When the earth shakes, seismologists get excited. There's nothing like a good set of observations taken right in the rumbling ground to get the patterns of strong ground motion. These kind of data are especially valuable to researchers developing computer simulations of earthquakes that help us understand and predict how quakes behave at fine scales.

Apostolos Papageorgiou, civil engineer at Rensselaer Polytechnic Institute, has used exceptionally good quality observations from two recent California quakes to refine models that recreate, with a very high degree of accuracy, the impact of these quakes at the level of local shaking. Papageorgiou runs his simulations on the IBM RS/6000 Scalable POWERparallel Systems (SP) supercomputer at the Cornell Theory Center (CTC), allowing him to incorporate local geological conditions into the multidimensional models.