What do Cedar Rapids, Iowa, and Tula, Mexico (a city outside Mexico City), have in common? Both have histories of communities experiencing unequal flood exposure, unfair recovery outcomes, and a limited ability to adapt to flooding. These inequalities represent what we call flood injustice, and they demonstrate how flood risk is shaped by politics and policy as much as, or perhaps even more than, by weather and climate change.
Cedar Rapids saw a major flood in 2008 that displaced more than 18,000 residents and incurred over $3 billion in economic losses. Flooding primarily occurred within affordable housing and other residential areas west of downtown. In subsequent years, when the city sought federal funding to improve flood infrastructure such as levees, removable floodwalls, and green space, these areas failed a cost-benefit analysis required to receive money for flood protections. Those same communities remain vulnerable to flooding today.
In 2021, a torrent of rain that fell on Mexico City severely flooded Tula, some 100 kilometers away, killing 15 people and damaging more than 31,000 homes. The precipitation flowed through a maze of government-funded and -controlled conduits connecting Mexico City to Tula. Thus, the 2021 flooding in Tula was not a product of extreme rainfall worsened by climate change but, rather, of infrastructure deliberately engineered to move water away from the wealthy capital city to its lower-income surroundings.
Political decisions regarding flood infrastructure often end up keeping the most disadvantaged in harm’s way, a lesson that research continues to highlight.
When a storm hits, water must go somewhere. These cases from Cedar Rapids and Tula teach us that political decisions regarding flood infrastructure often end up keeping the most disadvantaged in harm’s way, a lesson that research continues to highlight.
In the United States, racial minority groups are more likely to live in “flood hot spots” [Tate et al., 2021] and experience higher mortality and flood damage [Tellman et al., 2020]. Globally, the most important determinant of flood-induced mortality is inequality, not levels of economic development. Indeed, flood protection is concentrated for the wealthy through various structures of inequality, leaving underserved communities vulnerable.
Fortunately, momentum is gathering to address environmental inequities. The U.S. federal government is making historic investments in environmental justice. The United Nations Climate Change Conference approved and funded a “loss and damage” mechanism to hold carbon-polluting countries financially accountable to low-income countries that experience the brunt of climate impacts. An unprecedented opportunity now exists for researchers, community organizations, governments, lawyers, and activists to address the underlying social drivers of inequality contributing to flood injustice.
To bring attention to the importance of flood justice and opportunities for engagement, our author working group, in partnership with the Arizona Institute for Resilience and the Center for Climate Adaptation Science and Solutions, held the first Flood Justice Symposium at the University of Arizona in April 2023. During the 2-day symposium, participants identified five key priorities that could be helped by increased research and civil engagement: better flood exposure data, understanding mechanisms of urban development that contribute to flooding, flood impact metrics that promote equity, improved transboundary collaboration before and after flood events, and broadening participation and coproduction in research.
Better Flood Data
Knowledge of flood inequalities and opportunities to redress them is partially impeded by the availability of only spatially coarse resolution, or sometimes nonexistent, data on historical flood exposure. Most of our understanding of flood risk comes from flood models that map areas of estimated inundation frequency. Such maps often depict, for example, the area within the 100-year floodplain, meaning the area is considered to have a 1% or greater chance of flooding in any given year.
Think of the value of models versus historical data in terms of a retirement account. A model may tell you to expect an 8% return on the investments you’ve chosen. However, if at the end of the year, your actual return is just 4%, you can use this real, historical information to adjust your investments in the hopes of achieving a better return. And the more such historical data you have, the better you’re likely to fare.
The lack of finer-scale geographic detail in flood exposure data limits our ability to assess flood events’ extent and damage, inequalities in their impacts, and changes in outcomes over time.
In the United States, the best data on historical flood exposure come from the National Centers for Environmental Information (NCEI), which report flooding at the county level. Many other countries also maintain domestic flood databases at varying scales and different levels of specificity with respect to loss and damage, although none with the same length and detail as NCEI’s. Also, the Centre for Research on the Epidemiology of Disasters’ EM-DAT data set provides global information on flooding at country scales. These are important data sets, but the lack of finer-scale geographic detail limits our ability to assess flood events’ extent and damage, inequalities in their impacts, and changes in outcomes over time.
Participants at the Flood Justice Symposium discussed ways to address this knowledge gap through research. For example, there is good news in that higher-resolution satellite imagery—from image providers such as the European Space Agency’s Sentinel platforms and Planet Labs—and machine learning data analyses are improving our understanding of historical flood exposure. Researchers presented a newly developed Google Earth Engine app that facilitates comparisons of estimated flood extents from satellite-based observations of real events with those from hydrodynamic flood models that estimate probabilities of inundation. Such hydrodynamic models include First Street Foundation’s model and the Federal Emergency Management Agency’s (FEMA) National Flood Hazard model (Figure 1). Although neither type of estimate is perfect, the satellite-observed flood extents go beyond the modeled 100-year floodplain, demonstrating that satellite-based flood mapping has much to offer in refining understanding of flood risk.
Alternative ways of gathering data and understanding flood exposures could include household or business surveys, but such research efforts must be harmonized and consistent so that lessons can be gleaned from comparing flood events. Repeat surveys on flood exposure, impacts, recovery, and mitigation efforts in flood-affected locations will provide more comprehensive assessments of localized risk that changes over time. The National Science Foundation’s Long Term Ecological Research network, with its sustained, historical collection of ecology-centric data and site-specific monitoring, provides an example of this repeated survey approach. With a harmonized compendium of historical flood exposure data, policymakers and scientists can begin to understand how flood impacts are disparate and what policies—recovery funds, communication of flood risk, and infrastructure investments—promote more equitable outcomes.
Risky Development
Accelerated urbanization and development in floodplains play a bigger role in increasing flood exposure than the changing climate [Wing et al., 2022]. Yet little is known about the policies, political and economic incentives, and loopholes that promote development in floodplains or how we might prevent building in areas of greater flood risk.
Despite government incentives in the United States to reduce construction in regulated floodplains, urbanization in floodplains continues, even in the wake of major storms such as Hurricane Florence, which inundated parts of eastern North Carolina in 2018. In the Global South, the vast majority of new urbanization occurs in informal settlements in areas not zoned or permitted for residential land use and where limited capacities to enforce urban zoning lead to larger populations living in flood-prone areas.
Patterns of urban development and zoning with respect to flood exposure have frequently been unequal, resulting in low-income neighborhoods and communities of racial and ethnic minority groups facing greater risk.
Even where policies regarding development in flood zones exist, individuals and companies with access and capital can influence the rules and further unsafe patterns of urbanization. For example, real estate developers have at times successfully excluded their properties from FEMA’s flood maps to increase property values.
Historically, patterns of urban development and zoning with respect to flood exposure have frequently been unequal, resulting in low-income neighborhoods and communities of racial and ethnic minority groups facing greater risk. When floodplain maps are revised or redrawn, the changes are not required to be disclosed to newcomers in at least 18 states across the United States, leading to people buying or residing in homes without knowing they are at risk of flooding. Further, communities of color commonly have limited access to subsidized insurance that would aid in recovery from flooding, increasing their vulnerability [Flores et al., 2022]. Similarly, where urbanization is informal, such as in parts of Mexico City’s Xochimilco borough, households cannot access public funds for flood mitigation.
Research around the world documents cases of increasing urban flood exposure. To address issues of flood justice, we must dig further to identify the specific policies and incentives contributing to urbanization in flood-prone areas that unjustly distribute flood risk to the poor. Points of leverage, where progress can be made, include modifying permitting practices and policies for urban development to require flood impact analyses and building the capacity of environmental justice organizations to hold developers and city authorities accountable for what is permitted.
Unfair Flood Metrics
At the core of flood injustice are metrics used to allocate resources for flood mitigation or recovery. The cost-benefit analysis (CBA), a standard method used today, assesses the value of property protected compared to the cost of building new infrastructure. However, this method often restricts or denies flood mitigation infrastructure or recovery funds to low-income communities with lower property values [Hino and Nance, 2021].
When Cedar Rapids proposed to build back after the disastrous floods in 2008, a CBA was used to define a new flood control plan. Because of the CBA’s penchant to prioritize high-value real estate, the new levee system constructed now protects the city’s wealthier east side. Meanwhile, the lower-income residents on the west side remain vulnerable to future floods, even though they pay an equal share of flood management taxes [Aschbrenner, 2016].
To dismantle flood injustices, we must redefine what is considered “cost-effective.”
In another example of unjust decisionmaking, poor households in Lagos, Nigeria, were forced in 2012 to retreat from coastal areas deemed at high risk of flooding. The action was substantiated by the city as a flood adaptation measure based on concerns about rising sea levels. Those same areas, however, were subsequently developed for wealthier residents in a move motivated by economic profit [Ajibade, 2019]. This new coastal development, which sits on a human-constructed peninsula designed to be protected from flooding by a large seawall, is a stark reminder of the injustice that arises when economic value is placed at the center of flood mitigation decisions.
To dismantle flood injustices, we must redefine what is considered “cost-effective.” Adjusting CBAs for differences in income and wealth among communities is one way for these evaluations to create more equitable outcomes. Measuring broader benefits from flood prevention and incorporating aspects of well-being will also shift emphasis away from monetized damages. However, recognizing the wider socioeconomic, cultural, ecological, psychological, and health effects of flooding is not enough. We must also integrate these considerations intentionally and responsibly into tools, metrics, and measures that inform flood risk management policy.
Initiatives to do this are already underway. In the United States, for example, the Climate and Economic Justice Screening Tool (CEJST), launched in 2022, is the main tool used by federal agencies to prioritize funds under the Justice40 initiative, which applies to more than 400 new and existing federal programs. Given the vast amount of spending that will be guided by CEJST, its ability to identify those with the greatest needs accurately is key for producing equitable outcomes.
Transboundary Flooding
Increased flood risk caused by urban development, dams, and border walls and the tensions and complexities of local cross-border politics make transboundary flooding (across international borders or subnational jurisdictional lines) an important component of flood justice. Unequal access to flood data and limited coordination of transboundary responses after flood events pose two critical challenges to achieving flood justice across borders.
For example, in the Ganges-Brahmaputra-Meghna river basin, which spans parts of Bangladesh, Bhutan, China, India, and Nepal, sharing of data on precipitation and river discharge is inconsistent and fragmented. This arrangement leaves forecasters in downstream Bangladesh, which captures the majority of upstream waterflow, with incomplete observations about water crossing into their borders for flood models, hindering flood mitigation [Kibler et al., 2014].
Coalitions are needed to develop transboundary flood tools and data sets, promote transborder dialogue, and build multisector collaborations across borders.
Postflood responses are complicated even in a single country, but within the 279 river basins around the world that cross international borders, decisions in one country affect others. In 37 of these international river basins where neighboring countries do coordinate responses, average death tolls from flood events were found to be lower [Bakker, 2009]. Yet such mutually beneficial collaboration is hardly the norm.
How can we tackle these issues? Coalitions among academics, community members, and government agencies are needed to develop transboundary flood tools and data sets, promote transborder dialogue, and build multisector collaborations across borders.
Kinship networks in Uganda may, for example, offer a model for international cross-border cooperation. Research has found that these networks increased the effectiveness of flood early-warning systems across multiple regional borders in the country [Canwat, 2024]. To aid in a transboundary analysis of flood vulnerability, remote sensing may be a particularly effective data collection tool because it can provide crucial cross-border assessments of flood impacts and support flood justice across geopolitical lines. However, there is limited technical capacity to collect and use these data in particularly underresourced places. As a result, cross-border arrangements are needed to ensure that such information is available everywhere it is needed.
Centering Coproduction in Research
For research to be relevant in achieving more just futures for flood-affected communities, community members’ knowledge of local context is vital. Coproduction of knowledge by groups including community members along with academics and representatives from nongovernmental organizations, private businesses, and governing bodies is a valuable approach in this regard. Coproduction is particularly important for applied research on the social injustices of flooding, which are inherently place based and interwoven within local politics, society, and culture.
Successful coproduction starts with community stakeholders holding leadership roles and being included in project design. Identifying specific, desired policy interventions and obtaining long-term funding to ensure sustained community engagement in projects are also key [Davis and Ramírez-Andreotta, 2021]. Community members’ input and interests should be prioritized throughout the entire research arc, from generating research questions to disseminating and reporting of results [Emmett et al., 2009].
An example of putting communities first can be seen in Durham, N.C., where deteriorating infrastructure and increasing precipitation are worsening urban flooding. The city has an online system that allows residents to report flooding and advocate for flood mitigation within their communities. However, the lack of an evaluation of how historic and current flood risk co-occurs with use of the system limits understanding of the factors that lead to potential underreporting of flood events and disinvestment of flood mitigation in communities inequitably at risk to flooding.
In response, the locally based Ellerbe Creek Watershed Association and the Museum of Life and Science, supported through AGU’s Thriving Earth Exchange, initiated a project to identify barriers to online flood reporting. This effort exemplifies how community-based partners engage with scientists to identify gaps in community reporting of flood events to improve the accessibility and accuracy of the system and, ultimately, promote mitigation efforts in vulnerable communities.
Research agendas that forgo a coproduction approach may result in researchers making faulty assumptions about the drivers of inequitable flood risk and flooding in areas they’re studying. For example, assuming that a lack of available flood risk information is to blame for flood inequities—rather than, say, obstacles that limit equal participation in local zoning decisions—may result in research agendas that adopt an ineffective “deficit model” approach. In such cases, the research goal may be simply to disseminate more information on flood risk to a community when what is really needed is an appraisal of barriers to participation as well as community-sourced recommendations to enable more equal participation. To avoid these pitfalls and address underlying causes of flood injustice, researchers must ensure that local voices are engaged and that solutions are community driven.
Putting It All Together
Coproduction is essential across all components of the research agenda to ensure, for example, that improved flood exposure data empower marginalized communities.
Making progress on the research needs discussed at the Flood Justice Symposium and outlined above will involve overlapping efforts. Coproduction is essential across all components of the research agenda to ensure, for example, that improved flood exposure data empower marginalized communities. Better flood data will assist efforts to create fair flood metrics and to discourage risky urbanization. And uprooting current urban development practices through science-informed policy measures could help reduce downstream effects of flooding that often span borders.
In short, these research priorities are not piecemeal, and making headway on each will influence the progress of others. Concerted and coordinated efforts across this agenda will improve our understanding of and ability to redress flood injustice for the benefit of historically marginalized communities in Cedar Rapids and Tula and around the world.
Acknowledgments
The authors thank the International Resilience Lab and the Center for Climate Adaptation Science and Solutions, both within the Arizona Institute for Resilience, for their funding contributions to the first Flood Justice Symposium at the University of Arizona in April 2023. We thank the Social Pixel Lab—including Jonathan Giezendanner, Rohit Mukherjee, Zhijie Zhang, Elise Arellano-Thompson, Ariful Islam, Prashanti Sharma, and Ruixue Wang—as well as Aaron Flores for developing the Google App showcased at the Flood Justice Symposium and in Figure 1.
References
Ajibade, I. (2019), Planned retreat in Global South megacities: Disentangling policy, practice, and environmental justice, Clim. Change, 157, 299–317, https://doi.org/10.1007/s10584-019-02535-1.
Aschbrenner, J. (2016), Why Cedar Rapids’ flooding plan will probably never get federal funding, Des Moines Register, 9 Oct., www.desmoinesregister.com/story/news/2016/10/09/why-cedar-rapids-flooding-plan-probably-never-get-federal-funding/91269958/.
Bakker, M. H. N. (2009), Transboundary river floods and institutional capacity, J. Am. Water Resour. Assoc., 45(3), 553–566, https://doi.org/10.1111/j.1752-1688.2009.00325.x.
Canwat, V. (2024), Social innovations and drivers in flood early warning systems: A community-based transboundary perspective from Elegu flood plain in northern Uganda, J. Flood Risk Manage., https://doi.org/10.1111/jfr3.12930, in press.
Davis, L. F., and M. D. Ramírez-Andreotta (2021), Participatory research for environmental justice: A critical interpretive synthesis, Environ. Health Perspectives, 129(2), 026001, https://doi.org/10.1289/EHP6274.
Emmett, E. A., et al. (2009), Development and successful application of a “community-first” communication model for community-based environmental health research, J. Occup. Environ. Med., 51(2), 146–156, https://doi.org/10.1097/JOM.0b013e3181965d9b.
Flores, A. B., et al. (2022), Federally overlooked flood risk inequities in Houston, Texas: Novel insights based on dasymetric mapping and state-of-the-art flood modeling, Ann. Am. Assoc. Geogr., 113(1), 240–260, https://doi.org/10.1080/24694452.2022.2085656.
Hino, M., and E. Nance (2021), Five ways to ensure flood-risk research helps the most vulnerable, Nature, 595(7865), 27–29, https://doi.org/10.1038/d41586-021-01750-0.
Kibler, K. M., R. K. Biswas, and A. M. Juarez Lucas (2014), Hydrologic data as a human right? Equitable access to information as a resource for disaster risk reduction in transboundary river basins, Water Policy, 16(S2), 36–58, https://doi.org/10.2166/wp.2014.307.
Tate, E., et al. (2021), Flood exposure and social vulnerability in the United States, Nat. Hazards, 106(1), 435–457, https://doi.org/10.1007/s11069-020-04470-2.
Tellman, B., et al. (2020), Using disaster outcomes to validate components of social vulnerability to floods: Flood deaths and property damage across the USA, Sustainability, 12(15), 6006, https://doi.org/10.3390/su12156006.
Wing, O. E. J., et al. (2022), Inequitable patterns of US flood risk in the Anthropocene, Nat. Clim. Change, 12(2), 156–162, https://doi.org/10.1038/s41558-021-01265-6.
Author Information
Jonathan A. Sullivan (jasullivan@arizona.edu), Hannah K. Friedrich, Beth Tellman, Alex Saunders, and Lucas Belury, University of Arizona, Tucson