The long-term goals for our group are to decipher the physics of earthquakes and the ramifications for ground motions.

In FY2004 Earthquake Source Physics PI’s tackled problems in 4 areas:

Computationally simulating rupture dynamics to elucidate earthquake physics

Rice/Dmowska (geometry, friction, lab)	Lapusta (multi-cycle)
Beeler & Tullis (lab)	Oglesby & Xu (multi-cycle)
Archuleta (energy)	Shaw (multi-cycle)
Day & Harris (geometry, materials, friction)
Purvance/Anooshehpoor/Brune (lab)
Harris & Archuleta (code-comparison/validation workshop)
Harris/Aagaard/Ampuero/Andrews/Archuleta/Day/Dunham/Lapusta/Oglesby/Olsen/Pitarka (code-comparison/validation)

Reference Earthquakes Database

Mai (many earthquakes)

Beroza & Olsen (Landers)

Investigating large- vs. small-earthquake physics

McGuire (rupture velocity)

Using earthquake triggering observations to decipher earthquake physics

Brodsky/Felzer (stress shadows)

Highlights from this research effort are as follows:

As part of our research into rupture dynamics, we in the Earthquake Source Physics Focus Group have 3 groups investigating the multi-cycle fault problem, where researchers are modeling the long-term dynamics of earthquakes over multiple earthquake cycles. This research area overlaps with Fault Systems focus group goals. Each ESP group is examining the problem from a unique perspective, by including complex friction formulations, viscoelasticity, and fault geometry in varying amounts. Additionally, our researchers are continuing their work on the dynamics of single rupture events. Investigators are tackling a range of problems, including the effects of fault geometry, material complexity, and friction formulation. Our most visible current effort is our code validation/comparison exercise, which now involves more than 16 people, including a significant number of students and postdoctoral researchers. We hope to show that when we use similar assumptions about fault geometry, materials, friction, stress, we will produce the same earthquake source physics results, regardless of the computational method. Our goal is to produce synthetic seismograms that match each other, and are validated by experimental and observational studies. The hope is that our validated methods can then be used with confidence by the engineering community. So far we have met in 2 workshops (November 2003, >30 attendees; September 2004, >50 attendees), and our next meeting will be November 2004. We have compared results for two benchmark problems (The Problem, Versions 1 & 2), and are about to embark on The Problem, Version 3, before moving on to validation with both a foam rubber experiment (and other lab experiments as they become available) and Parkfield as 2005 goals. Our collaborative project has received worldwide attention, with scientists from other countries using our findings to benchmark their codes.

In ESP in 2004 we have had 3 groups comparing laboratory results with numerical simulations of dynamic rupture. These include the Brune group, which is simulating rupture in foam rubber that will be used as a validation exercise for the code-validation group; the Beeler/Tullis group, which is observing dynamic rupture in rock; and the Rice group, which is performing numerical simulations of the laboratory homolite rupture experiments of the Rousseau/Rosakis group. This is an exciting time in the overlap between lab and computational simulations of rupture dynamics.

In 2004 our focus group began its work on a reference earthquake database, whose goal is to provide observations and models derived from the observations in one easy-to-access location. This database will serve as a testbed for hypotheses about earthquake source physics. PI Mai, on shoestring funding, put together the first part of this, with a web-accessible database on finite-source models from worldwide earthquakes.

In the coming year, FY05, our group hopes to continue some of these subprojects, all with the goal of deciphering earthquake source physics. The collaborative code validation exercise will expanding to include more PIs around the world (at negligible SCEC cost), and our group aims to participate in the NGA-H (see Implementation Interface section of this report) project when it comes on line. Our multi-cycle simulations show much promise, and the occurrence of the 2004 Parkfield earthquake should provide ESP with much eagerly anticipated data. The ESP group expects to benefit greatly from this earthquake, which is the best-recorded event in history.