Of airplanes and ash clouds: What we've learned since Eyjafjallajökull

by Mary Caperton Morton
Friday, March 24, 2017

Boeing 757 landing Credit: ©iStockphoto.com/igmarx.

In 2010, Eyjafjallajökull, an ice-capped stratovolcano in southern Iceland, awoke from a nearly 100-year hiatus, erupting mildly on March 20. Starting on April 14, the eruption entered a more dramatic phase, producing an ash plume that rose 9 kilometers into the sky and drifted over the North Atlantic into the path of planes flying to and from Europe — the busiest airspace in the world. This development led European aviation authorities to institute a complete stoppage of flights through the affected area from April 15 to April 20, as well as several later stoppages, which in total affected more than 100,000 flights and millions of travelers.

By the fourth day of the closure, economic losses mounted into the billions, and with static weather patterns firmly planted over the North Atlantic, the ash cloud showed little sign of dissipating. The aviation community’s zero-tolerance policy for flying through volcanic ash — instituted following three ash-related engine failure incidents involving passenger jets in the 1980s — suddenly came under intense scrutiny, and it became apparent the policy wasn’t going to fly.

On April 21, European aviation regulators amended the zero-tolerance ash policy to allow planes to fly through ash concentrations of less than 2 milligrams per cubic meter, as measured by the London Volcanic Ash Advisory Center using satellites and dispersion models.

At the time, data on how volcanic ash actually affected engines were limited and the decision was based more on geography than hard science, says Rory Clarkson, an engineer with Rolls-Royce in Derby, England, the world’s second-largest maker of aircraft engines. The boundary “between 1 milligram and 2 milligrams per cubic meter lay across Heathrow Airport in London” and the runways at Gatwick Airport, just south of London, Clarkson says. A tolerance level of “1 milligram would have kept Heathrow closed and allowed planes to land but not take off from Gatwick,” he says, while “2 milligrams allowed both to open, so that’s where they drew the line.”

The gamble paid off: Most flights were restored within hours to days, and few engine issues and no catastrophes resulted. “We know of about 10 cases where ash deposits were found in various parts of engines, but there were no noticeable in-flight engine problems or damage,” says Marianne Guffanti, a volcanologist at the U.S. Geological Survey (USGS) in Reston, Va.

“Until Eyjafjallajökull, the ash avoidance strategy worked great; there were fewer of the dangerous encounters that had occurred in the ‘80s,” Guffanti says. But the 2010 eruption was disastrous for travelers, airlines and many industries that depended on trans-Atlantic and European flights. The incident “forced the question: How do we keep the aviation system going when volcanic ash can’t be avoided?”

Ash from the April 2010 eruption of Eyjafjallajökull in Iceland rose to a height of 9 kilometers, reaching the stratosphere and obstructing the paths of planes flying across the North Atlantic. Credit: Arni Frioriksson, CC BY-SA 3.0.

The eruption also cast a spotlight on everything scientists and aviation experts didn’t know about how volcanic ash interacts with airplane engines, which was a lot. Clearly, ash was bad for engines, but how ash gummed them up, or whether dilute amounts could be safely tolerated, was unknown. The Eyjafjallajökull eruption “shattered our complacency and showed us how much we needed to learn about what constitutes hazardous airspace when it comes to volcanic ash,” Guffanti says.

In the wake of the eruption, researchers found new avenues to investigate how and why ash is detrimental to airline engines. And seven years later, most researchers are confident that if a similar event were to happen, the response would be very different.

Searching for Answers in the Aftermath

Ash drifted from Iceland to England and all the way across Germany before dissipating significantly. Credit: left: NASA's MODIS Rapid Response Team; right: NASA image courtesy of Jeff Schmaltz, MODIS Rapid Response Team at NASA GSFC.

Prior to Eyjafjallajökull’s 2010 eruption, airlines and engine manufacturers had mostly sidestepped the issue of testing the effects of ash on engines. “Nobody wanted to be responsible for testing it,” says Ulrich Kueppers, a volcanologist at Ludwig-Maximilians University in Munich, Germany. “Airlines deferred to the turbine manufacturers, and turbine manufacturers said they don’t need to test the effects of ash because there is no regulation requiring it.”

European aviation authorities halted all flights through the affected area from April 15 to April 20, 2010. Credit: ©iStockphoto.com/Edin.

Since Eyjafjallajökull, “the volcanic ash and aviation story has really turned into one of collaboration between different countries and agencies,” Clarkson says. Engineers, volcanologists and modelers tackled the issue from all sides, investigating ash-resistant coatings for engines, potential safe exposure dosages for different types of volcanic ash — since Eyjafjallajökull had proved it was possible to fly through dilute ash clouds without catastrophe — and more accurate ash cloud dispersion models to better predict where ash clouds might enter flight paths.

Ash-Proofing Engines

Because airplane engines are so expensive — $20 million to $50 million each — only a limited number of ash tests have been done on actual engines. One of these was the Vehicle Integrated Propulsion Research (VIPR) program, a collaboration among NASA, the U.S. Air Force, USGS and others that took place in three phases (VIPR I, II, III) between 2011 and 2015.

The VIPR experiments were designed to test specialized instruments that monitor the health and efficiency of an engine. The first two phases involved putting benign particulates such as cereal and crayons through a working engine — similar to those used in commercial airliners — to test whether the monitoring instruments themselves functioned properly. In the third phase, two different concentrations of volcanic ash — 1 and 10 milligrams per cubic meter — were fed into the engine. At the lower ash concentration, the engine experienced maintenance issues, such as detectable erosion of the compressor blades and glassy buildup on the turbines, but showed no performance issues. At the higher ash concentration, the engine was negatively affected by higher pressure, decreased fuel-flow efficiency, and increased compressor and exhaust temperatures, but even after 14 hours of exposure to the ash, the engine did not critically fail.

The results of the third phase of the VIPR experiments have not yet been fully released to the public, but the initial findings are promising, Clarkson says. “It seems that volcanic ash isn’t as dangerous as many people assume.”

Other experiments have been run at universities and research institutions in Europe, the U.S. and Australia, mainly on laboratory rigs built to simulate certain parts of an engine. Overall, these tests have found that issues tend to arise when ash builds up on hot engine parts, especially at the combustion exit, reducing airflow through the engine. “That’s when you get engine surging, banging and malfunctions,” says Clarkson, who served as a consultant on many of the tests. But “even in the most severe encounters, the fans and rotating parts aren’t badly damaged,” he says.

Research by Kueppers and others has found that even minor buildups of ash can lead to corrosion of engine parts over time. “We’re not just concerned with the immediate safety concerns but also the long-term implications for the life of the engine,” Clarkson says. “Gumming up the cooling systems may not cause immediate failure, but it does lead to expensive maintenance issues and potential loss of the airplane for hours, weeks or months during repairs.”

One possible solution that engine manufacturers are investigating is applying ash-resistant coatings to engine parts where ash tends to build up. Airplane engines operate at temperatures up to 1,400 degrees Celsius, above the melting point of most metal alloys, so engine parts would have to be coated with ceramic coatings that can withstand the heat. “Unfortunately, the technologies aren’t mature enough yet,” Clarkson says. Engine manufacturers and materials scientists are working on it, but until that technology is developed, he says, “the best option is to understand the true tolerance levels of today’s engines and operate within that tolerance level.”

Another challenge, Guffanti says, is that “every engine is different, each type of ash is different, and every flight pattern is different. There are still a lot of unknowns, especially about sustained or repeated flights through ash clouds.”

Acceptable Doses of Volcanic Ash?

Engines aboard aircraft on runways in Ireland were covered to reduce exposure to ash during the 2010 Eyjafjallajökull eruption. Credit: AP.

Since the zero-tolerance policy was lifted, airplane ash encounters have demonstrated that engines can usually tolerate flying through dilute concentrations of ash for short periods. “In the last few years, [the industry has] been moving toward this concept of dosages: defining how long an engine can withstand a certain concentration of ash without too much damage,” Clarkson says. Ideally, ash could be avoided altogether, but dosage information gives airlines and pilots “an option for when they have to operate within it for some period of time.”

Dosages are calculated by considering incidents involving airplanes encountering ash clouds during flight, and then correlating the degree of damage with the concentration of ash — determined through pilot and satellite observations, as well as modeling — and the time spent in the ash cloud. Dosage curves tend to be nonlinear, rising quickly toward dangerous levels of damage as planes enter higher concentrations of ash for longer periods.

There are many variables to consider when calculating dosages, such as ash composition — some plumes carry more glassy particles, for example — and engine design and condition, Clarkson says. But, based on calculations by Clarkson and others, the maximum safe dose of ash seems to fall between 1 milligram per cubic meter for four hours and 4 milligrams per cubic meter for one hour. In real encounters, however, dosages are seldom constant, Clarkson says, adding complexity to the calculations.

“To make life easier for airlines and pilots,” Clarkson says, Rolls-Royce plans to issue guidelines suggesting that a constant-dose assumption is adequate when ash concentrations are between 0.2 milligrams per cubic meter and 2 milligrams per cubic meter. But, he says, at levels much above 2 milligrams per cubic meter, it’s important to factor in changing ash concentrations as engines are more affected. “How this is implemented within operational considerations gets complicated; we need to work with ash cloud forecasters and airlines to evolve an approach.”

Tracking Ash Types

The latest research suggests that a total avoidance policy for volcanic ash isn't necessary. Credit: ©iStockphoto.com/wead.

A challenge in calculating acceptable ash dosages is that the type and amount of ash produced by a volcano can change between eruptions, and even within a single eruption. “People think: ‘one volcano, one eruption, one kind of ash.' But that’s not true,” Donald Dingwell, a petrologist also at Ludwig-Maximilians University, told EARTH in 2016. “The cauldron of magma can change as it gets delivered to the surface.” That is why, Dingwell says, it’s important to continually monitor ash during an eruption, and to characterize how different types of ash affect engines.

To this end, more tests on both real and simulated engines are needed using a variety of ash samples collected in the field, Guffanti says. Creating ash for standardized laboratory testing is also important, Kueppers adds. His team has found that factors such as ash chemical composition and grain size affect how readily it is remelted when exposed to extreme heat. “We’re investigating how chemical parameters and grain size affect the efficiency of how a solid fragment is transformed into a sticky droplet by heating it,” he says.

Work by Kueppers, Dingwell and others is also casting light on outdated practices for engine testing, including using sand as a stand-in for ash and using lithified ash from older eruptions. Sand is a “terrible analogue for volcanic ash,” Dingwell said. “Sand is dominated by quartz, and when quartz melts, it’s one of the most viscous liquids known to man. It behaves differently from just about any other rock,” especially ash, he said. Old ash deposits are also poor analogues, Kueppers says. “People take deposits from ancient eruptions that are hundreds or thousands of years old, because it’s conveniently close to where they are working,” he says. “But we know that glass is chemically unstable and weathering significantly alters it.”

Instead, researchers could collect recently erupted ash, or manufacture it to particular specifications in the lab. “It’s good to look at natural ash, but it’s also good to be able to play with the ash properties in a controlled fashion,” Dingwell says. “We make [ash] in the lab using the same process that occurs in nature: rapid decompression melting of magma.”

Scientists are trying to determine guidelines for acceptable doses of ash for aircraft. The maximum safe dose seems to fall between 1 milligram per cubic meter for four hours and 4 milligrams per cubic meter for one hour. Using such dosages, researchers can create scenarios for potential flight plans for airplanes to stay within acceptable levels. Credit: Rolls-Royce, edited by K. Cantner, AGI.

Dingwell and his colleagues have been developing a database of ash types, characterizing which volcanoes are most dangerous to engines. So the next time an eruption happens, “the authorities can tap into this knowledge to make better-informed airspace closure decisions,” Dingwell said. Ideally, they will figure out a way to collect ash samples in real time, he adds, noting that drones may be an option to safely collect samples as an eruption is occurring.

Modeling a Moving Target

Another daunting task for researchers is to define where an ash cloud is in time and space. “Think of the classic image of an explosive volcanic eruption with a plume that’s visible for hundreds of kilometers,” says Fred Prata, an atmospheric scientist at the University of Oxford in England. “It seems like you’d be able to navigate a plane around the plume, no problem.” The problem comes as a plume disperses downwind and becomes less obvious to the naked eye.

After Eyjafjallajökull, regulators toyed with the possibility of developing an ash-avoidance policy based on either visible ash or discernible ash, which is detected using satellites or other instruments. But “relying on what flight crews can see outside the flight deck window is not a reliable means of avoiding ash. … You cannot equate ash visibility to concentration,” Clarkson says. Visibility often depends on other factors, such as daylight, water vapor in the air and the plane’s position relative to the cloud. “Even in good light and good visibility, the visibility of an ash cloud depends on your vantage point. They’re often more visible from the same altitude, but invisible if you’re above or below the level of the cloud.”

A better tactic is to avoid discernible ash clouds, those that have been picked up by satellites, radar or other instruments. “Satellites can detect ash at very low concentrations, and then you can do modeling and forecasting to see where it will move in the coming hours,” Clarkson says.

Volcanic ash is quite different from run-of-the-mill wood ash from fireplaces; it's filled with sharp, abrasive particles and glass with grain sizes ranging up to small pebbles. It also changes frequently during an eruption and as it disperses in the atmosphere. Credit: K. Cantner, AGI.

At the time of the Eyjafjallajökull eruption, ash cloud dispersion models were still being developed. When European regulators declared that planes could fly through ash at concentrations less than 2 milligrams per cubic meter, modelers scrambled. “As modelers, we weren’t too happy to be given that task because we didn’t really feel like we had enough confidence in our models to say this is the line [between] where it’s safe and unsafe,” says Larry Mastin, a hydrologist at USGS in Vancouver, Wash.

“We fretted about how accurate the model was that forecasted that ash cloud. … Were we being overly conservative in overestimating the size of the ash cloud? Or was there a chance we were underestimating it?” Mastin says. “Since then, there’s been a big push to make ash cloud dispersion models more accurate.”

The biggest limitation in 2010 was the lack of real-time data about Eyjafjallajökull’s initial eruptive plume and its subsequent movement that could be input into the dispersion models, Prata says. So, modelers have been working to create models that use data gleaned from eyewitnesses, satellites and remote sensing.

Satellites can track ashflow around the world, such as this track (left) from the 1991 eruption of Mount Pinatubo in the Philippines (right). Credit: above: NASA; right: USGS.

Top: A view of the ash plume from Eyjafjallajökull, taken from NASA's Terra spacecraft. Bottom: A computer-analyzed map of ash plume heights, corrected to compensate for the effects of wind. Reds denote higher altitudes, blues are lower altitudes. Scientists are developing better techniques to track volcanic ash. Credit: NASA/GSFC/LaRC/JPL, MISR Team.

When modeling ash dispersion, scientists first determine how high the plume extends into the atmosphere, either by direct observation or by using temperature measurements from infrared satellite imagery and mobile radar techniques, Mastin says. In the troposphere, which extends up to about 10 kilometers above the surface, and is where most small- to medium-size eruptions top out, cooler ash clouds generally imply greater height because air temperature decreases with height, Mastin says. If larger eruptions reach more than 10 kilometers into the stratosphere, “it’s more complicated,” he says. “Air temperature increases with height in the stratosphere. Plus, rising clouds continue to expand and cool, so they might not be at the same temperature as the surrounding air.”

The next step is to determine the grain size of the ash being erupted, which controls how long the ash will remain aloft and how widely it will spread. Grain sizes can be found by sampling the ash in the field, although the logistics can be daunting and dangerous, especially for more remote eruptions.

Ash from volcanoes in some parts of the world, such as Indonesia, frequently wreaks havoc on the airline industry and has caused several near-miss accidents. Credit: ©iStockphoto.com/manjik.

“In many cases, the ash particles will clump together and fall out faster than the models calculate,” Mastin says. Determining how rapidly these particles aggregate is “one of the biggest challenges we’re working on right now. We don’t understand the physics of clumping well enough to include it in a model, and the few models that try [to include clumping] run so slowly that they can’t be used operationally during an eruption. It’s a big challenge.”

Global flight paths (blue) are routed over regions of frequent (red) and occasional (yellow) ash-producing eruptions. Credit: Openflights.org, modified by K. Cantner, AGI.

Further complications come from the evolving nature of volcanic eruptions. “Ash clouds are constantly changing. No sooner do you get it figured out than it’s changed again,” Mastin says. Weather and wind patterns can speed up or slow down dispersal. In one instance, he notes, an aircraft sustained damage “from an ash cloud in the western Pacific that was tracked to an eruption of the Tungurahua Volcano in Ecuador two weeks earlier.” It was a case in which the “processes that cause ash clouds to be removed [from the atmosphere], for some reason, didn’t happen. So then we have to ask, is there any way to predict that?”

“Modelers have a vision of doing automatic comparisons between the model and the satellite image, and adjusting the model inputs to produce an output that better matches the satellite,” Mastin says. “The next step is to get next-generation satellites in the air to provide the data that we need.”

Another Eyjafjallajökull?

With everything researchers have learned in the last seven years since Eyjafjallajökull blew, the question that arises is whether aviation would be disrupted to the same degree if another similar eruption happened. Most researchers say no. “If another event were to happen tomorrow, it would be handled very differently, without a doubt,” Mastin says.

In 1989, a jet flew through ash from an eruption of Mount Redoubt in Alaska, causing all four engines to shut down and forcing an emergency landing in Anchorage. Credit: R. Clucas, USGS.

he modification of the zero-tolerance ash policy is a huge step forward for keeping flights in the air, Clarkson says. “If we had another event in Iceland, say at Katla Volcano, we would act much more quickly to get the proper protocols in place.” In Europe, where many airlines from different countries share the same airspace, authorities are testing the idea of safety cases, where each airline proposes to its national aviation authority how they plan to navigate through ash clouds. “An individual airline’s safety case may say that they have special instruments to monitor engine performance or real-time ash cloud info from a third-party provider,” Prata says. “Every case is different. A larger airline, with more resources, may fly into an ash cloud that a smaller airline would not go near.”

Meanwhile, volcano monitoring, ash cloud detection and dispersion forecasting are conducted around the clock all over the planet by various organizations in different countries, Guffanti says. “It’s mind-blowing how the world works together to deal with this issue.”

A jet takes off from Kagoshima Airport in Kagoshima, Japan, with Shinmoedake Volcano erupting behind it in February 2011. Credit: ©iStockphoto.com/wdeon.

When an eruption occurs in U.S. airspace, NOAA’s Volcanic Ash Advisory Centers in Anchorage, Alaska, and Washington, D.C., use pilot reports, USGS eruption reports and satellite data to generate specialized warnings about the whereabouts of ash clouds that are sent to pilots, dispatchers and air-traffic controllers. “NOAA tracks the ash cloud and starts making forecasts for where the ash is expected to go. This information is then used by aviation decision-makers about potentially rerouting flights,” Guffanti says. “Ash clouds are still avoided if at all possible. When it comes to flying, it makes sense to be very conservative.”

Preparation goes a long way, with certain volcanoes flagged as potential troublemakers based on their location and eruptive history, Guffanti says. “We are most concerned with volcanoes that have shown a potential for explosivity and can affect flight routes, such as the Cook Inlet volcanoes in Alaska and Mount St. Helens,” she says (see sidebar, page 35). “We try to keep our eye on all of them.”

In the future, improved monitoring of remote volcanoes will help track eruptions that occur far from people, but right in the middle of air-traffic superhighways. When it comes to aviation hazards, Guffanti says, “there is no such thing as a remote volcano.”


© 2008-2021. All rights reserved. Any copying, redistribution or retransmission of any of the contents of this service without the expressed written permission of the American Geosciences Institute is expressly prohibited. Click here for all copyright requests.