Fourteen years after the BP Deepwater Horizon disaster, would we fare any better at cleaning up another huge oil spill? Jocelyn Timperley examines the latest science of ocean clean-ups.

On April 20, 2010, a blowout caused a huge explosion on the offshore drilling rig operated by BP in the Gulf of Mexico. Eleven people were killed. Two days later, the rig collapsed. Oil began seeping into the sea, and it continued to flow for almost three months.

The Deepwater Horizon disaster is among the most lamented environmental catastrophes of the past century. It’s hard to comprehend how incredibly huge the spill was. It was the world’s largest ever marine oil spill, releasing an estimated 4.9 million barrels of crude oil (779 million litres, or over 300 Olympic swimming pools-worth). Up to a million seabirds were killed outright, and the human health and socioeconomic effects are still being felt today.

BP, rig operator Transocean, and several government agencies immediately tried to limit the damage, with BP’s chief executive saying the company was “determined to do everything in our power” to contain the spill. Booms were deployed to try to contain the oil, skimmer ships nibbled at the edges of the widening slick and fires were set to try to burn it off the sea surface. Various devices were deployed deep below the surface to try to contain or capture the oil. BP also began to spray the oil with enormous amounts of dispersants both on the sea surface and 1.5km (0.9 miles) underwater, where oil was gushing from the wellhead.

However, it is thought that these measures recovered or dispersed only around a third of the spilled oil. The BP spill sparked a huge amount of research into oil spills and their impacts. But 14 years on, what hope is there for better measures should another oil spill occur?

Jeffrey Short, an expert in oil spills and now-retired scientist from the US National Oceanic and Atmospheric Administration (Noaa), was working for Oceana, a marine conservation organisation when the BP spill occurred. When a colleague told him about the spill at lunchtime, he felt sick.

“I knew immediately that this would be ecologically and economically disastrous, that it would wreck tens of thousands of people’s lives, and that it would dominate my professional life for the next several years,” he says. “All of which proved true.”

Oil spills are the third largest source of oil in the sea, after land-based runoff (largely from cities and vehicles) and natural oil seeps. The problem with spills, of course, is the sheer volume of oil that enters the sea all at once. This means that oil spills – especially big ones – are “much, much more dangerous per unit oil released”, says Short.

While no spill has since surpassed Deepwater Horizon’s in sheer volume, Noaa responds to more than 150 oil spills every year. Just last month, oil began spewing from a submerged oil tanker and at least two other sunken vessels in Manila Bay, in the Philippines, after they were hit by monsoon rains and Typhoon Gaemi. Another oil tanker hit by projectiles from Yemen’s Houthi movement remains in a precarious position in the Red Sea. However, the number of oil spills from tankers is today far lower than in the 1970s, due to improved standards.

When oil spills occur, the first step is to control the source, “whether that be a ship, pipeline, or leaking well”, says Doug Helton, regional supervisor of the emergency response division at Noaa’s Office of Recovery and Restoration. “The second priority is recovering oil at sea.”

The major priority is to avoid the oil reaching the shoreline, where it can do the most damage. Shoreline cleanups can last days to years, depending on the type of oil and severity of contamination, says Helton.

Spilt oil tends to spread quickly into a thin layer on the sea surface. Within days, centimetres-thick layers become a film of a millimetre or less, spread in drifting patches over a wide area. Efforts to scoop up the oil from the sea surface therefore offer diminishing returns as time goes on. “Floating oil spreads very quickly and there is a limited window of time – days – when at-sea tools are effective,” says Helton.

Hundreds of skimmers were deployed to clean up the BP Deep Horizon spill. Skimmers are boats that scoop up spilled oil from the water’s surface, usually after the slick is first surrounded with floating booms to keep it from spreading. They do this in various ways – some, for example, suck up the oil like a vacuum cleaner, while others use oil-attracting “conveyor belts” or gravity to carry the spilled oil into a reservoir.

But hopes at the time that the skimmers could pick up oil “like a lawnmower cutting grass” proved to be overblown. They only recovered an estimated 3% of the oil. “At sea, the oil may spread more rapidly than the skimming vessels trying to capture oil,” says Helton. “Going faster is not an easy option because the bow wave from the ship will push the oil away.”

The satellite photos of the BP disaster “speak volumes”, says Short. “You’ll see a half a dozen surface skimming boats that, from the sea surface next to the boat, look quite large and quite effective. But from a satellite, you realise that you are […] just having a nearly negligible effect on the size of the spill.”

In fact, a 2020 review of 30 large offshore oil spills found only 2-6% of oil was recovered using mechanical methods like skimmers. Short says that mechanical recovery has improved in recent decades, with better booms to corral the oil and better systems to remove it from the sea surface. But even with improvements, mechanical methods can’t have much impact on a large spill, he says.

In recent years a plethora of studies and reports have emerged looking at different ways to soak up oil spills, from laser-treated cork and textiles based on leaves to graphene, magnets and even hair and fur. These mostly rely on the oil-attracting and water-hating properties of the material, with various forms of oil-attracting sponges a particularly common solution. But the difficulty of handling oil-soaked materials means these techniques are typically only useful for small spills.

When Guihua Yu, a professor of materials science at the University of Texas at Austin, and his team began considering whether a new material his lab was working on could be used to help clean up oil spills, he says he was surprised about the lack of focus on how these innovative materials could be used in practice.

One central problem, he says, is that most can only be used in a non-continuous way, requiring processing to remove the oil before the same material can be used again.

Yu and his team landed on a solution which he thinks could help. In a 2023 paper, his lab developed a prototype with a collection speed 10 times faster than current clean up rates.

The lab produced their own super oleophilic gel capable of 99% separation of oil from water, which they used to cover a mesh filter. But they also designed a continuous roller system, which Yu says would be attached to the front of a ship. This conveyer belt picks up oil from the water surface, then rolls it round to beside an induction heater, which heats the oil, detaching it and allowing it to drip down to a collector in the middle. The roller is freed up to be directly reused as it rolls down to the water again.

“The most important [innovation in our work] is probably higher throughput,” says Yu. “I personally felt it’s very unique and very different from what is conventional.”

The invention has so far only been tested on a small motor oil spill in a lake in China using a metre-scale prototype, but Yu says he has had conversations with industry potentially interested in scaling it up. The overall costs, he believes, would be reasonable. However, he admits his current design does not address the bow-wave issue of oil being pushed away from the ship, noting that how to balance collection and bow wave is “worthy of further investigation”.

But Short says that, for large spills which require more than a day to clean up, movement of oil during the night (when operations can usually not be carried out) will always limit the efficacy of collecting oil on the sea surface.

“The following day, the oil must first be located before response equipment can be effectively deployed,” he says. “For large spills, especially when response equipment is limited, these challenges may limit the amount of recoverable oil to less than 10% of the initial spill volume.”

Still, improvements have been made in tracking the oil too. Noaa now uses drones and satellites to help find and track oil spills, and tools for mapping and coordination have advanced. Undersea manned and autonomous tools that can tap into sunken vessels to extract oil have also been developed since the Deepwater Horizon spill, says Helton.

Burning is another, more controversial, way to remove floating oil at sea. An estimated 5% of the BP oil spill was burnt off the surface.

Burning requires concentrating the oil on the sea surface to at least 2-3mm – relatively thick for an oil spill. It also requires quick action, and lucky weather conditions. In the Exxon Valdez oil spill in Prince William Sound, Alaska in 1989, a storm dispersed the oil over a wide area into a film too thin to catch alight.

Improved boom designs to better corral oil have improved the effectiveness of burning over the years, says Short. But successful burning also has its own problems for the environment and human health in the form of air pollution.

The impacts of air pollution on the workers attempting to clean up the BP oil spill are still being investigated today. A major 2022 study found that workers involved in cleaning up the spill were 60% more likely to be diagnosed with asthma or experience asthma symptoms one to three years after the spill, compared with those who did not work on the cleanup.

Burning is not the only culprit for air pollution. The evaporation of the oil itself is also highly toxic, as is another controversial way to try to dissipate the impacts of oil spills: dispersants.

During the Deepwater Horizon disaster, BP sprayed roughly 1.84 million gallons (8.37 million litres) of the dispersant Corexit on the surface and deep into the water column – the largest volume of dispersant ever used for an oil spill.

Dispersants work by breaking down the oil into smaller droplets that can mix with the water below, which both helps it to degrade and removes it from the surface, where it tends to do most damage (especially to diving seabirds, surfacing marine mammals, turtles and young fish). But it needs to be added quickly after oil spills.

Little was known about exactly how this quantity of dispersant would affect the environment in the BP spill, but the hope was that it would stop the oil from reaching shoreline habitats. But the sheer volume used has been widely criticised as largely ineffective as well as harmful to the environment and humans. It’s thought just 8% of the oil was dispersed using Corexit.

In Short’s view, prior knowledge about oil spills meant that in the Deepwater Horizon spill “you can be quite certain in advance” that the continued application of dispersants on parts of the oil slick which had already emulsified was “a waste of time” beyond the first few days. “But it shows the public that you’re doing something.”

Environmentalists and scientists have a term for these kinds of reactions to oil spills – “response theatre“. It describes when companies responsible for a spill focus more on being seen to do something about the spill than necessarily doing the best thing.

Some researchers, however, say the dispersants were relatively effective and may have helped avoid further air pollution by getting rid of the oil. A 2019 report from the US National Academies found that dispersants can help cope with oil spills in some circumstances, but that limitations in the research make it hard to make conclusions about whether it improves the human health aspect compared to not using dispersants.

Still, counterintuitive as it may sound, there are occassions where some interventions could be a worse option than leaving an oil spill alone. In many places, ocean microbes have developed to eat the oil seeping naturally into the environment. These same bacteria and fungi can munch away at oil spills too – albeit relatively slowly and some more than others – but if they are impacted by chemicals, such as those in dispersants, this process could be disrupted.

Bioremediation – such as adding nutrients to encourage oil-degrading bacteria – has a long history of use in oil spills. But scientists are still at the beginning of understanding the complex interactions between microbial communities and chemical dispersants, as well as how these interact with environmental factors like temperature and sunlight. Research, for example, has shown that sunlight levels impact oil degradation in different microbes differently.

A study published in 2024 became the first to use an advanced microbiology technique to look at these interactions. Rather than look at the DNA of microbes, as previous studies have done, the scientists examined the protein expression of microbes in waters off the coast of Florida – a technique usually only used in medical or clinical science.

Using these techniques can show far more detail than looking at DNA alone, says Sabine Matallana-Surget, an associate professor of environmental and molecular microbiology at the University of Stirling, Scotland, who led the study. If she did a similar study on humans, for example, she would be able to tell when they had lunch by tracking the enzymes involved in food digestion.

Her team found that Corexit induces a high expression of proteins involved in oxidative stress in oil-degrading bacteria. “I have never seen so many proteins involving DNA damage [and] repair, [as] when you introduce the Corexit to your microbial community,” says Matallana-Surget. More sunlight also increased the toxicity of Corexit and oil in their experiment, creating a “double pill effect”, she adds.

The team plan to conduct similar experiments in other places with natural seepages of oil, with different microbes, temperature and sunlight levels. If there is another oil spill in one of these places, Matallana-Surget says, these findings could inform the optimum level of Corexit to use in that particular location for maximum oil recovery. “I’m hoping that in the near future, if there was an accident somewhere else, we would be able to say, ‘Well, listen, no, you shouldn’t apply Corexit in that region, or not as much, or maybe this concentration.'”

Dispersants aren’t the only intervention after oil spills that have caused concern. “We have found after lengthy research that aggressive cleanup of some environments can cause more harm than the oil,” says Helton. “Marshes and sheltered intertidal habitats, for example, are often treated very carefully.”

The high-pressure, hot-water washing used to clean the ecologically sensitive shorelines of Prince William Sound in Alaska after the 1989 Exxon Valdez spill, for example, sterilised the beaches, inadvertently killing bacteria as well as larger animals. Research has shown that areas not cleaned by the hot water recovered faster than the treated sites.

If a spill like Deepwater Horizon happened today, says Matallana-Surget, the reaction would be completely different. “There have been huge conversations around what happened with applying tonnes of [a] chemical [where] we have no idea what’s going in the environment. I think nobody in any part of the world would do that.”

Ultimately, since spills are so hard to clean up, avoiding them happening in the first place remains the most important thing. “Prevention is going to be the most fruitful line of approach,” says Short. “Continuing to implement safety measures and especially being vigilant.” The problem is that standards are expensive to maintain, he says. If years go by without a spill, they “tend to start slipping”.

Major changes have been made to US regulations governing offshore oil and gas operations, as well as advances in preventing blowouts in the first place. New performance measures and enforcement mechanisms have been introduced to improve pipeline safety. However, there are also new potential risks for oil spills: deeper drilling, ageing infrastructure, transport of new types of oil and through different routes such as the Arctic, and climate impacts like sea-level rise and more intense and frequent storms.

A report released by BP in September 2010 concluded that decisions made by “multiple companies and work teams”, including BP and others, had contributed to the spill. The unprecedented costs – over $65bn (£49bn) – to BP of the Deepwater Horizon has acted as an incentive to companies maintain the vigilance to avoid future disasters, says Short. “I think that’s really got a lot of attention in the industry, that this is not a trivial operating expense that you can just write off as business as usual.”

BP also quickly announced $500m (£380m) for a 10 year research programme, which has been credited with galvanising advancement in oil spill science.

But while risks can be reduced, so long as oil is being produced, “you’re not going to get rid of [spills]”, adds Short. Oil supply is set to reach a record high this year, with the US last year producing more oil than any country ever has before. Until oil dependence begins to fall, sadly the risks of another oil spill will stay with us.

Both BP and ExxonMobil declined to comment for this article.

* Jocelyn Timperley is a senior journalist for BBC Future. Find her on Twitter @jloistf

The Klamath River is free of four huge dams for the first time in generations. But for the Yurok tribe, the river’s restoration is only just beginning – starting with 18 billion seeds.

Brook Thompson has been fishing on the Klamath River ever since she could stand up in a boat. To Thompson and her family, who are part of the Karuk and Yurok tribes from northern California, fishing is second nature. “The river was our grocery store,” the 28-year-old explains. That was until a catastrophic fish die off happened in 2002.

“It changed everything,” Thompson remembers. “We’d always had plenty of food up until then. As a seven-year-old, the salmon were almost as big as me, and I saw thousands of their bodies piled up on the shoreline, I smelled their rotting flesh. It was apocalyptic.”

The Yurok Reservation sits on the final 44 miles (71km) of the Klamath River before it meets the Pacific Ocean – a remote strip of land where there is one convenience store attached to the local gas station. Pre-contact, their territory spanned more than one million acres (400,000 hectares). The tribe relies on the river and the land for sustenance. And in the oral history of the Yurok tribe, which extends back thousands of years, there was no record of anything like this ever happening before.

“Since time immemorial, nothing like this had ever happened to us. A whole generation of salmon died on that one day,” says Thompson. Low water flow from the Iron Gate Dam, one of four on the lower Klamath River, was found to be a “substantial causative factor”, a report from the Yurok Tribal Fisheries Program found.

The dams have long been a point of contention for the tribe, who have been campaigning for their removal since the 1990s. The river is the lifeblood for the Yuroks, and the salmon are family. “The death of salmon means the death of our entire way,” Thompson says. “Everyone is connected. Taking these dams down is a life-or-death situation for us.”

Finally, at the end of August 2024, after years of negotiating, and decades of activism, the last dam fell, reopening more than 400 miles (644km) of river, in what is the largest dam removal project in US history.

“This is decades and decades in the making,” says Thompson. “We were told it was never going to happen. That it was foolish to even ask for one removal. We were asking for four.”

The Klamath Basin covers more than 12,000 square miles (31,000 sq km) in southern Oregon and northern California, and was home to the JC Boyle, Copco 1, Copco 2 and Iron Gate dams, all owned by PacifiCorp, an electric utilities company. The Klamath was once the third-largest salmon producing river on the US’s West Coast before the construction of the dams blocked fish from accessing almost 400 miles (640km) of critical river habitat for almost 100 years.

Fall chinook salmon numbers plummeted by more than 90% and spring chinook by 98%. Steelhead trout, coho salmon and Pacific lamprey numbers also saw drastic declines, and the Klamath tribes in the upper basin have been without their salmon fishery for a century, since the completion of Copco 1 in 1922. The situation became so bad that Yurok tribe – who are known as the salmon people – began importing Alaskan salmon for their annual salmon festival, traditionally held to celebrate the first return of fall chinook salmon to the Klamath River.

The dams also had a severe impact on water temperature and quality – growth of toxic algae behind two of the dams resulted in health warnings against water contact.

“It was painful,” says Willard Carlson, a Yurok elder who is known as a river warrior and was part of the inter-generational campaign. “All those years seeing our river damaged like that. I remember as a kid we’d have other people from nearby tribes making fun of our river. ‘Oh, you’re Yurok, your river is dirty.’ For us, the dams were a monument to the [coloniser] people who conquered us.”

Removing the dams has been a rocky road, thanks to the number of tribal, local, state and federal stakeholders, and the cost of removal – estimated at $450m (£340m).

In 2022, the go-ahead was finally given to remove the dams, 12 years after the original agreement to pave the way for the project was signed in 2010. Oregon and California agreed to shoulder joint liability and in October 2023, the first dam came down.

“The removal is a victory,” says Carlson, although he isn’t fully celebrating yet. “We’re still having to be cautious, because our resources are still under threat.”

The water that initially rushed downriver was dirty and smelly, say the tribe, and debris that had piled up behind the dams for decades suddenly was dislodged and brought downstream. But even in the few months since the first dam was removed, they’ve noticed a difference.

“A couple years down the road, once the river has been able to repair itself, we’ll begin to see healthier fish runs,” says Oscar Gensaw, a Yurok tribal member and fisherman. “You can definitely see already the river is starting to do its own thing, and that’s the best thing for us – letting the river do what it needs to do, because it knows what it needs to do to repair itself.

By 2061, it is estimated that the chinook salmon population will have recovered by an average of 81%.

But something that does need a helping hand is the restoration of 2,200 acres (890ha) of land that is above ground for the first time in a century following the emptying of four reservoirs.

“Removing the dams is one thing, restoring the land is quite another,” says Thompson, a civil engineer and part of the crew working on the restoration project – which is being managed by Resource Environmental Solutions, an ecological restoration company.

Preliminary planning started back in 2011, and in 2020, the reservoir action management plan, a detailed 260-page document was published by the Klamath River Renewal Corporation – the non-profit entity formed to manage the entire dam removal project. ”The document gave us our targets, and has kept us on track,” says Joshua Chenoweth, senior ripariani ecologist for the tribe who was hired to manage the revegetation project.

Between 2018 and 2021 seed collection crews – many of whom are tribal elders – were hired to harvest native seeds, by hand, in preparation for the dam removal. They collected 98 species and around 2,000lbs (900kg) of seeds. The seeds were then dispatched to specialised nurseries, which propagated them en masse, and sent the seedlings to storage facilities where they were kept until the time came for them to be planted.

A total of 18 billion native seeds were propagated – more than 66,000lbs (30,000kg) worth – each species selected for a purpose: to retain sediment, to prepare the soil for other plants, for cultural uses, or to be a food source. Wheatgrass, yarrow, lupine and oak trees – an important cultural species for the Yuroks and a keystone species – to name a few.

“It was an extremely complicated process,” says Chenoweth. “Everything we introduced had to be genetically appropriate, sourced from nearby watersheds. It’s quite a process to work out what species you need, how many seeds you need, and how long you need to grow them for to get the seeds you need.”

The team also had to take into account what species could be found in enough numbers, whether there were nurseries who knew how to propagate them, and which plants would grow quickly enough.

“A lot goes into species selection,” says Chenoweth. “And we had a lot of limits. It’s definitely been a very challenging project, and we were lucky that there were delays in dam removal.” But in the end, the timing worked out. “We didn’t have enough seed until this year, when it was time for the last dam to be removed.”

Seeds from trees and shrubs were also collected, which was a challenge during 2021 and 2022, particularly hot and dry years that exacerbated widespread wildfires, Chenoweth says. “The collection was hard, there was just not much out there. Physically, it was really hard for the crew too – smoky because of the fires, and very hot.” Eventually, the team got the numbers they needed right before the removal – including 1,500lbs (680kg) of acorns.

The most important part about restoration is ensuring a high diversity, says Chenoweth, who had previously worked on the Elwha River dam restoration, which was demolished in 2014. “Most restoration projects are lucky to have six to eight species. We have as many as 22 species in our seed mixes.”

When it came to planting the seeds, Chenoweth has used a variety of methods, including seeding by helicopter alongside the old-fashioned way, by hand. Beginning straight after the first draining in late 2023, the team hand-sowed 500 acres of land, including 25,000 acorns. “It really is the most effective restoration tool. And the only reason we used a helicopter was when muddy conditions made it too dangerous for us to walk on the land.”

The results were “excellent”, Chenoweth says. “Despite a hot and dry summer, where we’ve seeded is green, we’re seeing flowering plants teeming with life. I counted eight different butterflies in the areas we hand-seeded. There’s moths, hoverflies, bees, birds. It’s been a really exciting first year.” The team will continue to seed and plant the area for another two years, prioritising the tributaries important for salmon habitat.

“The first year has definitely given me a lot of hope,” Chenoweth, “but it’ll probably get more challenging from here. Now we’re completely at the mercy of Mother Nature.”

After planting is completed, the area will be monitored and maintained for five years by Resource Environmental Solutions. There are four success criteria set out by the restoration plan that need to be met in order for the project to be considered successful: species richness, vegetation cover, scarcity of invasive species and stem count – applicable for forest areas.

For Thompson, it’s a moment that she has been waiting for almost her entire life. “I saw how nasty the river water was behind those dams – it was this neon, snot-nosed green – and now the water is already running clear,” says Thompson. “Seeing this transformation just within these last few months has been amazing. Growing up, the dams were these huge, immovable objects, 100ft [30m] tall. It all felt so impossible. But anything that can be built can be taken down, right? And watching it happen in real time, I just thought, ‘Wow, it was really this easy to remove them this entire time.'”

Being involved in the restoration has been healing for the tribe, too, and sends a powerful message. “For so long we’ve been forcibly kept off our land, and it’s really powerful to me that we’ve been able to participate,” Thompson says. “It’s so important that this is indigenous-led because usually you see three, five, 10-year timelines when it comes to restoration. When it comes to indigenous mindsets, it’s a seven-generation plan. How is this plant going to affect us 100 years from now? Taking care of a plant that your great-grandmother took care of puts a lot of perspective into life. And this is the first step in reclaiming our ability to take care of the Klamath River.”

But Thompson isn’t hoping to return to the past – she’s looking firmly to the future.

“There’s almost a bit of a fallacy in thinking like it will be returned to what it used to be,” Thompson says. “But I think with traditional ecological knowledge, tribal-led initiatives and current academic understanding of the landscape so you can almost make it better.”