Spectacular lightning streaks out from the Eyjafjallajökull volcano ash plume in the early morning hours of April 17, 2010.
With the crippling disruption of air traffic in Europe, the massive ash plume from Eyjafjallajökull in Iceland in 2010 focused worldwide attention on recurring hazards from active volcanoes. Few outside the volcanological community, however, are aware of the important connection between large, explosive eruptions and the occurrence of lightning. Recent spectacular photographs of volcanic lightning from Eyjafjallajökull and Chaiten volcano in Chile have now highlighted this phenomenon for the public.
Scientists have been collecting accounts of volcanic lightning for well over a hundred years. Indeed, a comprehensive report by the Krakatoa Committee of the Royal Society in Britain published in 1888 noted that the “abundant generation of atmospheric electricity is a familiar phenomenon in all eruptions on a grand scale.”
The first written account of volcanic lightning dates to the famous 79 A.D. eruption of Mount Vesuvius that buried Herculaneum and Pompeii in searing ash. Pliny the Younger, a Roman Senator and scholar, witnessed and later wrote to a friend and historian about the event: “Behind us were frightening dark clouds, rent by lightning twisted and hurled, opening to reveal huge figures of flame.” It has even been speculated that volcanic lightning of the Minoan eruption of Thera around 1500 B.C. was the inspiration for Zeus's thunderbolts in Greek mythology. Lightning has now been documented at more than 200 eruptions, including Mount Pinatubo in 1991 and Mount St. Helens in 1980.
Lightning crackles within the ash plume of Eyjafjallajökull volcano in Iceland in 2010.
So why are many volcanic eruptions accompanied by lightning, and is there anything in the characteristics of the eruption that might lead us to expect lightning development in particular cases? Despite the long anecdotal history of volcanic lightning—and recent documentation—impediments such as remoteness and rugged terrain, the inherent danger in getting close to explosive eruptions, and the lack of appropriate technology meant that even as recently as 2000 volcanic lightning remained a relatively understudied area of science. Modern research is only just beginning to get a handle on the phenomenon.
“Lightning has now been documented at more than 200 eruptions.”
Recent eruptions have led to a number of spectacular photographs being posted online. This online documentation, combined with the advent of lightning detection networks and weather radar, has meant that volcanic lightning is now receiving more media and scientific attention than ever before. This article follows one interdisciplinary strand of the recent spate of volcanic lightning research; at its center is Steve McNutt, a volcano seismologist at the University of Alaska and Alaska Volcano Observatory (AVO). His efforts are bringing together a combination of unprecedented remotely sensed observations, careful examination of information gleaned from more than 200 reported cases of volcanic lightning, and theoretical considerations at the interface of both volcanology and meteorology, to generate new insights into an old phenomenon.
Electronic detection is a crucial element of the modern study of volcanic lightning. Electronic detection devices work by monitoring “sferics,” or broadband radio signals emitted by the intense electrical discharge of a lightning bolt. Have you ever turned on an AM radio during a thunderstorm? It can serve as a crude detector: Whenever there is a lightning discharge, you will hear crackling static disrupt the normal signal for about half a second. Lightning sensors use an omnidirectional antenna and sophisticated signal processing technology to extract from the radio emissions time and directional data due to discharges. Some sensors can only detect cloud-to-ground lightning, others intracloud (also known as in-cloud) and cloud-to-cloud lightning as well. In the 1980s, the Bureau of Land Management (BLM) installed a lightning detection network as an alert for possible forest fires in interior Alaska, and this paved the way for an experimental deployment in 1990 around Cook Inlet, Alaska, near Redoubt Volcano.
Mount Redoubt had begun erupting December 14, 1989. On December 15, the resulting ash plume had nearly taken down a Boeing 747 passenger jet with 231 people aboard. The jet lost power in all four engines, losing 14,000 feet of altitude. Ultimately, the pilots were able to restart the engines, and the jet landed safely. However, damages to the plane were estimated at $80 million. At least three other commercial jets suffered damaging encounters with Redoubt's eruption plume during its first three months. Part of the problem was that the plume was seldom visible during the long Alaska nights and frequent episodes of bad weather.
The fire of the Eyjafjallajökull volcano provides a stunning contrast to the snowscape of the mountain.
Alaskan airspace is surprisingly busy because it lies along “great circle” routes (that is, the shortest distance on a sphere) between much larger population centers in North America or Europe, and Asia. Alaska contains more than 50 volcanoes that have been active in recorded history. For remote sites or mountains shrouded in cloudiness where visual confirmation of an ash plume is often impossible, “seismicity,” that is, the presence of volcanic earthquakes, can be an important factor in determining whether an eruption is occurring. Unfortunately, seismicity alone is not always a reliable eruption indicator. Therefore, because of the obvious danger to aviation in the area, Rick Hoblitt, a volcanologist with the U.S. Geological Survey, initiated the experiment at Cook Inlet to determine whether lightning detection would be useful for eruption detection in an effort to improve the warning infrastructure. In Alaskan coastal regions, natural thunderstorms are rare, so Hoblitt concluded from the Cook Inlet data that lightning detection with seismicity was an almost certain indication of the presence of a volcanic ash cloud. One day in 1998 the AVO received three calls from local residents wondering if Augustine Volcano in Cook Inlet, Alaska, was erupting. The calls were prompted by alert skywatchers who had spotted lightning above Cook Inlet; the last time they had seen such lightning was in 1986, when the volcano was erupting. In this particular instance in 1998, there were no earthquakes, and the lightning turned out to be the result of a rare natural thunderstorm.
An important pattern emerged from Hoblitt's research with volcanic lightning detection at Cook Inlet: a change in polarity of lightning strikes from negative (that is, lightning bolts lowering negative charge to the ground) early in the eruption, to positive later on. This behavior is similar to that of many thunderstorms.
An engraving of the eruption of Mount Vesuvius in 1779 depicts a dramatic ash plume lightning display.
A good scientist recognizes a research opportunity when it presents itself. By his own account, Steve McNutt just “stumbled into” volcanic lightning by accident. He had been at the Alaska Volcano Observatory for less than a year when Mount Spurr in the Aleutian volcanic arc began a series of explosive eruptions during the summer of 1992.
“What happened was we recorded lightning on our seismic instruments … normally we bury our cables to keep them away from animals,” he said. “This particular case it was a remote area and not too many animals, so the cables were just lying on the ground, acting like an antenna. Lightning makes a wide band [radio impulse] that was entering the system as a spurious signal.”
It turns out the static you can hear from lightning on your AM radio also can be seen on a seismograph meant to measure volcanic earthquakes. “It gave these sharp spikes that we could readily pick out, and because those stations were right under the erupting volcano, we actually got better data about lightning from our seismic system than we did from some other instrumentation [the lightning detection network near Cook Inlet].”
Lightning affecting seismic instruments was not a new concept to seismologists. “We've all seen these transient spikes that occur during rainstorms; we know what lightning looks like, and for a typical seismologist, it's noise. But the presence of lightning in the plume can tell you something about the distribution of electric charges and the ash concentrations. That was the hook, the realization that this was actually data and could be used for a different purpose.” Ultimately, with help of Cere Davis, an undergraduate physics student, results of this research were published in the Journal of Volcanology and Geothermal Research.
One of the peer reviewers of this particular paper was an atmospheric physicist and lightning expert from MIT named Earle Williams. This fortuitous pairing ultimately would lead to a productive collaboration between the atmospheric scientist and the volcano seismologist.
Williams had first been made aware of volcanic lightning as a graduate student at MIT through Professor Bernard Vonnegut (the older brother of author Kurt Vonnegut) of the State University of New York (SUNY) at Albany. They met when Earle was visiting SUNY, and discovered they had a personal connection—the Vonneguts were from Indiana and had spent boyhood summers in Earle's hometown, tiny Culver, Indiana. They were known by Earle's father. Vonnegut, a cloud physics expert and discoverer of silver iodide's efficacy in cloud seeding, served on Earle's thesis committee. He had been involved in studies at Surtsey Volcano, Iceland, with researchers from the New Mexico Institute of Mining and Technology's (New Mexico Tech) storied Langmuir Lab, renowned for its lightning investigations. Williams was not connected with that effort, but he credits Vonnegut with informal discussions on the subject of volcanogenic lightning.
Together, McNutt and Williams set about to try to answer some of the basic questions about volcanic lightning, such as whether its incidence is uniform with latitude, and whether it changes with the chemical composition of the eruption. Answers to these questions could help shed light on the relative importance of possible charging mechanisms.
Ascending eruption cloud like a “dirty thunderstorm” from Mount Redoubt in Alaska, during April 21, 1990, eruption, as viewed to the west from the Kenai Peninsula.
“An explosive volcanic eruption plume looks like a ‘dirty thunderstorm,’ that is, one with ash in it.”
Electrical discharges in the atmosphere require a phenomenon called “charge separation.” Positive charge builds up in one location and negative charge in another. When the potential difference between the two locations becomes great enough, “dielectric breakdown” occurs, an ionized channel in the air develops, and electrical current in the form of a flow of electrons ensues. The mechanism of plume electrification and charge separation had been a fundamental question about volcanic lightning since scientists had first observed the phenomenon. It seemed from the literature that plume electrification was mostly the province of geologists and volcanologists. The accepted explanations had involved combinations of fractoemission (charging by fracturing of volcanic rocks), triboelectrification (charging by collision of ash particles), and streaming potential (frictional charge accumulation as magma flows through the vent). These are so-called “dry” charging mechanisms, because they do not involve the presence of condensed water.
The collaboration between McNutt and Williams added another important piece to the puzzle of electrified volcanic plumes. To an atmospheric scientist, an explosive volcanic eruption plume looks like a “dirty thunderstorm,” that is, one with ash in it. It is a buoyant, vertically rising current of, initially, hot ash, gases, and air that can reach to the top of the troposphere and beyond. Ironically, it was the volcanologist Rick Hoblitt who originally introduced that phrase in describing one stage of the eruption of Mount St. Helens. He had noted behavior in that plume, and also with Mount Pinatubo, that strongly resembled traditional thunderstorms. The breakthrough that Williams and McNutt contributed was to recognize water substance as the fundamental common denominator both in explosive volcanic eruptions and in thunderstorms, and to demonstrate that the source of that moisture in a volcanic plume was not the ambient air, but water that had been dissolved in the magma (molten rock).
Any volcanologist can describe the essential role of magmatic water for large, explosive eruptions. In situations prior to an eruption where water-laden magma rises quickly from deep underground toward the surface, a rapid decompression can take place. This will cause the water to come out of solution and undergo a rapid expansion into the gas (vapor) phase, like popping a cork from a champagne bottle. The ensuing massive volume of steam will drive an explosive eruption.
Colleagues from New Mexico Tech install a VHF antenna, used to record radio signals from lightning, during the eruption of the Eyjafjallajökull volcano in Iceland last summer. This station, one of six Lightning Mapping stations installed around the volcano, was placed 20 kilometers northeast of Eyjafjallajokull on a small farm near Hvolsvollur.
Similarly, meteorologists can explain two key roles of water in thunderstorms: enhancing the updraft and ice particle-based charge separation. Thunderstorms begin as plumes of warm air rising from the surface. As that air rises, it expands and cools, causing the water vapor in the air to condense into liquid. The transition from a higher energy state (gas) to a lower energy state (liquid water) releases 600 calories of heat energy per gram of water. This “latent” heat release helps to drive the intense buoyant updrafts that are the signature of all thunderstorms. And the more water that is available, the stronger those updrafts can become. But as that buoyant air continues to rise and cool to temperatures below freezing, ice formation also becomes possible. The charge separation necessary to produce lightning is achieved when smaller, positively charged ice crystals get blown to the top of the cloud in the strong updraft, while larger, heavier, negatively charged ice particles like hail and graupel settle to the lower part of the cloud. Generally, the stronger the updrafts are, the more efficient the charge separation, and the more vigorous the lightning. This is why tornadic supercell thunderstorms often display nearly continuous flash rates.
The best research ideas often are those that seem obvious afterward but were apparently unnoticed prior to conception. Williams and McNutt applied some simple calculations that, in retrospect, it seems remarkable nobody had thought to work out before. Their paper explicitly addressed this disconnect: “The behavior of water in magma within the Earth is reasonably well understood in volcanology, and the behavior of water in the atmosphere is adequately understood in meteorology. The perceived gap in understanding lies in the transition from earth to atmosphere. This study is aimed at bridging that gap.”
“Bridging the gap” involved simply estimating a typical water content of magma for an explosive eruption, then calculating the resulting atmospheric water vapor concentration in the plume above the volcano. The literature suggested magmatic water content of 5 percent by weight. This proportion of water was assumed to emerge from the vent in a large, explosive eruption as part of an estimated typical quantity of magma expanding into a so-called “relaxation volume” defined by the size of the explosion. Results suggested meteorologically very large water vapor values, much more, in fact, than present in a typical thunderstorm. This was a jaw-dropping conclusion for meteorologists and volcanologists alike—neither had any idea that a volcanic plume was so wet by meteorological standards!
It also means that as the eruption plume rises to the upper atmosphere and cools to below-freezing temperatures, it will produce large quantities of ice or ice-coated ash particles, effectively shutting down the “dry” charging mechanisms high in the plume, but opening the door to very efficient ice charging mechanisms. One would expect standard thunderstorm electrification processes to subsequently occur. Many intriguing anecdotal accounts of unusual liquid and ice precipitation from eruption clouds support this possibility.
“Lightning seems to be considerably more frequent with volcanic plumes greater than 23,000 feet in height.”
Further investigation by McNutt and Williams served to reinforce the “dirty thunderstorm” hypothesis. Based on an idea called the Volcanic Explosivity Index (VEI), their work showed that lightning is reported much more frequently for high VEI values than for low values. When looked at in terms of plume height, they noted that lightning seems to be considerably more frequent with volcanic plumes greater than 23,000 feet in height, that is, tall enough and cold enough to invoke ice charging mechanisms, and therefore comparable to thunderstorms. Additional evidence for magma as the water source for lightning generation is provided by looking at the latitudinal distribution—lightning occurs as frequently for high-latitude volcanoes as for those of low latitude. If the volcanoes were drawing in water by incorporating the surrounding air, one would expect considerably less lightning from high-latitude eruptions than those in the tropics, where vapor in the atmosphere is markedly more abundant.
A third line of evidence involves seasonal occurrence. Although thunderstorms in most mid-latitude locations are much more common during the warm season, volcanic lightning is actually more common in winter, again pointing to the importance of magmatic water as the source of electrification in the volcanic plume.
While these discoveries were crucial to our understanding of volcanic lightning, Earle Williams turned out to be important to McNutt in another way. As a lightning scientist, Williams had worked with Ron Thomas, Paul Krehbiel, and Bill Rison of the New Mexico Institute of Mining and Technology, developers of the Lightning Mapping Array (LMA). LMAs are lightning sensors, each about the size of a picnic cooler, that can be quickly deployed to make lightning measurements. When Augustine Volcano in Alaska began intermittent explosive eruptions in January 2006, Williams convinced the New Mexico team to try out these systems there. The deployment of the LMA at Augustine would bring the study of volcanic lightning full circle, back to the Langmuir Lab.
The system was deployed with assistance from the AVO facility in Homer, Alaska, just in time for a new eruption. It took about four days to get two stations set up, and when the team members finished, they went to a celebratory dinner; by the end of their meal, an eruption was occurring. Ron Thomas says, “One of the stations was there at the base, so we went up there and we could look at the data right away, and instantly we saw that we had amazing lightning.”
The LMA senses the radio signals emitted by lightning, usually using VHF television channel 3, 4, or 2. As Thomas notes, “Usually television channel 3 is unused, so we can use that and not get any interference. Usually if there's a channel 3 [in a given area] then there's not a channel 4” and the LMA can be switched to that frequency. “We can usually find an unused television channel.” Unlike the previously mentioned cloud-to-ground lightning systems, the LMA can monitor in-cloud strokes as well. This is important, because in Hoblitt's study, visual observations had indicated that the eruptions often were electrically active with intracloud strokes, even when the sensors were not picking up any active cloud-to-ground lightning.
“We're measuring radio signals in the VHF at 60 megahertz,” Thomas said. “Those come from the formation of leaders in the lightning channel, so any time there's lightning extending, we're going to get signals. We get lots and lots of signals from each lightning flash, so we can map their pattern and structure. What you see [visually] at the bottom of the cloud is only a tiny piece of the lightning, and what's up in the cloud has all kinds of branch structures and we can see all that [with the LMA].”
This unprecedentedly detailed view of the volcanic lightning yielded a surprising result: The electrical discharges seemed to occur in two distinct phases. The first occurred immediately upon eruption, with numerous discharges near the vent as highly charged ejecta exited the volcano, and this appears consistent with dry charging mechanisms. The second occurred 4 to 12 minutes after the beginning of the eruption with a sequence of about 300 well-defined lightning discharges, apparently within the volcanic plume. The plume ultimately reached up to 10 km in height, giving credence to the view that this second phase may have been associated with the development of ice particles and charge separation similar to traditional thunderstorms. Thus, the new, highly detailed observations seemed to support both the hypotheses of so-called “dry” charging mechanisms near the vent, and also Williams and McNutt's version of the “dirty thunderstorm” hypothesis.
“Overall the LMA has been deployed successfully and collected lighting data at four colcanoes.”
The Augustine Volcano results ultimately led to a National Science Foundation grant for collaboration between the atmospheric physicists and McNutt, the volcano seismologist, and to support further refinements on the LMA system so that it could be left in the field longer with a minimum of attention. Sonja Behnke is a doctoral student in atmospheric physics working with Ron Thomas. Her research with Thomas, McNutt, and others on Redoubt's 2009 eruption corroborated previous findings of a correlation between the level of electrical activity and the size of the eruption plumes: The larger, taller plumes typically yield more lightning, again lending credence to the “dirty thunderstorm” hypothesis.
Overall the LMA has been deployed successfully and collected lightning data at four volcanoes now—Augustine, Chaiten in Chile, Redoubt, and Eyjafjallajökull in Iceland. Steve McNutt believes that this is a fortuitous assortment—each has a predominantly different eruption chemistry, and therefore analysis of those data may help to distinguish whether the different compositions yield differences in lightning characteristics and plume electrification.
Since 2000, it is fair to say that volcanic lightning has received more attention by researchers and the media than ever before. The American Geophysical Union 2010 Fall Meeting hosted a full session just on this topic. The development of ever more sensitive lightning detection devices and networks and the expanding presence of weather radar have extended the ability to examine volcanic plume behavior. Researchers in Europe gained a wealth of data from Eyjafjallajökull, logging lightning strikes and examining radar patterns. Researchers from Britain even sent up a balloon-borne instrument to measure plume electrification over Scotland.
With additional scrutiny come new explanations for some observed phenomena. Carrying the concept of a dirty thunderstorm one step further, a recent study by Pinaki Chakraborty and colleagues from the University of Illinois in Urbana proposed that large volcanic eruption plumes may lead to “volcanic mesocyclones.” Mesocyclones are the small-scale, low-pressure systems of rotating supercell thunderstorms, which are the parent storms of tornadoes. Satellite observations showed that the plumes of Pinatubo, Chaiten, and other volcanoes were, in fact, rotating. The researchers had been intrigued by a 200-year-old account by a sea captain of a volcanic eruption in the Azores in which the plume rotated “like an horizontal wheel,” accompanied by continual lightning flashes. Waterspouts also were observed, suggesting that this was, indeed, a volcanic mesocyclone. The researchers point out that this concept provides a unifying theory for previously unexplained observations of funnel clouds and “lightning sheaths” in volcanic plumes that also are common to supercell thunderstorms.
Supercell thunderstorms like this one often display almost continuous lightning flashes.
With the recent air traffic disruption in Europe from the eruption of Eyjafjallajökull and so many “great circle” routes bringing a disproportionate volume of flights near Alaska, aircraft safety continues to provide a strong motivation to study volcanic lightning and find ways to better incorporate it into the warning infrastructure. Many basic questions remain. Thomas and Behnke see volcanic lightning detection not only as an important tool for aviation safety, but also as a chance to look at atmospheric electrification in two different kinds of systems, that is, volcanic plumes versus thunderstorms. How are they similar? How are they different?
Nevertheless, the dangerous and hostile environment of an explosively erupting volcanic plume remains an impediment to research. Coming years will see further development of the lightning detection technology, additional analysis of the data that have been collected to date, alternative methods for remotely sensing plume characteristics, and the use of unmanned aerial vehicles, or other ways of making direct measurements of water, chemistry, and electrification within volcanic plumes. The direct measurements are important, because we still do not understand the detailed microphysics of one of our most basic questions: How do individual particles within volcanic plumes gain charge? (Surprisingly, this is true even for ordinary thunderstorms.) The future of volcanic lightning as a subject for scientific study rests not only on the improving technology but, most importantly, on the continuing attention by scientists with a passion to prevail in the difficult environment of an erupting volcano for the sake of science. Ron Thomas, who has spent more than his share of research time looking at meteorological thunderstorms, likens collecting such measurements to chasing tornadoes: “It's fun to get out there and try and get to the right place at the right time … and we got to Iceland at the right time.” Pliny the Younger, the Roman scholar, would have liked to see the results.
BRADLEY M. MULLER is a professor of Applied Meteorology in the Department of Applied Aviation Sciences at Embry-Riddle Aeronautical University in Daytona Beach, Florida. He lives in De Leon Springs, Florida.