The Dilemma of Scale
Ethan Coon
Columbia University
Los Alamos National Laboratory
Story by Jacob Berkowitz
As every rock climber knows, survival is in the details. You can’t think of the 3,000 foot-tall rock face while you’re clinging to it.
The only thing that matters is the half-inch-wide granite lip onto which your feet are wedged. It’s a perspective born of years of rock climbing that Ethan Coon brought to his summer practicum at Los Alamos National Laboratory in 2006. The DOE CSGF fellow took a page from his climbers’ manual to make more physically accurate computer simulations of Earth processes, putting hand-grip level details into the big-picture view. And it turns out that the attention to detail that keeps climbers safe can help us better predict the consequences of earthquakes and squeeze more precious petroleum out of aging reservoirs.
As a student in applied mathematics at Columbia University, Coon works with those at the forefront of merging computational modeling and earth sciences. That includes his thesis adviser, Marc Spiegelman, an associate professor jointly appointed to Columbia’s renowned Lamont-Doherty Earth Observatory and the Department of Applied Physics and Applied Mathematics.
In his doctoral research, Coon is creating a new mathematical model of the two-dimensional geometry of rock faults in order to improve computer models of earthquakes.
“Many standard physics-based computational models, including geological ones, are inherently bad at capturing small-scale effects,” Coon says. “Like a postcard view, they can only show the broad strokes, while features smaller than these strokes are blurred out or completely missed.”
Solutions for saturation of an oil-reservoir simulation on two different meshes. Click image for larger version and more information |
It’s not from lack of trying. Computational scientists continually work at putting the fine-scale into the big picture to create more detailed models. But they face a perennial hurdle: the increased time and cost it takes to run these more complex simulations.
To tackle this challenge, computational scientists turn to approximations — they average out fine-scale details over the big-picture view. In modeling geological formations, this could mean averaging out tiny layers of sedimentary rock that compose a miles-long and miles-deep oil reservoir. Put into a climber’s perspective, it’s like averaging the number and locations of rocky hand-holds, without knowing exactly where each one is — a situation that can render the model useless as a real-life guide.
