Application of Grflt12s (Excel-Ver.) programs:

1992 Landers EQ Simulation using the Wald and Heaton Model (1994)

Y. Hisada, Aug. 4, 2006

0. References

a. The source data of the 1992 Landers earthquake model

We used the source data of Wald and Heaton (Combined Model) to reproduce the strong motion at the Lucerne Valley Station.

b. Simulation results using this program

c. Theories of this program

for the fling step.

for Green’s function.

1. Input Data

As shown in Figure 1, the 1992 Landers earthquake model consists of the following three fault models. In order to obtain the strong motion at the Lucerne valley station, the three results have to be superposed.

a. grflt12sx1-CEF.csv: Camp Rock/ Emerson Fault Model

b. grflt12sx1-HVF.csv: Homestead Valley Fault Model

c. grflt12sx1-JVF.csv: Johnson Valley Fault Model


2. Run the Programs

a. Compile and run phs3sQx.f to obtain phase velocities of surface waves.

To plot phase and group velocities of the Love and Rayleigh waves, use the output csv files, Ldisper.csv and Rdisper.csv, respectively. You can also plot the medium response of the Love and Rayleigh waves, use the output csv files, Lmedres.csv and Rmedres.csv, respectively. See Figure 2, for examples for Love wave.


b. Compile and run grflt12sx1.f to obtain strong motions in the frequency domain.

See Figure 3 for the coordinate system and fault parameters.


c. Compile and run grfftspx.f to obtain strong motions in the time domain.

The Fourier amplitude data are stored in amp.csv, and the waveform data are stored in wave.csv.

3. Simulated Results

As an example, Figure 4 shows the velocities at Lucerne valley using grflt12sx1-CEF.csv (Camp Rock/ Emerson Fault Model). Figure 5 shows the displacements, which are integrated at the time domain.

4. Comparison of Simulation Results with the Observations

We obtain the final simulated strong motion at the Lucerne valley station by summing up the results from the three fault models (grflt12sx1-CEF.csv, grflt12sx1-HVF.csv, and grflt12sx1-JVF.csv; see Figure 1). Figures 6 and 7 show the comparison between the simulation and the observed records; the records were corrected by Dr. Iwan of Caltech, and can be downloaded from COSMOS Virtual Data Center (

Note that the data are rotated in the fault parallel (N40W) and normal (N130W) components (see Figure 1). And the observed records are original, while the simulations are high-cut filtered tapering from 0.8 Hz to 1.0 Hz. As shown in Figures 6 and 7, the fling step and the forward directivity pulse are dominant in the fault parallel and normal components, respectively.