Introduction

This seismic event occurred in the metropolitan Chicago region on November 4, 2013 at 18:35 UTC (12:35 CST). It was widely felt

Did you feel it map

The USGS/NEIC source parameters are

2013-11-04 18:35:33 UTC

Parameter Value Uncertainty
Magnitude 3.2 mb_lg 0.05
Location 41.790N, 87.882W Not Specified
Depth 0.0 km 2.0 km
Number of Stations Used Not Specified
Number of Phases Used 27
Minimum Distance 46.6 km (0.42)
Travel Time Residual 0.55 sec
Azimuthal Gap 32
Review Status MANUAL
Event ID usb000ksfx


Group Velocity Analysis

After converting all broadband recordings to ground velocity in m/s, I used the program do_mft to determine group velocities and spectral amplitudes. The stations that provided useful spectral amplitudes are shown in relation to the epicenter in the next figure.

na.png

These stations are well distributed in azimuth about the epicenter. The spectral amplitudes were measured with periods of 0.5 to 5.5 seconds.  This are very short periods. For similar studies of larger events, periods in the range of 5 - 50 seconds are typical.  The use of low frequency spectral amplitudes yields a robust estimate of the source mechanism.

The first step was to define a velocity model for use in the source inversion. This was difficult because the dispersion in the period range is very sensitive to variations in shallow geological structure. The dispersion to TA stations L44A and M44A was used to improve the velocity model.  These stations are roughly 45 km north and south of the epicenter and had very good surface waves.  Using the program surf96, I started with the CUS velocity model and replaced the upper layer by many thin layers.  I did not permit the lower part of the model to change for stability in the inversion and because long period dispersion were not available. The resultant velocity model is named modl.out. A comparison of this velocity model to the CUS model is shown in the next figure.
shwmod96.png


The significant difference between the two velocity models in in the upper 1 km. The next figure shows a comparison between the observed group velocities and the model prediction.

srfphv96.png

The Love wave dispersion is not fit very well. However since spectral amplitudes will be used for the source inversion, this is not a major problem, since spectral amplitude inversion is very robust.
On the other hand, an attempt at waveform inversion would be strongly affected.

Source Analysis


The srfgrd96 program of Computer Programs in Seismology was used to find the best shear-dislocation and depth that bet fit the observed spectral amplltudes. The results of the grid search over strike, dip, rake and depth show that the source was shallow.  The best fit as a function of depth is given in the next figure.

fmdfit.dat

The mechanism plots are the best fitting solution for each depth.  Since spectral amplitudes are used and since the source is modeled as a shear dislocation, the other equivalent solutions can be obtained by increasing the strike by 180 degrees and by exchanging the pressure and tension quadrants. A tabulation of the goodness of fit and moment magnitude for each depth is given in the file fmdfit.dat.txt.

The best fit solution had the parameters

H(km) STK DIP RAKE Mw FIT
0.1 125. 50. 75. 3.01 0.8041

For this solution, the comaprison of the observed and predicted radiation patterns as a function of period are given in the following links:

Love Wave Radiation
Rayleigh Wave Radiation

In addition to the radiation patterns two figures present compare the model predicted group velocity and anelastic attenuaiton coefficient dispersion to that obtained from the observations.  The observed dispersion is compared to the observed, and the predicted and observed spectral amplitudes are assumed relation by Aobs = Apr exp ( - gamma r ) where r is the epicentral distance, Apr is the amplitude predicted on the basis of the moment magnitude, mechanism and infinite Q at the station, and gamma = pi f / Q U.


The plots contain all observations and represent an average over the broad region shown on the map. 

Relocation

Since a new velocity model was derived, it was appropriate to use the Computer Programs in Seismology program elocate. Using a fixed depth solution with a depth of 0.1 km, the data and results are given in the file elocate.txt and the source parameters from that file are

 Error Ellipse  X=   0.5472 km  Y= 0.9280 km  Theta = 165.3228 deg

RMS Error : 0.079 sec
Travel_Time_Table: CHI
Latitude : 41.7931 +- 0.0052 N 0.5792 km
Longitude : -87.8679 +- 0.0110 E 0.9083 km
Depth : 0.10 +- 1.35 km
Epoch Time : 1383590133.993 +- 0.39 sec
Event Time : 20131104183533.993 +- 0.39 sec
Event (OCAL) : 2013 11 04 18 35 33 993
HYPO71 Quality : BC
Gap : 105 deg

The P-wave arrivals were difficult to read and there were very few reliable first motion data. Thus the specification of a preferred focal mechanism (e.g., one of the four possible from the radiation pattern fit) was not possible.

Conclusion


This event was shallow, had an Mw = 3.0.   The strong Love waves indicate either an earthquake source or a significant horizontal movement of mass as part of a mining operation.
Using a rough value of 3.0E+18 dyne-cm as the isotropic moment for a 1 ton explosion, this event near Chicago would have had an explosive yield of about 100 tons (Mw=3.0 = seismic moment of 3 E+20 dyne-cm).

I, Robert B. Herrmann, believe that this equivalent source size is too large for normal quarry operations and thus I prefer a shallow earthquake source.