Altered states (of lava domes) – Magma Cum Laude

Time for some shameless self-promotion – but also some research blogging. Last week I (finally) had a paper come out about my graduate modeling work on the hydrothermal systems and alteration in lava domes. (I’m sorry it’s not open access – I couldn’t afford it this time! But feel free to contact me if you want a copy.)

Basically, the rundown is this: Lava domes, like volcanoes in general, are big permeable piles of rock. This means that water can get into them, and (assuming the dome isn’t erupting and so hot that it immediately vaporizes all the water), the domes can form some sort of hydrothermal system. It’s not easy to see what this looks like, though. There are a few dissected domes we can look at, as well as domes that have collapsed recently, but not many, and people in the past were concentrating more on structure and eruptive processes than post-eruptive ones, which is what I’m interested in.

But why am I interested in them? My post on soggy volcanoes goes into this a little bit. Within a certain temperature range (around 100-250°C), hydrothermal alteration minerals like clays, alunite, kaolinite, etc. form in the presence of water (or steam) and acid. These minerals are a lot weaker than the original rock, and they can have different effects on an edifice depending on where they are and in what quantity. They could weaken it, making gravitational collapse more likely, or they could create impermeable zones that block fluids (liquids and gases, in non-volcanology parlance) from moving through it. Either of these things could happen at a dome that hasn’t been actively erupting for some time – perhaps even decades later.

Decades aren’t long in geological terms but that’s plenty of time for humans to forget the dangers that come of living near a volcano. Active eruptions are flashy and immediate and you can see what’s going on; aside from observing ground cracks or deformation, which are hard to do if you’re not actually on a lava dome or don’t have access to InSAR/Lidar/GPS, it’s a bit harder to monitor whether a dome collapse might be getting ready to happen.

So what does a geologist do when the domes are hard to get to, hard to look into, and just otherwise inhospitable field locations? We make numerical models based on what we know about material properties and heat and precipitation, try and think up plausible scenarios for all those things – and then let them run. The scenarios I investigated included a couple of different dome shapes (one confined by a crater wall, one perched at the top of a slope), and a couple of different ways heat might be delivered to the system. For reference, here are some examples of those types of dome “geometries” at Santiaguito, Guatemala and Unzen, Japan.

Unzen Volcano, Japan, 1991. Aerial view of northeast side of Unzen lava dome. Photo by T. Kobayashi, Univ. Kagoshima. Japan, November 2, 1991. (From the USGS Photo Library)

Figuring out material properties was one of the hardest things about this research. For one thing, it’s hard to characterize them in bulk, so we end up with a lot of measurements taken from hand-sample-sized bits of lava or pyroclastic deposits. For another, very few people have done this systematically for a variety of samples, which means you often have to pull things like density from one study, permeability from another, porosity from another, etc. Sometimes you get lucky and someone will do a whole suite of samples for, say, a region where geothermal exploration is being conducted, and that’s like hitting a gold mine. I’ve spent years building up my database of rock properties, and I’m still adding to it.

Modeling always involves assumptions and sometimes, in order to keep a model from crashing (or make it run efficiently enough that you can get results out of it in a useful amount of time), you have to make simplifications to your input information. In the case of my simulations, I had to assume that the heat was coming from a relatively cool source (200°C), that the interior of the domes had a relatively simple geometry, and that things like fractures were “averaged” into the material properties instead of explicitly defined (this is called taking a “continuum approach”). Basically, we ended up with domes that looked a little like this:

Schematic cross-section of the lava dome geometry we used as a basis for these models (this one is a figure from my dissertation). This is, of course, a simplification, and doesn’t account for things like fractures, faults, complicated feeder dikes, etc.

After setting up a number of scenarios that we thought were plausible for the systems we wanted to investigate, we ran them for a hundred years. In modeling terms, this means the model was solving equations for mass balance, heat transfer, and a whole slew of other things at every point on a predefined mesh for a number of timesteps, incorporating the effects of neighboring points as well as conditions along the boundaries of the model and things like precipitation. (Recreating reality takes a lot of computing power, and it’s really nice when you can run these things on large computing clusters.) Then we took the results of those models – particularly the temperature and liquid flux fields – and used them with an equation that gives an index of how likely hydrothermal alteration is to happen. (Basically, it’s most likely where you have a steep temperature gradient coupled with a high flux from one side of the gradient to the other – hot water flowing to a cool region or cool water flowing to a hot one.)

What did we find? In a nutshell, we found that alteration is most likely to occur where permeability contrasts exist in a dome, particularly between things like the relatively solid core and the blocky, broken-up talus. Now, this is kind of intuitive – water builds up when it hits impermeable things, like a hot but not-all-that-fractured core of a dome. If the core is cool enough that all the water doesn’t vaporize immediately, it’s available to participate in alteration processes. So, if you have alteration forming under the talus of a dome, that implies that collapses related to weakness in the rocks would mainly happen in shallow parts of the dome – limited to the talus and maybe a bit of what’s underneath. However – and this is speculative and a variation on what I’m working on for my postdoc – if you had enough alteration building up to seal the interior of the dome, and there was enough gas or liquid getting trapped inside, it could theoretically pressurize part or all of the dome and cause deeper, more catastrophic collapses.

A version of Figure 7 from my paper, with less-jargon-y annotations about where alteration potential is highest in my dome models.

Now, this would be really cool to prove, and that could be done with a combination of more complicated modeling, field methods and geophysical techniques. Altered rocks show up differently in geophysical surveys (they conduct electricity better because clays contain water, for example, and they’re less magnetic than unaltered rocks). And it is possible to go out and find dissected domes and map out the patterns of alteration they contain, although to my knowledge no one’s done this in a systematic way just yet (most mapping has been of dome structures). That would be something I’d like to keep working on the future, and especially using it as a way to gather more material properties (did I mention I’m still collecting those?)

But what I think it’s important to remember is that these models and calculations are some of the first steps in figuring out what’s going on in dome hydrothermal systems and how they affect stability. They can definitely be improved on, because models can always be improved on – that’s the nature of modeling, especially when people keep coming up with new and more flexible modeling methods. I’m doing that right now in my postdoc, and there are grad students back at the geology department at the University at Buffalo and other places working on it (from both a numerical and analog direction). One of the UB students will be presenting on it at AGU’s Fall Meeting, so you can even keep up with the progress that’s been made since I did this research.

And that’s what modeling in geology is all about – testing ideas, re-evaluating them when we have more information from those models or from other sources, and improving them. It’s a great reminder that science is work in progress, and that we always have more to learn about natural systems.