Moment Tensor 20061010080251

Location

2006/10/10 08:02:51 -56.10 -122.37 10

Moment Tensor Comparison

The following compares this source inversion to the USGS Rapid Moment Tensor Solution and to the Harvard CMT solutions, if they are available.

SLU

SLU Moment Tensor Solution 2006/10/10 08:02:51 Best Fitting Double Couple Mo = 8.32e+24 dyne-cm Mw = 5.88 Z = 5 km Plane Strike Dip Rake NP1 100 80 20 NP2 6 70 169 Principal Axes: Axis Value Plunge Azimuth T 8.32e+24 21 325 N 0.00e+00 68 126 P -8.32e+24 7 232 Moment Tensor: (dyne-cm) Component Value Mxx 1.69e+24 Mxy -7.40e+24 Mxz 2.87e+24 Myy -2.66e+24 Myz -8.72e+23 Mzz 9.73e+23 ##########---- ###############------- ### ############---------- #### T #############---------- ###### #############------------ #######################------------- ########################-------------- #########################--------------- #########################--------------- --########################---------------- -------###################---------------- -------------#############---------------- ----------------------####---------------- ------------------------################ ------------------------################ -----------------------############### - -----------------############### P ----------------############### ----------------############# ---------------############# -----------########### ------######## Harvard Convention Moment Tensor: R T F 9.73e+23 2.87e+24 8.72e+23 2.87e+24 1.69e+24 7.40e+24 8.72e+23 7.40e+24 -2.66e+24

Data set

The following broadband stations passed the QC and were used for the source inversion. AFI ANMO BBGH CTAO DWPF HNR JCT KIP LVC MBWA MIAR NNA NWAO OTAV PAYG PMG POHA RAO RAR RCBR SDV SNAA SNZO TAU TRQA TUC WMOK

Deviatoric Moment Tensor Inversion

All observed and Greens function waveforms are corrected to instrument response to ground velocity in meters/sec for the passband of 0.004 - 5 Hz. The traces were then lowpass filtered at 0.25 Hz and interpolated to a sample rate of 1 second.

For the deviatoric moment tensor inversion, the observed traces and Green's functions are read in an cut using the following commands

Processing parameters

#####
#	Driver script for the teleseismic waveform inversion
#
#	The depth HS must be of the form 0010 for a depth of 1.0 km
#       of 6700 for a depth of 670.0 km
#
#	The Filter_Band is an integer with the following meansing
#
#	FILTER_BAND	1/FH	1/FL
#	1		60	12	for Mw < 6.4
#	2		100	20	6.4 =< Mw < 6.8 
#	3		120	40	6.8 =< Mw < 7.2 
#	4		143	80	7.2 =< Mw < 9.3 
#####
#	Source duration - halfwidth of triangular function
#	this filters only the Green functions
#
#	halfwidth = 1.05 * 10-8 * M0^1/3 (M0 is dyne-cm)
#
#	MW	   half-width (sec)
#	5.0         0.75
#	6.0         2.45
#	7.0         7.7
#	8.0        24.5
#	9.0        74.3
#                       0.5*(MW - 9)
#  or half-width=74.3*10
#  or echo 5.87 |  awk '{print 74.3*exp(0.5*log(10.0)*(0150 - 9.0)) }'
#####
#  Processing window for P (use SAC variable A for P arrival
#  	Start   A  - 30
#  	End     A  + 2*HALFWIDTH + 0.03*HS + 30 + 1/FH
#  Processing window for SH (use SAC variable T0 for SH arrival
#  	Start   T0 - 60
#  	End     T0 + 2*HALFWIDTH + 0.06*HS + 30 + 1/FH
#  Processing window for SV (use SAC variable T1 for SV arrival
#  	Start   T1 - 60
#  	End     T1 + 2*HALFWIDTH + 0.06*HS + 30 + 1/FH
#	
#	The term involving HS serves to include the depth phases and
#	to exclude the PP or SS at most distanace
#####

The cut windows attempt to include the P, pP, sP, pS, S and sS arrivals. However, oen must be very careful about the fact that PP may be included in some distance ranges.

The waveforms are then bandpass filtered by the application of the following high- and low-pass stages (an optional microseism filter):

hp c 0.0167 2
lp c 0.0833 2
int
br c 0.12 0.25 n 4 p 2

The traces were next integrated to ground displacment in meters. Finally the observed data are interpolated to ahve the same sampling at the Green's functions.
NOTE: this was done for speed. The proper sequence is to read traces, filter and then cut - gsac will be modified to introduce a command CUTWR to define the cut upon a write.

The source inversion is a multipass operation since a lower frequency filter band is used for larger earthquakes and since a search is made over depth. Up to three passed of the outer loop are made, after which the moment magnitude is determined and filter settings readjusted. The inner loop over depth samples all depths from 0 to 800 km with 5 km increments in depth to 50 km, followed by 10 km depth sampling for the remaining range.

The following filter ranges are used according to the moment magnitude Mw:

   FILTER_BAND   1/FH(s)  1/FL(s)
       1           60     12      Mw < 6.4
       2          100     20      6.4 < Mw <= 6.9
       3          120     40      Mw > 6.9

The map displays the distribution of stations used for this source inversion.

Location of the earthquake (yellow star) and great circle path from the epicenter to each station (red) [created using GMT (Wessel, P., and W. H. F. Smith, New version of Generic Mapping Tools released, EOS Trans. AGU, 76 329, 1995.)]

For this data set the favored solution is

WVFMTD96   15.0   97.   90.  -22.   5.87     0.386 0.730E-06     0.388     0.621 0.326E-06  52.0

The following figures show the sensitivity of the goodness of fit parameter so source depth, the waveform comparison as a function of epicentral distance in degrees and the source to station azimuth

Depth Sensitivity

Goodness of fit as a function of source depth. The measure is 1 - SUM (o -p)² / SUM o². A value of 1.0 is the best fit. The best double couple mechanism for the solution depth is plotted above goodness of fit value to indicate how the mefhanism may change with depth.

Detailed Waveform Comparison

P-wave Z component

Comparison of the observed traces (red) and solution predicted traces (blue) ordered in terms of increasing epicentral distance. Each pair of traces is annotated with the station name, epicentral distance in degrees, source to station azimuth in degrees. Each pair of traces is plotted with the same scale and the peak amplitudes are indicated at the lect of each trace. Finally the time shift between the P-wave first arrival picked and the the theoretical P-wave first arrival in the predicted trace is indicated, with a positive sign indicating that the predicted trace has been shifted to the right by the given number of seconds. as a function of source to station azimuth in degrees (D). The purpose of this display is to highlight the azimuthal dependence on the first motion. The traces are annotated with the station name at the top.

SH-wave T component

SV-wave R component

Inversion Details

Output of wvfmtd96 for the best depth..

Grid Search

For the grid search, the observed traces and Green's functions are read in an cut using the following commands