From Neal and others (2009): "As fall arrived in Alaska, a phreatic eruption from a volcano not considered active in the Holocene surprised AVO and residents of south-central Alaska. Late on Sunday, September 17, AVO received several citizen telephone reports of a dark plume, fed by what appeared to be two sources, rising from the area near Cape Douglas in lower Cook Inlet. Satellite analysis indicated the source of the plumes to be roughly between closely spaced Douglas and Fourpeaked volcanoes. he drifting cloud produced by these rising plumes persisted throughout the night, reaching a minimum altitude of 20,000 ft (6,100 m based on radar data; Dave Schneider, USGS, written commun., 2006), but drifting only 20 km (12 mi) downwind in an unusually calm and clear atmosphere over south-central Alaska.
The next day, AVO issued an information release mentioning Douglas and Fourpeaked volcanoes, but stated that the exact nature and source of the plumes remained unclear. Neither volcano was monitored seismically, and data from the nearest stations at Katmai and Augustine initially showed nothing dramatic. Deteriorating weather conditions precluded visual observations, but an overflight was scheduled for the first clear window. On the afternoon of September 18, NWS relayed a pilot report of a strong sulfur odor 300 km (190 mi) northwest of Fourpeaked. This was an area likely to be impacted by drift of the September 17 cloud according to PUFF, a volcanic ash tracking model used by AVO (Searcy and others, 1998). Pilots in the Lake Iliamna area on the morning of September 18 also reported a strong sulfur odor, a locality also consistent with likely cloud drift on the 17th.
On September 20, AVO received NWS radar and other satellite data indicating particulate and sulfur-bearing properties in the cloud of September 17, confirming a volcanic source for the event. Based on this, AVO announced Level of Concern Color Code YELLOW for both Douglas and Fourpeaked volcanoes because the source was still uncertain. NEXRAD Doppler radar images from King Salmon showed a particulate cloud in the atmosphere from at least 12:00 p.m. to 9:45 p.m. ADT on September 17. The cloud appeared most energetic (dense) in the first 2 hours (Dave Schneider, USGS written commun., 2006). The bulk of the cloud remained over the point of origin for much of this time, reflecting very calm atmospheric conditions on that day. In addition, Ozone Monitoring Equipment onboard National Aeronautics and Space Administration's (NASA) Aura satellite indicated a pod of SO2 gas in the general vicinity of the eruption plume at 3:00 p.m. ADT. Subsequently, AVO received reports of a very fine-ash dusting at Nonvianuk Lake outlet (110 km or 70 mi west northwest of Fourpeaked) and near Homer (150 km or 90 mi northeast of Fourpeaked). The Nonvianuk report stated, 'We have a heavy sulfuric smell in the air and ash filling the air. The wind here is blowing from the ESE' No samples were retrievable from either location for analysis, but the Nonvianuk report is consistent with PUFF trajectories for a cloud reaching about 20,000 ft (6,100 m) on the afternoon of September 17. Interestingly, multi-spectral satellite images showed no ash signature in the plume on September 17, but AVO analysts did note a thermal anomaly in the vicinity of Fourpeaked (J. Dehn, University of Alaska Fairbanks Geophysical Institute (UAFGI), written commun., 2006). Following the initial event on September 17th, no further particulate clouds were imaged by radar or other means.
An AVO overflight on September 20 confirmed Fourpeaked as the source of volcanic activity. Despite cloud cover of the actual summit, observers in a fixed-wing aircraft circled several distinct vapor clouds rising through the cloud deck above Fourpeaked. Visible patches of discolored snow and ice, especially north and west of the Fourpeaked summit, suggested ash fallout. An AVO helicopter crew later on the same day photographed a lobate, dark debris-flow tongue that had emerged from glacial ice about 3,000 ft (900 m) below the cloud-covered Fourpeaked summit; patchy areas of grey ash on the glacial ice around the summit and to the northeast also were noted. Clouds and fuel limitations prevented any further exploration of the area.
Aerial inspection of the Fourpeaked area on September 23 finally revealed the source of the September 17 plume and continuing fumarolic emissions. Water vapor and volcanic gas billowed from a dramatic, linear series of pits or vents extending about 1,250 m (4,100 ft) across the heavily crevassed and disrupted glacial-ice cover on the north side of the summit region. Multiple sources of vapor from these vents explain the apparent double plume seen in the photograph of September 17. A light dusting of dark material surrounded some of the open pits and several elongate dark stripes leading down slope from at least one vent probably represented remobilized fragmental ejecta mixed with melted ice and snow (or plume condensate). A subtle series of snow-mantled rills leading down slope in the same area indicated several episodes of surface debris or water flow. Stormy weather between September 17-23, produced new snowfall indicating that the ejecta collars and small debris flows observed on September 23 had occurred after the initial event.
AVO scientists visited the area by helicopter on September 24 and 25. The vent area consisted of as many as nine discrete craters or pits, and between three and five were venting steam and volcanic gas at any one time. The vents occurred along a line trending north from the summit basin, inferred to be a glacial cirque, obliquely down-slope across a northwest trending, ice-covered ridge. The bottoms of non-steaming pits were covered with blocks of debris-mantled ice. The upper craters within the Fourpeaked summit cirque had coalesced creating a heavily disrupted ice zone. Most craters were surrounded by fine (?) debris collars that did not extend very far from their rims; there was no evidence of additional, significant ash emission since September 17. FLIR imaging of the pits indicated elevated temperatures as high as 75C (167F); however, these values are minimums due to steam obscuration. A strong sulfur odor was noted downwind of the vents as far as 50 km (30 mi).
The glacial outburst associated with the September 17 event originated from beneath a chaotic ice jumble on the unnamed north-trending glacier at an elevation of about 5,000 ft. This flood apparently scoured a steep-walled canyon more than 100 m (330 ft) deep in places. Blocks of ice up to 5 m (16 ft) or more across had been rafted in a mixture of water and fine-grained to cobble-boulder sized, heterolithologic volcanic debris at least 6 km (4 mi) down slope, where material spilled off the front of the glacier ice and traveled an unknown distance into the Douglas River drainage. Levees of ice, sediment, and rock clasts as much as 10-15 m (33-50 ft) high marked the margins of the deposit. Where traced into the Douglas River drainage about 3-4 km (2-2.5 mi) from the glacier snout, the deposit was thin (about 2 cm or 0.8 in) and fine grained with a maximum clast size of about 1 cm (0.4 in). The field crew flew the length of the Douglas River to Cook Inlet and noted evidence of a flooding event represented by recently emplaced fine-grained gray sediment on beaches and river banks. On the day of observation, there was no evidence of continuing discharge of meltwater or debris down this newly carved drainage system.
AVO scientists collected samples of the fine, gray ash-fall deposit from the September 17 event. Deposits were most impressive west of the vent area where they were estimated to be 1-2 mm (less than 0.1 in) thick. Other flanks of the volcano received a mere dusting (<1 mm); based on these few observations and the outlier reports of extremely light ash fall noted at two distant locations, the fall deposit likely represents well under 1 million cubic m of material. Preliminary petrographic analyses indicate the tephra consists of hydrothermally altered volcanic rock and crystal fragments, notably pyrite, and other accessory minerals (J. Larsen, UAFGI, written commun., 2006).
AVO geologists in the field on September 24-25 also reported loud rumbling sounds associated with sudden bursts of water from the ice-bedrock contact at an elevation of approximately 4,000 ft on the northwest flank of Fourpeaked; these flows lasted several seconds and then disappeared. An overflight of the cliff below these outbursts did not reveal anything unusual, and any relationship between these periodic outbursts of water with the events of September 17 is uncertain.
AVO geologists made further ground-based observations of the deposits and features related to this unrest in mid-October. Close helicopter passes of the vent area on October 14 revealed that several of the original craters had coalesced and the rims of the sheer-walled pits had retreated, enlarging most by perhaps tens of meters. The pit rims were blanketed by fresh snow indicating no additional ash emission of significance since the explosion on September 17. However, intermittent, vigorous fumarolic activity capable of entraining a small amount of locally derived material may have continued and gone undetected between overflights. Field observers saw no sign of large ballistics littering the surface, so any ongoing phreatic emissions were not very energetic. Yellow-stained (most likely sulfur) snow surrounded the upper crater.
Deposits related to the outburst flood into the Douglas River were examined more closely and consisted of gray, soggy, water-saturated, sulfur-smelling silty material containing cobble-sized clasts of dense, altered, volcanic rock and pyrite (and possibly marcasite). Preliminary results from x-ray diffraction and x-ray fluorescence analyses of a non-pyritic material indicate that the volcanic rock composition primarily is dacite, and that the fine fraction also contains minor gypsum and minor smectite (K. Bull, ADGGS, written commun., 2007). Interestingly, near the snout of the glacier impacted by this outburst flood, AVO geologists noted multiple layers of similar, sulfurous, heterolithologic material exposed in the ice stratigraphy, and postulated that these may represent prior (possibly historical in age) debris-flow events captured in the ice.
Eight airborne gas measurements were obtained between September 23 and November 18, 2006. Sulfur dioxide output was steady and high for a non-erupting volcano ranging between 820-2,940 ton/d (Doukas and McGee, 2007). For the same period, CO2 flux was between 340-834 ton/d. In contrast to Augustine Volcano (McGee and others, 2006), H2S output from Fourpeaked remained quite high, between 70-140 ton/d, likely reflecting the dominance of a wet hydrothermal system at this ice-clad volcano. In addition to these onsite, airborne measurements, Ozone Mapping Instrument (OMI) sensors occasionally detected SO2 clouds in the area. Beginning in October, low sun angles prevented good results and AVO stopped receiving reports from the OMI satellite team at the University of Maryland (D. Schneider, USGS, oral commun., 2006).
Seismic activity as recorded on the three new stations installed following the event on September 17 remained relatively low through the end of the year, typically with only a few volcanic earthquakes captured on most days. These three stations augmented coverage by regional seismic station CCDN about 17 km (~11 mi) northeast of Fourpeaked. On October 3, a swarm consisting of tens to hundreds of very small, non-locatable earthquakes occurred in the vicinity of Fourpeaked. A second swarm on November 5-6 occurred within the new Fourpeaked subnet and 75 events were located. Seismicity remained elevated with occasional small swarms of activity (10 located events per day or less) through the end of the year. Small explosion signals also began to be recorded in the spring; these signals may have reflected transient increases in fumarolic emission.
Further analysis of Katmai area seismic stations during the time period of the eruption cloud and opening of vents in the ice revealed a small swarm of earthquakes between 11:48 a.m. and 3:50 p.m. ADT on September 17 (M. West, UAFGI, oral commun., 2006) coincident in time with the onset and development of the plume seen in radar images. The University of Alaska Fairbanks infrasound array also detected a signal at about 20:50 UTC on September 17, likely an explosion source, at a time and location consistent with the plume sighting (S.R. McNutt, UAFGI, oral commun., 2006).
AVO concluded that the unrest at Fourpeaked volcano most likely involved the presence of new magma at fairly shallow (less than a few kilometers) levels, accounting for the seismicity and degassing, and providing a heat and gas source for a phreatic explosion, vigorous phreatic emission of gas and fine particulates, and a glacial outburst of meltwater, glacial ice, and hydrothermally altered debris on September 17.
A shallow, degassing intrusion of fresh magma also would account for the ongoing gas emissions (K. McGee, USGS, oral commun., 2006). The lack of a seismic network at Fourpeaked precludes exact determination of the onset of seismicity that may have been associated with an intrusion. However, the swarm detected on the Katmai network on September 17 likely captured the most vigorous phase of the event, including the onset of phreatic eruption.
Through the remainder of 2006 and into 2007, a variably robust plume of vapor and volcanic gas discharged from the linear chain of pits in the ice. Overflights into mid-November documented minor changes in the pit morphologies, primarily related to coalescence and widening. No further ash emissions of significance were noted, although an increasingly visible coating of a yellow, likely sulfurous deposit stained snow and ice cover around the Fourpeaked summit."

For detailed observations and photographs of this event, pelase see: Neal, C.A., McGimsey, R.G., Dixon, J.P., Manevich, Alexander, and Rybin, Alexander, 2009, 2006 Volcanic activity in Alaska, Kamchatka, and the Kurile Islands: Summary of events and response of the Alaska Volcano Observatory: U.S. Geological Survey Scientific Investigations Report 2008-5214, 102 p., available at pubs.usgs.gov/sir/2008/5214/ .

McGimsey and others (2011) report that throughout the first half of 2007, seismicity, steaming, and gas emissions decreased at Fourpeaked following the 2006 eruption. McGimsey and others (2014) continue the chronology of this event: "Steam and gas emissions were frequently observed in web camera images and during routine gas measurement flights through the first half of 2008, and were visible to an AVO field crew on Augustine Volcano in July 2008 (AVO internal log entries). Activity continued to decrease into, and through 2009.
"Steam plumes were visible in the web camera on February 8, 2009, and during gas measurement flights on June 6, 2009, and November 2, 2009; during the November overflight, no gas was detected (M. Doukas, U.S. Geological Survey, written commun., 2012). As the phreatic activity diminished, the melt holes begam filling with snow and fumarolic activity was observed only from a single vent.
"The seismic and infrasound networks, and the web cam, were serviced in the summer of 2008, and a year later, as the batteries drained, the instruments stopped recording data. On November 18, 2009, prompted by the network outage and inability to assess the level of seismic activity, AVO issued a Volcanic Activity Notice and an Information Release downgrading Fourpeaked from Aviation Color Code GREEN and Volcano Alert Level Normal to UNASSIGNED, thus removing it from the list of seismically monitored volcanoes."