Displacement seismograms recorded by TERRAscope for the 28 June 1992 Landers earthquake (MW 7.3) are deterministically modeled using a forward, point-source summation technique. Although the data set is sparse, it was possible to robustly determine important rupture parameters such as gross slip distribution, rupture velocity, rise time, and total source duration. The relatively simple approach lends itself to rapid application following large earthquakes, provided that a catalog of Green's functions appropriate for the region is available.
The fault used in the modeling of the Landers mainshock has a length of 70 km along strike and a width of 15 km along dip. A model was found in which the distribution and amplitude of slip at the surface matches the observed surface slip and provides a very good level of fit to the seismic data. Seismically, the Landers earthquake is characterized as two subevents. The peak slip of the first subevent is 10 km north of the epicenter and the second is 40 km northwest along strike from the epicenter. The seismic moment is distributed as 2 × 1026 dyne-cm to the first and 6 × 1026 to the second subevent, respectively. It was assumed in our modeling that the distribution of seismic moment along strike was the same at all depths. This assumption implies that slip at depth is 69% of that at the surface as a result of differences in the material properties in the velocity model. the sensitivities of the source model to rupture velocity and dislocation rise time were examined. A rupture velocity of 2.9 km/sec (80% of the shear-wave velocity) and a rise time of 1 to 3 sec were found to satisfy the data. The rise time is only a fraction of the total source process time of 24 sec, and implies that slip on the fault occurred within a narrow band (3 to 10 km), at any instant during the rupture.