Brian Stone Jr. is a professor in the School of City and Regional Planning and director of the Urban Climate Lab at the Georgia Institute of Technology. This essay is adapted from his recent book, “Radical Adaptation: Transforming Cities for a Climate Changed World” (Cambridge University Press, 2024).
In the months after Hurricane Katrina, New Orleans proposed a flood control program unlike any other in U.S. history. Developed by the Bring New Orleans Back Commission, a diverse group of stakeholders appointed by the mayor, the resulting plan called for large parts of the city to be converted from longstanding residential zones to floodable parks. Released to the public in the form of a map, large green circles were positioned over neighborhoods where owners would be forced into buyouts. These were some of the most historic districts in a very historic city — the Lower Ninth Ward, St. Claude, Gentilly — and almost exclusively in majority Black and marginalized neighborhoods.
Christened in the press as the “Green Dot” map, the proposal ranks among the most profoundly unsuccessful plans ever issued by a municipal body and would never be put to a vote in the city council. But it would give life to an idea that has only grown more compelling in the subsequent decades: We cannot engineer our way out of climate change; retreat is inevitable — and not just in coastal cities.
The Green Dot map’s remarkably brief tenure can be attributed in part to its proponents’ failure to adhere to the most basic rule of community planning: Never designate the where before building support for the what. It is possible that a proposal pairing the creation of floodable space somewhere in New Orleans, alongside the offer of generous and non-compulsory property buyouts, could have engendered community support — indeed, more than a decade after the city’s almost complete inundation, space for a large floodable park was assembled from publicly acquired land in Gentilly.
But because the commission launched into climate adaptation with a list of neighborhoods to be condemned, an opposing army immediately assembled. Today, the Green Dot map endures only as a precisely-how-not-to-do-it example in urban planning textbooks and, until recently, as a small but popular French brasserie situated in one of the neighborhoods targeted for abandonment: the Green Dot Cafe (4.6 stars).
Almost 20 years after the devastation of Katrina, neither New Orleans nor the United States national government has developed a policy framework for planned retreat. Much of New Orleans recovered and even thrived in that time, but some neighborhoods today are characterized more by land vacancy than occupation and often lack the most basic of community amenities: schools for neighborhood children, markets and restaurants within walking distance, churches and other community institutions.
Most importantly, enshrining a “right to return” everywhere in post-Katrina New Orleans resulted in a greater resilience to flood risk almost nowhere. The most serious risk confronting residents today is not a levee-topping storm surge but a bowl-filling deluge. New Orleans now floods annually from heavy storm events that overwhelm the drainage network of a city that sits at a lower elevation than the water that surrounds it. A more extensive network of floodable parks is arguably the most effective strategy to prevent regular flooding of streets, homes and businesses, but there is no plan yet in place to bring this outcome about, even in low-lying areas that remain largely depopulated.
New Orleans is not unique in the extreme risk of climate change it confronts. All major cities are now at an elevated risk of three climate impacts for which they are largely unprepared: extreme flooding, extreme drought and extreme heat. This is not a future risk but one that is daily unfolding across the U.S.
In the Southwest, rising temperatures are fueling elevated rates of evaporation and years of drought, forcing a growing number of communities to rely on the regular delivery of drinking water; in the Midwest, the intensity of heat and humidity may soon render basic municipal services, such as garbage collection, too hazardous to operate during heat waves; along the east coast, high tides pull oceanside homes into the sea in calm weather. To watch a video of barrier island homes breaking up and falling into the Atlantic is — for me, at least — to experience initially a sense of awe at the destructive potential of a gradual but incessantly rising ocean. The emotion that lingers, however, is not one of wonder but of shame.
Only the most scientifically and institutionally incapable of societies would experience such impacts from annual weather events that can be forecast years in advance. Perhaps more than the most violent of storms, it is the slowest-moving climate impacts that are most clarifying: The era of retreat is underway. It is an era for which we find ourselves remarkably unprepared.
Our governmental programs for managing a rising risk of flooding assume three principal forms:
- Engineered infrastructure to divert large quantities of rainfall and contain rising bodies of water;
- Emergency response protocols for evacuation and post-event recovery when those systems are overwhelmed; and
- Federally subsidized insurance and loan programs for assisting homeowners in rebuilding.
None of these is designed to relocate residents and property outside of high-hazard zones in advance of extreme weather events. Indeed, they make retreat less likely.
Homeowners residing outside of federally designated high-risk flood zones assume urban drainage systems and levees will prevent flooding, despite the fact that an estimated 40% of all claims made under the National Flood Insurance Program involved homes situated outside of these zones. The availability of federal assistance to rebuild, even for property owners in floodplains who elect to not carry flood insurance, enables residents to remain within increasingly hazardous zones for which no private insurance company would issue a policy.
The combined result is a steadily rising level of risk that is rooted as much in institutional negligence as in rapidly changing environmental conditions. Were it the stated aim of the U.S. government to maximize the human and economic toll of climate change on its citizens, its policy framework may not look much different from our current array of disaster response programs.
As I write, tens of thousands of property owners are rebuilding and repairing homes, frequently with public dollars, across an expansive area of Florida that was inundated by Hurricane Ian in 2022. A large percentage of these homes were previously rebuilt following one or more of the 10 hurricanes that have blown through southwest Florida since their initial construction — again, frequently with public dollars. It is a statistical certainty that many will again require governmental assistance for rebuilding in the coming years.
Were each of these homes required to carry a commercially available flood insurance policy, it is likely not one would be rebuilt — the risk of future flood or wind damage during the standard period of a mortgage is rated at 100%. But with few prohibitions on where homes can be constructed along the Florida coast and the policy commitment of the federal government to provide disaster recovery funds independent of projected climate risk, the debris fields of Florida’s next hurricane are at this moment being populated.
As high-risk climate zones expand, other national governments are restructuring programs for disaster recovery. Commencing in the 1990s, the Netherlands — a country perhaps more at risk to sea level rise than any non-island nation — undertook a program at sharp odds with the U.S. approach to managing climate risk. Rather than raise and reinforce levee systems along rising water bodies, as has been the strategy in post-Katrina New Orleans, the Dutch have pulled their levees back along several urban rivers, requiring homes and businesses in the restored natural floodways to be relocated to higher ground.
The compensation offered is generous and, in some instances, complete neighborhoods have been relocated. But the property buyouts are mandatory; residents must move. Known as the “Room for the River” program, the Dutch approach is a process of planned retreat in advance of the next flooding event; it enhances the long-term welfare of both the relocated property owners and the (far more numerous) people living adjacent to the expanded floodway. The Dutch do not enshrine for their citizens a right of return. What they do enshrine is a right of resilience.
For anyone uncomfortable with the compulsory relocation of residents and business owners from high-risk climate zones, consider what the Canadian province of Quebec did in the aftermath of destructive flooding events in 2017 and 2019 along the Ottawa River. During these floods, a large number of properties in Gatineau flooded twice; after the second inundation, the provincial government predicated disaster relief funds on a requirement that homeowners either use the money to relocate out of the expanding flood zone or, for those choosing to rebuild, consent to a permanent prohibition on any future public relief funds for both present and future property owners.
This “one-and-done” approach to disaster funds in the most hazardous areas alters the usual calculus of property owners considering a buyout option. Those choosing to remain in a high-risk climate zone must contemplate a diminished future resale value should they elect to rebuild. In contrast to post-Ian Florida, a decision to rebuild in the expanded floodplain of the Ottawa River is a decision perhaps to never sell a property that can neither be insured nor receive public disaster funds for future rebuilding.
Whatever the mechanism of retreat — mandatory (Dutch) or quasi-compulsory (Canadian) — success should be gauged in whether vulnerable residents are relocated outside of zones certain to flood again and whether contiguous land is amassed to enhance the resilience of the larger urban population. To justify the substantial public investment needed to acquire property in zones all but certain to flood again, to burn again, to fall into the ocean, the acquired land must be actively managed to mitigate climate threats.
A key distinction between these and the U.S. approach to disaster response is that retreat is understood to be fundamentally a process of land assembly rather than one of land abandonment. The Dutch and the Canadians are not fleeing — they are staying. Retreat is the mode of remaining.
Retreat as a process of land abandonment has its roots in the earliest policy documents on climate adaptation. In the first assessment report of the Intergovernmental Panel on Climate Change (IPCC) in 1990, retreat was identified as one of three options for climate adaptation and defined as the “abandonment of land structures in vulnerable areas and resettlement of inhabitants.” No mention is made in the report of repurposing abandoned land for enhanced climate resilience, and this omission persists into contemporary U.S. programs focused on climate adaptation.
In contrast to the IPCC formulation, which addresses retreat as the first of three options, the U.S. Environmental Protection Agency positions retreat as the last of three possible actions, following the installation of flood defenses without relocation (protection) and retrofitting development to allow for periodic flooding (accommodation). If the most basic aims of climate adaptation are to protect vulnerable populations in advance of the next destructive event and to minimize the number of households ultimately displaced, our retreat-last approach in New Orleans, the Florida coast and, to varying degrees, all over the U.S., is failing on both fronts.
The mindset that positions retreat as abandonment is sustained by an idea that remains central to our national identity. Fueled by a vast expanse of appropriated land and a limitless cache of cheap energy, the American project of the last two centuries has largely centered on a process of continuous land development. Initially in the form of a westward migration and more recently in the form of sprawling urbanization, the most essential engine of American growth remains rooted in an ongoing process of spatial expansion.
Now imperiled by too much water along its coasts and too little within its interior, the American frontier is for the first time contracting. Neither our national identity nor our governing institutions have yet reconciled with the inevitability of a spatial retrenchment. But it is in retrenchment that we may find a new unifying project — particularly in our cities.
The great urban challenge of our time is not simply climate change — it is how to tackle the multi-generational problems of social justice, affordable housing, meaningful employment and other dimensions of community well-being within the context of a rapidly changing climate. Rebuilding our cities, rendered imperative by a delayed response to climate change, will require a radically altered approach to managing flooding, to lessening heat exposure and to coping with drought — and it will require, for all of these purposes, a physical restructuring of our urban landscapes.
Moving forward, every building and land parcel will need to absorb a large fraction of the rainwater it receives, expand green cover for climate regulation and be integrated into a far more decentralized system of power generation and use. This adaptive rebuilding project cannot be separated from other longstanding urban challenges of equal import; it is a central mechanism through which expansive affordable housing, long-delayed environmental justice and a broader community revitalization can be realized. In this sense, the path forward is not to be found in climate adaptation but rather in an adaptive urbanism.
Retreat is the essential catalyst of adaptive urbanism; it is a process through which threatened or underutilized urban land is repurposed for enhanced climate, social and economic resilience. Such a movement is beginning to take shape in large cities worldwide. In Denmark, shipping containers are converted into affordable, amphibious housing. There and elsewhere, new floating communities expanding into the underutilized slips of deindustrializing dockyards offer a way to decouple life from land ownership — and they are perfectly positioned to harness cheap solar energy.
Other forms of amphibious housing constructed in expanding floodplains can rise and fall with periodic flooding, reimagining otherwise hazardous flood zones for low-cost housing. More than 10,000 floating homes are now in place in the Netherlands alone; the oldest have now successfully weathered more than 20 years of climate-intensified storm events.
In New York, as much as 50 miles of parking lanes have been converted to “parklets” and “streeteries” — gathering and dining spaces that have expanded from sidewalks into the street itself, reclaiming a small fraction of the vast amount of Manhattan’s land area that is set aside for vehicles. Better still would be if the parking lanes were removed entirely and replaced with an integrated network of bioswales (green areas along roadways for rainwater collection), bike lanes and outdoor dining spaces.
A larger-scale vision of adaptive urbanism includes repurposing surface parking lots and one-story buildings for multistory affordable housing integrated with rooftop renewable energy and stormwater collection. A recent analysis found that redeveloping underutilized land across New York could provide affordable housing to more than a million residents. Virtually all of these building sites are presently occupied by surface paving or low-rise buildings, so a rebuilding process aimed at addressing the city’s affordability crisis could be equally productive in lessening flood, heat and drought risk if designed for these purposes.
In New York and other cities, the where of climate adaptation is again equal in importance to the what. Through its highly ambitious campaign to add a million trees to the city’s canopy that began in 2007 — a campaign undertaken, in large part, to lessen heat exposure and flood risk — New York directed more than 80% of the new trees to neighborhoods already well endowed with public greenspace at the expense of lower-income and more racially diverse areas in which parks development has been underfunded. Moving forward, the opposite imperative should guide public investment in adaptive urbanism, with the most climate-vulnerable zones prioritized in time and proportion of funding — a principle I refer to as “least-first.”
The essential lesson from experiments in Copenhagen, Amsterdam, New York and a growing number of cities is that retreat is more than a process of retrenchment. One of an array of terms unhelpfully framing climate change as a mode of warfare, retreat can also be understood as a mode of transformation. Derived from the Latin infinitive retrahere, to retreat is to “pull back”; the old French word treat, derived from the same Latin root, translates as “deal with.”
Climate change is not a battle to win or lose but a dynamic set of environmental conditions that we now must deal with. Planned retreat can be both a process for relocating the most vulnerable out of harm’s way and of leveraging publicly acquired land for a greater urban resilience — one directed toward adaptive infrastructure and societal need. It is not overly ambitious to suggest that we can make our cities more resilient by making them more equitable, more beautiful and more tethered to their underlying ecology; the earliest experiments make this obvious.
Retreat, when framed as transformation, is not the end of the American frontier but its much-needed reimagining. The initial step is to make room for a broader resilience. To remain, we must first retreat.
