The format follows that used by Herrmann in his 2011 Seismological Society of America presentation in Memphis.

- SSA Presentation (pdf)
All available broadband data are reduced to ground velocity (m/s) and narrow band pass filtered at center frequencies,

*f*, of 0.25, 1.0, 3.0 and 8.0 Hz. Defining corners_{c}*f*=_{L}*f*/1.414 and_{c}*f*=1.414_{H}*f*, the filter consisted of an 8-pole highpass Butterworth filter and an 8-pole lowpass Butterworth filter, implemented using the_{c}**gsac**commandhp c ${FL} n 8 lp c ${FH} n 8

The peak filtered velocity is then saved for the plots.The observed values are then compared to the predictions of the Atkinson Boore (1995), Frankel (1996) and the Atkinson (2004) models for eastern North America. I used the random vibration theory program

**rvcal**which consists of a wrapper to provide input to David Boore's**smsim**program. The program is run using the command line:rvcal -V -B -MODEL ${MOD}.mod -MW ${MW} -FN ${FREQ} -R ${DIST} -OA -DT 0.01

where the SHELL parameter MOD is the propagation/source model file, MW is the moment magnitude, FREQ is the filter frequency, DIST is the desired distance. The*-B*and*-V*flags indicate that the peak filtered velocity is desired.Plots are made for each of the 3-components of ground motion, vertical (Z), radial (R) and transverse (T):

### Comparison of Z component

### Comparison of R component

### Comparison of T component

Propagation models

Source code

To compile using gfortran,gfortran rvcal.f -o rvcal

For the commandrvcal -V -B -MODEL F96.txt -MW 5.7 -FN 0.25 -R 100 -OA -DT 0.01

The output is100.000 5.700 0.126E-02 vmax 0.25 0.05 T 0.711E-03 0.191E-02 0.29

With annotation the output isDist(km) Mw Peak(m/s) Peak Freq Damp Butter Max05 Max95 Freq20 100.000 5.700 0.126E-02 vmax 0.25 0.05 T 0.711E-03 0.191E-02 0.29

where Damp is a place holder and has no meaning for a Butterworth filter, Max05 and Max95 are the 5% and 95% bounds on the peak value and Freq20 is the frequency of zero crossings.## Discussion

Several observations can be made from the SSA presentation and from the data for this earthquake:- Data are required in the 0 - 100 km distance range. This requires a concerted effort to instrument areas where there is an expectation of a future earthquake, e.g., the Trinidad Colorado region where M 4-5 occur annually (e.g., 2011/08/23 05:46:19 Mw=5.2 earthquake or the Wasbash Valley of Indiana and Illinois)
- Although not designed for diffenences in high frequency ground mmotion, the Frankel 1996 model does well for thrust (2010 Quebec and 2011 Virgina) earthquakes and strike-slip (2008 Illinois, 2010 Oklahoma and 2010 Arkansas) earthquakes.