This image shows on the left side 4 locations of GPS stations on Okmok volcano. The blue and green arrows show the long-term motion for each site in the horizontal (blue) and vertical (green) direction. For OKCE, inside the caldera, we show the deformation time series from 2004-2020 in the right panel. The three graphs show east (top), north (middle) and vertical (bottom) position changes relative to a far away GPS station (AV09). We can see that each component shows a large step – that was the 2008 eruption when OKCE moved about 150 cm east, 10 cm south and 200 cm down! Before and after the large step we see motion in the opposite direction for each of these directions – suggesting that the magma chamber is filling with magma. Figure: Ronni Grapenthin, UAFGI/AVO.
From Neal and others (2011): "Okmok Volcano, a 10-km (6.2-mi) diameter Holocene caldera system in the central Aleutians [fig. 14; imageid 13283], began a protracted explosive eruption on July 12. The opening explosions consumed a portion of intracaldera Cone D within the east-central sector of the caldera, reaming several new craters into the caldera floor. Eruptive activity began only a few hours after a subtle increase in the rate of earthquakes followed by a short swarm sequence, both noted only in retrospect. Over the next 5 weeks, several hundred million cubic meters of tephra and lahar deposits blanketed much of northeast Umnak Island. Within the caldera, nearly continuous hydrovolcanic explosions accumulated many tens of meters of wet, mostly fine-grained tephra. Explosive activity completely disrupted existing groundwater and standing water bodies within the caldera, formed new lakes, and constructed a new tephra cone about 100-200 m (330-660 ft) high. This eruption was the first dominantly phreatomagmatic volcanic event in the United States since the Ukinrek Maars eruption in 1977. The following summary is taken largely from Larsen and others (2009).
"Other than the seismicity recognized in hindsight, AVO noted no clear signs of precursory unrest at Okmok prior to the eruption onset. In fact, during the 2 months prior to July 12, Okmok produced only three earthquakes and no tremor episodes (intermittent tremor episodes had been recorded since the seismic network was installed in 2003 and had continued into mid-2005; Reyes and McNutt, 2008). Campaign and continuous GPS data had recorded nearly continuous inflation from 1997 to 2005, quiescence between 2005 and 2007, and notable (but not unprecedented) inflation in early 2008. Pre-eruptive displacements measured by GPS and InSAR indicated inflation of a pressure source about 2.6-3.2 km (1.6-2.0 mi) below sea level and underneath the approximate center of the caldera (Lu and others, 2005; Fournier and others, 2009).
"AVO was first notified of the eruption by the USCG who had been contacted by the caretaker of Bering Pacific Ranch at Fort Glenn [Fort Glenn is a former U.S. Army base that now houses a cattle ranch operation about 10 km (6.2 mi) southeast of the caldera rim]. The caretaker and his family reported wet volcanic ash falling as they were evacuating the island first by helicopter and later by boat. An immediate check of Okmok seismicity by the AVO duty scientist confirmed that an eruption was in progress. AVO issued a notice of the eruption declaring Aviation Color Code RED and Volcano Alert Level WARNING and commenced 24-hour operations to respond to the event.
"In addition to 24-hour staffing of the operations room in Anchorage, AVO mounted two helicopter-supported field responses to the eruption. The first operated from Unalaska between July 29 and August 5 and the second was based at Fort Glenn over a week in mid-September about 3 weeks after the eruption had ended. AVO received photographs taken by Fort Glenn ranch caretaker Lonnie Kennedy on several occasions. These photographs along with images from commercial and USCG aircraft, satellite imagery, and mariner accounts provided critical visual documentation of the eruption through time.
"The most energetic phase of the eruption occurred over the first 10 hours of activity on July 12. The first satellite images of the ash plume were geostationary operational environmental satellite (GOES) images starting at 20:00 UTC on July 12. By 22:12 UTC, the ash cloud extended east over much of Unalaska Island [fig. 15; imageid 14279]. Both geometric image analysis of GOES and comparison of cloud motion with the PUFF ash dispersion model indicated a maximum initial column height of approximately 16 km (52,000 ft) ASL. Photographs of the eruption column by crews of a USCG C-130 and an Alaska Airlines jet taken about 5-6 hours into the eruption show a vertical, gray, ash-rich column rising into meteorological cloud layers; the top of the eruption plume appeared white and was estimated visually to be 30,000-35,000 ft (9,100 m-10,700 m) ASL. The ground was obscured and the aircraft too distant to make out any detail at the base of the eruption column.
"The opening explosions and heavy tephra fall destroyed or disabled several AVO seismometers and continuously recording GPS instruments, however the remaining network density was sufficient to track the eruption. From July 13 through the end of the month, seismicity varied but remained well below the intensity of the opening eruption sequence on July 12. Eruption columns and clouds seen in satellite imagery and by passing aircraft varied significantly in altitude although these changes were not often in phase with recorded seismic amplitude (Larsen and others, 2009; table 5 in Neal and others 2011). Characteristics of most eruption clouds implicated the continuous involvement of water in the eruption process. On July 13, a Moderate Resolution Imaging Spectroradiometer (MODIS) satellite image showed two plumes -- one dark and ash rich and the other light in color and inferred to be very rich in water vapor -- emanating from the eastern portion of the caldera floor. Between July 13 and 21, photographs from Alaska Airlines and aerial observations by AVO staff from a USCG plane showed a light-colored plume with a wide base and multiple potential sources of ash explosions [fig. 16; imageid 14436].
"Evacuated Fort Glenn ranch caretaker Lonnie Kennedy returned to the island on July 23 and, over the next several days, photographed eruption impacts and continuing ash emission and ash fall in the vicinity of the caldera. Kennedy documented ongoing muddy water flow across the lowlands surrounding the ranch; lahars in several drainages north of the ranch had been sufficiently energetic to destroy pre-existing wooden bridges and culverts and cause severe bank erosion. Dramatic new deltas had formed at the mouths of a number of creeks draining the northeast and southeast flanks of Okmok [fig. 17; imageid 15446]. The exact timing of lahar activity during the first days of the Okmok eruption is uncertain. It is also not clear if lahar formation was due to rain-remobilization of tephra, syn-eruptive condensation of water vapor entrained in the eruption cloud (W. Scott, USGS, written commun., 2008), dewatering of wet tephra fall, melting of snowpack, or some combination of these or other processes. Overbank deposits and the presence of large boulders atop the surface of the 2008 lahar fan at the mouth of Crater Creek (Crater Creek drains the caldera northwestward into the Bering Sea) suggest temporarily high discharge rates possibly caused by a sudden release of water from the caldera early in the eruption.
"Kennedy's aerial photographs of the caldera from August 1 show the upper Crater containing an active, braided channel of muddy water indicating some drainage from the caldera. The terrain immediately east of the caldera was thickly covered in light brown to gray tephra. Deep rills and dendritic drainage networks existed on most surfaces; at higher elevations, the pre-eruption snowpack was visible beneath the 2008 debris, and water flowed from the base of the snowpack in many places. A partially clear view into the caldera on August 1 showed ash and water-vapor-rich clouds boiling from at least two point sources on the northwest flank of Cone D and just to the west of Cone D [fig. 18; imageid 15666]. Dark collars of debris enclosed each locus of venting. The pre-eruption lake northeast of Cone D had been significantly modified: standing water covered a much smaller area and what had been the lake was now a surface of tephra and scattered ponds.
"The first AVO crew on scene in late July was unable to land near the caldera due to active ash emission. They focused on documenting the extent and character of ash fall and lahar deposits outside the caldera, taking observations of the ongoing eruption, repairing a key data repeater site on Makushin Volcano on Unalaska Island, and collecting samples and eye-witness accounts. They obtained some close-up views into the caldera and also distant views of the eruption column from the Fort Glenn ranch. On August 2 and 3, the eruption column had increased in intensity, height, and ash content [figs. 19; imageid 14700 and 20; imageid 14718]. This change was coincident with an increase in amplitude of seismic tremor. AVO crew observations, photographs, and film footage during this time of heightened activity suggest a migration of the location of active venting on the caldera floor over the span of minutes. On overflights near the eruption site, the field crew observed a ground-hugging cloud of tan-colored ash covering the caldera floor and obscuring views of the immediate area. In glimpses of the caldera floor near the site of the pre-eruption lake near Cone D, they noted chaotic, disrupted terrain and channels of flowing water.
"Due to the renewed intensity of the eruption, AVO elevated the Aviation Color Code and Volcano Alert Level again to RED/WARNING early on the morning of August 2 [see table 5 in original text], and the caretaker and family at the Fort Glenn ranch decided to evacuate for the second time. During the last days of July and the first days of August, prevailing winds shifted to be out of the northeast.
"Subsequently, over the first 2 weeks of August, eruption intensity and cloud height generally decreased and ash emission ceased altogether by August 19. A USGS helicopter crew working in the Aleutians entered the caldera on August 13 during the waning phase of eruption and photographed a single active vent enclosed within a steep-sided tephra cone [figs. 21; imageid 15120 and 22; imageid 15119]. Dark ash boiled out of the tephra cone surrounded by a collar of white water vapor; winds were from the northwest sending the ash and water vapor cloud over the summit of Cone D and the caldera rim. A significant lake was now present near the site of the pre-eruption lake and the landscape was completely covered with dark gray ash. The surface north of Cone D was pocked with craters several meters to several tens of meters in diameter. A series of scallop-margined basins and coalesced craters, some hosting standing water, extended in a line west of Cone D.
"On August 23, about 1 week after the end of the eruption, Lonnie Kennedy again photographed the eastern caldera from the air. Although Crater Creek was open and flowing just inside the caldera, a through-going surface connection between Crater Creek and the growing lakes had not been established. Wind re-suspended a tan-colored ash in the vicinity of the largest of the new vents.
"AVO's week-long September expedition to Okmok gathered reconnaissance information about the eruption deposits and impacts, repaired some seismic and GPS instruments, and deployed additional GPS recording stations (some of which were retrieved in the summer of 2009 by the Plate Boundary Observatory field crew). Most tephra sections excavated within and outside the caldera exposed planar to slightly wavy-bedded, fine-grained fall and surge deposits. Northeast of the vent region where tephra accumulation was thickest, the basal unit from the July 12 opening phase was a coarse ash-lapilli fall deposit; individual clasts were coated with a very fine ash [fig. 23; imageid 31902]. Evidence of significant water interaction throughout the 2008 tephra sequence includes (1) very high porosity of individual beds reflecting post-emplacement de-watering; (2) plastering texture on perpendicular surfaces facing the vent; (3) abundant accretionary lapilli; (4) overall fine-grained nature of the deposit. Outside the caldera, excavated sections contained mostly fall deposits with thin and discontinuous aeolian horizons; no clear evidence for energetic, far-traveled, extra-caldera surges was noted.
"Field observations in September combined with analysis of photographs and satellite images indicate that the eruption occurred from a series of vents that opened during the first 2 weeks of the eruption. These vents extended in a roughly linear zone about 2 km (1.2 mi) long across the caldera floor [figs. 24; imageid 15480 and 25; iin original text]. One crater formed next to, and eventually captured and drained, the pre-existing lake northeast of Cone D. A tephra cone ('New Cone') had been constructed atop the longest-lived 2008 vent [fig. 26; imageid 15476]. By mid-September, the explosion and collapse craters to the west of Cone D had filled with water and formed a new lake ('New Lake') about 0.6 km2 (0.2 mi2) in area [figs. 24; imageid 15480 and 25; in original text].
"Eruptive products from the 2008 sequence are basaltic andesite in composition, slightly more silicic than the range of Okmok chemistry represented by other post-caldera (last about 2,000 years BP) eruptions (Larsen and others, 2009). The fine-grained nature of most 2008 tephra and the lack of an effusive phase pose a challenge to understanding changes in eruption chemistry with time. Coarse juvenile lapilli from the opening phase of the 2008 eruption were collected on the caldera rim just above the outlet of Crater Creek in a notch through the caldera wall informally called 'The Gates' [fig. 14; imageid 13283]; this location was near the main northeast axis of deposition. Clast types in the July 12 opening tephra range from dense to vesicular scoria and pumice, and crystal fragments (Mariah Tilman, AVO/UAFGI, written commun., 2008). Some dense clasts may represent accidental lithics incorporated during the opening explosions through the caldera floor and lava flows of Cone D. Larger frothy pumiceous clasts were found along the shoreline of one of the shallow lakes within the caldera; because they are not in place, they cannot be confidently assigned to the 2008 event, but they were fresh-looking, fragile, and most likely represent material erupted on July 12. These clasts are up to 5-10 cm (2-4 in.) across and are light to dark brown, vesicular, scoriaceous pumice with irregular, fractured chill rinds.
"Following cessation of eruption in mid-August, seismicity remained relatively low with occasional bursts of higher amplitude tremor. AVO downgraded the Aviation Color Code and Volcano Alert Level to YELLOW/ADVISORY on August 27 [see table in original text] after a week with no ash clouds discerned in satellite images. Occasional thermal anomalies were visible by satellite and are likely attributable to the still-warm vent area, fumaroles, and/or lakes within the caldera. By mid-November, after 3 months of steadily decreasing seismicity, AVO changed the Aviation Color Code to GREEN and Volcanic Alert Level to NORMAL. In the accompanying remarks, AVO indicated that despite the cessation of eruptive activity, hazardous conditions persisted. Unstable, muddy surfaces and slopes of new volcanic debris within the caldera could collapse at any time. New and rapidly changing lakes, ponds, and multiple steep-walled craters through the new tephra blanket would present a hazard to anyone visiting the caldera. Magmatic gases and areas of high temperature could persist around the new tephra cone. All drainages leading downslope from the rim of the caldera are susceptible to remobilization of ash and other loose debris during heavy rains and spring melt. The Crater Creek drainage on the north-northeast flank of Okmok was considered especially vulnerable to sudden flooding events if tephra dams within the caldera were to fail suddenly and release impounded water.
"AVO maintained 24-hour staffing from July 12 through August 28. Over the course of the eruption, AVO issued 17 Volcanic Activity Notices and two Information Statements.
"Eruption impacts were modest except for the significant disruption to the ranch caretaker family on Umnak Island who evacuated twice from the island including the first time under great duress during the most energetic phase at the start of the eruption. Primitive roads on the east and south flanks of the caldera were cut by lahars and water floods and rendered at least temporarily impassable. Ash accumulation suppressed grass growth that resulted in diminished over-winter feed for the livestock that roam the island. According to the Kennedy family, it is possible that an increased number of cattle perished in the winter of 2008-09 because of this (Susan Kennedy, written commun., 2009). The island also hosts a large number of caribou, although we are unaware of any systematic population counts to gauge the impacts of the eruption. Offshore Umnak Island, volcanic sediment delivered to the coastline built significant new lahar deltas and fishermen reported dramatic changes to bottom conditions in the weeks after the eruption (Lonnie Kennedy, oral commun., 2008) Several boats received minor to trace ash fall with no ill-effects reported other than a single collapsed air filter (Dustin Dickerson, oral commun., 2008). Out of concern for the effects of ash fall, the U.S.Coast Guard closed Umnak Pass for several days in the immediate aftermath of July 12.
"Over the course of the eruption, trace amounts of ash fell on several occasions in Unalaska-Dutch Harbor 120 km (75 mi) northeast of the volcano. The airport closed briefly to allow for clean-up of the runway and taxi ways. Cannery workers and other residents were concerned about impacts of the ash fall, however slight, on their health, and AVO worked with local health care providers and cannery management, and with the Alaska Department of Environmental Conservation Air Quality Division to issue health related information. AVO and UAFGI staff installed a 3 stage DRUM impactor air sampler in Unalaska to sample volcanic particulate from the ash fall events (Peter Rinkleff, AVO/UAFGI, written commun., 2008). AVO staff also traveled to Unalaska in late July to meet with the local Director of Public Safety, the resident Department of Environmental Conservation employee, U.S. Coast Guard Station Chief, Native Health clinic supervisor, cannery management, contract weather observer, and airport supervisor and maintenance manager about their concerns and accounts of the eruption. AVO staff participated in local radio interviews, gave a public lecture on the eruption, and met with members of the community to gather eyewitness accounts and answer questions.
"AVO received both ash and pumice samples from citizens in Unalaska. By July 24, some Unalaska residents reported pea to gravel sized light to dark brown pumice washing ashore on several beaches in Unalaska; based on the timing and physical characteristics of the clasts, it is possible that these represented marine transported pumice from the July 12 eruption onset. Whether pumice of this size fell at sea or was washed into the sea by lahars or other flowage processes on Umnak is unknown.
"Flights across the North Pacific were impacted for a period of several days in mid-July as the July 12 eruption cloud drifted north and east over the Gulf of Alaska. During the week following the eruption, aircraft over the lower 48 States observed and photographed the remnant of the Okmok aerosol cloud as it transited across North America at elevations in excess of 30,000 ft (9,100 m) ASL. NWS maintained a nearly constant SIGMET for the area impacted by ongoing ash production during the event; SIGMET boundaries were modified over time on the basis of pilot reports of ash cloud drift as well as satellite images showing the cloud combined with forecast motion."