1999 – 2001 Eruption Sequence in the Gakkel Ridge

Gakkel Ridge across the Arctic Ocean basin https://www.researchgate.net/figure/284359093_fig1_Fig-1-The-study-area-is-the-Gakkel-ridge-Dotted-line-divides-the-Gakkel-Ridge-into

The Gakkel Ridge is an extension of the Mid-Atlantic Ridge (MAR) across the Arctic Ocean between Greenland and Siberia.  The ridge separates the North American and Eurasian Plates.  It is the slowest spreading center known.  Prior to 1999, it was thought to be non-volcanic primarily due to the slow speed of the spreading and the lack of seismic events.

The location of the ridge was initially suggested by Soviet Arctic explorer Yakov Gakkel.  It was confirmed by Soviet Arctic expeditions in the 1950s.   The ridge is located some 3,800 m below the Arctic Ocean icepack.  There are no permanent residents within hundreds to thousands of kilometers of the ridge.

There have been three western expeditions to Gakkel, the first was a US Navy submarine sonar survey 1998 – 1999, the second in 2001 and the third in 2007.  The last two used submersibles.  The most recent found large surprises in terms of volcanic activity, hydrothermal activity and life around the hydrothermal vents.  The ridge was mapped by the US Navy between 1993 – 1999 via a series of submarine towed sonar sensors.

Any time something interesting happens in the Arctic, we here in Alaska get inundated with all manner of “We’re all gonna die!” stories.  Activity along the Gakkel is no different, contributing to scares of massive methane releases at it’s Laptev Shelf end, and massive melting of the Arctic Ocean ice pack following the discovery of a plume of warmer and chemically different water following the 1999 – 2001 series of eruptions.  The size and longevity of the plume was was interesting and unexpected.

Arctic Ocean floor showing major ridges and basins https://en.wikipedia.org/wiki/Gakkel_Ridge

 

Overview

The Arctic Ocean floor is divided into two major basins – the Eurasian Basin and the Amerasian Basin.  The dividing line between the two basins is the Lomonosov Ridge, a major ridge running mostly in a straight line from north central Greenland (Lincoln Shelf) to the New Siberian Islands north of eastern Siberia.

The Lomonosov Ridge is a major feature, rising 3,300 – 3,700 m from the ocean floor.  It tops out in places within 400 m of the surface.  The ridge is 60 – 200 km wide and 1,800 km long.  As with most things in this part of the world, ownership is in dispute with territorial claims by Russia, Denmark and Canada.  All three nations claim the ridge is an extension of the continental shelf.

The Lomonosov Ridge is not considered to be an active tectonic rift.  It is thought to be part of the Eurasian continental crust that rifted from the Barents – Kara Sea margin and subsided some 60 ma ago.

On the European side of the Lomonosov Ridge we have the Gakkel Ridge which runs mostly parallel with the Lomonosov.  This subdivides the Eurasian Basin into two pieces.  The westernmost and deepest piece is the Fram Basin.  The easternmost shallower piece is the Nansen Basin.  The Gakkel Ridge is the northernmost extension of the Mid Atlantic Ridge (MAR) system.

http://arctic-news.blogspot.com/2014/02/mantle-methane.html

 

On the Canadian side of the Lomonosov Ridge we have the largest portion of the Arctic Ocean Basin, the Amerasian Basin.  It is also subdivided by a pair of ridges and basins.  The Alpha Ridge which transitions into the Mendeleev Ridge as it connects into the continental shelf offshore western Siberia.

If you look at a map of the ocean floor, you see three ridges that are mostly parallel.  The Gakkel and Lomonosov are parallel throughout their entire length.  The Alpha is also parallel with the other two while the Mendeleev veers away at a 45 degree angle westward.

USS Hawkbill and friends at the North Pole, 1986 http://www.public.navy.mil/subfor/uwdc/asl/pages/1980.aspx

 

Expeditions

Three major western expeditions surveyed the Gakkel Ridge.  The first was the SCICEX program carried out by US Navy submarines from 1993 – 1999.  The USS Hawkbill used a SCAMP instrument package for its 1998 and 1999 cruises.  These got nearly complete bathymetry, sidescan and gravity coverage to 50 km from the ridge axis for 600 km.  Axis coverage went on for another 23 degrees of ridge length.

SCAMP is an interferometric instrument that does a great job producing high-resolution sidescan.  Unfortunately it produces low resolution bathymetry.  Navigational noise forced the data to be filtered in such a way that small seamounts smaller than a few km across were for all intents and purposes removed from the data.

USCGC Healy, Coast Guard icebreaker supporting the 2001 AMORE expedition https://www.nsf.gov/od/lpa/news/03/pr0367.htm

 

The second expedition was the Arctic Mid-Ocean Ridge Expedition (AMORE).  This used a pair of icebreakers to carry the mapping, sampling and seismic refraction in 2001.  Two icebreakers are needed for this because the lead ship is used to break the Arctic pack ice while the second tows the sonar array.  As this was primarily a rock sampling expedition, bathymetry is much higher resolution, capable of resolving features a few hundred meters across.  Coverage was limited to the rift valley floors and portions of the walls.  The two datasets complement one another.

 

Pillow lava dredged from Gakkel Ridge by AMORE expedition in 2001. https://www.mpg.de/research/mechanisms-of-sea-floor-generation

 

I have not yet found scientific results of Soviet / Russian expeditions to the Gakkel Ridge.

A third expedition in 2007 used a pair of autonomous unmanned submersibles.  This one was chartered thru Woods Hole Oceanographic Institution and took advantage of surplus Swedish equipment.  It looked for hydrothermal vents, life around those vents and evidence of recent volcanic activity.

Bathymetry of all three active zones of Gakkel Ridge https://www.researchgate.net/publication/10690686_Magmatic_and_amagmatic_seafloor_generation_at_the_ultraslow-spreading_Gakkel_ridge_Arctic_Ocean

 

Rift Valley

Geologists divided Gakkel Ridge volcanic activity into three sections based on bathymetry and rock recovery.  These are the Western Volcanic Zone, the Sparsley Magmatic Zone, and the Eastern Volcanic Zone.  These three zones occupy half of the length of the Ridge closest to Greenland.

The Western Volcanic Zone is some 220 km long from the western edge of the ridge.  It has a relatively shallow rift valley, abundant basalt and numerous volcanic landforms.  There are numerous volcanic high points including nearly circular volcanic cones a few km across and 100 – 150 m high in the rift valley floor.  This portion of the Gakkel Ridge operates most like the MAR.  The seamounts are smaller than those in the MAR and represent significantly less erupted magma than volcanoes in the MAR.

The Sparsley Magmatic Zone extends for some 275 km between the Western and Eastern Volcanic Zones.  The valley floor is generally very deep (up to 5,400 m) and is generally either flat or concave downward.  There are few obvious volcanic features.  The floor is also very narrow (less than 5 – 8 km across).  There is a portion of the zone that thickens near what would be a typical oceanic crust.  While there are seamounts in the Sparsley Magmatic Zone, they are only half the height of those in the Western Zone.

Schematic of underlying crust, mantle, hydrothermal action, and magma melt below Gakkel Ridge volcanic / magmatic zones http://www.nature.com/nature/journal/v535/n7611/full/nature18277.html?WT.feed_name=subjects_geophysics

 

The Eastern Volcanic Zone extends for another 580 km eastward.  The far eastern end has been covered in sediments and was not closely surveyed.  The rift valley is nearly buried in sediments to the east of a volcanic center at 69 degrees.  There are at least two additional volcanic centers along the ridge.  These extend for some 30 km along the axis.  One of the differences between the Eastern and Sparsley zones is that the Eastern Volcanic Zone shows what might be recent volcanic activity.  There is a tendency for volcanic structures to cluster on the large volcanic centers.  Those that are not associated with volcanic centers are located near the edge of the rift valley floor.

All three zones are considered magma-poor when compared with the output of the MAR.  Finally, it appears that small magma bodies generated between the crusts may use deep faults as a conduit to the surface.  Western and Sparsley zones appear to be above a major difference in mantle composition.

Detailed bathymetry of Gakkel Ridge recently active portion. This is in the East Volcanic Zone as imaged during the AGAVE expedition. https://volcano.si.edu/volcano.cfm?vn=377020

 

Activity

Scientists originally thought that the ridge was quiet with no volcanic activity.  It had been for 100 years.  A flurry of earthquakes in 1999 on the East Gakkel Ridge led to further investigations.  Sonar mapping in 1999 identified a line of volcanoes along the ridge with one cone at the location of earthquake epicenters.  A survey in 2007 named the cones and discovered that two of the cones were covered with a light dusting of pyroclastic materials.  There are also hydrothermal vents in the ridge.

237 seismic events defined submarine eruptions in 1999.  The events were scattered at the southern rift wall and inside the rift valley.  Analysis of arrival times allowed the location of the conduit to be postulated close to a chain of recently imaged volcanic cones.  Seismicity was focused at the crust-mantle boundary at 16 – 20 km depth.  Seismicity also ascended through the possible conduit indicating a vent opening on the sea floor .

Images of basalt (center) and pyroclastic (right) materials produced during the 1999 eruption sequence. http://www.climate4you.com/ClimateAndVolcanoes.htm

 

Activity was once again observed in 2001.  It was detected acoustically.  It is thought to be ongoing degassing of a magma chamber that started to discharge in 1999.  The lack of microquakes between 1999 and 2001 indicate that the magma chamber / dike system is not refilling with fresh magma.

Analysis of the earthquake swarm using a recent data model (2010) allowed researchers to refine the location of the signals and recorded sounds.  They are now interpreted as submarine Strombolian eruptions with bursting gas bubbles from a magma reservoir along a major fault at the southern rift valley wall.

Detail of SCAMP bathymetry following 1999 eruption in East Volcanic Zone http://www.nature.com/nature/journal/v409/n6822/fig_tab/409808a0_F2.html

 

High resolution mapping and sampling first took place via an Arctic expedition in summer 2001.  Before taking a close look, geologists expected that magmatic activity along a spreading ridge would progressively decrease as the speed of opening decreased.  As the Gakkel is the slowest spreading ridge known, it was not expected to have much in the way of active volcanic / hydrothermal activity.

But this is why we take a look. What the expedition found was a wide variation in volcanic activity.  There is a 300 km long central zone with mantle peridotites and little activity.  On either end, abundant, continuous volcanism in the west and large, widely spaced volcanic centers in the east.  Hydrothermal activity is unexpectedly high.

Pillow lava photo taken during 2007 Woods Hole expedition http://polardiscovery.whoi.edu/expedition2/postcard9.html

 

This suggest that mantle melting is not a simple function of spreading rate.  Additionally, the presence of pyroclastic materials along with the expected basalt eruptions means that mantle temperatures and / or chemistry varies significantly along the ridge.

Before Gakkel, pyroclastic deposits were never observed below 3,000 m ocean depth.  This is generally because volatile content in the usual suite of spreading center basalts is too low to produce gas required to fragment magma at such a high hydrostatic pressure.  A large area of the axial rift valley (> 10 km2) is blanketed with unconsolidated pyroclastic fragments.  In order to produce this sort of coverage and pyroclastics at a 4,000 m depth, the CO2 content of the erupted magma must have exceeded 13.5% by weight.    It demonstrates that pyroclastic activity is possible along the deepest portions of the mid-ocean ridge volcanic systems if there is sufficient gas in the magma.

http://arctic-news.blogspot.com/2015/10/

 

A water column plume above the central rift valley was observed in 2001.  The plume was quite large and extended as far as 175 km to the west.  Plumes are not uncommon above and down current from submarine volcanoes.  But this one was very large and very long-lived.  The plume was most likely connected to variable and continuous volcanic activity during the period 1999 – 2001.  This also means that the two events – the pyroclastic eruptions in 1999 and the Strombolian activity in 2001 were not separate events.  Rather they were portions of a continuous eruptive sequence.  The plume was still observable by 2007 but had diminished and aged, no longer showing signs of fresh input.

Schematic of Arctic Ocean showing cross-section of major basins and ridges https://earthscience.stackexchange.com/questions/4307/what-is-the-tectonic-explanation-for-parallel-ridges-in-the-arctic-ocean

 

Tectonics

The ridge is the slowest spreading center known varying from 12.8 mm / yr near Greenland to 6.5 mm / yr near the Siberian continental margin.  Despite the slow spreading rates, organized seafloor spreading appears to be occurring.  A 2005 paper by Jokat and Schmidt-Aursch moves the spreading rate down to 1.3 – 0.65 mm / year.  The slowest spreading rate is found at the termination of the ridge close to the Laptev Sea continental shelf offshore Siberia.

Crustal thickness along the Ridge varies greatly with crust generally not much thicker than 3.5 km.  In areas where there is no magmatic activity the crust thins to 1.4 – 2.9 km.  This thin crust measurement is also found in the Western Volcanic Zone.  The presence of volcanic structures in the rift valley increases the thickness at their locations.

Comparison of crust – mantle composition models for various spreading ridges. Models are based on spreading rate with the fastest to the left and the slowest (Gakkel) to the right. Also note the presence of mantle rocks at the surface on the Gakkel. https://www.nap.edu/read/13232/chapter/4#17

 

There is a problem with the crustal thickness estimates, mostly due to the lack of data and measurements throughout the length of the ridge.  Some papers have crust thickness in portions of the Western Volcanic Zone as thick as 6 – 7 km, the thickness of normal oceanic crust.

In non-transform portions of the ridge, some seismic returns suggest the presence of mantle rocks exposed on the ocean floor without the normal basalt or oceanic crust coverage.  I have found no explanation for the surprising thinness of the crust along the ridge or the exposed mantle rock.   These returns show up in the Sparsley Magmatic Zone along with measurements as thin as 1.4 km thick.

 

Distribution of small earthquakes along Sparsley Magmatic Zone 2011 – 2015 https://www.researchgate.net/publication/284359093_New_data_about_small-magnitude_earthquakes_of_the_ultraslow-spreading_Gakkel_Ridge_Arctic_Ocean

 

Slow spreading rates of the Gakkel Ridge are found elsewhere such as the Southwest Indian Ridge structure.  So are the very thin oceanic crust thickness.  What is not found on the Gakkel Ridge are transform fault lines which are found elsewhere.  The Jokat paper suggests that magma supply along the Gakkel Ridge was never sufficient to allow the development of transform faults with significant offset.

Slow spreading ridges also tend to be characterized by slow upper mantle velocities which in turn are described by higher temperatures or serpentinization of upper mantle rocks.  In the Western Volcanic Zone, this is thought to be due to hotter than normal mantle as compared to eastern portions of the ridge.

Recovery of Jaguar remote vehicle after getting stuck under an ice flow. From 2007 expedition. http://www.chrislinder.com/portfolio_arcticseafloor.html

 

Conclusions

The Gakkel Ridge is one of the most remote tectonic regions on the planet.  Yet it has volcanic activity and life.  It is a spreading center that connects the active Mid Atlantic Ridge to the non-spreading Eurasian Plate.  Not unexpectedly, volcanic activity is found closer to the MAR, though hydrothermal vents are found in the half of the ridge closest to Greenland and the MAR.  It is home to the deepest pyroclastic activity ever observed.  It also has some of the thinnest oceanic crust ever observed and may even have mantle rocks exposed to the ocean.  Like all our volcanic regions, the more we look at this, the more we find things that we never expected.  Which is why we do this.

Dumbo Octopus viewed by 2007 Gakkel Ridge expedition submersible http://www.whoi.edu/oceanus/feature/arctic-voyage-tests-new-robots-for-ice-covered-oceans

 

Additional Information

http://www.divediscover.whoi.edu/expedition11/

http://news.nationalgeographic.com/news/2001/11/1129_icebreaker.html

https://volcano.si.edu/volcano.cfm?vn=377020

https://www.nature.com/nature/journal/v423/n6943/full/nature01704.html

https://www.nature.com/nature/journal/v453/n7199/full/nature07075.html

http://onlinelibrary.wiley.com/doi/10.1029/2009GC002706/full

http://link.springer.com/article/10.1007%2Fs10950-009-9179-6

http://www.sciencedirect.com/science/article/pii/S0264817201000411

https://www.nature.com/nature/journal/v409/n6822/full/409808a0.html

http://www.people.fas.harvard.edu/~langmuir/Jokat&Schmidt-Aursch%2707.pdf

https://academic.oup.com/gji/article-lookup/doi/10.1111/j.1365-246X.2008.03860.x

http://geology.com/articles/arctic-ocean-features/