Peering inside volcanoes: A guest post by Oliver Lamb

Picture yourself here. You are a scientist monitoring an active or erupting volcano. In the past, this volcano and others like it have produced lava domes which are prone to suddenly and violently explode or collapse to produce devastating pyroclastic flows, often without warning. Unfortunately, it's this sort of behaviour by Sinabung in Indonesia that recently took the lives of 14 people (see the video below for a spectacular example from Sinabung). So what can you, as the scientist, use to work out when the growing dome is going to explode or collapse next? What is already known is that the transition from quiet to violent activity during dome growth is because of a complex range of processes, particularly before explosions.

Fortunately for scientists like me, active volcanic systems produce a wide range of seismic signals prior to and during an eruption and these signals are easily picked up by seismometers deployed around the system. For my PhD at the University of Liverpool I am carefully analysing the seismicity from recent or ongoing dome-forming activity, focusing particularly on any signals that can tell us something about the conditions within the volcano before explosions or dome collapses. Hopefully by the end of this project, I will have looked at seismicity from at least four different active volcanoes: Volcan de Colima, Santa Maria, Mt Unzen and Mt. St Helens.

'Drumbeats' recorded over a 24 hour period at Mt St Helens, 15 November 2004

To give you an idea of the size of the dataset for this project, I've included the snapshot above of the seismicity from Mt St Helens during its eruption from 2004 to 2008. What you're looking at is the seismograph from a single day during this eruption, and each of those peaks is an individual volcanic earthquake. Now imagine trying to analyse each event, every day, every month,  for well over a year, then you have a gigantic number of events to look at! Thankfully, at least for my sanity, we have tools which we can use to automatically count all these events and measure different traits about each event. These traits include how big the event was, how long it lasted for, how long was it since the last event, the frequency content and much more. With these tools, we can now begin to track how these events change over time and begin to get a better idea of what's going on as the eruption progresses. I have already used the tools on seismicity from  Mt St Helens, and by my latest count the volcano experienced well over 500 000 seismic events from November 2004 to March 2006, an average of around 50 events per hour. I hope that emphasises how large the dataset is!

Left: Volcán de Colima, Mexico, during more peaceful times, taken August 2012.
Centre: Mt Unzen, Japan, nearly 20 years after it ceased erupting. (Credit: J. Kendrick, Liverpool).
Right: Santiaguito, Guatemala, taken December 2007. (Source: photovolcanica.com) 

But the project doesn't stop there. There are plans afoot for a whole raft of other analytical methods to be used. Methods such as locating each event as the eruption progressed, looking for events which may have come from the same source, and then some statistical analysis on top of that. Eventually, armed with all the geophysical results, my project will then begin the experimental phase. During the experiments, I will be placing samples into similar pressures and temperatures as we think the seismicity is being produced in, and break the samples apart and record the resulting acoustic emissions. The current idea is that by the comparing the experimental and volcanic waveforms, we can then help figure out what's going on inside the volcano during a dome-forming eruption.

The author, on the south flank of Volcán de Colima in July 2012

There you go, there's my project in short. For those of you who don't know me, I'm Oliver Lamb and I am a first year postgraduate research student with the Department of Earth, Ocean and Ecological Sciences at the University of Liverpool. If you have any questions or comments about what I've written here, then please feel free to get in touch with me via e-mail, olamb245@gmail.com, or find me on twitter, @olamb245. Thanks for reading!