A volcanic cruise through the Mariana Islands: Part 1

This week I welcome Dr. Ed Kohut as a guest blogger here on Eruptions (while I am off in the Sierras doing some field work). I've known Ed for 10 years now - we were both graduate students in igneous petrology at Oregon State University - and we are both Massachusetts natives. Ed was in the Coast Guard before getting degrees at University of Rhode Island and Boston University before heading of to Oregon State for a Ph.D., where he worked on melt inclusions in minerals. One of his major research areas is magmatism in the Mariana Islands and he was nice enough to put together a look at the Marianas for Eruptions - this is Part 1. Enjoy!

Big thanks to Erik for inviting me to post a bit on the Marianas. I have been working in the Marianas since 2001 and I have to extend gratitude to Sherm Bloomer at OSU and Bob Stern at UTDallas for not only getting my involved to begin with, but also for keeping me involved. I hope to show that there is a lot more to the Marianas than just the trench.

What, Where
If you look at a map or a globe, you will see many long chains of nearly regularly spaced islands, particularly in the Pacific. Some of the chains are curved to form an arc shape, and are the origin of the term island arc. Since these islands are largely volcanic, they are also the source of the geologic term "volcanic arc". These arcs are the volcanic expression of subduction zones (more on that later). The Western Pacific is home to several of these arcs, including the Marianas.

General location map of the Mariana Islands in the Pacific Basin. Click on image to see a larger version.

The Marianas Islands lie south of Japan and run ~750 km from Guam to the Bonin Islands. The larger southernmost islands (Guam, Rota, Saipan, and Tinian), are uplifted volcanic and carbonate rock and are not part of the active volcanic arc. The more northern islands are active volcanoes that have built up from the seafloor. Several of the larger submarine volcanoes were historically noted as shallow banks. The Marianas are part of an extended series of three subduction zones, Izu, Bonin and Mariana, that are known as the IBM system and are the focus of ongoing research.

First, a little human history. Geographically the islands are part of Micronesia and the original inhabitants are the Chamorros. The Chamorro way of life dramatically changed on March 6, 1521 when Ferdinand Magellan landed in Umatac Bay, Guam during his expedition's circumnavigation (crazy as it sounds, Guam was the first land they touched since leaving South America). Spain eventually formally claimed the islands in 1667, ushering in 231 years of Spanish rule. Defeat in the Spanish-American War forced Spain to cede Guam to the U.S. in 1898 and sell rest of the islands were to Germany. As a result of Germany's defeat in WWI, Japan was given the northern islands as a League of Nations Mandate. Guam was taken by Japan on Dec. 8, 1941. Major battles in 1944 resulted in the destruction of Japanese naval air power and bloody amphibious assaults on the Guam, Saipan and Tinian. Tinian became a major U.S. airbase and the atomic bombing missions in 1945 originated from there. After the war, Guam returned to U.S. territorial status and the northern islands eventually became the U.S. Commonwealth of the Northern Mariana Islands. During the period of recorded history, eruptions have occurred on Asuncion, Agrigan, Pagan, Guguan, Sarigan and Anatahan Islands.

The reason for the islands and the volcanoes is the process of subduction. To get subduction you need to have two plates converging, plates being the crust and uppermost mantle, called the lithopshere . The lithosphere is above another layer of mantle called the asthenopshere, where the rock is at just the right pressure and temperature to flow over time and help move the plates. One plate then begins to go down under (subducts) the other. The leading edge of the continental plate must be basaltic oceanic crust, continental crust is less dense and too buoyant to go under the mantle. In the Marianas the Pacific Plate is subducting westward under the Philippine Plate. As the subducting plate (or slab) sinks, it slowly warms. Water in crust, in minerals, and in any sediments along for the ride is released in to the mantle of the overlying plate. This water forms new minerals that are then dragged deeper as the process continues. Around 80-120 km deep, these minerals breakdown and release the water into the surrounding mantle rock. At this depth, water promotes melting of the rock and we have magma. Since this occurs at same depth along the subduction zone a chain of volcanoes develops. The magmas start off with abundant water and CO2, and are often thicker andesitic and dacitic magmas that will hold onto the volatiles until reaching the surface. As a result, subduction zone arc volcanoes often produce explosive Vulcanian to Plinian events.

Schematic cross section of the Mariana Islands subduction zone. Click on the image to see a larger version.

Looking at the cross-section, you'll notice a couple of other features. Where the plates meet a depression forms in the seafloor - a trench. The sides of the trench are not as steep as imagined since drawings often have vertical exaggeration, they are usually around 8 to 10x wider than they are deep. The Marianas Trench is famous as the deepest point in the Earth's oceans (10,916 meters or 35,814 ft). Beyond the trench is the forearc, an area of compression, folding and thrust faulting. Huge megathrust quakes such as the 2004 Indonesian quake and resulting tsunami originate in these areas. Then we have the arc itself. Behind that there may be an area of rifting appropriately named the back-arc basin. Back-arc basins operate much like mid-ocean ridges, but exist for different reasons. In the Marianas, the Phillipine Plate itself is subducting on its NW edge and being pulled in that direction at the same time it is converging with Pacific Plate:

Plate motions at the Philippine Sea. Click on image to see larger version.

The difference in movement creates tension within the Marianas Subduction Zone, resulting in a rifting of the arc not once but twice. Rifting of the old arc left fragments in the present fore-arc, these bits have been uplifted to form the Guam, Saipan, Tinian and Rota. The arc began rifting again ~7 million years ago and eventually opened to form the current Mariana Trough back-arc basin (not to be confused with the Mariana Trench!)

Subduction history of the IBM.

The Marianas, together with the rest of the IBM, is largely free of interaction with continental crust and sediments. As a result, the cycling of water, CO2, chlorine, sulfur and silicate material between the mantle , crust, oceans and atmosphere in subduction zones can be examined and all stages of back-arc development are present in the IBM as a whole. The Marianas then are much more than a deep spot in the oceans and are an exciting place to visit and work.

Next I'll give you a tour of the volcanoes!

More like this

Thanks, doc Ed!

That was a very informative condensation of what, where and why of the oceanic subduction mechanics; 'specially it explains why the deepest point in oceans is connected with subduction.

By Kultsi, Askola, FI (not verified) on 13 Jul 2010 #permalink

We are all very happy to welcome you, Dr. Ed. And bringing to your opening night such an illuminating subject.

By Renato, Rio (not verified) on 13 Jul 2010 #permalink

Dr. K, thank you for informative post. You've given me an explanation of the asthenopshere that explains what it is and why it exists. So, do the pressure/temperature gradients vary from subduction zone to zone? can it vary within a subduction zone? Is there a fixed range of pressure/temperature? Thank you.

By parclair NoCal USA (not verified) on 13 Jul 2010 #permalink

Thank you, Dr. K for introducing an interesting subject. I have been hearing about the Marianas and I have wanted to know more about them and get an idea of not only where they are (I didn't know that Guam was part of them), but how they formed. As I was looking at the map of plate motions in the area, I was thinking what a mess it is with all the different directions of plate movements. It is no wonder why there is so much volcanic and earthquake activity in the area.

I look forward to your next post.

By Diane N CA (not verified) on 13 Jul 2010 #permalink

Hello to all,
Welcome Dr Kohut.Subduction and the ever changing face of planet Earth has been a source of fascination for me for a long time and I read your first post regarding The Marianas with great interest.
What will this planet look like in the next 100 million years or so ? Fundamentally different to how she is now.
I look forward to further posts with anticipation.

By Adrian,Dorset, UK (not verified) on 13 Jul 2010 #permalink

Let me explain.After the controversy yesterday,I think that what I saw on the left of d9tRotterdam's time-lapse video from Mulakot cam was the "ghost" or even an "echo" from Fimmvörduháls.
I believe that the "plume" that I saw was in fact a large amount of steam rising from the Fimmvörduháls site,as explained by the article itself and what is said by the Hikers interviewed.
So,no new activity from anywhere,just a "blast from the past".

By Adrian,Dorset, UK (not verified) on 13 Jul 2010 #permalink

Please post your off-topic comments to the Open Thread.

Thank you for that explanation. I knew about volcanic arcs arising from subduction zones, but was confused by the term "back-arc". I need to go and re read some previous links now with a fresh eye.

Congratulations, you have won a spanking brand new bubblegum-machine filled with scout-cookies.

Thank you for a beautifully written, interesting and informative article.

As Paul says, thank you for an article so beautifully written, interesting and informative that even Passerby will enjoy reading it.


By Henrik, Swe (not verified) on 13 Jul 2010 #permalink

Hey!! EKoH!! Great to see you stepping in for Erik. I was wondering where you were and then you serve up this delicious course for us! Funny, how even when I read the basics again, I still learn new stuff (all credit to you I might add). For instance I didn't realize the upper asthenosphere moved in line with the plates. I had always assumed the lithosphere was decoupled from the asthenosphere at the Moho and that the plates were moving across it as this explains hot spot-related volcanic chains... oh dang it, must have got it wrong again. How do hot spots work if the aesthenosphere itself moves? Is it because the asthenosphere itself is not very thick?

BTW the way a volcanic forms at the 100km depth contour of the subducting plate is beautifully illustrated in Japan in this movie by Ross Stein et al that I chanced upon two days ago:


Another question, the Kermadecs and the Havre Trough interest me, particularly if you follow the ridges down into New Zealand as they line up beautifully with the TVZ on the one hand (which is line with the arc of the Kermadec Islands) and Taranaki and Mayor Island on the other hand which is in line with the ridge marking the western border of the Havre Trough. My question: this western margin of the Havre Trough, is this also formed by rifting as you illustrate above for the IBM and if so, is there any connection between the associated faulting and the western arc of volcanism in New Zealand (i.e. Taranaki, Mayor Island?

My point being, that given that volcanos form above the approx. 100km depth contour of a subducting slab, is this evidence for subduction of a microslab west of the TVZ, it's margins being delineated by the rifting found in the Havre Trough? ok, I'm postulating wildly here, guess I better rein myself in a bit.. ;-)

By bruce stout (not verified) on 13 Jul 2010 #permalink

me again, re that last little postulation, I just checked out the seismic records from Geonet again and I can't see any evidence for a subduction of a microslab.. which begs the question: why is Taranaki there?

By bruce stout (not verified) on 13 Jul 2010 #permalink

Hi Bruce,
Tonga Kermadec is also a complex little world of subduction and multiple rifts. I'll check on your questions and see what I can find.
The asthenosphere is a particularly weak part of the mantle that is a few 100 km thick. It does flow, not only from back arc to arc and the down, but parallel to the arc in some places. It also upwells under mid-ocean ridges and back-arcs. The mantle plumes that a presumed to be the source of many hotspots are supposed to originate deeper in the mantle.

Erik should do something on hotspots and the plume idea and debate - that'll teach him to confuse BU with BC ;)

Ed Kohut

To all readers, I will try to answer any question through the comments section.
I should add that Dr. Yoshi Tamura at IFREE in Japan is also an important investigator in the Marianas and has done a lot to propose and schedule research cruises to the area.

Ed Kohut

Thanks EKoh!

I take it that the circulation in the asthenosphere from back arc to arc then down then upwelling in the back-arc basin is also at least partly driven by friction against the submerging plate (i.e. not only does the collision throw up an accretionary wedge but it also drags asthenosphere down with it. Is this right?

By bruce stout (not verified) on 14 Jul 2010 #permalink

What an excellent volcanic cruise. I look foward to subsequent installments!

@Bruce stout: the movie by Ros Stein is amazing. Thanks for posting. Wish @Lurking could take a look at this.
And thank you very much EKoh for sharing your knowledge with us.

By Renato Rio (not verified) on 14 Jul 2010 #permalink

hi everyone

Today there was a earthquake at laki volcano. the depth was at 1.1 km and the mag. was a 2.2. This could be tectonic more than magmatic but something to keep watching.

I think the article was fantastic. I had heard the back arc / fore arc terminology used in discussions of the Sumatra Shindig, and had murkily worked out what they meant. This provided "back-fill" that explained what I was reading. THANKS!

OT: @jack [20] Laki and the Technonics⢠I'm not that sure that a tectonic event at Laki/Eldgjá would be a very good thing. 934 and 1783 didn't turn out so well.

Here is a cross section pretty much perpendicular to the ERZ. Lat/Long coordinates omitted to eliminate some confusion. The view angle is approximately 307° (looking to the North West)


Oops... those are all July quakes to present.

By chance I have been doing some introductory reading about geodynamics. It is a fascinating story that is becoming as confusing as it is enlightening.

Kultsi(1)says that EKoH has explained why the deepest parts of oceans are subduction zones.

I don't easily see it but assume that such deep places are caused by subduction in contact between two or more 'dense', thin, structurally resilient oceanic plates.

There are many other processes involved and I allow myself the opportunity to go perceptually dyslexic and fall back into confusion.

Most everything in volcanology could be attributed to state change and the annealing/quenching brought about by flux in the respective factor. (factor = temperature, gas concentration, chemical concentration, pressure, ? )

'State change' as in melting/solidifying and 'flux' as in temperature and thermal transport keep emerging as the main factors. Numerical simulations show that rifting is particularly sensitive to establishing and maintaining a thermal flux.

This leads me to consider some fanciful questions as a means of rapidly setting context.

- How deep does an ocean have to be before the pressure does away with the lithosphere altogether and problem is reduced to a liquid/liquid water/mantle phase boundary? Obviously the water acts as a calorimeter sink and the existence of a liquid/liquid interface is implausible, yet as the lithosphere transformation is described in the topic introduction, the deepest part of the deep oceans of 2 miles of liquid. ... Let's just say there seem to be limits as to how deep an ocean can become. Remember that the situation is a liquid floating on top of a plastic lithosphere floating on top of a fluid mantle.

- The question is more interesting when one considers ice ages and polar ice caps. I don't know if ice sheets of a 100 miles thickness are implausible but the scenario must have interesting consequences for plate tectonics and continental drift.

- 80% (??) the bulk of the heat flux is derived from radiogenic activity in the lithosphere (natural radioactivity).

Here are some crude heat flux values in units of watts per square meter.

Sunlight falling on the earth = 60 watts/m^2

Non 'hot spot' ( i.e. most of the earth's surface) derived from the earth's interior + radiogenic activity = 10's of milliwatts/m^2

Tectonically active regions = 100 milliwatts/m^2

"Restive" volcanoes on active faults with aqueous lakes = a few 10s of watts/m^2

It is intriguing to see that thermal fluxes that determine tectonic activity are of order(s) of magnitude less than the thermal fluxes involved in climatological process.

Or rather that factors and processes involved in attenuating or augmenting thermal fluxes at a scale of .1 W/m^2 are everything.

Perhaps it explains why most of the the plate boundaries are submersed and that the rising and continued evolution of an eruptive fault line above the water table has a big influence on the future evolution of faults.

Given a low globally averaged subterranean derived thermal flux of ~ 30 mW/m^2 there would seem to be ample opportunity for coupling with climatological, biological and extraterrestrial activities in regard to 'hot spot' and fault line evolution.

It certainly is a complex system. There are coupled interactions at length scales ranging from the microscopic to the astronomical.

For example, it would seem that even a small asteroid impact could change the future evolution considerably. Alternately the many many scales of coupled interactions ( many phase states are involved at many scales) might serve to buffer and re-establish a dynamic 'quasi equilibrium'

Ocean crust = 5 mi
Continental crust = 25 mi average thickness.

Now who do you think will win a pissing (convergent plate) stress contest?

On the theory of subduction with mechanics explanation, see 'Origin' section, with recent references:


Talking about subduction, looks like aftershocks on Chilean EQ have risen since yesterday: 6 EQs over 5+

By Renato Rio (not verified) on 14 Jul 2010 #permalink

At the bottom of the 'Subduction' wikipage, you may have seen Tatsumi's humorous cartoon diagram of ocean crust processing. Two papers worth a read:

Manufacturing Continental Crust in the Subduction Factory. Tatsumi, 2006 Oceanography 19:104-112.


Original 2005 Tatsumi classic, www.geosociety.org/gsatoday/archive/15/

An Overview of the Izu-Bonin-Mariana Subduction Factory (2003)


Not sure I follow your logic on 'how deep an ocean can be'. The constraints lie in the formation of the oceans and large -scale hydrologic cycles over time (evaporation vs condensation, which can be evaluated by isotopic ratios, see Rayleigh Fractionation).

Ice sheets during the last glaciation cycle: 3.5-4 Km, northern hemisphere.

Thickness of the Antarctic Ice Sheet

That icecap as quite a bit to do with both salinity and thermal behavior of the deep oceans.


The lithosphere is dense, but not as strong as the lithosphere. I do not think 'ocean depth weight' is an issue.

The answer to your question then, is found in the process by which the oceans formed:


The rate limiting step would be the mass and composition of the early atmosphere and the rate of cooling that allowed condensation to form.

And of course, it would require a gravity field.

Wow,the mantle certainly doesn't seem to be the viscous undifferentiated fuzzy warm body I learned about 40 years ago.

Are the interior of stars still viewed as warm fuzzy fluid balls being too hot for differentiated structure formation? I.E. ..such as http://library.thinkquest.org/27930/media/sundiagram3.gif

Continuing to read. ...

Thanks Passerby

I would be seriously remiss if I didn't add a truly elegant paper by Udipi Ramachandra Rao that explains the importance of the magnetic field in maintaining the water and ozone balance in the atmosphere while limiting NOx formation.

Big Picture perspective is always useful.

Crucial role of the magnetic field in the evolution of life. www.ias.ac.in/jarch/pramana/15/00000038.pdf

Minor wording correction: The relatively rigid lithosphere is dense, but not as strong as the aesthenosphere.

Sorry, can't answer your question on modern theory of star formation. Not my bailiwick.

>the mantle certainly doesn't seem to be the viscous undifferentiated fuzzy

We're hoping our guest host will shed some explanatory light on this complex topic.

But first, island arcs and island volcanics.

@Passerby - Re: Stars

Current ideas are that no, Stars are just as differentiated as our Sun. Via Helioseismology, they have deduced that the Sun has two different rotations. The upper Differential rotation area where the poled rotate slower than the Sun's equator, and an inner region that rotates as a solid body.. not that it's necessarily solid, just that it rotates uniformly from pole to pole. It's also though that the turbulence where these two regions interact is responsible for the Solar dynamo, and the sunspot cycles. Since stars have been seen to exhibit similar periodicity, the general idea is that they operate in a similar manner.

Sorry... meant to be @Raving. My bad.

A good read:

"Sunquakes: Probing the Interior of the Sun" - J. B. Zirker

The mantle is made of a family of rocks called peridotites. These are made of dense silcate minerals with lots of magnesium and iron. There are variations in places due to some material being removed by partial melting. There are some changes in minerals with depth as one mineral becomes unstable and changes to a more stable one. Then there is subducted oceanic crust which metamorphoses from basalt and gabbro to a rock called eclogite. There are also global isotopic differences that make mantle types called things like HIMU and FOZO by that strange group, mantle petrologists ;)
But compared to continental crust, mantle rock is remarkedly homogeneous.

The Wikipedia pages are not bad:

Ed Kohut

@Ekoh I haven't looked at those references yet but what you have written clears up and/or adds confusion immediately ... oh the joy and the confusion ... solid/solid exothermic phase transition ... compressive (kinetic heat) exothermic phase change of subducted material originating from convergence and subduction of plates (more release of kinetic energy)... deeper down close to the outer core an endothermic state change. ...

... "mantle rock is remarkedly homogeneous" ....



Dr. Kohut, welcome, and thanks for the informative post and especially for making geochemistry interesting and sensible. That last section, about why the Marianas are so helpful to study, is doubly enjoyable for me and here's why.

Back in the 1980s, I tried to get an undergraduate degree in geology but just could not grok geochemistry - even mineralogy was tough! It is science articles like this, written in plain English, that help me see how the chemistry fits in with all the rest of it (which I truly love and did well in academically).

Great design first thing⦠Second thing the prespective at which you see the sport is quite neutral quite in my words wikipedia style. This is a welcome change from people who are either racist or support a particular agenda.!

I admire what you have done here. I like the part where you say you are doing this to give back but I would assume by all the responses that this is working for you as well. Respectfully, Sean.