Observatories and Research Facilities for EUropean Seismology
Volume 2, no 1 March 2000 Orfeus Newsletter

Application of Regional Moment Tensor Inversion

Rami Hofstetter 1, Gonca Örgülü2, Mustafa Aktar2,3
1Seismology Division, Geophysical Institute of Israel, P.O.Box 182, Lod 71100, Israel
2Bogazici University, Kandilli Observatory and Earthquake Research Institute, 81220 Cengelkoy, Istanbul, Turkey
3TUBITAK Marmara Research Center, Earth Science Research Institute, P.O.Box 21, 41470 Gebze-Kocaeli, Turkey

Introduction - Methodology - RMT studies in Turkey - RMT studies in Israel - Future prospects - Acknowledgements - References


The Eastern Mediterranean Region (EMR) is known to be seismically active over a period of more than 2000 years, based on historical records and documents of eyewitnesses on one-hand and instrumental records on the other hand. The seismic activity and the tectonic setting of the major faults of the EMR, i.e. the North Anatolian fault, the Cyprean Arc and the Dead Sea Transform, are recognized and studied by numerous researchers (several early studies out of many, Montessus de Ballore, 1906; Quennell, 1958; McKenzie, 1970; Ambraseys and Finkel, 1987). Despite all those studies, we still know little about the seismotectonic of these significant geological faults. Recently, the strong earthquake of Gulf of Aqaba (22 November 1995, Mw = 7.2) and the catastrophic events of Izmit (17 August 1999, Mw = 7.4) and Duzce (12 November 1999, Mw = 7.2) have demonstrated the potential of those faults in creating large destructive events

Although localization of earthquakes are done for many years, since the beginning of the 20th century, only recently the source parameters of strong earthquakes are calculated on a semi-routine basis. Analysis of seismic moment tensors at regional distance is more challenging than at teleseismic scale, due to the strong effect of crustal and upper mantle heterogeneity on seismograms. Nevertheless, this kind of analysis is important for studying smaller events, in a lower magnitude range (4<M<5.5) and which occur frequently. The recent availability of broadband stations covering seismically active regions is the basis for dedicated research in this field. Below we briefly describe applications of the RMT inversion that are independently implemented in Turkey and in Israel.


At regional distance, moment tensor solutions of moderate-sized earthquakes (4.0< M<5.5) are not easily determined due to the complexity in wave propagation at short-periods. In recent years, the estimation of the source mechanism has significantly improved due to advances in broadband station coverage. In this context, the Regional Moment Tensor (RMT) inversion technique developed by Dreger and Helmberger (1993) is used based on the inversion of long-period data (20-100s). The method makes use of the full waveform from a single or preferably many broadband stations recorded at regional and near-regional distances. Thus the RMT method facilitates the understanding of the role that large number of moderate-sized earthquakes play in the regional tectonics and that usually remain unsolved.

Synthetic seismograms are computed with a frequency-wavenumber integration code (Saikia, 1994) which is a hybrid one consisted of two algorithms: Filon and Bouchon Integration algorithms. They are calculated for each path between source and station using simple 1-D velocity models for different source depths. The synthetic seismograms are compared with the observed long-period data to find the best fitting double-couple solution. Comparison of the inversion solutions based on several velocity models shows that the inversion results are not strongly model dependent for short propagation paths to near-regional stations.

Figure 1. Broadband stations (squares) located in western Turkey and medium-band stations (triangles) around the Sea of Marmara. The epicenter of the aftershock (black star) and its source mechanism is shown on the map as well as the location of the Izmit and Duzce earthquakes (red star).

RMT Studies in Turkey

The installation of 3 broadband (>100 s) and 5 medium-band (~40 s) stations in the western part of Turkey (Fig. 1) has enabled systematic inversion of moderate sized earthquakes occurring. The robustness of the inversion is often dependent upon the azimuthal coverage as well as the relative homogeneity of the travel path, which is the case for the region of the Marmara Sea. The method is widely applied to study the aftershock activity of the recent destructive earthquakes of Izmit (Mw = 7.4) on 17 August 1999 and Duzce (Mw = 7.2) on 12 November 1999.

An inversion example is given for an aftershock (30 August 1999, M = 5.2) of the Izmit earthquake (Fig. 2) located within close distance of the epicenter of the mainshock. The well fitting between long-period data and synthetic seismograms provides right-lateral strike-slip mechanism having a strike direction of N76E slightly different than the general character of the North Anatolian Fault Zone. A similar solution is provided using a different methodology by Braunmiller et al., 2000 within this issue.

Click for Figure

Figure 2. Comparison of three-component long-period (20-50s) displacement data (solid) and synthetic seismograms (dashed) for the aftershock (M = 5.2) that occurred on 31 August 1999.

Additional inversion solutions, among others, the 9/20/1999 earthquake, located at the north of the Marmara Island in the Sea of Marmara, are provided on the Kocaeli web page of KOERI.

RMT Studies in Israel

The locations of the broadband stations BGIO, Israel, and KEG, Egypt, are shown in Fig. 3 relative to one of the observed aftershocks of the Gulf of Aqaba sequence. The same above-mentioned procedure of determining the synthetic seismograms was executed, however using a different crustal model. The solution of waveform inversion is plotted in Fig. 4. The solution suggests a left-lateral strike slip mechanism, which is in good agreement with the tectonic setting of the Dead Sea rift in general and the Gulf of Aqaba in particular.

Figure 3. Broadband stations (squares) that were used in the waveform inversion for the November 22, 1995 event (red star).

Click for Figure

Figure 4. Comparison of three-component long-period (20-50s) displacement data (solid) and synthetic seismograms (dashed) for the 22 November 1995, M = 4.4, event in the Gulf of Aqaba.

Future Prospects

In conclusion, we showed that the above-described method can be used, either manually or semi-automatically, to provide useful results for better understanding the seismotectonic setting of a given region. In the future, as the the mean distance between permanent broadband stations is decreased, the RMT solutions will be increased quantitatively as well as qualitatively.


We wish to thank Seismological Laboratories of KOERI (Turkey), GII (Israel), NRIAG (Egypt) and MEDNET (Italy) for providing data.


Ambraseys, N., and Finkel, C., 1987. Seismicity of Turkey and neighbouring regions, 1899-1915, Annales Geophysicae, 5, 701-726.
Dreger D., and Helmberger, D., 1993. Determination of Source Parameters at Regional Distances with Single or Sparse Network Data, J. Geophys. Res., 98, 8107-8125.
McKenzie, D., 1970. Plate tectonics of the Mediterranean region, Nature, 226, 239-243.
Montessus de Ballore, F., 1906. Comte de 1851, les treblements de terre, Paris, Colin.
Quennell, A. M., 1958. The structural and geomorphic evolution of the Dead Sea rift, Quart. J. Geol. Soc, 114, 1-24.
Saika, C., 1994. Modified Frequency-Wavenumber Algorithm for Regional Seismograms using Filon~ Rs Quadrature: Modeling of Lg waves in Eastern North America, Geophys. J. Int., 118, 142-158.

page 6
Copyright © 2000. Orfeus. All rights reserved.