TMTS, the interactive moment tensor review interface, and the Berkeley Moment Tensor Catalog

Jennifer Taggart, Angela Chung, Rick McKenzie, Peggy Hellweg, and Doug Dreger

Research Objectives

Since 1992, the BSL has been calculating moment tensor solutions for selected events in Northern California (Romanowicz et al, 1993). At present, moment tensor solutions are calculated in real time for most events with $M_{w}$ 3.5 and above in Northern California (Pasyanos et al, 1996). The newest version of the program used for reviewing them is a Web-based interface, TMTS. The implementation at UCB is based on a package developed at Caltech (Clinton et al, 2007) and uses the waveform moment tensor inversion code developed by Dreger (Minson and Dreger, 2008). This interface allows analysts to calculate solutions from any computer with Web access and a browser. The interface relies on recent waveforms in the DART and archived data at the NCEDC (see Chapter 3, Section 6). In addition, TMTS allows the calculation of deviatoric moment tensor solutions and optionally the full moment tensor. In the past, solutions have been stored in the Berkeley Moment Tensor Catalog, a text file with the strike, rake, and dip of the fault plane solution as well as moment, $M_{w}$, and a listing of the stations used. The new moment tensor review interface connects to a database where all the information needed to recreate the moment tensor solution is saved (see Figure 2.81). Thus, porting the catalog to the new database required us to recalculate each moment tensor using TMTS.

Figure 2.81: Tables and relationships in the database where the moment tensor information is stored.
\epsfig{, width=7.5cm}\end{center}\end{figure}

Figure 2.82: Comparison between selected moment tensors in the UCB catalog and deviatoric moment tensors calculated with the new program (right).
\epsfig{, width=7.5cm}\end{center}\end{figure}

Data Set

We selected earthquakes in the Berkeley Moment Tensor Catalog from within the reporting region for recalculation with the new interface and inclusion in the database. Geysers events and events outside the region were set aside. Events from the years 2004 and 2005 were assigned to the USGS for training purposes.


When recalculating the moment tensors with the new program, we made an effort to match the original solution as closely as possible. Provided three or more stations were used, the same stations were used for the recalculation. For earthquakes with $M_{w}$ between 3.5 and 4.0 and with an original solution of high quality, we tried to match the strike, rake, and dip of the original solution to within 10 degrees and the $M_{w}$ to within .1. For earthquakes with $M_{w}$ above 4.0, the strike, rake, and dip were to match within 5 degrees and the $M_{w}$ to within .05. We also attempted to match the value for $M_{0}$.

TMTS has a number of features that make calculation of moment tensors faster. When an event loads, the program automatically selects a set of stations based on distance and azimuth. Stations can be changed using the buttons on the Station Normal page, accessible by a tab in the program. The Station Advanced tab takes the user to a page where other parameters can be changed. Users can choose between three filters, and, for each station, the correlation length between the seismogram and the synthetics can be changed to any value, facilitating analysis of different-sized earthquakes at different distances. Three sets of synthetic Greens functions are available to account for the velocity structure of different areas of California. The Zcorr, a value that determines how far to advance the synthetic seismogram and depends on station-event distance, can be set to any value for each station to allow fine-tuning and correcting misfits. The depth can be fixed once it is known, allowing quicker recalculations when refining the solution. On the left hand side, next to the tabs, a plot of the variance reduction, a measure of fit, is shown at different depths. This plot can also be shown in another window or tab of the Web browser. A plot of variance reduction over percent double couple and plots of each full solution at the program's 12 chosen depths can be shown in this way as well. In contrast to the previous arrangement, where parameters had to be changed in text files and multiple programs had to be run at the command line, TMTS makes each option available as a button or box.

Significance of Findings

In Figure 2.82, selected mechanisms from the UCB catalog are compared with the deviatoric moment tensors calculated with the new program. At the top is the August 18, 1998 Bolinas event, an earthquake with an unusual moment tensor for that area. Next is an example of a shallow offshore event. An event from the Parkfield area and the October 31, 2007 Alum Rock event both show a high degree of double couple. In contrast, the Geysers earthquake at the bottom is an event where the compensated linear vector dipole component of the deviatoric solution is important. Also shown is the smaller of the two January 2008 Red Bluff events. Non double couple elements in the solution may be real (Dreger et al, 2000, Dreger et al, 2008, Ford et al, 2008), as is likely to be the case for the Geysers event. For small events, they may indicate a relatively poor signal-to-noise ratio.

Approximately 350 moment tensors for earthquakes in Northern California are now in the database with their complete waveforms and all of the parameters involved in their calculation. Table 2.2 shows the year by year breakdown of recomputed events. In Figure 2.83, a comparison of $M_{w}$ from the old and new moment tensor catalog shows no systematic bias toward higher or lower values with the new program. Plans are in place to make the new Berkeley moment tensor database searchable so that researchers can call up full solutions of events that meet their criteria.

Table 2.2: Table of events in old and new catalogs.
Year Old Catalog New Catalog
1995 68 50
1996 51 34
1997 68 54
1998 46 39
1999 42 38
2000 29 22
2001 28 24
2002 23 19
2003 41 34
2004 59 USGS
2005 20 USGS
2006 27 23
2007 34 22

Figure 2.83: Comparison of values of $M_{w}$ for moment tensors calculated with both programs.
\epsfig{file=taggart_08_03.eps, width=10.5cm, bb=80 158 632 632, clip}\end{center}\end{figure}


Implementation at UCB would not have been possible without the programming support provided by Alexei Kireev, Pete Lombard, and Doug Neuhauser.

Work on this project was supported by the CISN funding of the California Governor's Office of Emergency Services under contract 6023-5 and the United States Geological Survey project 07HQAG0013.


Clinton, J.F., E. Hauksson, and K. Solanki, An Evaluation of the SCSN Moment Tensor Solutions: Robustness of the $M_{w}$ Magnitude Scale, Style of Faulting, and Automation of the Method, Bull. Seism. Soc. Am., 96, 1689 - 1705, 2006.

Dreger, D. S., H. Tkalcic, and M. Johnston, Dilational processes accompanying earthquakes in the Long Valley Caldera, Science, 288, 122-25, 2000.

Dreger, D., S.R. Ford, and W.R. Walter (2008), Source analysis of the Crandall Canyon, Utah mine collapse, Science, 321, 217, 2008.

Ford, S. R., D. S. Dreger and W. R. Walter, Source Characterization of the August 6, 2007 Crandall Canyon Mine Seismic Event in Central Utah, submitted to Seism. Res. Lett., 2008.

Minson, S. and D. Dreger, Stable Inversions for Complete Moment Tensors, Geophys. Journ. Int., in press.

Pasyanos, M.E., D.S. Dreger, and B. Romanowicz, Toward Real-Time Estimation of Regional Moment Tensors, Bull. Seism. Soc. Am., 86, 1255-1269, 1996.

Romanowicz, B. D. Dreger, M . Pasyanos, and R. Uhrhammer (1993). Monitoring of Strain Release in Central and Northern California Using Broadband Data, Geophys. Res. Lett., 20, 1643-1646, 1993.

Berkeley Seismological Laboratory
215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760
Questions or comments? Send e-mail:
© 2007, The Regents of the University of California