On March 24, 1989, at four minutes past midnight, the Exxon Valdez ran hard aground on Bligh Reef in Alaska’s Prince William Sound, tearing open eight of its eleven cargo tanks and releasing nearly eleven million gallons of North Slope crude into one of the richest fishing waters on the Pacific coast. The reef had been named for William Bligh, the sailing master on James Cook’s third voyage who would later become synonymous with maritime disaster when his crew mutinied on the Bounty. The irony was not lost on the people of Cordova, fifty miles away, who had spent the previous decade warning that exactly this would happen. Cordova had no road.
It still does not. You reach it by ferry or by air, and that isolation shaped everything about the town’s economy. In the 1980s, Cordova lived on fish, and the first big paycheck of the year always came in April when the Pacific herring arrived to spawn. Biologists measured the run in mile-days of milt, flying the coastline in small planes and counting the miles of shoreline where the water had turned opaque and milky white.
For the people of Cordova, that white water was the calendar. The herring were fished for their roe, which the Japanese market bought at extraordinary prices as kazunoko. A sane crew could work an opening measured in minutes and come out with a season’s income. The herring check paid the insurance, the boatyard bill, the net web and diesel that let a crew fish salmon in June, which paid the mortgage and got a family to the next April.
It was not one industry among several. It was the hinge the whole year turned on. Cordova had lost its one thing before. The town had been organized in 1906 to support a railroad built to haul some of the richest copper ore ever found down from the Kennicott deposits.
For twenty-seven years, Cordova was a copper town. Then in November 1938, the grade ran thin, the price ran low, and the last train came down from the mine. The company left, and the town changed crops, rebuilding itself on salmon and herring and the canneries that packed them. So Cordova already knew the shape of this story: one resource, one buyer, a boom, and then a decision made in an office two thousand miles away that ends it.
That local history explains why so many people in that town spent the 1970s shouting about oil tankers, and why almost nobody in the rest of America listened. The Trans-Alaska Pipeline needed a warm-water port, and Valdez was the obvious choice: ice-free, deep, at the head of a fjord. But the only way out of Valdez was through Prince William Sound, past Bligh Reef, and into the Gulf of Alaska. The Cordova District Fishermen United went to court and lost.
What the fishermen got instead was a set of promises folded into the Trans-Alaska Pipeline Authorization Act of 1973: state-of-the-art tanker safety, double hulls, radar coverage of the whole sound, and a response force at Valdez able to put boom around a tanker spill within hours. On the strength of that, Alaska granted the rights of way, and in 1977 the oil began to move. Then the promises came apart quietly, one at a time, over twelve years in which nothing went wrong. The double hulls were never built.
Full radar coverage was never installed. The vessel traffic service no longer reliably tracked ships all the way out past Bligh Reef, and it did not announce the change. Crew sizes were allowed to shrink. Harbor pilots began stepping off the tankers just outside the narrows.
Every one of those decisions saved money, and every one was defensible against a spotless record. Between 1977 and 1989, roughly 8,700 loaded tankers went out through the sound. Nothing had ever happened. That is the most dangerous sentence in industrial safety.
Sixteen hours before the grounding, a Cordova gillnetter named Ricky Ott stood up in a Valdez meeting room and told the mayor’s oil action committee that the fleet was playing Russian roulette. She had a doctorate in marine pollution and had sat on the board of Cordova District Fishermen United since 1987 with one assignment: fix Alyeska, the oil company consortium that ran the terminal. Then everybody went home. The Exxon Valdez was three years old, 987 feet long, and one of the newest ships in the fleet.
Her master was Joseph Hazelwood, a Sun Maritime graduate who had been the youngest captain Exxon ever appointed. He was also a man whose New York driver’s license had been suspended or revoked three times over alcohol, who had gone through treatment in 1985, and who had been arrested for drunk driving in September 1988. Exxon knew, and he was still taking loaded tankers out of Valdez. There was ice in the shipping lane that night, and Hazelwood asked the Coast Guard for permission to cross into the inbound lane to get around it.
That part was ordinary. Then at about 11:00, he handed the bridge to third mate Gregory Cousins with instructions to begin the turn back into the lane when the ship came abeam of Busby Island, and went below to his stateroom. Cousins was alone on the bridge. He had been on duty most of the day and had slept perhaps four hours the night before.
The ship’s raycast collision avoidance radar had not worked in over a year, and Exxon had not repaired it. A working set, the National Transportation Safety Board later concluded, would have shown him the reef coming. He ordered the turn. Nothing happened.
He ordered more rudder. Still nothing. By the time the bow finally began to swing, the ship was over the reef. For three days after the grounding, the weather was flat calm, exactly the condition in which oil can be boomed, skimmed, and burned.
Everyone who has studied the response agrees this was the window. The contingency plan had promised the spill would be encircled with boom within three hours and seven skimmers on scene within five. The reality was that Alyeska’s entire response inventory in Valdez amounted to thirteen skimming systems, about five miles of boom, and a single storage barge rated at 500,000 gallons. The oil eventually touched something on the order of 1,300 miles of shoreline.
And the barge was in dry dock, pulled for repairs without telling the Alaska Department of Environmental Conservation, which its own plan required. Roughly ten hours passed before it reached the ship. Instead of seven skimmers in five hours, three arrived at twelve. The Coast Guard and the state argued with Exxon about dispersants.
A helicopter dropped a load of Corexit and largely missed. There was one experimental burn that worked and was not repeated. For three days, the crew sat there in perfect weather while men in offices in Valdez, Houston, and Anchorage failed to agree on anything. On the night of the 26th, the wind arrived.
Gusts near seventy miles an hour came down the sound and whipped the crude into a brown emulsion that response crews call moose, heavier, stickier, and almost impossible to skim. In a single night, the oil became a thing on a coastline. It ran southwest out of the sound entirely, down past the Kenai Peninsula and the Alaska Peninsula, and eventually reached Kodiak, 450 miles from the reef. The herring were already gathering.
In two or three weeks, they would move into the shallow bays and lay their eggs in the kelp and rockweed of the intertidal zone, the exact band of coast where floating oil comes to rest. That is not a coincidence of bad luck. It is a structural fact about herring. The scale of what died was documented: between 100,000 and 250,000 seabirds, around 2,800 sea otters, some 300 harbor seals, 247 bald eagles, and 22 killer whales from two groups that researchers had been photographing individually for years and could therefore count exactly.
Fish and invertebrate losses were never countable at all. Those numbers ran on the evening news for a month, and they are why most Americans remember this event as a wildlife story. It was never only a wildlife story. In Cordova, the phone started ringing before dawn.
Exxon moved fast on the public relations front and slowly on everything else. Within days, the company’s senior man in Alaska, Don Cornett, flew into Cordova and stood up in front of a room of fishermen and their wives and children. He said the company would do whatever it took to keep them whole. He put his own name on it out loud: I said it, Don Cornett.
And added that Exxon did business straight. They believed him. The liability was obvious, and the man had said it in public with the Cordova Times in the room. Alaska closed the sound: herring, salmon, crab, shrimp, rockfish, and sablefish all shut for 1989.
The April herring openings never happened. So Exxon hired the fleet. The company chartered fishing boats to lay boom and ferry crews, and paid wages and day rates that a good year of fishing could not match. Over 11,000 people worked that cleanup at its peak, and a great many of them were Alaskan fishermen whose own seasons had been cancelled.
The biggest beneficiary was not from Cordova. Within weeks, Exxon named Veco International its lead contractor, and Veco belonged to Bill Allen, an Anchorage oil field man with a gift for being useful to powerful people. Veco built Exxon upwards of $750 million, and its own profit came to around $32 million. Allen sank it into the Anchorage Times and a pro-development campaign, burned $50 million in two years, sold the paper, and in 2007 pleaded guilty to bribing Alaska legislators.
And it split the town in half. The people who got Exxon contracts made more money in 1989 than they had ever made, and the people who didn’t watched them do it. Cordova invented a word for them: spillionaires. It was not affectionate.
In a place of two and a half thousand where everyone’s grandfather knew everyone’s grandfather, the company had introduced a line straight down the middle of the harbor, and it never entirely went away. The cleanup itself was enormous and, by the measure that matters, a failure. Less than 10 percent of the spilled oil was ever recovered off the water. The signature method on the beaches was high-pressure hot water washing, deployed on a scale never attempted before or since.
It moved a great deal of oil, and it also cooked the shoreline. The heat and pressure killed the mussels, barnacles, worms, and bacteria that live in the sediment, destroying both the food web and the natural machinery that breaks crude down. On some beaches, later studies suggested the washing did more lasting damage than the oil. The worst of it had names: Sleepy Bay, Snug Harbor, the Bay of Isles, Northwest Bay.
One of the beaches on the list is called Herring Bay, after what used to happen there every April when the water went white. Crews with hot water wands were still working in 1992, three years on. Exxon spent more than $2 billion on the response in the first year alone. In 1991, it settled with the federal government and the state of Alaska: $900 million in civil damages paid out over ten years, plus a criminal plea.
The revised plea came to $150 million, of which $125 million was forgiven in recognition of Exxon’s cooperation, leaving a $25 million fine and $100 million in restitution. Inside that 1991 civil settlement sat one clause that almost nobody outside the lawyers noticed at the time. It was called the reopener for unknown injury. If the governments later found substantial damage that nobody could reasonably have anticipated in 1991, they could go back to Exxon for up to another hundred million.
It was insurance against being wrong about how bad this was. Because the strange thing about Prince William Sound in the early 1990s was that it appeared to be recovering beautifully. In April 1990, the herring came back, and the numbers were good. The fishery opened again in 1991 and 1992, and those years produced some of the biggest herring harvests the sound had ever recorded.
The 1988 year class, the fish spawned the spring before the tanker hit, was a monster and was coming of age right on schedule. Exxon could point at record harvests and say the sound had bounced back. In Cordova, a fisherman could look at three good springs in a row, decide the disaster had been a scare rather than an ending, and borrow against his permit to buy a bigger boat. In 1990, the Prince William Sound salmon purse seine permit hit its all-time peak: $557,000 in today’s money.
That is what people believed was still in this water. Then came April 1993, and the fish did not show up. The planes flew the same shorelines and found almost nothing. The adult population had collapsed.
The 1993 harvest came in at roughly 14 percent of the previous years, and the 1989 year class, the eggs that had been sitting in the intertidal zone in the weeks after the spill, turned out to be one of the smallest cohorts ever recorded in the sound. Alaska shut the herring fishery. It stayed shut through 1996, opened briefly in 1997 and 1998, closed again in 1999, and did not open for a commercial season again for twenty-six years. The pink salmon failed the same season.
Two collapses in one year in the same water in a town that had just spent three years being told everything was fine. The disaster was not the worst part. The recovery was. Three good springs bought Exxon a settlement, bought the state a clean conscience, bought the fleet a round of optimistic debt, and then the bill arrived for a class of fish nobody had been watching.
By 1994, the same permit that had been worth $557,000 four years earlier was worth $63,300. By 2003, it was $19,700, a fall of something like 96 percent in real terms. If you owned one, you did not lose a season. You lost the retirement and still owed the bank for it.
In August 1993, Cordova did the only thing left. Between sixty and a hundred fishing boats moved up the Valdez Narrows and blocked the channel, stopping the tankers from getting out. They were not asking for damages. They were asking for money to study what had happened to the fish, because four years on, nobody could tell them.
Interior Secretary Bruce Babbitt flew up and met them on the 23rd and promised federal help. Out of it came the Sound Ecosystem Assessment Program, the largest research project the Trustee Council ever funded. Three months before the blockade, Bob Van Brocklin killed himself. He had been Cordova’s mayor through the spill, the man who stood at the front of the room while Don Cornett made his promise.
He left a note. In it he wrote that the stress from Exxon, which had brought on his financial stress, was too much to deal with alone. His ashes were scattered over Bligh Reef. He was not the only one.
The sociologist Steven Picou followed Cordova for years and reported that around a third of the fishermen he surveyed showed signs of clinical depression, and roughly 37 percent carried symptoms consistent with post-traumatic stress. Alcohol use rose, divorce rose, domestic violence rose, bankruptcies rose. Picou made a point that still gets quoted. The communities did not realize what was happening to them until the suicides and the divorces started.
A natural disaster ends. This one never ended, because the cause was a company, and the company was in court, and the court took nineteen years. About 32,000 people were certified as a class seeking punitive damages from Exxon. Their lead attorney was Brian O’Neill of the Minneapolis firm Faegre and Benson.
The case went to trial in Anchorage in May 1994. On September 16, 1994, the jury awarded $287 million in compensatory damages and $5 billion in punitive damages, at the time the largest punitive award in American history. In Cordova, people cried in the street. Not one dollar of it was paid for fifteen years.
Exxon appealed, then appealed again, then again. The Ninth Circuit sent the award back as excessive in 2001. Judge Holland reinstated it at $4 billion. It went back up, came back at $4.
5 billion, then was cut to $2. 5 billion. And there was a piece of the machinery the class did not learn the full shape of until much later. In 1991, Exxon had quietly settled with seven seafood processing companies, known in the litigation as the Seattle 7, for around $64 million.
The processors agreed that if they were ever awarded punitive damages, they would hand that money back to Exxon. Seven of the plaintiffs sitting inside the punitive class had been bought a seat, and the seat paid out to the defendant. On June 25, 2008, the Supreme Court of the United States settled it. In a 5-3 decision, the court held that for maritime cases of this kind, punitive damages should not exceed compensatory damages by more than a one-to-one ratio.
The relevant compensatory figure was $57. 5 million. So the punitive award became $57. 5 million, about a tenth of what the Anchorage jury had ordered nineteen years and three months after the grounding.
ExxonMobil’s profit for 2008, announced seven months after the ruling, was $45. 22 billion, the largest annual profit any American company had ever reported. The punitive damages for the worst tanker spill in the country’s history came to roughly four days of that year’s earnings. There was one more turn that went the plaintiffs’ way: in June 2009, the courts ruled that Exxon owed interest on the judgment, running from 1996 at 5.
9 percent, close to another $500 million, which roughly doubled the payout. Exxon paid the principal in December 2009. The average share went from something near $15,000 to something near $30,000. The plaintiffs’ side put the number who died waiting at close to a fifth of the class.
Heirs collected instead. The reopener clause came to nothing. In August 2006, the Justice Department and the state of Alaska formally invoked it, presenting ExxonMobil with a plan to clean up the oil still sitting in the beaches at an estimated cost of $92 million. Exxon did not pay.
The claim sat. And on October 15, 2015, with both governments having concluded they could no longer make the case that the injuries were unanticipated, Judge Holland closed the file. Twenty-six years of litigation ended not with a verdict but with a claim being withdrawn. The oil was still in the beaches because it had never left them.
In the summer of 2001, a federal chemist named Jeff Short ran a systematic survey of shorelines across the sound and found far more crude than anyone at Exxon had estimated, around 90 tons still buried at concentrations orders of magnitude above the company’s figures. Much of it lay in the intertidal band, the same strip of gravel and rockweed where the water goes white every April and the herring leave their eggs. Estimates put the loss at something like 4 percent a year. Surveys in the 2010s still put 16,000 to 21,000 gallons in Prince William Sound beaches, and researchers who have sampled it describe material nearly as toxic as the day it came out of the ship.
Some of what did not come back is countable, and some of it is not. The sea otters were eventually declared recovered around 2013. Bald eagles recovered. Harbor seals recovered.
Pigeon guillemots did not. Marbled murrelets fell by something close to 70 percent over the decades that followed and have not come back. And then there is the AT1 pod. Killer whales in Prince William Sound come in two kinds: residents who eat fish and transients who eat marine mammals.
The AT1 transients were a genetically and culturally distinct group that lived in the sound year round. Researchers had photographed and named 22 of them and counted the same 22 every year from 1984 through 1989. Whales from the group were photographed swimming through oiled water near the wreck. Nine of them were gone within a year.
The group has not produced a calf since. When the last one dies, a lineage with its own dialect and its own hunting method ends. Not a species, a culture inside a species. Joseph Hazelwood was tried in Anchorage in 1990 on charges including criminal mischief and operating a vessel while intoxicated.
Twenty-one witnesses said he seemed sober around the time of the grounding. His blood sample had been taken almost ten hours late and without the required preservative. The jury acquitted him of everything except a misdemeanor, negligent discharge of oil. He was fined $50,000 and given a thousand hours of community service.
He spent eight years appealing before he gave up and served it, picking up trash along the roads around Anchorage and working shifts at a soup kitchen across five summers. He died on July 21, 2022, at 75. He was the only person ever prosecuted over any of it. The ship survived longer than most of the people in this story could stand.
She was refloated, towed to San Diego, and rebuilt at a cost around $30 million. Then Exxon changed her name. She sailed as the Exxon Mediterranean, then S/R Mediterranean, then simply Mediterranean. In 2008, she was cut down into an ore carrier and renamed Dongfang Ocean.
In 2012, she was run up onto the sand at a shipbreaking yard in India and taken apart for scrap by hand. She was never allowed back into Prince William Sound. The Oil Pollution Act of 1990, which Congress passed largely because of her, barred any vessel that had spilled more than a million gallons after March 1989 from ever operating in these waters again. That statute forced the phase-in of double hulls across the tanker fleet and created industry-funded citizen oversight bodies with permanent seats at the table.
In the sound itself, the numbers turned around completely. Thirteen skimming systems became 90. Five miles of boom became roughly 40. Half a million gallons of storage became 37 million.
Under the ship escort and response vessel system, every loaded tanker now leaves Valdez with two escort tugs alongside it the whole way out. Every time. Cordova, meanwhile, did what Cordova does. It found another fish.
The kings and sockeye that run up the Copper River in May are exceptionally fatty, and in the years after the spill, the fleet and its marketers turned that biological accident into one of the most successful premium seafood brands in the country. By the 2000s, the first catch of the season was flying south the day it landed. But it was a narrower town than the one that went into 1989. The processing industry consolidated and thinned out.
A cannery is not just a building. It is the buyer. Without one, a fisherman with a full hold has a boat full of something that is about to spoil. In 2024, Trident Seafoods put four Alaska plants up for sale.
Peter Pan Seafoods stopped that April. And then quietly the herring began to come back. Not to the old numbers, nothing like the old numbers, but the direction changed. A strong 2016 year class recruited into the spawning population.
After a record low aerial count in 2018, the white water started spreading again in 2019, and it kept spreading. By 2024, the estimate for the spawning biomass in Prince William Sound was 35,485 short tons. The management plan sets a floor of 22,000 tons below which nobody fishes. On October 25, 2024, the Alaska Department of Fish and Game announced that the Prince William Sound herring fishery would open in November for the first time since 1998, with a guideline harvest of 500 tons and an exploitation rate set at a cautious 10 percent.
And then the town discovered the problem. There were still on the order of 130 permits attached to this fishery on paper, but there was nobody left to sell the fish to. The canneries were shut for the season, and several were shut for good. Ken Jones, a commercial fisherman and a member of the Cordova City Council, put it about as flatly as it can be put.
It was really exciting, but nobody was sure anyone out there would be buying, because the canneries are all closed. The fishery opened on November 12, 2024. Seven vessels registered for it. Of those seven, exactly two had a market.
The others could not find anyone who would take the fish. In the end, a buyer did step up: Copper River Seafoods, an Alaskan processor named for the river that had rescued the town, took the herring. That is the reopening. Twenty-six years of waiting, a biomass above threshold, a 500-ton quota, and two boats.
The fish are coming back slowly, but coming back. The fishery is not, because a fishery is not a population of fish. It is a fleet and the men who ran it and the permits they held and the crews they trained and the plant on the waterfront that opens its doors in April and takes the delivery and cuts you a check. All of that was disassembled over thirty-five years by a closure, then a court docket, then a generation of young people who went to Anchorage instead.
The boats still tie up along the same floats they did in 1989. Every spring, a plane still goes out over Montagu Island and Naked Island and the Green Island shoreline, counting the miles of coast where the water has turned pale with milt and writing the number down. In 2024, there were more miles than there had been in years. The planes fly in early April, the season the town used to run on.
The pilots make their count, and the harbor stays quiet, and the number goes into the file.
