The return of the dreaded "giant magma chamber" of the Cascades


The southern Washington (and northern Oregon) Cascades, with Hood (foreground), St. Helens and Adams (middle) and Rainier (background).

Remember a few months ago when an abstract from a meeting got people are ruffled because the study said there was a "giant magma chamber" feeding all three of St. Helens, Adams and Rainier in the Cascades? Well, Nature Geosciences has been kind enough to publish the work (the article) and yes, yet again, we can say it is still most likely not true at all.

Graham Hill and others claim that their magnetotelluric work in the Cascades show that a vast magma chamber (at only 12-15 km depth!) connects all three of the southern Washington volcanoes of St. Helens, Adams and Rainier. Hill goes as far as to claim "our results speak for themselves," which, in my world, is code for "we don't have a foot to stand on." Anyway, I (along with others) have already picked apart the thermal, geochemical and geophysical reasons why this isn't feasible - the magma compositions at all three volcanoes each tell there own story, which is evidence enough that this magma body isn't feeding all three. However, the most damning indictment comes from Dr. Steve Malone (UW) - a man who knows his Cascades and the structure of the arc - who said: "The geothermal evidence for what Graham suggests is nonexistent. There is also no seismic evidence for such a magma pool." Well, then.

So, although the authors of the paper claim the take-home message is "there is evidence of a primary magma pool that feeds the chambers underneath the volcanoes", I think the real take home message is that maybe there is evidence that there is some amount of melt (5%? 1% 0.1%) under the southern Washington Cascades - heck, I'm sure there is, but I would bet dollars to donuts that there is no "vast swirling cauldron of magma" that is the single source for all these volcanoes. (Or maybe the take home message is that Nature and Science will always choose controversy over good science).

UPDATE 10/25/09 5:45 PM: Hidden midway through a Seattle Times article on the paper: "The team estimates only 2-12 percent of the rock is actually molten." (note: my emphasis) If that is their conclusion, that is is absolutely nothing new. Partially molten crust - as an area of interconnected (or not) lenses of melt is nothing new - just their dramatic spin as a "common magma chamber" is ... headline-grabbing "science" at its worst.

More like this

Bad science gets the headlines. Good science always triumphs in the end.
Now on the basis that most rumors are loosely based on fact, what perhaps DID they find that might be useful?
The Discovery Ch will surely show a graphic of a half-filled cavern of orange bubbling goo, with a flaming current winding thru stalagmites.
I'm guessing its more like a linear region of broadly related melting that follows the subduction boundary. *yawn*

Chris Rowan and I were just tweeting back and forth about this, and I've got a question: wouldn't seismic tomography be a better technique for looking at this? I thought the disappearance of S waves was the ultimate way to recognize a large body of liquid, and there are seismic stations in the area (run by the Pacific NW Seismic Network, http://www.pnsn.org//). Between the Cascadia subduction zone and Mt St Helens, there are both deep and shallow sources of earthquakes - it ought to be a place where natural seismicity could be used for that kind of study.

I mean, I realize this would hardly be the first time that someone drew outlandish conclusions from limited types of data, but in this case, there ought to ways of testing the conclusion beyond the magnetotellurics.

Oh, good lord.

Kim - Seismic tomography is (as far as I know) the method that volcano seismologists usually use to look at the shapes of magma chambers, but there's definitely some arm-waving involved in the calculations associated with that - I had a volcano seismology class last semester that talked quite a bit about the subject, and there are a number of uncertainties involved that means no one is going to say things like "our results speak for themselves". We looked at some studies on Mount St. Helens as well, and there is definitely no giant magma chamber involved - more like one shallow and one deep one. (This is a link to a 1992 paper that we read, although I'm sure there are newer versions.)

You know, just thinking about it as a mass question, the area underneath the three volcanoes - being very very conservative - is something like 800 km2. Now, assuming that the "magma chamber" is at least a 0.5 km thick - I mean, that might be the smallest to visualized at 12-15 km depth - thats 400 km3 of magma just sitting there. That would translate to something like 40 km3 magma erupting (assuming a 10:1 intrusive to extrusive). However, to stay molten for that long and so show up as a hot, liquid body you'd have to guess it was a basaltic body - and when was the last time you saw a big basaltic eruption of any of those volcanoes (maybe Simcoe field behind Adams, but thats not 40 km3 of basalt at once). This doesn't even get into the fact that (a) the Cascade crust is hot to begin with, meaning there could be lots of discrete pods of melt - tiny things that might not be connected and (b) how much basalt input from the mantle is needed to keep a molten body like that active. I think Dr. Malone's comments back up the idea that PNSN has never visualized anything close to what is being suggested by the Hill group.

I do have some difficulties with all those things super and gigantic in volcanology, which not only are the problems that should concern a bit less than everyday questions (because they are highly unlikely and/or, such as a "super" eruption at Yellowstone, extremely rare), but also because, on the educational level, they can be a source of significant, and unpleasant, misunderstanding in the public. So each time some colleague comes up saying this or that is a super volcano and a mega magma chamber, the media jump on it and contort the information further. This is not what we should facilitate.
Some colleagues see some magmatic relationship between the different volcanic areas here in Italy, some even speculate there's some mega plume linking all volcanoes of Europe. While I am in not in the position to say that all of this is not true, it does concern me relatively superficially because we've got issues with half a million people living in the high-danger zone at Vesuvius, and hundreds of thousands having property in areas at risk from lava flows at Etna. Still we're far from understanding just the most superficial dynamics of these volcanoes. If Vesuvius behaves like Redoubt earlier this year - hesitating for several months before erupting - and half a million people are in evacuation camps, that's far more than an academic problem. And we all know that similar conditions exist at a vast number of volcanoes elsewhere on this planet.

Erik: is this a peer-reviewed article? I can just about accept the idea of a zone of partial -with the emphasis on partial- melt underlying the Oregon Cascades region, and I can accept, obviously, the idea of magma bodies at 12-15km depth. It's putting the two together that sticks in the craw, since the relevant stratovolcanoes have very different eruptive histories; compare St Helens and Adams for example

How good is the method then? I hear it is used quite a lot. For example this comes from an abstract by Heise and Caldwell et al. on magnetotelluric imaging of the TVZ:
"An unusual feature of the MT data at periods of 3-30s is the large phase tensor skew angle values that coincide with the margins of a localized gravity high in the centre of the survey area. This feature appears to be caused by the interaction of a thick near surface layer of high conductive volcaniclastic material with conductive structures at greater depth".

Is this referring to melt or infill? (the word volcaniclastics threw me) .. and if it is referring to melt, isn't this similarly scary? Or is the method to be taken with a large grain of salt?

By bruce stout (not verified) on 25 Oct 2009 #permalink

Hey Erik,

This discussion about whether or not there is a "giant magma chamber" under the Cascades got me wondering about your comment from a few days ago.

"The northern Cascades of the US are filled with calderas."

Where are the calderas and what are their ages?

By Thomas Donlon (not verified) on 25 Oct 2009 #permalink

I stopped reading Nature about 17 years ago since there were too many outright ridiculous articles in fields I know something about - and yet Nature maintains its cult following.

A colleague was recently complaining about articles published in other journals though; perhaps reviewers are just getting slack and there's too much nonsense published? I've knocked down the 2 papers I've been asked to review so far this year; I just can't recommend the publication of a paper which shows laziness and a fundamental misunderstanding of basic physics. In place of "our results stand for themselves" I had to contend with phrases like "it is obvious" when I thought the opposite was painfully obvious even from undergrad textbooks. (Although I indicated that the one paper may be suitable for publication if the authors care to develop their ideas properly.)

By MadScientist (not verified) on 25 Oct 2009 #permalink

I love a bit of doom mongering..so long as its "factual". Most programmes on TV regarding "Super Volcanoes" when talking about Magma Chambers tend to miss out the point that usually only a small portion of the "Chamber" is Melt and eruptable. Infact I think really only the Super Volcano 2 part program/DocuDrama about Yellowstone actually tried to present some kind of factual science in with the fiction, it was at least entertaining.

I have a question and I hope its not too trivial.
Mapping a magma chamber is tough enough, and granted probably impossible to verify.
But does anyone actually know why magma rises to the surface?
Is it "bubbles" of heat rising and melting the rock above it til it broaches the surface? Or less dense rock "floating" upwards and melting as the pressure falls? Or is it the gas trapped in the rock fizzing upwards and carrying the magma along? All the above?

Well that's a relief. Discovery Channel did a few specials on that idea and their documentaries/what if stories basically boiled down to "it could really happen, even tomorrow!!" I'm not fond of sensationalism, especially when it distorts the actual probabilities of such an event.

This is an interesting example of where the âplate tectonic paradoxâ and true volcanology literally collide. Of course there is a huge magma body below the Cascades, they are entirely volcanic are they not? Yes, hence the validity of the huge magma chamber as well as the massive evidence of volcanism on the crust. The paradox is that according to âplateâ tectonics this is supposed to be a âsubduction zoneâ at the same time a âspreading centerâ, and therefore we have a paradox. So which is it? Perhaps neither. Itâs just a massive magma chamber, period, and the plumbing for these edifices are probably connected very deep. The bathymetry in the Pacific render no indication of a âsubduction trenchâ, rather a series of canyons and therefore âsubductionâ is only assumed as the science is damned to the theory, which does not work. The orthodox approach with the PT theory needs to be abandoned as the reality is there are massive magma chambers scattered all over the globe that have at one time in the Earthâs history âreared their headsâ and made a huge mess of things. Therefore I believe in the concept of Vertical Tectonics vs. orthodox âplateâ tectonics as being what is observed in the form of giant rising magma chambers. The âbig pictureâ of the Earth system truly needs to be re-scrutinized and revised. I'm a PT "holdout" if you didn't notice...

Hi Fitz,

I am also interested in what drives volcanic eruptions and I am hoping to learn much more within a year.
From what I have learned so far, there is a little bit of truth to each scenario that you put forth although it would be described differently by a scientist.

However, the main driver of most volcanic activity is subducted water. Most volcanoes take place near subduction zones. Plate tectonics shows new ocean surface which forms at mid-oceanic volcanic ridges being conveyed or convected outward to where it will collides with and and then dive under a continental plate. The oceanic plate has become water-logged through seepage and chemical binding and when it is subducted it starts to heat up. Now, free water and chemically bound water is in the rock - but the heat at depth causes the water to come out of the rocks. One type of rock can change through a process called metamorphosis into another rock called metamorphic rock. Heating acts on the free water and also chemically unbinds water from minerals - changing them into metamorphic minerals + free water.

The presence of water then lowers the melting point of neighboring rocks changing it to magma. The rising steam in turn helps melt the rocks above it - changing them to magma.

There is also some volcanic activity that forms from hot upwelling of magma (currents or hotspots and a splitting continental plate).

This article talks about a newly discovered chemical reaction that can form volcanoes.
http://www.physorg.com/news134893218.html
-------------------------------------
Geologists Discover Magma and Carbon Dioxide Combine to Make 'Soda-Pop' Eruption

This discovery overturns a longtime belief by geologists, who thought that carbon dioxide was incapable of dissolving in magma, said Calvin Barnes, professor of geosciences and lead investigator.
-------------------------------------------------

There are many or several mechanisms for volcanic eruptions and subducted water is a most prolific driver of volcanic activity.

By Thomas Donlon (not verified) on 26 Oct 2009 #permalink

To arch,

I don't see why a subduction zone and a spreading zone are mutually exclusive. In fact you will see that a lot of plate boundaries are bordered by microplates where the overriding plate has fractured behind the plate boundary and this extensional scenario created by the appearance of a microplate is often the location of vigorous volcanism. (Havre Trough, Toba graben to name just two.) I imagine the leading edge of the overriding plate is dragged down into the subduction zone to some extent due to the weak coupling of the plates on the boundary. This results in stresses and extension behind the boundary which is where you get your volcanism popping up.

By bruce stout (not verified) on 26 Oct 2009 #permalink

Fascinating article, but only just the first of a series of new puzzle pieces emerging which concern the cascades directly and volcanology/PT indirectly.

I think one thing people need to keep in mind is the significant changes that take place in both physical and reactive chemistry at depth. At 10KM we're talking about 1000's of bars of pressure, and temps frequently in excess of 1K kelvin. Throw continental sediment, oceanic crust and underlying mantle material into that (all three of which have differing chemistry), add oceanic moisture and other volatiles from depth, and you create a massive chemical reactor/distiller.

Time factors greatly into this as well, as given enough time and the differential density of materials, and based on chemistry what temp/pressure boundary at they will crystalize, you can get remarkably different melt chemistry and composition from what might be very similar original material. You can see a certain amount of this on a small scale with basalt flows where the melt has been sequestered long enough, and remained hot enough that some minerals could precipitate out and settle at the bottom of the flow. I would hypothesize that this could in part explain differences in chemistry seen in the output from the three volcanoes in question, assuming a common source of melt.

I actually would be surprised if there was *not* some cross-connected plumbing at depth in the Cascades, considering how massively fractured they are at depth by faults minor and major. Whether this would translate into a hypothesized single massive magma chamber is more doubtful.

Regarding Arch's comment, I don't see isolated magma reservoirs invalidating classic PT. There are a myriad of possible theoretical causes ranging from (just scratching the surface...) relict impact hot spots to mantle up-welling to isolated concentrations of radioactive material. Reading his comment, I have also never heard previously that there was a spreading center as well as a subduction zone under the cascades. There any references to this?

By Jeff Allen (not verified) on 26 Oct 2009 #permalink

Posted by: Fitz @ 1 "Bad science gets the headlines. Good science always triumphs in the end."

Unfortunately the public tends to remember only the headlines.

By Katkinkate (not verified) on 27 Oct 2009 #permalink

Erik -

I just read the Nature Geosciences paper and see nothing that claims a "giant magma chamber" nor a "vast swirling cauldron of magma" as you seem to be insinuating. Perhaps you should read the paper itself before claiming it is saying something it clearly is not.

From your "update" it is obvious you didn't even read the sensationalized media articles thoroughly before blogging away about the paper, touting that you (and others) have clearly debunked the paper and that it is "bad science".

How was the statement "the team estimates only 2-12 percent of the rock is actually molten"..."hidden"? It wasn't in fine print. It was right there in the article. The Seattle Times reporter wrote this article, not the authors of the paper. To insinuate that the authors of the paper were somehow trying to "hide" the conclusions of their paper is ridiculous! Are you even aware that in most cases, editors write the headlines for articles, not the reporters?

You claim that you (and others) have "picked apart" the paper, but clearly you haven't even read it! Otherwise, why would you be trying so hard to refute something Hill et al. haven't even concluded!

Furthermore, you claim that Hill "doesn't have a foot to stand on". Yet you clearly slant your blog to favor your own opinion on a paper you clearly haven't read...

For example, Steve Moran does claim that "the geothermal evidence for what Graham suggests is nonexistent" and that "there is also no seismic evidence for such a magma pool."

However, Olivier Bachmann, an extremely well-respected petrologist/geochemist from the University of Washington points out, "the lack of surface evidence is no reason to discount the possibility of a southwest Washington magma pool....adding that the geysers and hot springs in Yellowstone, Iceland and New Zealand are over much shallower magma pools than the Southern Washington Cascades Conductor."

George Bergantz, another well-known and well-published volcanologist from UW called the study the "best of its kind yet".

Also, if you had actually read the paper...it would be clear to you that the authors point out the MT data presented in the paper actually AGREES with seismic data from the area...not the opposite as Steve Malone claims.

Lastly, although it is certainly true that the idea of "partially molten crust - as an area of interconnected (or not) lenses of melt" has long been touted by volcanic petrologists and geochemists, it is also true that actual GEOPHYSICAL evidence of such zones of partial melt is, in fact, relatively new in the world of volcanology and entirely new for this particular region of the planet.

May I kindly suggest you actually become familiar with the latest (including geophysical) literature about partial melts and crystal-mush zones before claiming that something is "nothing new"?

I have to say that, to me, seeing a so-called "educator" and "scientist" spout off about things he hasn't even taken the time to read and subjectively presenting only one side of an argument doesn't look like "science" at all...but simply attention-seeking blogging at its worst.

Liquidhotmagma

By liquidhotmagma (not verified) on 27 Oct 2009 #permalink

Lets clear up a few misconceptions here:

(1) I did indeed read the article.

(2) What my problem is here is that every single article I read on this article says the study calls for "giant pools of magma" or "vast magma chamber", with quotes from the authors to back it up - there obviously was this implication in the paper even if not stated explicitly. Even the 2-12% melt comment wasn't exactly prominently displayed in the first paragraph of the articles. I'm not trying to imply that the authors of the paper were hiding the results, but clearly they were letting the media run with this idea that a big body of magma connects all the systems, which isn't what the paper ends up saying (at least not in those words exactly).

(3) Of course I have a slant on this - we all do (thank you postmodernism) and to claim otherwise is foolish. If you can convince me that their findings are more than just speculation on some fascinating data they've collected (and it is fascinating), I'm more than happy to change my view. That's science, in a nutshell. Controversial papers are good with that - a recent example would be something like Eichelberger et al., 2000 that spawned quite the discussion in the field - something we need every once in a while. That shouldn't stop me from being irritated by the way it is being presented in the media and the way the authors are touting it in said media.

(4) Before throwing Bachmann & Bergantz at me (two people I know quite well and respect greatly), I'm sure they have a much more intimate knowledge of this data set than I do - with all of a paper to examine. Bergantz' comment of it being "provocative" is just that - the research, likely technically done quite well - will provoke a response. It really doesn't pass judgement about whether he agrees or not. As for Bachmann's comments, again, he implies that the lack of geothermal activity doesn't preclude a pool, but rather that it doesn't negate it.

(4) I'll point you to Schmitz et al. (1997) in Tectonophysics that concludes that the Central Andes have exactly this type of midcrustal melt seen in seismic and gravity data. Again, the findings for the Cascades are new, but the concept is not. I'm perfectly happy with the idea of a network of partial melts - however, it is the presentation in the media of a vast and eruptable pool of magma that is troubling.

(5) Telling me to "keep up with the literature" is sort of laughable. I think about these problems every day and I would welcome any new data sets. However, if you talked to most people who work on magmatic systems, the popular conception that this work represents a giant pool of magma would be seen as just wrong. A network of lenses of (maybe) interconnected melt at depth? Yes, that seems believable, but a reservoir that feeds all three volcanoes? Not so much.

(6) A note to all: You know, I don't really care if you come after me about something I say on the blog. It is what it is: a platform for me to talk about volcanoes. These reports on this study definitely got me riled up, and maybe I should have been more careful with my words and criticism, I will give you that. However, I come out here using my real name, with all the good and bad that comes with it. If you want to disparage my ideas, character, cat, whatever, I would appreciate it if you did the same. It is much easier to trash someone when hiding behind the cloak of anonymity than face-to-face.

Also, if any of the authors of this work are out there and would like to be interviewed on the blog, I welcome you to contact me! I'd love to hear more and try to clear up any misconceptions that exist due to the media coverage of the article.

liquidhotmagma,
I saw the story on the front page of Yahoo news with the headline: Controversial Study Suggests Vast Magma Pool under Washington State.

If the authors did not imply that their study indicates a giant magma pool, then it is up to them to correct the misconception being presented to the public. It would be professionally irresponsible otherwise.

Dr. Klemetti dows a fine job in sorting out the fact from speculation in popular science articles. One only needs to look at the damage done by speculation in the media during the current pandemic. When the new H1N1 swine flus first became noticed in Mexico, media reports had the public thinking that it could be some type of impossible Hollywood epidemic that would kill on cantact and sweep around the globe in days. Because such unrealistic scenarios did not unfold, many in the public now misunderstand and underestimate the real dangers from this bug.

We just couldnt leave your website before telling you that I really enjoyed the quality information you offer to your visitors... Will be back soon to check up on new stuff you post!