining a state’s Clean Water Act Sect. 303(d) list. The U.S. EPA requires states to submit reports every 2 years to identify impaired and threatened waterways under the Clean Water Act. States are also required to report specific known contaminants that contribute to impaired status. Depending on the state, the years for these data varied for counties.

2.5 Collaborative program analysis

After we quantified the ecological ceiling and social foundations metrics using the DM approach outlined above, we analyzed the extent to which different counties within our study collaborated with one another by scanning for sustainability programs through the 180 different county websites in 2022. We identified cross county collaborations by specifically focusing on one ecological ceiling metric and one social foundation metric used in our DM analysis of regional sustainability. We focused on whether counties explicitly mentioned working with other counties on programs associated with climate change (ecological ceiling metric) and public health (social foundation metric). We focused our search for collaborative programs for climate change and public health as we believed these types of programs to be the most prevalent. To search for these collaborative programs, we scanned each county’s website for explicit language regarding programs that addressed climate change or public health. We also considered renewable energy programs as collaborations aimed to address climate change. To identify potential cross county collaboration, we searched for the following keywords: “climate,” “renewable energy,” “health improvement,” “health,” and “collaboration.” Counties without websites were not included in the analysis.

We also searched for collaboration between counties beyond climate change and public health by searching county websites using other key terms including air pollution, economic development, or transportation. To be included as an example of cross-county collaboration in this study, we included counties that explicitly mentioned collaboration between one or more counties. Following the collaborative program analysis, we selected two example case studies of cross-county collaboration to highlight the promise of increasing regional sustainability by bringing rural, peri-urban, and urban counties together.

2.6 Statistical analysis

We used multiple one-way analysis of variance (ANOVA) to quantify whether or not there were significant differences in our ecological ceiling and social foundation metrics across the urban, peri-urban, and rural county gradient. To analyze whether county type was a predictor for cross county collaboration, we performed a Pearson’s Chi-squared test. We assessed assumptions of the statistical tests using plots in R [64]. When our data violated assumptions of the statistical tests, we log transformed data and performed analyses on these log transformed data. When we observed differences across county types in dependent variables, we did multiple pairwise comparisons using the emmeans package in R [65]). We plotted our data using the ggplot2 package in R [66].

3 Results

3.1 Doughnut model analysis of regional sustainability: ecological ceiling metrics

Across our urban, peri-urban, rural gradient we found variation in metrics for climate change, freshwater use, air pollution, and air quality (Fig. 2). Carbon dioxide emissions were higher in urban counties compared to peri-urban (F = 2.85, d = 177, p < 0.05) and rural counties (F = 5.13, d = 177, p < 0.001), but we did not find a difference between peri-urban and rural counties. We found high variation of freshwater use across county types and freshwater use per capita was slightly higher in rural areas compared to urban areas, (F = 2.58, df = 177, p < 0.05), but there was no difference between urban and peri-urban or peri-urban and rural counties. We found no significant differences across county types for biodiversity, nitrogen and phosphorus use, and ozone (Fig. 2).

Fig. 2
figure 2

Ecological ceiling metrics across the rural urban continuum. Orange bars represent urban counties, yellow bars represent peri-urban counties, and green bars represent rural counties. The boxes denote interquartile range, median is denoted by the dark line in the box, and whiskers represent the lowest and highest values. Brackets above the boxes represent mean pairwise comparisons with Bonferroni correction, where ns = not significant, * = p < 0.05, ** = p < 0.005

Full size image

3.2 Doughnut model analysis of regional sustainability: social foundations

Across our urban, peri-urban, rural gradient we found variation across all county types for 9 out of the 12 metrics we assessed (Fig. 3). We found higher rates of food insecurity in rural areas compared to urban and peri-urban areas (F = 4.10, df = 176, p < 0.0005; F = 4.30, df = 176, p < 0.0001). The percentage of the population without bachelor’s degrees was significantly higher in rural areas compared to urban (F = 9.08, df = 177, p < 0.0001) and peri-urban areas (F = 6.10, df = 177, p < 0.0001), but did not differ between urban and peri-urban areas. Urban counties had lower levels of per capita income deficit (higher incomes) compared to peri-urban (F = 2.90, df = 177, p < 0.005) and rural counties (F = 7.00, df = 177, p < 0.0001) and peri-urban counties had lower per capita income deficits compared to rural areas (F = 3.10, df = 177, p < 0.005). With respect to political voice, levels of voter non-turnout were significantly higher in urban (F = 3.36, df = 170, p < 0.001) and rural counties (F = 2.60, df = 177, p < 0.01) compared to peri-urban counties but did not differ significantly between urban and rural counties. We found that rural (F = 5.33, df = 177, p < 0.0001) and peri-urban (F = 3.53, df = 177, p < 0.0005) counties had higher measures of social equity as measured by Gini index (rural counties had lower measures of Gini index) compared to urban counties, but not peri-urban counties. We also observed significantly lower levels of gender equality in elected representatives in peri-urban (F = 2.82, df = 177, p < 0.05) and rural counties (F = 2.54, df = 177, p < 0.05) compared to urban counties. Urban counties had a higher homeless population compared to rural areas (F = 2.73, df = 86, p < 0.05), but did not differ compared to peri-urban areas. We also found no difference in peri-urban counties and rural counties in homelessness. Access to broadband internet access (networks) also varied significantly across county type (Fig. 3). Rural counties had higher percentages of the population that did not have broadband internet access compared to both urban (F = 6.08, df = 177, p < 0.0001) and peri-urban counties (F = 6.13, df = 177, p < 0.0001), but peri-urban and urban counties did not differ. Finally, we found higher levels of waste generation per capita in rural areas compared to urban (F = 2.77, df = 137, p < 0.05) and peri-urban counties (F = 2.45, df = 137, p < 0.05), but no differences in urban and peri-urban waste generation. We found no significant differences across county types for health (deficit years compared to US mean lifespan), peace (homicide rates), and water quality (the number of impaired waters).

Fig. 3
figure 3

Social foundation metrics across the rural–urban continuum. Orange bars represent urban counties, yellow bars represent peri-urban counties, and green bars represent rural counties. The boxes denote interquartile range, median is denoted by the dark line in the box, and whiskers represent the lowest and highest values. Brackets above the boxes represent mean pairwise comparisons with Bonferroni correction, where ns = not significant, * = p < 0.05, ** = p < 0.005, *** = p < 0.0005, **** = p < 10–5

Full size image

3.3 Cross-county program analysis quantitative summary

Our analysis of cross-county collaborative programs advertised through county websites revealed that 79.4% of the counties (143 counties out of the 180 in the study) do not explicitly mention collaborative programs to address public health (social foundation) or climate change (ecological ceiling) challenges across county types. We found that 8.8% (16 out of 180) of counties did explicitly mention programs to address these broad complex problems across county boundaries. We also found 11.6% of counties did not have websites (21 out of 180). We did not find differences between county types and whether county websites addressed regional collaboration with other county types. In other words, county type was not a predictor of whether counties participated in collaborative efforts for increasing regional sustainability (x2 = 1.95, df = 2, p > 0.05).

3.4 Case study I: cross-county collaboration for climate adaptation: Puget Sound, WA

Puget Sound is in the pacific northwest of the United States and spans 13 counties in northwestern Washington state. As of the 2020 US census, the population of Puget Sound is roughly 4.3 million. The majority of the population is white (62.6%), followed by Asian (14.0%), Two or more races (6.9%), Black or African American (5.9%), and white (Hispanic; 5.0%). In 2019, the median household income level of the region was $92,000 and 1/5 of the population of the region were considered to be low income or impoverished.

Over 2/3 of Washington state’s population lives in the Puget Sound region. With continued development pressures and increasing impacts from climate change, various stakeholders formed the Puget Sound Climate Preparedness Collaborative to increase climate change preparedness. There are considerable concerns about climate change and land use change from development negatively impacting the Puget Sound watershed [67, 68] and important fisheries [69]. The goals of the regional collaboration focus on implementing climate adaptation efforts through joint research efforts, informational workshops, and increased cross-county and regional collaboration (Fig. 4). The collaborative also explicitly acknowledges the need to address climate impacts in a socially and racially equitable manner. This case study highlights that counties of the Puget Sound region have developed shared goals and a framework for increasing climate change adaptation.

Fig. 4
figure 4

Case study of cross county collaboration in Puget Sound, WA, USA. The case study highlights the work of multiple counties and tribal groups within Puget Sound on climate adaptation planning. Figure image from https://pugetsoundclimate.org/

Full size image

The Puget Sound Climate Preparedness Collaborative is funded by various foundations such as The Bullitt Foundation and The Kresge Foundation of the Institute for Sustainable Communities (Fig. 4). These foundations recognize the importance of coordination between cities, counties, and tribal communities. Regional collaborations such as the Puget Sound Climate Preparedness Collaborative can serve as an illustrative model for creating and financing plans that address climate change across regions.

3.5 Case study II: cross-county collaboration for regional sustainability: Chicago, IL

As of the US 2020 census, the population of the Chicago Metropolitan Statistical Area was 9.5 million people. Population demographics include 45.3% white, 29.2% Black or African-American, 28.2% hispanic or latino,7.4% two or more races, 6.8% Asian, and 0.5% Native American, Native Hawaiian, or Pacific Islander. The median household income for the Chicago region as of the 2020 US census was $66,000 and 17.1% of people were considered low income or impoverished. The Central Metropolitan Agency for Planning coordinates regional planning across the region and all of the Chicago area Council of Governments have adopted the Greenest Region Compact.

Development pressure, changes in land use, and climate change represent fundamental vulnerabilities the Chicago metropolitan region faces over the next few decades [70]. Elected officials with the City of Chicago explicitly acknowledge that climate change is driving warmer and wetter winters, heavy precipitation in the spring, and hotter, drier summers in the region. Furthermore, climate change driven decreases in air quality, higher temperatures, and a changing agricultural landscape in the Chicago region negatively impact human health outcomes [71]. To address these vulnerabilities, the Chicago region has adopted a regional collaborative plan for addressing these possible threats due to climate change (Fig. 5). Participating counties include both urban counties such as Kane and Will, and peri-urban counties such as Lake and McHenry counties.

Fig. 5
figure 5

Case study of cross county collaborative in Chicago, IL, USA. This regional collaboration focuses on broad strategies for creating sustainable communities. Figure image from https://mayorscaucus.org/initiatives/environment/rec/

Full size image

With the creation of the Greenest Region Compact (GRC), the Metropolitan Mayors Caucus explicitly acknowledged the importance of collaboration to affect change to address broad sustainability challenges. The plan aims to achieve net zero greenhouse gas emissions by 2050 and outlines a framework for regional collaboration. An extension of the original Greenest Region Compact, Greenest Region Compact 2, was created in 2016 by the Metropolitan Mayors Caucus. The GRC developed 10 categories that are used to improve the sustainability of the communities including climate, economic development, energy, land, leadership, mobility, municipal operations, sustainable communities, water, and waste and recycling (Fig. 5). These categories provide a framework to promote cross regional collaboration and sustainability transformation through strategies, tools, and objectives.

The GRC is funded by the Searle Funds of the Chicago Community Trust (https://2019annualreport.cct.org/searle-funds/). The Chicago Community Trust is one of the largest community foundations in the United States and focuses on awarding grants to initiatives that seek to increase equity, opportunity, and prosperity for all. The Searle Funds within the Chicago Community Trust support: (1) research efforts in science and medicine among other areas; (2) educational initiatives; (3) economic development; and (4) environmental initiatives. In 2019, the Chicago Community Trust awarded nearly $400 million in grants. The GRC highlights the potential for private trusts to support regional collaborative sustainability initiatives.

4 Discussion

4.1 Variation in ecological ceiling and social foundation metrics across county types

Our results revealed considerable variation across county types in DM metrics for both ecological and social parameters. Interestingly, we found more differences across county types in social parameters compared to the ecological parameters. For example, we mainly found differences across county types for 2 of the 8 ecological ceiling metrics—air pollution and climate change—and 9 out of the 12 metrics for social foundations, suggesting that human development and management shape an urban region’s ecological identity similarly across county types [72,73,74].

The degree of variation in social metrics across county types reflects broader trends in urbanization in the United States following World War II. That is, as the relative economic contribution of agriculture decreased in rural areas, urban areas experienced rapid population growth that drove urban development and economic opportunity [75, 76]. These changing economic tides from urbanization may explain differences in education levels [77,78,79], food insecurity[80], income level and early achievement [81, 82], and access to the internet [83] between urban and rural counties.

Additionally, because of rising inequality within cities, levels of inequality are now greater within cities than between urban and rural areas [82]. Inequality within cities can be partially explained by cities simultaneously providing opportunities for high income earners with college degrees and resources for homeless individuals [84]. Furthermore, those experiencing economic hardship have lower rates of voter turnout [85], which may increase the wealth gap between elected officials and private citizens. Our results in ecological and social parameters across county types suggest that urban and rural areas offer convergent visions for how ecological measures are managed and divergent visions for social measures. These findings also highlight the potential for urban and rural areas to work together to improve social outcomes and decrease the environmental overshoot of a region.

4.2 The case for increasing regional collaboration across urban and rural areas

Increasing political polarization in the United States threatens the viability of national policy changes to increase national sustainability. Thus, regional collaboration across diverse regions may represent a more feasible approach to navigating the disparities across regions as highlighted in the previous section. Furthermore, collaboration is an essential strategy to increase environmental governance for sustainability [77]. Our analysis of 8 ecological and 12 social parameters of sustainability across 180 counties in the United States highlight large disparities across the 32 metro regions in this study. These disparities underscore the importance of collaboration to leverage human capital and funding for sustainable transformation [86, 87]. Calls for increasing regional and cross-county collaboration are not new, but long-term challenges have inhibited regional collaborative efforts [88, 89]. Several models and mechanisms have been proposed to increase and foster regional collaborations including (1) creating regional governmental agencies that consolidate smaller local government activities; (2) focusing on shared specific goals on a policy-by-policy basis; and (3) creating an umbrella layer of government across a broader region area that oversees local agencies [90]. Ultimately, increasing collaboration for regional sustainable transformations depends on sound theoretical frameworks and coalitions of diverse stakeholders affecting change locally [91, 92].

Our collaborative program analysis highlighted several cases where local and regional agencies served as innovative laboratories for increasing regional sustainability. Agencies in both Puget Sound and in the Chicago metropolitan area coordinated and leveraged multiple funding opportunities to address perceived sustainability challenges in specific geographic contexts. These collaborations represent important models for developing new collaboration initiatives and offer blueprints on how local agencies can form partnerships to form regional coalitions. Recent research has highlighted how cities and urban areas can serve as important models and drivers for urban transition and living laboratories for sustainable transformation [93,94,95]. Furthermore, scenarios and visioning for sustainable transformation can provide pathways for diverse stakeholders to imagine ways to increase regional sustainability [96, 97]. Conceptualizing urban areas as social-ecological-technological systems [98] and grassroots efforts in urban food systems transformation [99] have also provided theoretical framing to govern sustainable transformation, though there is concern that there is more sustainable transformation theory development compared to on the ground sustainability transformation [100]. Our case studies provide examples of how diverse socio-political regions can harness diversity of thought and human ingenuity to increase regional resilience to stressors such as climate change (Puget Sound, WA) and across a broad scale of sustainability metrics (Chicago, IL).

4.3 Limits of local and regional collaboration to increase global sustainability

Several barriers remain for effective implementation of regional collaborations to scale sustainability efforts. Beyond limited funding opportunities, cross disciplinary and cross agency initiatives face many logistical and structural challenges including assigning shared costs, especially between urban and rural counties, as well as incentive structure for third-party participation [90, 101]. Additionally, many organizational, psychological, and sociological factors may influence the outcome of collaborative projects [102,103,104]. The difficulty of finding common visions, aligning goals, implementing programs, and assessing the extent to which regional collaborations progress toward increasing regional sustainability all represent barriers that can dissuade broader participation [105]. Political divisions within government, between policymakers, and an increasingly divided electorate represent fundamental limits to the willingness of different counties to collaborate on regional sustainability initiatives[106,107,108], though political divisions may represent an opportunity for sustainable transformation[109]. Broad-scale sustainability issues are metaphorically and geographically larger than the scale in which most counties operate and this spatial disconnect represents both a structural and logistical challenge to cross-county collaboration.

Beyond structural and logistical challenges to foster cross county collaboration, our collaborative program analysis likely underestimated the extent to which different local and regional agencies work toward shared sustainability goals. In this study, we reviewed county websites for select keywords in an effort to identify ways in which counties may work to increase public health outcomes (social foundation metric) and resilience to climate change (ecological ceiling metric). It is possible that there are counties across our study region that actively collaborate with one another, but do not report these activities on their county websites. There were also counties that did not have websites, limiting our ability to assess whether those particular counties were engaged in cross county type collaborative projects. Furthermore, there may be instances of collaboration to increase regional sustainability across county types that our specific keywords did not find. In other words, we may be underreporting cross-county collaborations based on our specific keywords. Despite these limitations, our study provides an approach to assess regional progress toward sustainability transformation across the United States. Our case studies also provide concrete examples of regional sustainability collaboration, challenging the notion that capacity for sustainable transformation is rare on the ground [100].

5 Conclusion

Global social and ecological systems are under increasing duress from land use change and biodiversity loss driven by global agricultural expansion. Continued fossil fuel burning, the arrival of the Anthropocene, and climate change threaten the earth system processes that provide humans with essential life support systems. The interconnectedness of ecological and social systems requires a framework that marries ecological and social outcomes such as the DM. By applying the DM and its suite of metrics for ecological and social sustainability across regions, we highlight considerable variation across the United States, particularly in the social foundations. This variation in regional sustainability highlights the promise of bringing diverse stakeholders together to solve shared challenges. We argue that sustainable transformation guided by county level collaboration has the potential to not only decrease social and ecological disparities across urban and rural regions, but also provides a meaningful roadmap to meet global targets to limit climate warming. Future collaborative work should leverage institutional capacity, networks, and public and private funding to increase regional sustainability through sustainable transformation theory development and local application through the creation and maintenance of programs in and across urban areas.