The speedy magma of Chaiten (and it's still going!)

Chaiten has made it back into the news in the past couple days, both with new events at the caldera and with findings from the initial blast in May 2008. Here goes:

Chaiten erupting in 2008.

Third Dome Spotted
The latest USGS/SI Weekly Volcanic Activity Report mentions that over the last week, Chaiten experienced what was likely a significant dome collapse of one of the two domes growing in the caldera. People living close enough to the volcano to see the ash plume noticed it became larger and darker on September 29th. Afterwards, visual observations of the caldera by air confirmed that a dome collapse likely occurred and a third dome has begun to form in the SW part of the caldera. A new series of blast craters were also noted on the original two domes. Seems that the eruption is still going strong and will likely continue to fill the caldera with new rhyolite for some time to come. You can see all the details of the last few days events on the Volcanism Blog's translation of the latest SERNAGEOMIN Chaiten update.

The Rapid Ascent of Chaiten Magma
The other news this week was the report in Nature that the rhyolite magma from the initial eruption of Chaiten on May 1, 2008 travelled much faster than most geologists thought could be possible for rhyolite magma. Rhyolite is a very sticky, viscous magma that tends to, when erupted, travel quite slowly. The analogy to understand how slow it moves (when flowing) is that you can outrun a basaltic lava flow coming at you, but with a rhyolite, you could likely build a house and live comfortably for many years without worrying about being overrun by the rhyolite lava (*however, the explosive nature of rhyolite and its propensity to flow pyroclastic flows makes this scenario, well, a little less ideal). The stickiness of rhyolite magma suggests that when it travels through the crust, it should travel slowly - or, at the very least, slower than most basaltic lavas.

This study by Dr. Jonathan Castro and Dr. Donald Dingwell found that the rhyolite magma from Chaiten travelled through the crust at extremely rapidly. From the initial precursor earthquake, it was only ~24 hours before the eruption started. Based on their analysis of samples of the rhyolite lava, they have determined that the rhyolite was moving from ~5 kilometers below the caldera. By studying the minerals in the rhyolite and their compositions, you can determine the pressure and temperature of the formation of the lava. These results can be supported by running experiments - can you reproduce the same mineral and compositions in your lab by heating and pressuring samples of the same rhyolite. Castro and Dingwell were able to constrain the depth of the magma chamber from whence this rhyolite came to 5 +/- 0.5 km depth.

Now, you could just say that, alright, it started at 5 km depth and made it to the surface in 24 hours, so it was traveling at ~0.2 km/hr. Of course, is there evidence in the minerals to back this up? When lavas erupt, the pressure is released and minerals will tend to grow rims of different compositions and shapes (due to this depressurization). You can use this to estimate how quickly the magma ascended - and that is what Castro and Dingwell did. They found that the ascent rates from their experiments on Chaiten rhyolite were even higher - closer to ~1.8 km/hr. This rate is a veritable cheetah when it comes to sticky rhyolite magma. It suggests that the magma took ~4 hours to reach the surface from the magma chamber, which is a shorter interval than the precursor earthquake might imply.

What is the upshot of all this? Well, it means that we could have little notice for another Chaiten-like eruption at a rhyolite volcano. These eruptions are, luckily for non-volcanologists, very uncommon (which is why Chaiten is so fascinating). Worldwide, Chaiten is like the only rhyolite dome eruption of this magnitude in the last 200 years. Castro and Dingwell suggest that we might want to monitor some of these large rhyolite systems a little more closely (e.g, Medicine Lake or Newberry Caldera in the U.S.) because we might not have much time to react when the signs of eruption start.

Do you have a burning question about the Chaiten magma you'd love to be able to ask Dr. Castro? He has kindly offered to answer some questions about Chaiten and his research for Eruptions readers. Send me your questions at


and I'll choose some of them for Dr. Castro to answer. I'll post the interview and the answers to your questions here on the blog.

More like this

5 km in 24 hours is 0.21 km/h not 0.02 km/h as you've stated.

5 km at a speed of 1.8 km/h takes 2.78 hours, much less than the 4 hours you state


Thanks for catching the decimal point issue for the 0.2 km/hr, Nick. However, the 4 hour estimate is taken from the Castro and Dingwell paper itself, so I'll stick with that value.

This might seem bit of an amateur question, but what are the chances of Chaiten eventually forming theclassic volcanic cone shape over the calsera given the composition of the lava?

(Have you noted the seismic activity in the Solomon Islands? What could be happening there)

By Stephen Tierney (not verified) on 08 Oct 2009 #permalink

Worldwide, Chaiten is like the only rhyolite dome eruption of this magnitude in the last 200 years.

I thought that Katmai/Novarupta was rhyolitic. Isn't Novarupta a rhyolitic dome? I did read that there is a mixed magma composition at Katmai.

I got that partly from this abstract:

Katmai volcanic cluster and the great eruption of 1912

which says:

"Rhyolite erupted only in 1912 and is otherwise absent among Quaternary products of the cluster."

I will have to point out that finding this could make me sound like an expert: I remembered that Katmai was rhyolitic but it also says there's some "crystal rich dacite" and "crystal rich andesite" and some rhyodacite -- which of course makes perfect sense erupting from a from subjacent andesite-dacite mush, right? Goes without saying.

To Nick´s point--Our deompression experiments best reproduced what we see in the natural Chaiten pumice (ie., the crystal textures and compositions) when we ran them for about 4 hours; this is a ¨simulation¨ and marks a minimum time that the magma needed to travel from storage to the surface. It is quite possible that the magma moved faster, but not likely slower.

There could be many reasons the EQ duration exceeded the inferred ascent timescale, things like prepatory fracturing do to gas release, or volcano-tectonic deformation.

Thanks Erik--you gave a great summary of the work!


By Jonathan Castro (not verified) on 08 Oct 2009 #permalink

Volcanoes have a nasty habit of changing their magma chemical composition at 'no notice'. The Maderia volcano, now dormant, is over a hotspot and has changed from basaltic to andesitic in its life.

By John Marshall (not verified) on 09 Oct 2009 #permalink

Oakden: you've trumped me, I was going to follow Erik's invitation and ask Dr Castro if his work might have relevance to the Novarupta event, and help to unravel the Devil's Own magmatic plumbing system under the area -in particular, how the (mixed-but-mostly-rhyolite) magma moved laterally underground from Katmai to Novarupta some seven km (I think) distant.

And there's been at least one other rhyolite eruption in recent times which might be worth re-examination, St Andrew Strait in 1953-7, which produced a spectacular obsidian dome/lava flow complex (wonderful pic on the GVP site)

Mike--good idea to look for parallels b/t Chaiten and Katmai/Novarupta. I believe there was a report at last year's AGU meeting that suggested magma flowed up quickly at Novarupta (or sideways?).

In addition to the St. Andrew Strait (or Tuluman islands) eruption Mike mentions, there may have been a rhyolite eruption at Daddahu, Africa in Sept. 2005; This event was a small fissure eruption that produced minor ash fall and a small dome. We are still waiting to see a chemical analysis that proves it's rhyolite, however.

The Islands of Vulcano and Lipari, Aeolian Islands, were also the site of rhyolitic activity in early historical times, but certainly no scientific observations were made.

All that said, since Katmai, I believe there has been no significant (subaerial) explosive rhyolite eruption until Chaitén woke up.


By Jonathan castro (not verified) on 10 Oct 2009 #permalink

It is true that Novarupta/Katmai was a rhyolitic eruption but nearly all of it was explosive, the Novarupta lava dome is a tiny thing in comparison with Chaitén's huge dome. AND Chaiten's is the first rhyolite eruption in all of history to be observed with modern scientific methods and knowledge. So in any case it is pretty much unprecedented; I believe that few dome eruptions have been as vigorous as this one in terms of lava extrusion rate.
But one other quite massive recent dome-building eruption is that of the Soufrière Hills on Montserrat (Caribbean), which has been a bit out of the news lately; this Saturday 9 October 2009 it was confirmed that new dome growth is occurring on the south side of the dome formed in the past 14+ years, above the White River. That's the area where a huge piece of the dome plus the underlying volcano flank came down in a debris avalanche and major pyroclastic flows on 26 December 1997. The Montserrat Volcano Observatory has a brief report and one single photograph:

Questions for Dr. Castro: In the Chaiten research, what challenges were the most limiting, and how did you overcome those challenges? What were the advantages of your methods to overcoming the challenges, and how might your methods benefit future work?

Maybe Chaitén doesn't fit in the "normal scheme" of volcanic behaviour: The viscosity of rhyolithic magma depends largely on its water content and temperature. Basaltic lava, ascending on fractures interconnected with hot Lower Crust/Upper Mantle material, normally flows faster due to its temperature of ~1,200°C and a more or less well defined freezing point, while rhyolithic magma, stemming from crustal material submitted to decompression, tends to be relatively "cool" (~700 °C). Its content of overpressurized water (subduced with the ancient sediments and released by crystallites changing their composition during anatexis and pressure change) occupies the role of a flux agent.

So far we find it in the standard theories of Crustal Dynamics.

To explain the high flux rate, four hypotheses can be assumed:

- The magma is hotter than the "normal" ~700 degrees Celsius, which implies an effective heat source further below (a basaltic magma chamber, f.i., or a chamber with differentiated magma - basaltic or andesitic composition in the lower part, composition enriched in silicic acid in the upper part, the latter feeding the chambers above at ~5 km below surface).

- The magma contains more hypercritical water than "usual" (but I suppose there is an optimum for lowering the melting point so that a surplus of water won't have any impact on the FRP) - maybe it could change fluid dynamics of the magma if there's no possibility to ascend separately forming hydrothermal fluids. At least, the observations of the ongoing eruption noted large quantities of steam in Chaitén's column.

- The pressure gradient between chamber and surface is unusually high (which I think is less probable - pressure corresponds to the thickness of overlying crust and some tectonically induced horizontal components, which would have to be transmitted by the Liquiñe-Ofqui Fault System and could be traced by a certain amount of high pressure xenoliths in the magma).

- The friction caused by the ascension through fractures and the vent is lower than "normal" - a contribution that would be outweighed by the inner friction of ascending magma.

Here's my simple question: Which of the above hypotheses (or others) is actually the favoured one, and why?

By Jared J. Myers (not verified) on 19 Oct 2009 #permalink

What type of volcano is Chaiten, type of magma, and did any one get injured or died?

Hrmm that was weird, my comment obtained eaten. Anyway I wanted to say that itâs nice to know that someone else also talked about this as I had trouble finding the same info elsewhere. This was the primary place that advised me the answer. Thanks.