Journal of Geophysical Research: Atmospheres

Aviation response to a widely dispersed volcanic ash and gas cloud from the August 2008 eruption of Kasatochi, Alaska, USA

Authors


Abstract

[1] The extensive volcanic cloud from Kasatochi's 2008 eruption caused widespread disruptions to aviation operations along Pacific oceanic, Canadian, and U.S. air routes. Based on aviation hazard warnings issued by the National Oceanic and Atmospheric Administration, U.S. Geological Survey, the Federal Aviation Administration, and Meteorological Service of Canada, air carriers largely avoided the volcanic cloud over a 5 day period by route modifications and flight cancellations. Comparison of time coincident GOES thermal infrared (TIR) data for ash detection with Ozone Monitoring Instrument (OMI) ultraviolet data for SO2 detection shows congruent areas of ash and gas in the volcanic cloud in the 2 days following onset of ash production. After about 2.5 days, the area of SO2 detected by OMI was more extensive than the area of ash indicated by TIR data, indicating significant ash depletion by fall out had occurred. Pilot reports of visible haze at cruise altitudes over Canada and the northern United States suggested that SO2 gas had converted to sulfate aerosols. Uncertain about the hazard potential of the aging cloud, airlines coped by flying over, under, or around the observed haze layer. Samples from a nondamaging aircraft encounter with Kasatochi's nearly 3 day old cloud contained volcanic silicate particles, confirming that some fine ash is present in predominantly gas clouds. The aircraft's exposure to ash was insufficient to cause engine damage; however, slightly damaging encounters with volcanic clouds from eruptions of Reventador in 2002 and Hekla in 2000 indicate the possibility of lingering hazards associated with old and/or diffuse volcanic clouds.

1. Introduction to Aviation Protocol for Volcanic Eruptions

[2] During July and August of 2008, three volcanoes in Alaska's Aleutian Islands chain (Kasatochi, Okmok, and Cleveland; Figure 1) erupted, causing widespread impacts to aviation operations as the volcanic ash and gas ejected into the stratosphere dispersed into major North American jet airways. The eruptions of Okmok in July 2008 and Kasatochi in August 2008 both produced SO2 clouds notable for their dispersion over the northern conterminous United States. Flights were diverted north of normal air routes to avoid Okmok volcanic clouds, and a nondamaging encounter with suspected ash from the Okmok eruption occurred near Kodiak Island on 15 July 2008 (the flight crew reported seeing a strange halo and brown spots on the windshield, but no further effects or damage were noted). However, Kasatochi produced the largest of the three eruptions and the most extensive volcanic cloud, and the overall impact on aviation of Kasatochi's eruption was greater than Okmok's. In this paper, we focus on the effects of Kasatochi's eruption on aviation operations and discuss some of the real-world aspects of using scientific data to respond to hazards to aviation from volcanic clouds.

Figure 1.

Location map showing Kasatochi, Okmok, and Cleveland volcanoes in Alaska and some representative air routes. Black filled triangles indicate active and potentially active volcanoes [from Simkin and Siebert, 1994].

[3] Volcanic ash clouds dispersed into air routes have damaged aircraft in the past [Miller and Casadevall, 2000]. The severity of damage from encounters of aircraft with volcanic ash clouds has varied widely [Guffanti et al., 2004]. In the worst cases, encounters have caused in-flight losses of engine thrust power and serious failures of avionics and other aircraft operating systems. Even in cases when no degradation of aircraft performance or evidence of damage was detected during flight, airlines have grounded planes suspected of flying through volcanic ash for inspection and flight recertification. Because of concerns about safety and costly damage, aviation authorities and the airline industry worldwide have sought to avert ash encounters by avoiding entry into ash clouds.

[4] The protocols to avert encounters are coordinated on a global scale by the International Civil Aviation Organization (ICAO). Following dramatic encounters of aircraft with ash clouds in the 1980s that involved in-flight losses of engine thrust power [Miller and Casadevall, 2000], ICAO formulated a body of operational protocols for eruption reporting, ash-cloud detection and forecasting, and dissemination of specialized warning messages worldwide to air traffic controllers, dispatchers at airline operation centers, and pilots. Using information about the occurrence of explosive eruptions and the movement of associated ash clouds provided in the warning messages, as well as observations of anomalous phenomena made by en-route flight crew, the aviation sector employs in-flight diversions, preflight route modifications, and flight cancellations to avoid ash-contaminated airspace.

[5] A major achievement of ICAO was to define the role of nine Volcanic Ash Advisory Centers (VAACs; Figure 2) to detect the presence of ash in the atmosphere primarily using data from civilian meteorological satellites, pilot reports, and volcano observatory notices. VAACs predict movement of ash clouds several hours into the future using various numerical weather forecast and atmospheric dispersion models and provide volcanic ash advisories; the advisories in turn are the basis of formal ash hazard warnings (called Volcanic Ash SIGMET for SIGnificant METeorological events) that are issued by the worldwide system of Meteorological Watch Offices.

Figure 2.

Map showing areas of responsibility of the nine Volcanic Ash Advisory Centers (VAAC) worldwide (taken from ICAO Web site; please check http://www.icao.int/anb/iavwopsg/ regularly for updates). Hachured areas not covered by any VAAC.

[6] In the United States, the ICAO protocols are the basis for a formal operational plan that specifies the coordinated responsibilities of the Federal Aviation Administration (FAA), National Oceanic and Atmospheric Administration (NOAA), U.S. Geological Survey (USGS), and Air Force Weather Agency in carrying out the ash avoidance program in the National Airspace System [Albersheim and Guffanti, 2009]. In the case of Kasatochi, eruption reporting was provided by the Alaska Volcano Observatory (AVO). Cloud tracking and dispersion forecasting were conducted by the three VAACs into whose areas of responsibility the volcanic cloud dispersed, the Anchorage VAAC and the Washington VAAC in the United States and the Montreal VAAC in Canada. Volcanic ash SIGMETs were issued by three meteorological watch offices in the two countries: the Alaska Aviation Weather Unit in Anchorage and Aviation Weather Center in Kansas City, Missouri, both within NOAA, and the Canadian Meteorological Aviation Centre in Edmonton. The FAA's Anchorage Air Route Traffic Control Center and the Command Center in Herndon, Virginia, were responsible for issuing international notices to airmen to inform pilots of impending and ongoing volcanic hazards.

2. Tracking Kasatochi's Volcanic Cloud With Comparative Remote Sensing Techniques

[7] Kasatochi erupted on 7–8 August 2008, producing volcanic clouds to altitudes of 14–18 km asl and sending an impressive amount of sulfur dioxide (SO2) gas (approximately 2 Tg according to Yang et al. [2010]) as well as volcanic ash into the stratosphere. For the purpose of our analysis of aviation impacts, we track volcanic cloud movement over a period of about four days after the main eruptive activity, based on a series of time-coincident images from two different satellite-based sensors. (1) The ozone monitoring instrument (OMI) is a hyperspectral spectrometer on NASA's polar-orbiting Aura satellite that detects SO2 in the atmosphere at ultraviolet wavelengths [Yang et al., 2007]. SO2 is measured throughout the atmosphere with high resolution and sensitivity during periods of sunlight, and data are available once a day at low latitudes and at least once a day at high latitudes. The OMI data are made available by NASA and NOAA within 3 h of acquisition. (2) Thermal infrared (TIR) sensors on NOAA's geostationary operational environmental satellites (GOES) provide data for the brightness-temperature-differencing (BTD) technique (also known as band 4 minus 5) that is used to detect ash particles [Prata, 1989]. GOES data are available within approximately 30 min of acquisition by the sensors.

[8] Three main explosive events comprised the 2008 Kasatochi eruption. The first two explosive events (at 22:01 UTC on 7 August 2008 and 01:50 UTC on 8 August 2008, according to seismic data recorded by AVO) were water-rich, likely due to the eruption passing through and eliminating a 0.5 km diameter crater lake [Waythomas et al., 2010]. These events were relatively short-lived (less than 1 h in duration each), and the resulting clouds did not show thermal infrared BTD signals indicative of fine ash particles, probably due to masking by abundant magmatic and crater water. The third explosive event started at 04:35 UTC on 8 August 2008 (according to seismic data recorded by AVO) and produced a dark-colored umbrella cloud that transitioned into a narrow eruption plume. Emission of eruptive material was sustained for at least 16 h to an altitude of about 10 km asl and was characterized by a robust thermal infrared BTD signal indicative of abundant volcanic ash.

[9] Over the next several days, the volcanic ash and SO2 cloud dispersed over the Gulf of Alaska, Canada, and the northern conterminous United States. Initially, during the first 12–18 h after the first eruption began, a low pressure system located just east of Kasatochi transported the ash cyclonically around the low, keeping the cloud near the volcano. At 01:00 UTC on 9 August 2008, about 20.5 h after the start of the third, ash-producing, explosive event, a GOES BTD image shows the ash cloud just south of the Aleutian Islands; an OMI image at the same time shows an area of SO2 gas congruent with the ash signal (Figure 3). At 00:00 UTC on 10 August 2008, about 43.5 h after the start of major ash production, the ash and gas cloud had moved over the Gulf of Alaska, again with congruent areas of ash and SO2 (Figure 4).

Figure 3.

(a) Kasatochi's ash cloud as indicated by GOES thermal infrared BTD image at 01:00 UTC on 9 August 2008, about 20.5 h after the start of major ash emission. (b) Kasatochi's SO2 gas cloud as indicated by OMI image at 01:00 UTC on 9 August 2008. One Dobson unit (DU) is equal to 2.68 × 1016 ozone molecules per square centimeter.

Figure 4.

(a) Kasatochi's ash cloud as indicated by GOES thermal infrared BTD image at 00:00 UTC on 10 August 2008, about 43.5 h after the start of major ash emission. (b) Kasatochi's SO2 gas cloud as indicated by OMI image at 00:00 UTC on 10 August 2008. One Dobson unit (DU) is equal to 2.68 × 1016 ozone molecules per square centimeter.

[10] By 23:00 UTC on 10 August 2008, more than 2.5 days (66.5 h) after the start of major ash emission, the area of ash indicated by the BTD signal from GOES data had become noticeably smaller than the area of SO2 indicated by OMI data (Figure 5), suggesting that most of the ash had been removed by gravitational settling [Rose et al., 2008]. On 12 August 2008, about 4 days after the start of major ash emission, OMI data show a sprawling feature over Canada and parts of the conterminous United States (Figure 6); at that point, no BTD signal indicative of ash was detected.

Figure 5.

(a) Kasatochi's ash cloud as indicated by GOES thermal infrared BTD image at 23:00 UTC on 10 August 2008, 66.5 h (nearly 3 days) after the start of major ash emission. (b) Kasatochi's SO2 gas cloud as indicated by OMI image at 23:00 UTC on 10 August 2008. Location of aircraft encounter over northern British Columbia is indicated by a red star. One Dobson unit (DU) is equal to 2.68 × 1016 ozone molecules per square centimeter.

Figure 6.

Kasatochi SO2 gas cloud as indicated by composite OMI image on 12 August 2008. One Dobson unit (DU) is equal to 2.68 × 1016 ozone molecules per square centimeter.

[11] This comparison of GOES TIR and OMI UV data shows generally congruent areas of ash and SO2 gas in Kasatochi's volcanic cloud in the 2 days following the start of major ash production at the volcano and diminishing areas of ash relative to the area of gas after about 2.5 days. After about 4 days from the start of major ash production, the volcanic cloud was characterized by an extensive OMI (SO2) signal and no BTD (ash) signal, indicating substantial ash fall out had occurred with continued dispersion of the predominantly gas, ash-depleted cloud.

3. The Aviation Response

[12] The flow of information to the aviation sector about volcanic cloud hazards associated with Kasatochi began with AVO. In addition to an alert level scheme for ground hazards, AVO employs a ranked, color-coded alert scheme for aviation users (green, yellow, orange, red) that focuses on ash-producing activity at a volcano [Gardner and Guffanti, 2006]. AVO also carries out a telephone call down that has the FAA's Anchorage Air Route Traffic Control Center as the first external call and issues a specially formatted, ICAO-endorsed message called a VONA (Volcano Observatory Notice for Aviation) [Albersheim and Guffanti, 2009], aimed at dispatchers, pilots, air traffic controllers, and aviation meteorologists. AVO raised the alert level at Kasatochi from Green to Yellow on 7 August 2008 at 03:17 UTC after earthquake activity in the greater vicinity of Kasatochi volcano had increased rapidly. (An archive of AVO notifications is available at http://www.avo.alaska.edu/activity/avoreport.php/.)

[13] As the eruption began on 7 August 2008, critical aviation warning messages were issued over a period of an hour and a half. AVO gave the aviation sector the warning of impending eruptive activity when it raised the aviation color code for Kasatochi from yellow to orange at 21:57 UTC on 7 August, conducted a call down, and issued an accompanying VONA. The FAA's Anchorage Air Route Traffic Control Center in turn quickly issued an international notice to Airmen at 22:26 UTC informing pilots that they should remain alert for a possible eruption of Kasatochi. NOAA satellite data at 22:30 UTC showed an ash plume to an altitude of at least 35,000 ft. in the vicinity of Kasatochi Volcano (from the eruption that seismic data later determined to have started at 22:01 UTC), and the first SIGMET warning was issued at 22:55 UTC by the Alaska Aviation Weather Unit. AVO raised the color code to Red and conducted a call down, then issued an eruption VONA at 23:26 UTC; another Notice to Airmen was issued concurrently by the FAA to inform pilots of ongoing ash hazards.

[14] On the basis of these initial warnings and numerous subsequent updates, airline dispatchers and pilots coordinated with FAA air traffic controllers to reroute flights in progress, modify routes during preflight planning, and sometimes cancel flights; the total numbers of these actions are not known. That effort continued for at least 5 days as Kasatochi's ash and gas cloud dispersed over the Gulf of Alaska, Canada, and northern conterminous United States. On 10 August 2008, as the cloud began to approach the west coast of Canada and the United States, flight disruptions were acute. Most notably, more than 40 flights from Anchorage to Seattle, Portland, San Francisco, Denver, Los Angeles, and Vancouver, B.C., were cancelled, stranding more than 6000 passengers according to accounts in the Anchorage Daily News and Associated Press.

[15] By 11 August 2008 when Kasatochi's cloud had begun to move over Canada and parts of the northern conterminous United States, dozens of pilots were reporting sulfurous odors in the cockpit and layers of visible haze at altitudes above about 8 km asl (about 26,000 ft km) that were thought to be ash clouds. For example, at about 01:00 UTC on 11 August 2008 the flight crew of a 757 over northern British Columbia, Canada, smelled sulfurous odor and saw a yellowish brown haze layer from the flight deck at approximately 10 km asl (34,000 ft). The aircraft descended through the haze layer to about 8.5 km (about 28,000 ft); no changes in performance of engines or other aircraft systems were noticed. Upon landing at its destination, the aircraft's engines were inspected, and samples were taken of particles adhering to the leading edge, engine cowl, and air conditioning pack inlet. The inspection showed no discernible damage to the airframe or engines. Sample material was given to the USGS for analysis; the results are summarized in the following section.

[16] This nondamaging encounter occurred nearly three days after the start of major ash emission and at the northern fringe of the area of SO2 indicated by an OMI image two hours earlier (at 23:00 on 10 August; Figure 5b). The significance of the corresponding GOES BTD image (Figure 5a) is inconclusive, as there are small regions of slightly negative BTD values near the area of the encounter, but similar areas are observed throughout image field of view that are not related to volcanic ash.

[17] Pilot reports of sulfurous odors in the cockpit are not made or taken lightly, although the significance of the presence of such odors varies greatly. In ICAO's encounter severity index [Guffanti et al., 2004], the smell of sulfur alone at cruise altitudes qualifies as a lowest-severity encounter, although the uneven nature of reporting the occurrence of sulfurous odors reduces the utility of documenting the number of incidents in that class. Notwithstanding difficulties of “counting” low-severity encounters, sulfurous smells are a notable indicator of possible hazards. Volcanoes are the only sources of large quantities of sulfur gases at cruise altitudes, and both SO2 and H2S are detectable by smell. Accordingly, the smell of sulfur gases in the cockpit may indicate volcanic activity that has not yet been detected or reported and/or possible entry into an ash-bearing cloud. In some cases when sulfur gases are smelled, there may be little ash in the cloud owing to ash fallout during prior dispersion of the cloud or separation of the ash and gas components of the cloud, but a flight crew does not have the means to determine directly that the cloud is “safe” and thus is advised by ICAO guidance (http://www2.icao.int/en/anb/met-aim/met/iavwopsg/IAV Handbook/IAVW Handbook (Doc 9766).pdf) to report the observation to air traffic-control centers and airline dispatch centers and seek to exit the cloud. Pilot reports sometimes are forwarded to the VAAC, in which cases the observations may be useful in validating remotely sensed data.

[18] On 12 August 2008, Cleveland volcano in the Aleutians erupted with an ash cloud to about 7.5 km (25,000 ft), while low-level ash emission and seismic unrest occurred at Okmok and Kasatochi's cloud was still being tracked well beyond Alaska. With such notable eruptive activity and widespread effects on aviation in U.S. and Canadian airspace, the FAA Command Center in Herndon, Virginia, initiated a conference call at about 17:00 UTC on 12 August 2008 to share information about the eruptive activity, the evolving nature of the volcanic cloud, and the effects on the National Airspace System. The call involved more than 50 people from dozens of FAA Air Route Traffic Control Centers in the conterminous United States, the Washington VAAC, the Montreal VAAC, the Anchorage VAAC, and several of the National Weather Service's Central Weather Service Units.

[19] The extensive dispersion of the Kasatochi cloud required that the three VAACs closely coordinate their advisories as the cloud traversed across their jurisdictions. VAACs issue advisories that are used to produce SIGMET warnings; the advisories usually include 18 h forecasts of expected cloud movement and are available to air carriers as well as to the Meteorological Watch Offices that issue SIGMETs. A VAAC issues advisories as a volcanic cloud enters its area of responsibility. With the Kasatochi cloud, the Anchorage VAAC issued the first volcanic ash advisory early on 8 August 2008, and the Montreal VAAC issued the final advisory early on 13 August 2008. A total of 54 advisories were issued by the three VAAC over the 5 day period.

[20] The Anchorage VAAC issued the first Kasatochi advisory at 01:45 UTC on 8 August 2008 for the first volcanic emission at 22:01 on 7 August 2008 (an archive of Anchorage VAAC advisories is located online at http://vaac.arh.noaa.gov/list_vaas.php/). The Washington VAAC issued 16 advisories and forecasts during 8–12 August 2008 that referenced first the Anchorage VAAC on 8–9 August and then the Montreal VAAC on 11–12 August as the primary sources of information as the cloud moved into and then out of the Washington VAAC's area of responsibility. (An archive of Washington VAAC advisories in 2008 is located online at http://www.ssd.noaa.gov/VAAC/ARCH08/archive.html.) The Montreal VAAC began to issue advisories at 00:03 UTC on 11 August 2008. Montreal VAAC advisories initially incorporated pilot reports that used the terms “ash” or “ash cloud” to describe their observations of haze, so the advisories showed large areas of ash. Based on observations in the pilot reports, the Montreal VAAC advisories put the base of the haze layer near 8.5 km (28,000 ft), with a couple of exceptions closer to 6 km (20,000 ft). By including this information, airlines had some useful basis for determining if they could safely fly under the haze layer.

[21] On 12 August 2008, OMI data indicated a sprawling SO2-rich feature extending over Canada and extending into the United States (Figure 6). By this time, the consensus had grown among many scientists sharing data and ideas over the Internet that the haze seen at cruise altitudes was SO2 gas (which is colorless) reacting with moisture in the atmosphere to form a layer of sulfate aerosols made visible to pilots by the tiny droplets' scattering of sunlight. Advisories issued by the Montreal VAAC included information about visible “ash” received in pilot reports, but also contained remarks that the amount of ash present in the volcanic cloud was decreasing with time. As a result, the area of ash defined in the advisories gradually diminished over time, except to reflect specific pilot reports that cast some doubt as to the absence of ash in specific areas. In the final advisory issued for the eruption at 03:52 UTC on 13 August 2008, the Montreal VAAC included a remark to indicate that volcanic cloud over North America was believed to contain generally SO2 and aerosols but little ash and that it likely would linger several days. Indeed, filaments of the Kasatochi SO2 cloud dispersed over the northern hemisphere for about a month, and a concomitant stratospheric sulfate aerosol layer continued to be observed [Carn et al., 2008].

[22] Even as recognition grew that the volcanic cloud was primarily SO2 gas and related aerosols and was ash-depleted, operational uncertainties remained for the meteorologists responsible for forecasting the location of ash-contaminated airspace and for the dispatchers, pilots, and air traffic controllers making routing decisions. How much, if any, ash was associated with the cloud characterized in remote sensing images primarily by its SO2 signature; was there enough ash to damage an aircraft? What was the possibility of cumulative adverse effects from repeatedly flying through sulfate aerosols? Procedures for inadvertent entry of aircraft into volcanic ash have been formulated by ICAO, but procedures are not as well specified for flying into large areas of volcanic gas and related aerosols. In the midst of an uncertain situation, airlines were careful and coped by flying over, under, or around the observable haze layer.

4. Ash Sample Data

[23] Samples from the nondamaging incident on 11 August 2008 over Canada were given to the USGS and analyzed using the Environmental Scanning Electron Microscope (ESEM) at the University of Alaska Fairbanks Advanced Instrumentation Laboratory. The ESEM analysis documented the presence of volcanic ash in the samples, although the majority of the particles collected could not be confirmed as being volcanic in origin, and we can only speculate that these are particles that routinely collect on aircraft such as dust from the runway, paint chips, pollen, etc. Figure 7 shows several representative photomicrographs of particles that clearly are volcanic in origin, including glass shards with very fine particles of adhered ash and fresh volcanic mineral crystals such as hornblende and albite. These types of particles are consistent with the composition of erupted deposits on Kasatochi Island, namely crystalline-rich, hornblende-bearing andesite with glassy matrix material [Izbekov, 2009].

Figure 7.

Scanning electron photomicrographs of material collected from the aircraft that encountered the Kasatochi volcanic cloud on 11 August 2008 over northern British Columbia, Canada. Images were taken on the Environmental Scanning Electron Microscope (ESEM) at the University of Alaska Fairbanks Advanced Instrumentation Laboratory. (a) Bulk material collected from the leading edge of a wing of the aircraft; the image shows both organic and clastic material, most of which is nonvolcanic in origin. (b) Material collected from the leading edge of a wing of the aircraft; the image is of relatively large vesicular glass shard with fine-grained ash particles filling the shard vesicles. (c) Volcanic hornblende crystal in material collected from the leading edge of a wing of the aircraft. (d) Material collected from the leading edge of a wing of the aircraft, the image shows an aggregate glass shard with fine particles adhering to its surfaces. (e) Mineral grain coated in volcanic glass collected from a pack inlet of the aircraft. (F) Albite crystal collected from the leading edge of a wing of the aircraft; based on the pristine condition of this crystal, it likely was derived from the Kasatochi ash cloud.

[24] Analyses of ash samples from aircraft encounters are rare. The results of the ESEM analysis confirm the widely held assumption that some ash is present in volcanic clouds that are discernible primarily by their SO2 gas content. In this case, the aircraft's exposure to ash (ash concentration and time in cloud) was not sufficient to cause damage to the aircraft. However, slightly damaging encounters on two previous occasions highlight the uncertainties in generalizing the outcome of the Kasatochi incident to future encounters with ash-depleted clouds.

[25] On 28 February 2000, a NASA research plane flew through the diffuse edge of a volcanic cloud from an eruption of Hekla in Iceland one and a half days earlier [Grindle and Burcham, 2003]. The presence of the volcanic cloud was detected by sensitive atmospheric sampling instruments onboard. Elevated SO2/H2SO4 gas concentrations and aerosol number density were recorded for a period of about 7 min; no changes in cockpit readings, odor, corona discharge (St. Elmo's fire), or other indicators of ash were noticed. In-flight checks and post-flight inspections revealed no damage. The aircraft was flown for another 68 hours, then a borescope analysis was conducted that revealed damage in one engine, including turbine cooling holes plugged by ash and erosion of leading edge coating on turbine blades. SEM analysis of particles collected in cabin air heat exchanger filters showed evidence of basaltic particles between one and ten microns in diameter and clumped together in the fibers [Pieri et al., 2002]. The most likely interpretation is that the Hekla cloud at the time of the encounter was gas- and ice-rich with minor amount of fine ash particles [Rose et al., 2003].

[26] Another relevant case involves encounters with an old, far-traveled volcanic cloud. On 23–24 November 2002, two aircraft over Micronesia northeast of Papua New Guinea encountered a volcanic cloud that Tupper et al. [2006] concluded was not from a nearby source but had advected over a great distance. One aircraft reported typical indicators of entry into an ash cloud, intense St Elmo's fire and light white “smoke” with “burn smells.” Three pilot probes were later found to contain ash particles and were replaced; some light abrasion was found on the engine air inlets, but no damage on the windscreen or internal engine damage was reported. The flight crew of the second aircraft observed the cloud and smelled a slight odor but no damage was found. On the basis of backward and forward wind trajectories and dispersion forecasts, as well as analysis of satellite imagery, Tupper et al. [2006] proposed the probable source of the volcanic cloud as the 3–5 November eruption of Reventador volcano in Ecuador. If so, that eruption cloud would be the oldest (20 days) and furthest traveled (14,000 km) known to have caused damage to an aircraft.

[27] The Kasatochi, Hekla, and Micronesia incidents show that a gas-rich or old volcanic cloud can contain enough fine ash particles to be noticeable during flight, identifiable by subsequent close inspection of an aircraft and/or damage an aircraft.

5. Conclusions

[28] Overall, the aviation response to the extensive dispersion of the ash and gas cloud from the 7–8 August 2008 eruption of Kasatochi was successful. Numerous volcanic ash warning products were issued: seven Volcano Observatory Notices for Aviation, 54 VAAC advisories, hundreds of SIGMET, and several Notices to Airmen. Effective coordination among various aviation-weather agencies in the United States and Canada occurred. Air travel continued, albeit with numerous route modifications and flight cancellations along North Pacific, Central Pacific, Canadian, and continental U.S. routes for a period of 5 days. No damaging encounters of aircraft with volcanic ash are known to have occurred.

[29] Our comparison of time coincident GOES TIR data and OMI UV data characterizes the evolving nature of a volcanic cloud and puts time constraints on the process of ash depletion. In the case of Kasatochi, the comparison shows congruent areas of ash and gas in the volcanic cloud in the 2 days following the start of ash production at the volcano. After about 2.5 days, the area of SO2 detected by OMI was more extensive than the area of ash indicated by a BTD signal, suggesting ash depletion by fall out had taken place. However, samples of particulate matter from a nondamaging encounter of an aircraft with Kasatochi's nearly 3 day old cloud contained volcanic particles, confirming the widely held assumption that some fine ash is present in volcanic clouds that are discernible primarily by their SO2 content but not by their BTD signature. The aircraft's exposure to ash (ash concentration and time in cloud) was not sufficient to cause damage to the aircraft, but we cannot reliably generalize that outcome for all future encounters with ash-poor clouds. Continued research on aged volcanic clouds is needed.

[30] The capability of OMI to detect Kasatochi's volcanic gas cloud for many days provokes the question: under what conditions should a stratospheric SO2 feature form the basis for an aviation warning? The record of known encounters shows that the most severe ones (those resulting in engine damage) have occurred within a day of ash emission (M. Guffanti and T. Casadevall, unpublished data). It might seem justified to dismiss the hazard potential of a volcanic cloud several days old, but the slightly damaging encounters on 23 November 2002 over Micronesia with what is thought to have been a very old remnant of an ash cloud from Reventador and on 28 February 2000 with the diffuse Hekla cloud offer cautionary counterpoints. Moreover, the cumulative effect of repeated transit through sulfate aerosol clouds on engines, avionics, and airframes is not well known.

[31] The widespread, week-long closure of airspace over Europe and the North Atlantic in April 2010 as a result of the eruption of Eyjafjallajökull volcano in Iceland renewed interest in understanding the potential risks of encounters of aircraft with volcanic ash, particularly as to whether there is a “safe concentration” of ash in a volcanic cloud that poses minimal or no hazard to aircraft. This is a complex, threefold problem: knowing the tolerances of different types of engines to ingestion of ash and related sulfate aerosols; accurately determining the constituents and density of a volcanic cloud at particular times and locations, and tracking the operational conditions of a particular flight including duration spent flying in ash-contaminated airspace. Until these issues are thoroughly understood for the various combinations of volcanological, meteorological, and operational factors that can occur during any given flight, the response by the aviation sector to volcanic cloud hazards will continue to be attended by uncertainty as to the exact nature of potential hazards at specific times and locations.

Acknowledgments

[32] This work was supported by the Volcano Hazards Program of the U.S. Geological Survey. The manuscript was improved by reviews by Rosalind Helz, Gari Mayberry and three anonymous reviewers.

Ancillary