December 2020, a group of scientists suggested that the islands of the Four Mountains (IFM), a volcanic archipelago in the central Aleutians may actually be evidence for a large, previously unrecognized caldera in the Aleutians.
There are two groupings of islands. The westernmost includes Amukta, Chagulak, Yunaska. The eastern includes Herbert, Carlisle, Chuginadak, Kagamil and Uliaga. They are all volcanic, with Mount Cleveland on Chuginadak Island the most recently (and currently) active. The islands are located some 1520 km SW of Anchorage down the Aleutian chain, between the Andreanof Islands to the west and the Fox Islands to the East. The islands have a total area of 550 km2 and no permanent population. The two largest islands are Yunaska and Chuginadak.
The name is derived from the Russian translated into “Island of Four Volcanoes.” Early Russian explorers named them based on the four prominent volcanoes on the islands. There was some confusion about the names of the islands because only four of the five showed up on early maps and charts.
We wrote about volcanoes in this part of the world with posts on Makushin and Okmok to the NE, Cleveland, and Kasatochi to the SW. These would be good places to start for a review of local tectonics and an overview of the islands.
While the Islands of the Four Mountains (IFM) were not permanently settled, there is evidence of human visitation, generally hunting birds and sea mammals ranging back at least 3.7 ka. The Aleuts settled the Fox Islands NE of the Four Mountains around 9.0 ka. They settled the Andreanof Island group to the SW some 6-7 ka. It is reasonable to assume they visited the Four Mountains during that expansion SW. There is possible evidence of human impact on bird populations by 7.3 ka. Why the 3,000-year long gap in settlement?
That gap could be both climate and volcanically related, as there are sediment fans on two the islands from debris flow deposits covered by around 3 m of layered tephras which were in turn covered by other erosional debris. The highly eroded Tana volcano suffered a flank collapse and subsequent glacial erosion of hydrothermally weakened rock. The islands were almost completely glaciated as far down as the shoreline during the last ice age which contributed to both erosion and unstable local surface geology.
The oceanic climate of this portion of the Aleutians is wet, with significant rain during the summer and snow in the winter. Mean annual temperature is 1.3° C with persistent, strong winds. Vegetation of the islands is dense shrubs, grasses and sedges.
Other than Cleveland, the islands are not monitored. While Cleveland does not have a webicorder, it does have a webcam.
Possible Caldera?
An interdisciplinary group of scientists working together since 2014 presented research at the Dec 2020 meeting of the American Geophysical Union (AGU), suggesting that evidence exists for the presence of a caldera measuring at least 20 – 25 km in diameter in the ocean between the islands and volcanoes of the Islands of Four Mountains (IFM). The current islands and volcanoes are located on the rim of the proposed caldera. Such a structure would explain the unusual tight grouping of the multiple volcanoes in this part of the Aleutians. AGU members can find the presentation linked. Sadly, it is paywalled to AGU members only.
The proposal was an attempt to explain the tight grouping of similar volcanoes and the intense activity of Mount Cleveland. The first publication I found suggesting the presence of an underlying caldera structure was a Jan 2021 paper by Power, et al investigating seismicity under Cleveland and Tana volcanoes. This paper summarized seismic results gathered 2014 – 2019. As of this writing, ignimbrite samples from Chuginadak Island have not been analyzed, so the date of a possible caldera-forming eruption is as yet unknown. Presence of the ignimbrites on the other IFM islands is as yet unknown.
As expected, media coverage of the proposal was widespread and lively. The story was picked up by GeologyIn, Live Science, SciNews, SciTechDaily, TechTimes with similar stories. The AGU and Carnegie Science also had press releases.
What did the investigators find? From the presentation linked above, the researchers listed the following:
- Circular vent distribution in IFM. This includes circular ridges below sea level connecting several of the islands.
- Location of a very active Mount Cleveland.
- Satellite-derived gravity anomaly suggests a source 1.7 – 10 km depth. Similar gravity anomalies have been identified in a number of other calderas worldwide. This one is similar to the one at Aira caldera in Japan.
- Welded ignimbrites were found at several locations on Tana. They are dacitic and locally sourced, distinct from any other known tephras found in and Aleutians.
- A local seismic net on Cleveland 2015 – 2016 observed quakes 2 – 8 km. Earthquakes to the NE were at 5 – 10 km deep. Those under Cleveland were generally thrust and strike slip earthquakes. Those under the isthmus were generally normal faults, with inconsistent alignment similar to those of other Aleutian calderas like Fisher and Korovin. There may be a low velocity zone 20 – 30 km below Cleveland.
- There are widespread hydrothermal sources on Herbert, Tana and Kagamil. Those on Tana have isotopic signatures that suggest a magmatic source. Kagamil has not been sampled as yet.
- A large caldera can be expected in this portion of the Aleutians.
As of December, there is simply not enough data as yet to decide if a caldera or multiple calderas exist at IFM. As with all initial proposals, there is much more unknown than currently known, so a lot of additional data is needed to flesh this one out. This includes:
- Detailed bathymetry of the IFM and surrounding seafloor.
- Improved, widespread seismic analysis of the IFM islands.
- Improved gravitational analysis of the IFM.
- Sample gathering and analysis from Uliga, Kagamil and Carlisle volcanoes.
- Comprehensive gas and fluid sampling / analysis of hydrothermal areas on all the islands.
Seismic, gravity, gas and hydrothermal data are particularly important, as that data may be able to start tying the multiple islands and volcanoes into the larger system if one exists. Basically, the more similarities between gas emissions, hydrothermal systems, ignimbrites, earthquake depths, and underlying low velocity zones the additional research finds on the multiple islands, the more likely the islands are part of a single volcanic system. If these islands end up being completely separate, discrete volcanic systems, that will be a different solution. In either event, we will have learned something about the IFM.
This portion of the Aleutians has multiple, large calderas. IFM is located in a gap between neighboring calderas. If IFM ends up being a caldera, it will be the largest one in the Aleutians. Multiple cones around a large caldera is not uncommon. In the Aleutians, Fischer is one such example. Highly active volcanoes like Cleveland located on or within caldera rims are also common. Examples would include, Sakurajima, Aria, Anak Krakatau, and Bromo – Tenggar.
Volcanoes
Chuginadak Island is the largest of the IFM, measuring 23 x 10 km. The island has a rugged shoreline with some protective harbors and inlets. The island is constructed by two volcanoes connected by a narrow isthmus. Mount Cleveland is the most recently active of the two. Tana is heavily eroded. Cleveland is not and has been historically and currently active. There are no historic reports of activity at Tana, though it does have an active hydrothermal system on its N and NE flanks including hot springs, fumaroles and mud pots. Debris fans on the flank of Tana appears to predate human habitation. There are proximal moraines on Tana mostly of hydrothermally altered rock. Several other stratovolcanoes are located on nearby islands, Carlisle, Herbert, Kagamil and Uliga. These closely spaced volcanoes are unusual as compared with more widely spaced volcanoes in adjacent areas of the arc.
Mount Cleveland is an open vent, basaltic andesite to andesite system. It tops out at 1,730 m. The shallow active conduit extends at least 1.3 km below sea level and is fed from a deeper magma system. Explosions occur in the conduit from sea level to the summit. The shallow system overlies a significant low velocity zone 10 – 30 km deep in the lower crust. Cleveland’s magma system may extend through much of the crust. The top of the magma column is considered to be within hundreds of meters of the vent. Eruptions are not preceded by deformation, indicating magma moves easily to the surface.
Eruptive products are basaltic andesites to andesites. Historic ash plumes have reached as high as 9 km. Dome building is not uncommon, with the central portions of the dome subsiding after emplacement. Lava domes are generally destroyed by subsequent explosions.
Tana is highly eroded volcanic complex sharing Chuginadak Island with Cleveland. It has at least two volcanic centers along an E-W line. This includes a small lake-filled summit crater. The western peak is not heavily eroded and may have been active over the last 10 ka. Younger eruptive fissures and cinder cones are present on the western flank of Tana and extend into the isthmus between the two volcanoes. Low silica lava flows with flow levees constructed peninsulas on the NW and S sides of the complex. It erupts mainly andesites to dacites and today has an active hydrothermal system. Tana has not been surveyed with seismographs as yet but needs to be. There is also what appears to be a flank collapse amphitheater scar on the E – SE flank of the easternmost Tana eruptive center.
Volcanic cones in the isthmus between the two volcanoes are relatively new. There are as many as seven of them, thought to be active late to mid-Holocene. There are more than 60 visible tephras from these cones. They erupt more basalts than andesites, different from current magmas erupted from Cleveland. These cones are older and less recently active than Cleveland. Earthquakes centered 5 – 10 km below these cones indicate that magma is still active beneath these though there have been no eruptions for perhaps thousands of years.
Magma erupted from the isthmus cones comes from 10 – 30 km. There does not appear to be intermediate storage and evolution of these magmas prior to erupting like Cleveland magmas. This means that these magmas come via a completely different plumbing system than Cleveland.
The difference in earthquake locations beneath the isthmus and Cleveland itself may be due to interaction between the two volcanoes – Cleveland and Tana. Alternately, there may be a larger structure responsible such as a caldera. The orientation of the isthmus craters themselves may be driven by the presence of a possible caldera. Caldera structures are known to produce more chaotic stress directions in the Aleutians as observed at Okmok, Fisher, Atka and Emmons Lake. Many of these also support closely spaced eruptive centers.
There is no information on possible vents beneath the ocean due to lack of detailed bathymetric data. The orientation of the isthmus cones are unusual in the Aleutian arc. Similar satellite vents elsewhere generally align NW – SE, parallel to the greater volcanic arc stress. The isthmus satellite cones align nearly due north on the Tana side and 45° off that line on the Cleveland side.
Uliaga is a triangular shaped island with a single eroded stratovolcano. There is no evidence of a caldera. Little is known about the volcano. There are no reports of historic eruptions, though it is considered to have been active during the Holocene.
Kagamil volcano occupies most of the southern half of Kagamil Island. It is 3.5 x 2 km at the base, with the long axis generally NW – SE. The larger cone at the SE end tops out at 893 m. The NW cone is 610 m with two summit craters. The volcano has an active hydrothermal system with hot springs and fumaroles near the SE shore of the island. The cones are uneroded, indicating a post-glacial age. An early explorer reported the volcano flamed and smoked in 1840, but the only official report is of unspecified activity in 1929.
Carlile Island is a single symmetric cone, 1,524 m and 6.5 km in diameter. The steep slopes are generally snow-covered year-round. The lower slopes are a bit more eroded than the summit. Like neighboring Herbert, it has a rugged coastline with cliffs and deeply cut ravines. There are no bays or harbors. The island may contain two closely spaced volcanic cones. Eruptions have been reported since the 18th Century but are poorly documented. Historic reports of eruptions from Carlisle Volcano may instead be from neighboring Cleveland, as summits of the two volcanoes are separated by 10.6 km with frequent low clouds and fog. There is a single Coast Guard report of steam and some ash emitted from a volcano believed to be Carlisle Nov 1987. Steam but no ash was observed earlier in Aug 1987. A peat bog on the NE flank of the island has two significant tephras. The 2,000-year-old tephra came from Okmok. The 1,050 lapilli is thought to have come from Cleveland. There is an 18 cm thick tephra dated 6.03 ka that has yet to be sourced, thought this one may have come from Carlisle. The majority of eruptive products from Carlisle are basaltic andesites to dacites.
Herbert is a stratovolcano with a 2.1 km summit caldera. Topographic mapping suggests the caldera is breached to the NW, though this is not supported by radar images. There are no geologic studies yet published of this volcano. There are no reports of historic eruptions. It is primarily constructed by andesites. It is the westernmost of the IFM, around 10 km in diameter, with its volcano topping out at 1,300 m. The island has rugged shoreline with high cliffs and few bays.
Tectonics
The Aleutian Islands are a very old structure, created by the subduction of the Pacific Plate under the North American Plate. This collision may have begun as long ago as 70 Ma, though the currently accepted date is 55 – 50 Ma. The collision formed the Aleutian Trench at the point of subduction. The oldest exposed rocks are at Finger Bay, 37 Ma basalts. There have been three main pulses of arc-wide magmatism, 38 – 29 Ma, 16 – 11 Ma, and 6 Ma – present. These coincide with periods of intense magmatism in other western Pacific Island arcs.
Today, the convergence rate at IFM is roughly 6.1 cm/yr, which is enough to power massive megathrust earthquakes, and provide sufficient melt for caldera formation along this segment of the arc. Mount Cleveland and Tana volcanoes are roughly 65 km above the subducted Pacific Plate.
One of the interesting things about IFM is that it currently represents a gap in a line of calderas along this portion of the volcanic arc. This gap is one of the reasons that the piqued the interest in investigative team members of the possibility of one or more calderas associated with IFM.
The most recent massive earthquake in the region was a M8.6 in the Andreanof Islands to the E of IFM in 1957. This put an 18 m tsunami on IFM, greater than 27 m tsunami on the Pacific side of Umnak Island. The tsunamis caused millions of dollars of damage, with waves as high as 16 m in Kauai. The previous Andreanof Islands Great Earthquake in 1946 was also listed as a M8.6, released a bit less energy, though tsunamis were larger, 42 m which destroyed the newly-built Scotch Cap lighthouse on Unimak Island. The 1946 earthquake killed around 170 people and caused $26 million in damage.
Conclusions
In my opinion, the continuing research into IFM since 2014 represents the way scientific inquiry ought to proceed – identify something odd, ask some questions, gather a lot of data, propose an explanation, and then figure out you don’t have nearly enough data to reach a conclusion, which sends you back around the circle once again. It is a great, great detective story regardless of what the results end up being. Thank you to John Power, USGS, AVO for his assistance in getting access to the team’s presentation and for his patience in answering my uninformed questions.
From a volcanic standpoint, given the current activity of Cleveland, the apparent recent activity on Carlisle, Herbert, and active hydrothermal systems on at least four of the islands, there is no reason to believe that this system(s) will do anything but continue to erupt for the foreseeable future.
Additional information
Alaska Guide, information about Islands of Four Mountains
NASA Earth Observatory, Islands of the Four Mountains, Jul 2007
agimarc
Spica
Granyia
volcanohotspot.wordpress.com 