A view across the Santorini caldera. The newest eruptions in the caldera can be seen on the right on Nea Kameni.
This appears to be a week of media interest in new journal articles. Earlier, I discussed a study that claimed that volcanoes were the cause of the onset of the Little Ice Age. Now, we have a study in Nature that discusses the magmatic events that lead up to the Minoan eruption at Santorini – a fairly timely topic considering the rumblings there – that has gotten the media’s attention.
Now, I’m not going to pick apart this paper by Timothy Druitt and others as such – the study, called “Decadal to monthly timescales of magma transfer and reservoir growth at a caldera volcano“, is actually quite solid. The long-and-short of the study is that they examined plagioclase feldspar crystals looking at the zoning of different elements in these crystals (see below).
There are two main pieces to the study. First, if a crystal grows in a certain magma, it will suck in certain amounts of different elements – some are major constituents of the minerals. In plagiolclase feldspar, we can define the “An” of a crystal by looking at proportions of Ca and Na in the crystal (high “An” means high Ca – closer to the perfect feldspar endmember anorthosite). The “An” can then tell us if a crystal came from one type of magma or another (see figure below). If there are low abundance elements in the mineral, like strontium, magnesium and titanium in plagioclase feldspar, then the amount of the element is controlled by the partitioning of the element between the liquid magma (melt) and the crystal. This is what geologists call the “partition coefficient” – or how likely is an element to want to be in the crystal or melt. The partition coefficient will change depending on the overall composition, pressure and temperature of the magma and crystal, so crystals in different magmas will suck up different amounts of these elements. This gives them distinctive compositions depending on the magma in which they grew – a “compositional signature” so to speak. (Note: I looked that is in zircon from the Okataina Caldera in my Earth and Planetary Science Letters study from last year).
Part of Figure 1 from Druitt et al. (2012) that shows the zoning of plagioclase feldspar from the Minoan eruption of Santorini.
The second piece is diffusion. Elements in crystals will diffuse back into the melt (or vice versa) if there is a large compositional gradient between the crystal and the melt. So, throw a crystal of one composition into a new magma of another, the elements will begin to exchange over time starting at the rim of the crystal. So, assuming specific thermal parameters and compositional gradients, you can use diffusion as a clock – how long has the foreign crystal been exposed to this new magma based on how much diffusion of certain elements has occurred. Now, different elements have different abilities to diffused based on their size and charge, so you need to choose wisely.
The Druitt et al. (2012) study used these two petrologic characteristics of minerals and melt to determine two main conclusions: (1) the magma erupted from Santorini during the Minoan eruption in ~1600 BC was a mixed magma and (2) the intrusion that “got the ball rolling” towards the Minoan eruption and the subsequent mixing happened geologically quickly – in the the timescales of a century to a few months. Now, there is a big caveat not mentioned in the study to this second point. One quandary we have in petrology is that when we look at timescales of processes inside magmatic systems, diffusion profiles like the kind used in this study imply events occur much faster than if you try to date mixed crystals using radiometric elements (such as Ra, Th and U). This disconnect has not been resolved, so I would say that the timescales suggested by Druitt et al. (2012) are minimum timescales for the intrusion and mixing, not maximum. This will be important later on.
You might have noticed a lot of the media coverage about this study is claiming things like “supervolcanoes offer 100 year early warning” and “they may be predicted”. That is never said in the study. The authors do discuss some of the ways that this recharge/mixing might be manifested once the events have begun – interestingly not as “bulging” but rather “sagging” of the bottom of the magmatic system as the magma fills in, so uplift at Santorini might have been minimal. They actually predict that sinking of the land surface might be more likely rather than the classic St. Helens-1980-style bulge.
However, what I see as the biggest problem in this “early warning” claim is that it might still not be easily detectable – what if their timescale is off by even a factor of 2, so it takes 2 centuries to lead to an eruption? Human monitoring of an event 200 years in the making might be very problematic. Secondly, this intrusion isn’t a big event at 100 years than than over, it is growth and mixing over that century with a rapid culmination only months before the eruption according to Druitt et al. (2012). Whether or not this is detectable by current monitoring methods is unclear as well. The authors are right about one thing: “Long-term monitoring of large, dormant caldera systems, even in remote areas of the world, is essential if late-stage growth spurts of shallow magma reservoirs are to be detected well in advance of caldera-forming eruptions.” However, as usual, the many in the media has boiled down their research into meaningless copy that both misses the point of the research but also recklessly mischaracterizes the ramifications.
Image 1: Santorini caldera. Image by Navin75/Flickr.
Image 2: Figure 1 from Druitt et al., (2012)
No comments:
Post a Comment