When Hurricane Harvey struck Houston in August 2017, Giselle Mendoza was coaching gym class at Earl & Hazel Harris Elementary School on the city’s northern edge, about a mile from her house. As more than thirty inches of rain filled creeks, culverts, streets, and empty fields, creeping up to the doors of the school and rushing through nearby neighborhoods, Mendoza thought of her family’s one-story brick bungalow on Cypress Grove Lane.
Through sheer luck, Harvey spared the house, even as it wrecked much of the surrounding area. As Mendoza looked around her front yard more than two years later, on a warm afternoon last November, life seemed to have returned to normal. Her mother and sister were inside making dinner, her niece and nephews played tag on the front lawn. She was still coaching gym at Harris Elementary. There was just one thing missing from the picture: neighbors.
The lot next to Mendoza’s house was empty except for a few trees. So was the next one down, and the next one, and the next one. In fact, all but a dozen of the hundred-plus lots that once made up the Arbor Oaks subdivision were vacant. Their inhabitants had moved out one by one since Tropical Storm Allison in 2001, allowing the Harris County government to purchase and demolish their homes so they could escape one of the most flood-prone pieces of land in the continental United States. The result is a neighborhood that feels more like an unclaimed frontier than a subdivision in one of the nation’s fastest-growing metropolitan areas. Dogs run free across acres of empty land, the sound of a falling acorn echoes for blocks, and if you listen closely, you can even hear the soft gurgle of the creek on the neighborhood’s eastern boundary—the creek that, if another Harvey struck, could overflow and wash what’s left of Arbor Oaks away.
“Don’t get me wrong, we’ve thought about moving,” Mendoza told me. “Every time it rains now, we get nervous, we wonder if we’re going to get flooded—the [county] came around and they told us about the risks. We understand why everyone left, but for us it’s not that simple. This was always our neighborhood—I work around here, I’ve lived here my whole life.”
It’s no secret that climate change is already displacing people across the country, forcing coastal communities to move inland as the sea threatens to swallow them. The federal government is already sponsoring the wholesale relocation of a sinking Native American village in southern Louisiana and another Native settlement in Alaska. An island fishing town in the Chesapeake Bay may soon need to be evacuated as well. But there is another kind of climate migration happening alongside these high-profile evacuations. In areas like Houston, growing flood risk has induced a patchwork, household-by-household retreat from vulnerable floodplains and beachfront towns.
Funded by the federal government, local governments in coastal states are buying out thousands of homes in vulnerable areas every year, reshaping and breaking up communities as they go. In their wake, the departed residents of these communities have left what may be the country’s first climate ghost towns, abandoned places made uninhabitable by the warming of the planet. The vacant lots of Arbor Oaks and neighborhoods like it provide a stark warning about the shortcomings of the government’s haphazard, market-driven response to climate change, and raise difficult questions about the rights of people who live in the path of climate disaster.
In the wee hours of December 26, 1802, as the sleeping residents of Portsmouth, New Hampshire, digested their Christmas dinner, a fire started in a wooden building downtown. The blaze roared through the city, destroying more than a hundred buildings and leaving many of Portsmouth’s residents homeless. Because most of them had no insurance, recovery assistance came almost exclusively from wealthy local benefactors—until the thirteen-year-old U.S. Congress stepped in to help. In January of 1803, legislators passed a bill to temporarily suspend the “the collection of bonds due by merchants of Portsmouth, in New Hampshire, who have suffered by the late conflagration of that town”—the first disaster relief law in the nation’s history.
The federal government’s approach to such disasters only got more complicated from there. By the time Congress finally passed a comprehensive disaster policy in 1950, there were more than 128 different laws on the books. Even after that, Congress revised its mega-law multiple times—prompted by the likes of Hurricanes Agnes, Betsy, and Carla, plus an earthquake in Alaska—before it settled on the 1988 Stafford Act, which still largely regulates the government’s response to “major disasters” and provides statutory authority to the Federal Emergency Management Agency (FEMA). The Stafford Act represented the first major attempt by the U.S. government to plan for disasters before they happened: the law allowed FEMA to set aside a small percentage of its annual spending for disaster mitigation.
But is climate change a “major disaster”? The Stafford Act mentions “any hurricane, tornado, storm, high water, wind-driven water, tidal wave, tsunami, earthquake, volcanic eruption, landslide, mudslide, snowstorm, or drought . . . or, regardless of cause, any fire, flood, or explosion.” Although scientists agree that rising temperatures have already intensified flooding and drought in the United States, it’s difficult to pinpoint the exact relationship between incremental trends and an individual storm—a tidal wave would seem to qualify as an “emergency” (from Latin emergere, “to rise up”) more clearly than an uptick in the monthly heat index, for example. As a result, the federal bureaucracy tasked with handling response to these disasters still treats them as individual events rather than constituent parts of a larger crisis—despite the provision for mitigation strategies established by the Stafford Act, FEMA usually only steps in with a checkbook after the damage is done, and rarely devotes serious resources to planning for the next storm, the next drought, the next fire.
Nearly twenty thousand homes built since 2010 are at a significant risk of chronic flooding by the mid-twenty-first century.
Even as FEMA has started to invest in disaster mitigation over the past thirty years, its efforts have proven woefully insufficient for fighting climate change, the consequences of which seem to grow more extreme and unpredictable every time we forecast them. The agency has funded backup power generators in coastal New Jersey, elevating homes on beaches in North Carolina, and drainage projects in Florida; the city of New York and the Army Corps of Engineers, meanwhile, have proposed storm surge barriers in New York Harbor and sea walls around Staten Island and Lower Manhattan. No one would argue that these projects are useless, but they represent an outmoded way of thinking, one that assumes Mother Nature wreaks havoc in a relatively stable and predictable manner. Recent storm events in particular have disproved that assumption, becoming so monstrous and erratic over the past decade as to make designing the flood barriers of the 2030s and the 2040s a task akin to dressing for the weather on this day ten years from now. Hurricane Katrina overwhelmed the levees in New Orleans, the storm surge from Hurricane Sandy raised the sea level in New York by more than nine feet, and Harvey flooded two of Houston’s reservoirs to historic levels; engineers could learn from these lessons, but what about the lessons no one sees coming? The other question, of course, is why invest in a barricade that you know will someday burst, if not in the 2040s then in the 2050s?
It’s not only construction projects that are foundering in the face of ever-greater storms. FEMA’s National Flood Insurance Program (NFIP) also lacks the wherewithal to cope with the climate crisis. By collecting insurance payments from families living in flood zones, the program was once relatively stable—premiums collected from a family in Iowa would help a family in Louisiana one year, and vice versa the next. While the program often had to borrow money from the Treasury, the sums were once small and quickly repaid. But as the pace of natural disasters has escalated, the cost of flood recovery has outstripped the amount the program’s five million participating households can afford to pay into the central pot, and it has fallen into unprecedented debt. The 2005 hurricane season forced the NFIP to take on over $200 million in debt; by the time Sandy hit seven years later, it was over $17 billion in the red. As of last year, the program was more than $20 billion in debt, a hole so large Congress has hesitated to plug it, instead demanding that FEMA raise their rates, particularly on homes prone to repeated flooding—like one house in Houston that flooded on twenty-two separate occasions between 1979 and 2017, filing an insurance claim each time.
It doesn’t help, of course, that FEMA and other local authorities are stuck working to undo the mistakes of the shoreline mayors and governors who offered tax breaks to subsidize massive development along the Atlantic seaboard. This gold rush is still ongoing, as evidenced by Miami’s white-hot real estate market, among others: according to Zillow, nearly twenty thousand homes built since 2010 are at a significant risk of chronic flooding by the mid-twenty-first century. The people who continue to live in these areas are doing something so risky that the government will at some point no longer be able to subsidize it, whether by repairing homes after a flood or fortifying them against the next one. As the rains get heavier and the tides get higher, many coastal residents and the politicians who govern them are already facing a choice—batten down the hatches or pack up and leave.
A.R. Siders, an assistant professor of public policy and geography at the University of Delaware, likes to say that for every shoreline, there are two coasts. “You have some places where there are high property values, high development, and people are incentivized to build there,” she says—think South Beach or the Hamptons. “But then you have these other communities in the same states, where [people are] not living there because it’s prime real estate or they’re wealthy, they’re living there because they historically can’t afford to live anywhere else.”
Siders is the author of “Managed Coastal Retreat,” a compendium of all the tools that governments can use to encourage retreat from land exposed to rising seas and historic floods. She first got interested in the topic while living in New York after Sandy when, she says, “there were a lot of conversations . . . about how to rebuild, but the conversation that was missing was, should we rebuild.” Since then, she has become one of the country’s foremost experts on climate retreat policies—and perhaps their foremost advocate.
Many people find the idea of retreating from the coasts to be hasty or overly pessimistic, Siders says, but it’s not just Houston neighborhoods like Giselle Mendoza’s that are following her lead. FEMA buyouts have been concentrated in flood-prone areas such as New Jersey and southern Mississippi, and the Department of Housing and Urban Development has ramped up its own buyout program, including in regions of Florida that were struck by Hurricane Irma in September 2017. Some states have instituted laws like California’s Coastal Act, which requires new developments to build with adequate “setbacks” from coastal bluffs to account for future shoreline erosion. Even memelord presidential candidate Andrew Yang is on board with retreat: “We need to do everything we can to start moving the climate in the right direction,” he said at a debate last year, “but we also need to start moving our people to higher ground.”
Siders’s research has focused on teasing out the implications of a statement like Yang’s. After all, what is “need,” and who are “our people”? In the eyes of the Harris County Flood Control District, the retreat calculus is pretty much black and white: if the county can’t mitigate flooding, they do a buyout. But for Siders it’s more complex than that. When a government buys out a home, she says, it’s acting on the results of an economic analysis, a comparison between the costs of a flood barrier and the benefits—political, social, and economic—of keeping a community where it is.
In coastal North Carolina, for instance, Siders and one of her fellow researchers found in a forthcoming study that high home values, high income levels, and low racial diversity were all strongly correlated with “shoreline armoring” projects like seawalls. In low-income communities, meanwhile, property acquisitions were far more likely. Other parts of the country reflect the same trend: in the affluent beachfront community of Del Mar, California, city officials recently rejected a state commission’s managed retreat proposal, instead opting to rely on developing levees and fortifying sandbars so that the community could stay put; their argument was that buyouts wouldn’t work because local property values are too high. Similarly, in New Jersey, beachside homeowners have successfully sued for as much as $380,000 in compensation after the state built protective dunes on their property. In Houston neighborhoods like Arbor Oaks, though, where property values are lower, the government may conclude that fortifications are not worth the money, regardless of how residents feel. Where buyouts do occur, they lead to different outcomes for the rich and the poor, disproportionately benefiting wealthier and whiter families who have the means to sustain themselves as they buy new houses and settle in new communities.
As the rains get heavier and the tides get higher, many coastal residents and the politicians who govern them are already being faced with a choice—batten down the hatches or pack up and leave.
The biggest challenge of managed retreat may be that it requires breaking a community apart in order to ensure the survival of its members. Although it’s been estimated that a majority of residents who take buyouts from Harris County end up staying in the Houston area, the social and spatial bonds that linked the former residents of a neighborhood like Giselle Mendoza’s will never be recreated. Even if New York City one day builds a seawall around Manhattan, protecting the East Village forever from the kind of dissolution that took place in Arbor Oaks, hundreds of smaller and less affluent communities will fall apart piece by piece in the coming decades as the storms continue.
The only contemporary counterexamples, says Siders, are the aforementioned pair of Native American villages in Louisiana and Alaska that the federal government may soon relocate in their entirety. But special cases notwithstanding, community relocation is not well-supported by U.S. law. Despite the fact that the United Nations Guiding Principles on Internal Displacement, which outlines the rights of people forced to migrate within their own country, holds that governments must safeguard the “life, dignity, and security” of the internally displaced, outside of these special cases, Americans displaced by weather events often have at most the right to material compensation for whatever property they lost during a natural disaster. Such a system, pegged as it is to real estate valuations, will always reproduce the inequality that existed prior to the catastrophe, if it doesn’t make that inequality even worse.
There Will Come Heavy Rains
Perhaps more than any other major city, Houston embodies the short-sighted development that Siders’s research describes. As early as the 1950s, when developers began pushing the boundaries of the city onto the surrounding prairie and wetlands, there was no mystery about the vulnerability of the land they built on (buyers, of course, were rarely told). But Houston let them build anyway, greenlighting construction in and around the Brazos and Cypress Creek floodplains, along the edges of the Addicks and Barker Reservoirs, and right up to the banks of the San Jacinto River that flows out into Galveston Bay. Millions of homes were built on former bayous and marshes that sit just fifty feet above sea level and slope at a rate of about one foot per mile.
At first, the Houstonians who bought homes in these swampy suburbs were mostly spared of major weather events, but heavy downpours in the area have increased by 167 percent since the 1950s, and the county started a buyout program in the 1980s to clear out homes on a few vulnerable stretches of land beside Cypress Creek. The real wake-up call, though, didn’t come until 2001, in the form of Tropical Storm Allison. Allison parked over Houston for days, dumping nearly forty inches of rain and flooding some seventy thousand homes. In the aftermath of the storm, hundreds of homeowners participated in buyouts.
“The best-case scenario is to buy the properties and let it go back to what it was before it was developed,” said James Wade, the director of Harris County’s flood buyout program. “There’s really nothing from an engineering standpoint that we can do to reduce flooding.”
The biggest challenge of managed retreat may be that it requires breaking a community apart in order to ensure the survival of its members.
Harris County’s buyout program is the largest in the country by far—they have purchased more than three thousand homes, leaning on the program whenever bayou engineers can’t find a cost-effective way to defend nearby homes. But the process is far from simple. The county sends fliers to all the houses it considers endangered, then comes around and knocks on their doors, but even nearly twenty years after Allison, there are still holdouts here and there, like the Mendozas and their neighbors. If people don’t want to move, the county’s not likely to force them: FEMA only funds voluntary buyouts. The result is what Wade calls a “gap-tooth” or “checkerboard” buyout pattern, where occupied lots are scattered around an area with vacant lots in between. That’s the case in Arbor Oaks, where remaining houses sit together in little clumps of three or four, separated from each other by silent streets as long as football fields.
“We try to educate people, but there are some places where they aren’t as interested,” Wade says. “Maybe they’ve gotten lucky the past few hurricane seasons, or they have a misconception that maybe the problem isn’t that bad . . . but the problem is not fixable.”
Wade doesn’t think of his work as directly related to climate change—after all, he has his hands full evacuating homes that never should have been built in the first place. In the nearly two decades since Allison, the department has mostly pursued buyouts on houses that it considers “hopelessly deep”—those sitting at least two feet deep in FEMA’s hundred-year floodplain. But the map the county uses right now to measure that floodplain is over ten years old and relies on rainfall data from the 1960s, according to Wade. In 2018, when the National Oceanic and Atmospheric Administration (NOAA) finally released new rainfall predictions for Texas, the agency found that storms which were previously considered once-in-a-century events could now be expected every twenty-five years. A bigger hypothetical storm means a greater risk of flooding, which in turn has led FEMA to redraw its flood insurance rate maps, widening the hundred-year floodplain. 
When the new maps take effect in the early 2020s, Wade says, the floodplain will get much larger. The early projections he’s seen suggest that in Mendoza’s section of the city, whole subdivisions that the county now deems relatively safe will soon become buyout targets. In some parts of the city, the floodplain area will likely expand from a thin stripe on either side of a creek into a swath of land a mile wide, covering thousands of homes and businesses. But the county will not be able to start the buyout process for five years or more, until it has worked its way through a mountain of federal paperwork.
Even once a household has agreed to a buyout, the process can drag on for years: after Harvey struck in August 2017, FEMA didn’t allocate its first tranche of buyout money until the following June. While families waited for the federal paperwork to clear, Wade says, private investors swept in with cash in hand and bought several properties the county was targeting, flipping them to new owners and rewinding the buyout process back to square one, with new families. The county could use state funds to expedite the purchasing process, and maybe even invoke eminent domain, but Wade estimates that it would cost $1 billion to buy the remaining properties in his target areas, and only FEMA, underfunded as it may be, has anything like that kind of money. If the next five years are as bad as the last five, many of Wade’s future buyout targets may not be around by the time he has the ability to buy them.
“Some of these places are ten feet deep in the floodplain, and their [residents’] lives are at risk,” he says. “There’s been a lot of talk about, at what point are we putting them in harm’s way by letting them stay there?”
A Pocket Prairie Companion
Memorial Park is Houston’s marquee green space and an important place in the city’s history. The site of one of the first homesteads in the area, the land that now makes up the park also served as a training ground during World War I; in 1917, it was the site of an infamous race riot when a black infantry regiment retaliated against the racist city police department. After the war, the area was purchased by the oil-rich Hogg family and sold to the city at cost with the intent that it would remain green space in perpetuity.
It’s here that conservationist Jaime Gonzalez and his colleagues at The Nature Conservancy are charting out the first steps of a plan to make Houston a climate-resilient city—not by trying to keep the water out but by letting it in. The patch of weeds he shows me in a corner of the park, dripping with wildflowers and humming with insects, is a time capsule, an approximation of what the ground we’re standing on looked like for millennia before Houston was settled. It turns out that these prehistoric grasses are a lot better at absorbing water than a concrete parking lot or even a close-cropped lawn: when Harvey hit, the restored patches of “pocket prairie” here in the park and elsewhere are what sopped up a lot of the water, filtering it down into the earth.
If the government’s future approach to retreat is anything like the one it has taken thus far, families who can’t move or don’t want to will become the living collateral of the retreat process.
“A lot of our job, really, is to physically uncover the environment so we can have a more resilient place,” says Gonzalez. Normally, he says, the city mows the disused grass under its power lines multiple times a year, which generates carbon emissions and prevents the ground from holding water in the event of a flood. Wild grasses like these, on the other hand, store water as well as carbon, don’t need to be mowed, and also help improve Houston’s notoriously bad air quality.
“For some people, that tall grass [in Memorial Park] seems like being a bad neighbor,” he says. Many homeowner associations in the city, as a matter of fact, require owners to keep their lawns mowed at a certain length. “But my counterargument to that is: Is having a lawn that comes with so many ecological problems, that doesn’t allow water to infiltrate, is that being a good neighbor? We’re going to have to make the city function more like an ecosystem.”
This Memorial Park project, which has turned the disused land under a power line into a resilient stretch of grassland, is just one of several initiatives Gonzalez has spearheaded around Houston—a series of small suggestions that the best way to recover from Harvey might not be to rebuild, but rather to de-build. The Nature Conservancy has created similar ancient prairies at Rice University and the Texas Medical Center, and Gonzalez suggests they could do the same in subdivisions where the county has finished buying out all the homes. The empty lawns around Mendoza’s house could turn into a natural drain that would bear the brunt of the next storm. But first, of course, everyone who lives in these areas has to leave, and until they do, the county is doomed to mow hundreds of lawns month after month, protecting a conception of land use that Gonzalez says needs to change if a city like Houston is to survive climate change.
Later on, though, as we drive through the low-income neighborhood of Gulfton, where the vast majority of the population are immigrants, Gonzalez expresses sympathy with families like Mendoza’s, those who can’t imagine picking up and starting a new life. “We can’t just say, well, we’re going to buy people out at random,” he says. “[There] are entire neighborhoods that are now within the hundred-year floodplain,” and many residents are living there because that’s what they can afford. They aren’t likely to find comparable housing elsewhere for what the county can offer them in a buyout.
Growing up not far from Arbor Oaks, in the predominantly Hispanic suburb of Aldine, Gonzalez says that in his childhood, the area was “still very wild,” and he spent a lot of time in “the woods, prairies, creeks, and bayous” around his house. When he came back from college, though, that wildlife had vanished: new subdivisions had sprung up around Aldine, and the forest where he used to play had been cut down and turned into a junkyard. The argument he wants to make now, he says, is that those bygone creeks and swamps were not a threat that needed to be contained and controlled, but the solution to the existential crisis of climate change. “Our future is part of what we lost,” he says. “When the pioneers got here, this was an incredibly naturally rich, very ecologically stable place, and in less than two centuries we’ve made it unstable.”
Gonzalez’s vision of a resilient urban ecosystem is a refreshing one, but the scale of a storm like Harvey sometimes makes it hard to feel like he’s doing anything more than tinkering at the margins. Houston can try as it might to retrofit itself for the unforeseen dangers of climate change, yet the fact remains that its residents are in danger, as Giselle Mendoza says, “every time it rains.” Even if the federal government ever faces down the full extent of this threat, it will no longer be fiscally responsible or even logistically possible to protect neighborhoods like hers with levees or seawalls or sandbars. The only option will be to get out of the way. But if the government’s future approach to retreat is anything like the one it has taken thus far, families who can’t move or don’t want to will become the living collateral of the retreat process, their neighborhoods emptied and lives uprooted because of forces well beyond their control: not only the weather itself, but the policies of the people who govern them. When the last house in Arbor Oaks finally comes down and there are no homes left for the next flood to destroy, some will count it as a victory that residents got out of the way in time. But even that will only be a small consolation for the larger battle, the one we have already lost.
 Since even the most advanced engineers can’t predict exactly where a river will overflow or a levee will burst, FEMA analyzes landscape topography and weather patterns to calculate the likelihood that a given property might flood during the average storm season. If there’s a more than 1 percent chance of a house flooding, then that house is in the “hundred-year floodplain” and must purchase insurance—not because it will flood, but because it could.