Monday, July 27, 2009


As evidenced by the vague conference-y posts on my Twitter account a few weeks ago, I was indeed at a conference from 22-26 June: the Numerical Modeling of Crustal Deformation and Earthquake Faulting workshop, in Golden, Colorado.
(I started writing this up right away, and then my computer crapped out. It took me this long to feel motivated to rewrite the post. Oy.)

This is a small conference (they capped participation at 60 people), and is just as oriented around collaboration on modeling problems using Computational Infrastructure for Geophysics software as it is around science talks. (It was not poster-oriented at all, though a bunch of us did bring posters. They were taped on the walls around the conference room, but not many people seemed to look at them; they instead used poster session time for software tinkering.) There were people with a wide range of specialties in attendance, from people focused specifically on earthquake physics to engineers working in plasticity to mathematicians who had only recently started delving into earth science applications, not to mention all the code-focused people. Among the people I spoke with, it seemed like I was one of the few there who were working specifically on rupture dynamics, which was an interesting change from the groups of people I spoke most with at SCEC and AGU.

I admittedly felt pretty out of place for the code-focused parts of the workshop, which filled the afternoons of the first four days, and the entirety of the fifth day (though I spent much of the fifth day doing airport things instead). Though I have been running lots of fault models over the course of the past year, they have all been with only one mesher and one finite element code, both of which were written from the ground up by a former student of my advisor's. It's a very good code for what I'm doing, and part of the reason we've been using it exclusively is to get it out there more. But I have no experience using other code yet, nor in more than the slightest tweaking of the code we have. My only codewriting experience was a C++ class that I took in 1999 (I'll be taking more in the future, though). As a result, all the nuts and bolts coder discussion went right over my head and made me doubt myself about being at the conference at all. I did want to participate in the tutorials for the code more actively, though, even if none of the things presented actually included friction and dynamic rupture at the time being. (I understand these will be included in future versions.) At this point, I ran into the problem that neither mesher would work on my laptop, nor would one of the physics codes. Turns out that I needed to be running Leopard, and I was still on Tiger, so I had to look on at other people's progress instead of poking around on my own. (I decided to upgrade to Leopard when I got back to Riverside, and as luck would have it, I got a bad disk. This ended in me having to wipe my hard drive and install from a different copy.)

But for all the code stuff left me feeling like the newbie that I am, the science talks totally made up for it. In general, it was very useful to see how many other ways people are using finite element models to examine earthquakes and crustal deformation, since I was taught about them specifically in the context of rupture mechanics. One that was particularly exciting to me involved a different fault setup in terms of type and geometry, but ultimately a similar sort of stress barrier condition to the models I've been running all year. The shape of the curves delineating whether or not the rupture propagated through the barrier for a given stress case was very similar to the shape of the curve I found for whether or not rupture propagates through a given type of stepover. (Not going to say more about that, though, since neither paper is in press yet!) I was also particularly intrigued by a talk involving geologically-derived information on interactions between the San Andreas, Garlock, and Eastern California Shear Zone; this is the kind of material I'm already reading up on in anticipation of what I'll be doing after the Master's stuff. In addition to these, there were several talks on subduction, several on effects of plasticity, one on the East African Rift, one on Mt. Redoubt, one on compliant zones around Mojave faults, and a couple more specific to small-scale rock mechanics. Definitely a good eye-opening representation of what people are modeling beyond the coseismic part of the earthquake cycle!

So, on the whole, a good conference, even for the stress of not being a codehead. I'm hoping that, if the department thinks I should go again next year, I'll also have more experience with code-specific matters. Even if I'm not actually messing with the code, I should at least have experience with several more programs by then, and I know I'll be able to run their code now that I've upgraded my OS.

No comments: