Under a White Sky

By Elizabeth Kolbert

9 min read

Why does it matter? Because the experts fixing the environment are all cleaning up after the experts who fixed it before them.

The assumption most readers bring: environmental problems are discrete. Asian carp threatening the Great Lakes — that's an invasive species problem. Louisiana losing a football field of land every ninety minutes — a climate problem. A fish that fits in a teaspoon — a conservation problem. Each crisis has its own experts, its own bureaucracy, its own potential fix.

What Kolbert found is stranger. In chapter after chapter, the experts solving today's crisis are intellectual heirs of the experts who caused it. The canal that reversed Chicago's river to fix its sewage problem is the exact pathway now carrying the carp. The levees that saved New Orleans from flooding starved the delta of sediment that was holding the land up. The same impulse that broke things is the only tool available to repair them.

You Can't Fix a Nature Problem Without Creating the Next One

Chuck Shea, an Army Corps engineer in Chicago, extends his hand across the table. His fingertip is the nose of a fish. His palm is its tail. He wiggles it. This is the physics of the electric barrier on the Chicago Sanitary and Ship Canal: pulse enough current into the water, and the voltage difference between nose and tail drives a shock through the body. Asian carp, some growing past a hundred pounds, are the target. "Public enemy number one," Shea says. Kolbert asks about people, who are also fairly large. "All people react differently to electricity," he replies. "But the bottom line, unfortunately, is that it can be fatal."

What the demonstration doesn't show is how a Corps engineer ended up spending his career on a fish. That story is a cascade.

In the late nineteenth century, Chicago's sewage — human waste, slaughterhouse runoff, stockyard filth — drained into the Chicago River and from there into Lake Michigan, which was also the city's only drinking water source. Typhoid and cholera were routine. The solution: dig a canal that reversed the river's direction, flushing waste away from the lake. The project excavated 43 million cubic yards of rock. It worked. The water got safer.

The canal also quietly joined two aquatic worlds that had never touched. The Mississippi drainage basin covers 1.2 million square miles. The Great Lakes basin holds 80 percent of North America's fresh surface water. Before the canal, no fish could cross between them. For most of the twentieth century, the canal was too toxic to matter. Then the Clean Water Act cleaned it up, and creatures started slipping through.

Asian carp arrived as a direct consequence of Rachel Carson's Silent Spring, a book whose working title, with painful irony, was The Control of Nature. Carson's attack on chemical pesticides pushed agencies toward biological controls. Carp were an obvious choice: filter feeders efficient enough to clean nutrient-heavy water without chemicals. In 1963, the U.S. Fish and Wildlife Service imported the first documented Asian carp shipment. By the 1970s, Arkansas was stocking silver carp in sewage lagoons under Clean Water Act pressure. They escaped. "No one is quite sure how," Kolbert writes, "because no one was watching." Silver carp now make up nearly three-quarters of the fish biomass on the Illinois River.

Hence Shea and his hand. Hence the electric barrier, first estimated at $275 million, now approaching $775 million. Hence the permanent fix (severing the canal link) that every expert agrees would work and no expert believes will actually happen. The price: $18 billion and 25 years.

First you reverse a river. Then you electrify it.

The Engineering That Saved New Orleans Is What's Destroying Louisiana

The Army Corps levees saved New Orleans from flooding. The same Army Corps, a different river, the same logic. They also guaranteed Louisiana would disappear.

That argument is made most vividly not in the bayous but in a building at Louisiana State University, where researchers built a 1:6,000 scale replica of the Mississippi delta: two basketball courts of machined foam, from Baton Rouge to the Bird's Foot. When the model runs, a vat injects jet-black plastic pellets as simulated sediment. During a simulation of the record 2011 flood, the pellets tell the whole story. Instead of spreading across the marshes the way pre-levee floodwaters would have, they stream straight through the engineered channels, past New Orleans, around the Bird's Foot, and out to the Gulf, where they vanish off the continental shelf. Someone in that building runs this simulation for visiting engineers, for funders, for state officials. The pellets always do the same thing.

For millions of years, spring floods deposited fresh layers of sand and clay across the delta, building land faster than it could compact and sink. The 1928 Flood Control Act nationalized the levee system. No more flooding meant no more sediment, and land that needed annual replenishment started subsiding instead. Since the 1930s, Louisiana has lost more than 2,000 square miles, and the rate hasn't slowed.

The engineering response is to dredge sediment from the riverbed and pump it miles through steel pipe to re-create marsh. The new land starts subsiding immediately and will vanish within a decade. Louisiana would need to complete a project of that scale every nine days just to match the current pace of loss.

New Orleans presents the sharpest version of the trap. The city's 120 drainage pumps drain the marshy soils beneath it; those soils compact as they dry; the city drops — in some neighborhoods, half a foot per decade — which demands more pumping. The Melpomene Pumping Station, built in 1920 to expand the city into the old swamps, is a machine that keeps worsening the problem it was designed to solve. The levees hold the water out. The pumps pull the ground down. The engineering that kept the city dry is what's swallowing it.

We Have Accepted Permanent Custody of Every Species We Nearly Killed

On a summer morning in Nevada, five government biologists assemble at a fenced cavern in 105°F heat to count fish. Two scuba-dive into a pool descending more than five hundred feet — two divers explored it in 1965 and never came back. The others kneel on a catwalk and count fish on an underwater limestone shelf. The final tally: 195. High-fives. A biologist does what he calls "a little happy dance." Someone later calculates that the entire wild population of Devils Hole pupfish weighs about 100 grams — the weight of a fast-food fish sandwich.

These inch-long sapphire fish live in a single subterranean Nevada pool, the smallest range of any vertebrate on Earth. In the 1960s, a developer pumped the aquifer for alfalfa farming. The water level dropped, the breeding shelf was exposed, the population crashed. The Park Service won a Supreme Court case nine to nothing, but the aquifer never fully recovered. Since 2006, staff have been delivering supplemental meals of brine shrimp to the pool.

A mile away sits the backup: a $4.5 million concrete hangar housing an exact replica, with a Styrofoam shelf, a louvered ceiling calibrated to seasonal sun angles, a heating system holding water at 93°F, and an imported food chain. Four full-time staff manage roughly fifty adult fish.

The imported food chain turned out to be a problem. A poppy-seed-sized beetle arrived with the other invertebrates, developed a taste for pupfish larvae, and bred out of control. Staff now tweeze them from traps every single day. Import the ecosystem, and the ecosystem starts eating your fish. Manage that, and you've added another layer of management.

Kolbert calls species like this "Stockholm species" — creatures entirely dependent on their persecutors. If those four employees stop coming to work, the backup population is gone. The backup pool is paradoxically shielded from human disruption by being wholly a human artifact. That is the structure now: we didn't just nearly kill the pupfish. We became the apparatus keeping them alive. Protection was always the wrong word. There is no version of this story where we hand the custody back.

Gene Drives Don't Just Modify Organisms — They Rewrite Entire Wild Populations

What makes a gene drive different from every other intervention in this book? Not the biology — the permanence.

Normal inheritance is a coin flip — fifty-fifty, every generation. Some genes load that coin, passing themselves on at rates of sixty, seventy, even ninety percent. CRISPR, the gene-editing tool now used in medicine and agriculture, makes it possible to engineer this cheating deliberately. The trick is self-referential: because the CRISPR machinery is itself encoded in DNA, you can build an organism that carries instructions to reprogram itself. Release it into a wild population and it rewrites its species, generation by generation, without further human action.

Paul Thomas, a mouse geneticist in Adelaide, is working on what he calls an X-shredder. A normal male mouse has one X and one Y chromosome, and his sperm carry one or the other in equal measure. The X-shredder mouse has been gene-edited so that all his X-bearing sperm are defective — they simply fail. Only Y-bearing sperm reach the egg, so all offspring are male. The shredding instructions travel on the Y chromosome, so those male offspring are also X-shredders, who also produce only sons, and so on. The female population shrinks every generation until there are no females left to reproduce. Mathematical modeling suggests that releasing a hundred such mice onto an island could bring a population of fifty thousand ordinary mice to zero within a few years.

A hundred animals released. Fifty thousand eliminated. That ratio is the difference between a gene drive and everything else in this book. The electric barrier in Chicago, the sediment pumps in Louisiana, the daily beetle-tweeze in the pupfish tank: all of those require continuous human effort. A gene drive requires one release. After that, it propagates itself.

There is no version of this where you change your mind. The same logic that produced the canal, the levee, the backup tank (do something, fix the damage, manage the consequences) has arrived at a tool that cannot be recalled. Proposed fail-safes exist on paper. Whether they work in a wild population, across continents (drives follow whatever migration route their host species uses), over decades, is a question that can only be answered after the fact.

The Climate Fix That Cannot Be Turned Off

In April 1815, Mount Tambora erupted, injecting more than a hundred million tons of gas and fine particles into the stratosphere. Within 250 miles, day turned to night. The particles drifted on high-altitude winds for years, scattering sunlight back to space. The following summer, frost killed crops in Massachusetts in August. Harvests failed from Ireland to Italy. Lord Byron, trapped indoors on Lake Geneva by relentless rain, proposed a ghost-story competition. Mary Shelley's contribution was Frankenstein.

Harvard's Solar Geoengineering Research Program invokes Tambora not as a cautionary tale but as proof of concept. If a single eruption can cool the planet, a managed program of stratospheric aerosol injection could do the same. The logic is direct: spray sulfur dioxide into the stratosphere, it oxidizes into fine droplets, those droplets scatter sunlight, temperatures stabilize.

Frank Keutsch, one of the program's leading chemists, started out thinking the idea was "entirely crazy." He came around when he worked through what 3°C of committed warming does to food systems by 2080. The April 2020 COVID lockdowns cut global emissions by 17% (the largest drop ever recorded) and atmospheric CO2 still hit a record high the following month.

What the Tambora frame conceals is the difference between a volcano and a program. Tambora's particles settled after a few years. The cooling ended, the world warmed back toward where it would have been, and life resumed. A natural exit, built into the physics. No decision required. A managed geoengineering program would suppress warming not for years but for decades, long enough for CO2 to accumulate to catastrophic levels beneath the artificial shade. Then, if the program stopped for any reason — political collapse, war, funding failure — all that masked warming would arrive at once, a phenomenon scientists call termination shock. Possibly within a decade.

Every other intervention in this book can be abandoned. The electric barrier in Chicago needs electricity. The sediment pumps in Louisiana need diesel. Stop them and the consequences, however bad, are finite. Geoengineering has no equivalent off-ramp. Stopping the program doesn't just end the cooling. It releases the warming the program was holding back. Any government that starts has committed, in the same moment, to never stopping. Or to handing whoever comes next the full bill for everything deferred, at a moment no one can predict or control.

Civilization Only Happened Because the Climate Was Briefly, Freakishly Stable — And We're Ending That

Imagine Greenland's last 110,000 years of climate as a patient's heart monitor. For a hundred thousand years, the line spikes and crashes — temperatures lurching upward by 14°F in fifty years, then plunging almost as fast, over and over, twenty-five times. Then, roughly ten thousand years ago, the line goes flat. Everything we know — every city, every written language, every harvest — happened during that flat line.

Glaciologist J.P. Steffensen, running a drilling operation on top of three kilometers of Greenland ice, posed the question to Kolbert during a midnight conversation: why didn't humans build civilization fifty thousand years ago? Their brains were the same size. The capability existed. His answer: the climate wouldn't allow it. Every time a culture began to form, people had to move — no stable growing season to plan around, no predictable river to build beside. Then the Holocene arrived — ten thousand years of quiet. Around six thousand years ago, civilizations in Persia, China, and India crossed a threshold at roughly the same moment, developing writing, building cities, founding religions. Not because human intelligence suddenly advanced. Because the climate finally held still long enough.

The evidence came from an unlikely place: a classified Cold War military drilling project in Greenland whose accidental legacy was the first complete ice core. Inside, Danish geophysicist Willi Dansgaard confirmed what the temperature layers showed — the Holocene's stillness is not Earth's default. It is a geological accident.

That's what current warming is actually gambling with — not an interruption of conditions you've always known but the end of the only conditions under which civilization proved possible. Over the last hundred and ten thousand years, only one period was as stable as the last ten thousand. We're in it.

The Binary Is Already Gone — What Remains Is Governance

J.P. Steffensen, the Greenland glaciologist, has a Danish saying he shared from his geodesic dome atop three kilometers of ice: pissing in your pants will only keep you warm for so long. That's the position. Not a metaphor for failure — a description of where the warmth is coming from, and how long it lasts. The choice between managing nature and leaving it alone was settled before anyone alive made it. What's open now is the narrower, harder question: whether the institutions governing electric barriers and sediment pumps and aerosol injections can operate at the speed this moment requires. Because the interventions we're designing don't come with off switches. Or rather, they do — and using them is the catastrophe. You can't unknow that. The question is what kind of institution is capable of governing an intervention it cannot afford to stop.