1. Introduction
The transport of fine-grained dust by strong winds has been observed and reported over a broad range of spatial scales (see online at www.osei.noaa.gov/Events/Dust). Two major sources of the dust are Asia (Gobi and Mongolian deserts) and Africa (Sahara). Dust events occur when friction from surface winds (>5 m s−1) entrains and lifts the dust particles into the atmosphere and transports the dust across either the North Pacific or the tropical Atlantic Oceans, respectively (Gillette 1978). Dust from these large-scale events can affect radiative forcing and climate (Myhre and Stordal 2001), biogeochemical cycles (Chadwick et al. 1999), public health (Schwartz et al. 1999), and aviation safety (Simpson et al. 2003).
Airborne volcanic ash from eruptions is a major threat to aviation safety at all scales (Casadevall 1994; Miller and Casadevall 1999; Hufford et al. 2000; Simpson et al. 2000). Like fine-grained mineral dust, volcanic ash affects radiative forcing and climate (Pollack et al. 1976), public health (Bates and Beggs 1997), vegetation (Kobayashi et al. 1988), and can cause property damage and disruption to community infrastructure (Warrick et al. 1981; Blong 1984).
Resuspension of ash even after it has fallen to the ground can be as hazardous as new ash from an erupting volcano (Sparks et al. 1997). Most studies of resuspended ash involve recent eruptions and usually examine only local affects. However, on 20–21 September 2003, a unique set of conditions produced a very large resuspension of relic volcanic ash (and probably some dust) from an area around and including the Valley of Ten Thousand Smokes in Katmai National Park and Preserve, Alaska. The ash cloud extended from the valley south-southeastward off the coast, over Kodiak Island, and into the Gulf of Alaska. This paper describes this event and examines the atmospheric processes that generated it. Impact on regional aviation is also cited.
2. Source region of 20–21 September 2003 resuspended ash event
Katmai National Park and Preserve, located on the Alaska Peninsula, is one of the most active volcanic regions in the world (Fierstein and Hildreth 2001). There are eight volcanoes in the park that are known as the Katmai volcano cluster: Snowy Mountain, Mount Griggs, Mount Katmai, Mount Martin, Trident Volcano, Novarupta Volcano, Mount Mageik, and Alagogshak (Fig. 1). These volcanoes form a 25-km-long line of contiguous stratovolcanoes on the drainage divide of the Alaska Peninsula. Typical elevations of these volcanoes range from 1830 to 2320 m. Major eruptions from these volcanoes have deposited volcanic ash in the Katmai region at least 15 times in the past 10 000 yr. The 1912 Novarupta eruption, the largest of the twentieth century, produced at least 17 km3 of ash fall deposits and 11 km3 of ash flow (pumiceous pyroclastic flow) in about 60 h (Hildreth 1983). The Novarupta vent is located on the north foot of Trident Volcano. This 1912 eruption is virtually unique among major historical eruptions in that it generated a large volume ash flow that all came to rest on land. The ash flow moved like a sheet northwestward from Novarupta and formed the Valley of Ten Thousand Smokes, which covers an area of about 120 km2 (Fig. 2). The thickness of the ash in the valley varies up to about 250 m in depth. Griggs, Katmai, Trident, and Mageik partially surround the head of the flat floored valley (Fig. 1), which still remains largely vegetation free (Hildreth 1987).
Ground-based observers (tourists/park rangers) have reported seeing from time to time the resuspension of some volcanic ash from the ground in Katmai National Park and Preserve. However, resuspension on the scale of the 2003 event described herein is rare. The only potential evidence of a similar past event in Alaska is suggested by Riehle et al. (1987) for the Aniakchak Volocano. A pilot reported volcanic ash over Port Heiden, Alaska, near the volcano. There was no coincident volcanic eruption at the time so the conclusion in the report was that it was resuspended volcanic ash. On 20– 21 September 2003, weather satellite imagery suggested strong winds entrained and lifted ash from the surface of the Valley of Ten Thousand Smokes high into the atmosphere (>1600 m). This event appears to be much larger than any previously reported event: the length of the cloud extended in excess of 230 km from the source and the downstream cloud exhibited both a south and a eastward component (Fig. 3).
3. Meteorological conditions
a. Background
An analysis of the large-scale patterns in the National Centers for Environmental Protection (NCEP) sea level pressure charts during the 20–21 September ash cloud event shows 1) a weakening quasi-stationary cyclonic center located just south of Prince William Sound in the Gulf of Alaska [992 mb at 0000 UTC 21 September (not shown) to 1010 mb at 0600 UTC 22 September]; and 2) two anticyclonic centers (1024–1028 mb), one in the Bering Sea and the other on the Alaskan North Slope (Fig. 4). This pressure field produced northwest surface winds over the Katmai area during the entire resuspended ash cloud event. Examination of the 850-mb wind and temperature analysis from the NCEP Eta Model for 1200 UTC 21 September, shows the extent of cold advection and 30 kt (15.4 m s−1) winds (Fig. 5). For the flying community, large-scale winds of 20 kt (10.3 m s−1) or greater are the criteria for classifying wind as strong in mountainous areas (Carney et al. 2000).
Upper-air observations from National Weather Service (NWS) Alaskan stations, King Salmon (∼100 km upstream of the Katmai area) and Kodiak (∼150 km downstream), provide vertical profiles of the atmosphere from 0000 UTC 21 September to 0000 UTC 22 September (Fig. 6). Both King Salmon and Kodiak show a very dry surface layer; a rare condition in northwesterly flow for this time of year. The temperature profile below 850 mb for King Salmon has a superadiabatic lapse rate. A very shallow moisture layer is confined just below a capping inversion (1380–1500 m) throughout the King Salmon soundings, and winds are generally greater than 20 kt (10.3 m s−1) and consistently from the northwest. A shallow moisture layer is not present at Kodiak until 0000 UTC 22 September when it forms below a strong capping inversion at about 1500 m. Winds at Kodiak are from the west in the lower troposphere at the beginning of the resuspension event, becoming northwest by 1200 UTC 21 September.
The wind speeds observed in the soundings from King Salmon appear too weak to produce the dynamic vertical forcing necessary to lift the relic volcanic ash well into the atmosphere (>1600 m). However, there are several important factors that likely had a strong local influence on the winds that lifted the relic ash into the atmosphere: 1) The Katmai Cluster and the Valley of Ten Thousand Smokes are located in very complex terrain (Fig. 2); and 2) there are numerous mountains interspersed with valleys, channels, and gaps through the mountain barrier. In addition, the sounding at King Salmon shows that the lower troposphere is very stable, except below 850 mb where the atmosphere is super adiabatic. In this case, the synoptic-scale pressure gradient of stable flow is both nearly perpendicular to the mountain barrier with its channels and gaps and is in phase with local pressure-driven channeling. With the presence of the low-level superadiabatic lapse rate and a strong subsidence inversion to deflect the parcel downward, we have factors that can contribute to acceleration of downslope winds that would entrain and lift the relic volcanic ash from the ground and form the ash cloud observed on 21–22 September 2003 (Fig. 3). The air descending the lee slopes of the mountainous terrain in a superadiabatic atmosphere will become immediately colder than the surrounding environment and accelerate toward the earth. Impending turbulent mixing will then occur up to the inversion level.
b. Confirmation and validation of conditions
To entrain the relic volcanic ash and probably some dust from the surface of the Valley of Ten Thousand Smokes, the surface must have been dry for several days prior to the event. Sandi Fowler (National Park Service, Katmai 2003, personal communication) states that at Brooks Camp, the level of the lake lowered throughout the summer of 2003 and boats had difficulty using the docks because of the unusually low water level. There are no temperature or precipitation observations taken inside the park and preserve. An examination of National Climatic Data Center (NCDC) precipitation records for King Salmon showed precipitation above normal through the summer until mid-August at which time the departure through September was 51.6 mm (2.03 in.) below normal [58.4 mm (2.30 in.)]. These observations, along with park ranger observations, indicate that surface conditions in the region, including Valley of Ten Thousand Smokes on 20–21 September 2003 were going through a rare dry spell for the area. Simpson et al. (2002a) conducted a climatological study of Alaska and found that the Katmai area has maximum precipitation in October with September the second wettest month. This differs from King Salmon where the wettest month is August.
Three pilot reports (PIREPS) over Kodiak reported the presence of the ash cloud that reduced visibilities (Table 1). In addition, two of the pilots reported that the ash was being blown up from the ground. One pilot reported the top of the ash cloud at 5500 ft, and another at 6000 ft. This information confirms entrainment from the ground and that the top of the ash cloud was constrained by the presence of the temperature inversion.
We postulate that winds over the mountains and through the channels and gaps that surround the Valley of Ten Thousand Smokes accelerated because of the factors discussed earlier. The Alaska Marine Ferry Tustumena (call sign WNGW) at 57.9°N and 154.3°W (just offshore of the Katmai cluster in Shelikof Strait) reported a surface wind of 39 kt (21.8 m s−1), about double the speed of the wind observed upstream of the Katmai area.
A synthetic aperture radar (SAR) image of derived surface winds from the RADARSAT satellite was obtained for the coastal waters off the Katmai area at 0342 UTC 22 September (Fig. 7). See Monaldo (2000) for details of SAR-derived winds. The SAR image shows strong surface winds both north and south of Kodiak Island blowing from the Alaska Peninsula. A ship observation at 0000 UTC 22 September from the Tustumena at 58.8°N and 152.1°W confirms that the surface winds at the northern tip of Kodiak were 37 kt (19.0 m s−1) in good agreement with the SAR image. A unique feature in the SAR image is the zone of little or no wind that extends from the Katmai coast across Kodiak Island and into the Gulf of Alaska. This zone is a “shadow” of the resuspended volcanic ash cloud at that time. It appears that the ash cloud was sufficiently dense to attenuate the SAR signal from the RADARSAT satellite so that no surface winds below it could be measured. A search of the literature shows no studies on the attenuation of the SAR signal by airborne volcanic ash over water. The National Weather Service, Alaska Region, looks at SAR-derived coastal winds daily, and did not observe this pattern of attenuation either before or after this relic ash event. Unfortunately, observations from the King Salmon Weather Surveillance Radar-1988 Doppler (WSR-88D) weather radar were not useful for this event. The 1.5° elevation beam of the Doppler radar cannot “see” below about 1700 m over the Katmai Mountains.
4. Discussion
a. Winds and the superadiabatic lapse rate
Peltier and Clarke (1979) describe a severe downslope windstorm at Boulder, Colorado, in January 1972 that involved resonant amplification of the lee wave through the depth of the troposphere. Modeled surface winds closely fit the actual winds that exceeded 3 times the mean flow speed. However, the Katmai case involves the presence of a nonstandard atmosphere with a superadiabatic lapse rate below 850 mb, and a strong subsidence inversion between 740 and 850 mb, which likely rules out the possibility of wave amplification due to resonance through a deep layer. Under such conditions, Carney et al. (2000) reported downslope winds in excess of 100 kt (51.4 m s−1). Because the atmosphere is superadiabatic below 850 mb, the air is buoyant to both upward and downward perturbations.
Gap winds have been known to exceed the mean wind speed by a factor of 2 or more because of the Bernoulli effect (Carney et al. 2000). In the Katmai case, this means that wind speeds could have reached 60 kt (30.8 m s−1) to 90 kt (46.3 m s−1). Indeed, if the gap winds that flowed between the summits of Katmai, Trident, and Mageik Volcanoes began a descent through a superadiabatic atmosphere, then it could have been possible for those wind speeds to accelerate in excess of 100 kt (51.4 m s−1) before impeding on the relic, dry ash deposits.
b. Split window detection of relic volcanic dust
Operational detection of airborne volcanic ash by Volcanic Ash Advisory Centers (VAACs) often relies on satellite remote sensing using the “split window method.” This method evaluates the Advanced Very High Resolution Radiometer (AVHRR) or equivalent 11-(T4) and 12-μm (T5) brightness temperature (BT) difference (ΔT = T4 − T5). Meteorological clouds are associated with positive ΔTs (Yamanouchi et al. 1987) while volcanic plumes/clouds should have negative ΔTs (Prata 1989). Class separation becomes increasingly difficult as ΔT → 0 (Simpson et al. 2002b). Other factors [e.g., atmospheric water vapor, ice coating of the airborne ash, particle size, chemical composition, particle shape (departure from spherical), errors in satellite sensor radiometric calibration, surface emissivity and temperature, ground water and juvenile water in the magna, optically active coatings] may compromise accurate detection using the split window method (see Simpson et al. 2000, 2001; Pieri et al. 2002 and the reference contained therein). Likewise, because of its chemical composition (silicon, iron, aluminum, and calcium) and its particle size distribution (mean aerodynamic diameter 2–4 μm), both airborne Asian desert dust (e.g., Gobi Desert) and Saharan dust produce a negative T4 − T5 signal, comparable, if not stronger, to that often produced by airborne volcanic ash (Simpson et al. 2003).
A time series of AVHRR T4 − T5 scenes over the Katmai area for 20–21 September 2003 (Fig. 8) shows a strong negative T4 − T5 signal associated with the resuspension of relic volcanic ash from the 1912 Novarupta eruption. This uniquely strong resuspension event combines attributes associated with both airborne volcanic ash and airborne Asian dust. The chemical structure of the relic volcanic ash, coupled with its small particle size, contributes to its straightforward detection by the split window method. Moreover, for resuspension of the relic volcanic ash to occur, the atmosphere has to be dry and the winds must be strong, conditions also favorable for the atmospheric transport of desert dust. A very dry atmosphere likewise mitigates several of the problems associated with the accurate detection of airborne volcanic ash by the split window method.
c. Impacts on aviation
The National Weather Service Alaska Aviation Weather Unit issued a SIGMET warning to the aviation community on the presence of the ash/dust cloud at 1915 UTC 21 September 2003. The Kodiak Federal Aviation Administration (FAA) Flight Service Station received reports of some very light ash fallout in the Kodiak area. The FAA also reported that a regional airline cancelled a couple of flights into Brooks Camp until after the ash event ceased. The airlines based their decision on past experiences with ash where one of their aircraft suffered minor engine damage (scoring of the cylinders). Fortunately for this event, the Kodiak–Katmai region had mostly clear skies and the ash/dust cloud remained below 1700 m in the atmosphere. Pilots that did fly could easily see the ash/dust cloud and avoid it. If significant meteorological clouds had been present, then this could have compromised the ash/dust detection technique (Simpson et al. 2000, 2002b) and thereby increased the potential for a serious aircraft encounter with the relic volcanic ash.
5. Conclusions
The rare resuspension of relic volcanic ash and dust from the Katmai area on 20–21 September 2003 appears to be the result of a superpositioning of several factors. The resulting ash/dust cloud created a serious hazard and caused disruption to the aviation community, especially at the Kodiak airport. Although the occurrence of this event appears rare, there are signs of climate change in Alaska that may produce conditions that increase the likelihood of resuspension of volcanic ash and dust from Katmai. The occurrence of drought seems to be more frequent in the last several years (G. Weller 2003, personal communication) and the dry period observed during September 2003 may reflect this change. It is critical to aviation safety that accurate and timely warnings be issued for such future events. The forecaster is encouraged to use the following decision tree to aid in the detection and rapid alert of another event:
Has there been an extended period (several days) of little or no precipitation? Monitor the local climate data from King Salmon and check regularly with the rangers at Katmai National Park about local conditions.
Are strong (>20 kt) northwest winds forecasted upstream of Katmai and will they persist longer than 12 h? Monitor the local conditions and the numerical prediction guidance.
Is the lower level of the atmosphere superadiabatic? Monitor the King Salmon soundings.
Is a strong inversion of synoptic origin below 800 mb forecasted? Monitor the numerical prediction guidance and the King Salmon soundings.
Has a suspect cloud been generated and is the troposphere relative dry? Use the satellite AVHRR difference technique to determine if there is a strong negative signal present for airborne volcanic ash and dust.
Recognizing this type of event and nowcasting it correctly is probably the best we can do in the near future for the Katmai region. This area is very remote with poor spatial and temporal sampling. In addition, a survey needs to be conducted to determine the presence of other ash fields in Alaska where resuspension might occur.
Acknowledgments
Support for JJS was provided by the Scripps Institution of Oceanography. Ben Tsou, James Biskey, and Jared Berg assisted with manuscript preparation. Support for DH and GLH was provided by the National Weather Service. Data and satellite imagery were provided from the National Weather Service, Alaska Region database. The authors wish to thank J. Fierstein and W. Hildreth for use of their graphics from their 2001 report (see the references).
REFERENCES
Bates, M. A., and Beggs J. , 1997: Volcanic emissions and the risk to public health. N.Z. Public Health Rep, 4 , 81–83.
Blong, R. J., 1984: Volcanic Hazards: A Sourcebook on the Effects of Eruptions. Academic Press, 424 pp.
Carney, T. Q., Bedard A. J. Jr., Brown J. M. , McGinley J. , Lindholm T. , and Kraus M. J. , 2000: Hazardous mountain winds and their visual indicators. Federal Aviation Administration Tech. Doc. AC 00-57, 91 pp.
Casadevall, T. J., 1994: The 1989–90 eruption of Redoubt volcano. Alaska: Impacts on aircraft operations. J. Volcanol. Geotherm. Res, 62 , 301–316.
Chadwick, O. A., Derry L. A. , Vitousek P. M. , Huebert B. J. , and Hedin L. O. , 1999: Changing sources of minerals during four million years of ecosystem development. Nature, 397 , 491–497.
Fierstein, J., and Hildreth W. , 2001: Preliminary volcano-hazard assessment for the Katmai volcanic cluster, Alaska. U.S. Geological Survey Rep. 00-489, 50 pp.
Gillette, D., 1978: A wind tunnel simulation of the erosion of soil: Effects of soil texture, sandblasting, wind speed, and soil consolidation on the dust production. Atmos. Environ, 12 , 1735–1743.
Hildreth, W., 1983: The compositionally zoned eruption of 1912 in the Valley of the Ten Thousand Smokes, Katmai National Park, Alaska. J. Volcanol. Geotherm. Res, 18 , 1–56.
Hildreth, W., 1987: New perspectives on the eruption of 1912 in the Valley of Ten Thousand Smokes, Katmai National Park, Alaska. Bull. Volcanol, 49 , 680–693.
Hufford, G. L., Salinas L. S. , Simpson J. J. , Barske E. G. , and Pieri D. , 2000: Operational implications of airborne volcanic ash. Bull. Amer. Meteor. Soc, 81 , 745–755.
Kobayashi, T., Ishiharta K. , and Hirabayashi J. , 1988: A guide book for Sakurajima Volcano. Kagoshima International Conference on Volcanoes, Kagoshima, Japan, 88 pp.
Miller, T. P., and Casadevall T. J. , 1999: Volcanic ash hazards to aviation. Encyclopedia of Volcanoes, H. Sigurdsson et al., Eds., Academic Press, 925–930.
Monaldo, F., 2000: The Alaska SAR demonstration and near-realtime synthetic aperture radar winds. Johns Hopkins APL Tech. Dig, 21 , 75–79.
Myhre, G., and Stordal F. , 2001: Global sensitivity experiments of the radiative forcing due to mineral aerosols. J. Geophys. Res, 106 , 18193–18204.
Peltier, W. R., and Clark T. L. , 1979: The evolution and stability of finite-amplitude mountain waves. Part II: Surface wave drag and severe downslope windstorms. J. Atmos. Sci, 36 , 1498–1529.
Pieri, D., Ma C. , Simpson J. J. , Hufford G. , Grindle T. , and Grove C. , 2002: Analyses of in-situ airborne volcanic ash from the February 2000 eruption of Hekla Volcano, Iceland. Geophys. Res. Lett.,29, 1772, doi:10.1029/2001GL013688.
Pollack, J. B., Toon O. B. , Sagan C. , Summers A. , Baldwin B. , and Van Camp W. , 1976: Volcanic explosions and climate change: A theoretical assessment. J. Geophys. Res, 81 , 1071–1083.
Prata, A. J., 1989: Infrared radiative transfer calculations for volcanic ash clouds. Geophys. Res. Lett, 16 , 1293–1296.
Reihle, J. R., Meyer C. E. , Ager T. A. , Kaufman D. S. , and Ackerman R. E. , 1987: The Aniakchak tephra deposit, a late Holocene marker horizon in western Alaska. Geologic Studies in Alaska by the U.S. Geological Survey during 1986, U.S. Geological Survey Circular 0998, T. D. Hamilton and J. P. Galloway, Eds., USGS, 19–22.
Schwartz, J. G., Norris G. , Larson T. , Sheppard L. , Claiborne C. , and Koenig J. , 1999: Episodes of high coarse particle concentrations are not associated with increased mortality. Environ. Health Perspect, 107 , 339–342.
Simpson, J. J., Hufford G. L. , Pieri D. C. , and Berg J. , 2000: Failures in detecting volcanic ash from satellite-based technique. Remote Sens. Environ, 72 , 191–217.
Simpson, J. J., Hufford G. L. , Pieri D. C. , and Berg J. , 2001: Response to “Comments on ‘Failures in detecting volcanic ash from a satellite-based technique.’”. Remote Sens. Environ, 78 , 327–357.
Simpson, J. J., Hufford G. L. , Fleming M. , Berg J. S. , and Ashton J. B. , 2002a: Long-term climate patterns in Alaskan surface temperature and precipitation and their biological consequences. IEEE Trans. Geosci. Remote Sens, 40 , 1164–1184.
Simpson, J. J., Hufford G. L. , Pieri D. , Servranchx R. , Berg J. S. , and Bauer C. , 2002b: The February 2001 eruption of Mount Cleveland, Alaska: Case study of an aviation hazard. Wea. Forecasting, 17 , 691–704.
Simpson, J. J., Hufford G. L. , Servranckx R. , Berg J. , and Pieri D. , 2003: Airborne Asian dust: Case study of long-range transport and implications for the detection of volcanic ash. Wea. Forecasting, 18 , 121–141.
Sparks, R. S. J., Bursik M. I. , Carey S. N. , Gilbert J. S. , Glaze L. S. , Sigurdsson H. , and Woods A. W. , 1997: Volcanic Plumes. John Wiley and Sons, 574 pp.
Warrick, R. A., Anderson J. , Downing T. , Lyons J. , Ressler J. , Warrick M. , and Warrick T. , 1981: Four communities under ash— after Mount St. Helens. Program on Technology, Environment and Man, Monogr., No. 34, Institute of Behavior Sciences, University of Colorado, 1–150.
Yamanouchi, T., Suzuki K. , and Kawaguchi S. , 1987: Detection of clouds in Antarctica from infrared multispectral data of AVHRR. J. Meteor. Soc. Japan, 65 , 949–961.
(top) Location of Katmai volcanic cluster on the Alaskan Peninsula. (bottom) Ash fall (red contours) from the 1912 Novarupta eruption. Figure from Fierstein and Hildreth (2000). (Reproduced here courtesy of J. Fierstein)
Citation: Weather and Forecasting 19, 5; 10.1175/1520-0434(2004)019<0829:RORVAA>2.0.CO;2
Valley of Ten Thousand Smokes ash flow sheet (yellow), which erupted from Novarupta in Jun 1912 overlaid on local topography. (Figure provided courtesy of U.S. Geological Survey)
Citation: Weather and Forecasting 19, 5; 10.1175/1520-0434(2004)019<0829:RORVAA>2.0.CO;2
Resuspended ash cloud extending from Katmai over Kodiak Island into the Gulf of Alaska as seen in MODIS 250-m visible data at 2123:05 UTC 21 Sep 2003
Citation: Weather and Forecasting 19, 5; 10.1175/1520-0434(2004)019<0829:RORVAA>2.0.CO;2
NCEP North Pacific surface analyses for dates and time indicated. Highlighted wedge shows area of interest
Citation: Weather and Forecasting 19, 5; 10.1175/1520-0434(2004)019<0829:RORVAA>2.0.CO;2
NCEP Eta Model 850-mb wind and temperature forecast chart for Alaska area 0000 UTC 22 Sep 2003
Citation: Weather and Forecasting 19, 5; 10.1175/1520-0434(2004)019<0829:RORVAA>2.0.CO;2
Simplified skew T–log p plots at 0000 UTC 21 Sep, 1200 UTC 21 Sep, and 0000 UTC 22 Sep 2003 for King Salmon and Kodiak, Alaska, respectively. See appendix in Simpson et al. (2000) for detailed discussion of skew T– log p plot
Citation: Weather and Forecasting 19, 5; 10.1175/1520-0434(2004)019<0829:RORVAA>2.0.CO;2
Derived synthetic aperture radar surface wind image for Alaska Peninsula–Kodiak coastal area for 0342 UTC 22 Sep 2003
Citation: Weather and Forecasting 19, 5; 10.1175/1520-0434(2004)019<0829:RORVAA>2.0.CO;2
Time series of AVHRR T4 − T5 images. Katmai source region and Kodiak Island are shown (green). (a) The airborne relic volcanic ash appears over the source region (Katmai) as black and in (b)–(m) is advected south-southeast over Kodiak Island and into air routes
Citation: Weather and Forecasting 19, 5; 10.1175/1520-0434(2004)019<0829:RORVAA>2.0.CO;2
Airborne volcanic ash sightings by general aviation pilots of Katmai event
