Seismicity Parameters of Seismogenic Zones

By

Avi Shapira and Abraham Hofstetter

 

Formulation and b-value

The seismogenic zones in the region are identified by Shamir et. al., 2001 (see map of seismogenic zones). Their work is a product of a joint venture of the Geological Survey of Israel and the Geophysical Institute of Israel, assisted and approved by geologists of the Natural Resources Authority of Jordan (NRA).

The seismicity of a seismogenic zone is quantified in terms of the frequency-magnitude relationship. That relationship is a key element in estimating the probability that a magnitude M earthquake will occur in a certain seismogenic zone within a predefined time interval. The formulae used in this study are:

 

Where, n(M) is the annual frequency of earthquakes of magnitude M±dM (here, dM=0.1) and N(M) is the annual number of events with magnitude M or greater.

 

 

The completeness of the earthquake catalogue applies to the territory within the area shown on the map in Fig. 1.  This area comprises those parts of the Dead Sea transform system within the JSN and the ISN stations. Hence, assuming that Mmin=2.2 and Mmax=7.5 and following Wiechert (1980), we estimated the optimal values a, a and b that yield the least sum of squared residuals for log n(M) and log N(M).

Following the magnitude relationships above all magnitude values in the earthquake catalog are converted to the unified magnitude that is analogous to the moment magnitude Mw (when ML³5) or Mm (when ML<5). Here the seismic moment is determined from short-period recordings.

The analysis for that part of the area with the most complete instrumental and historical data  (Fig. 2) yields b=0.96 .

 

 

 

Seismicity parameters of the seismogenic zones

1. b-values

The amount of information available for each individual seismogenic zone may be found insufficient for accurate statistical assessments of the seismicity parameters. However, the b-value is indicative of the tectonic characteristics of a region and thus, we may assume that the seismogenic zones that constitute the Dead Sea Rift (DSR) or are branching off the Dead Sea Rift, will have the same value of b=0.96. This assumption is considered to better represent the tectonic characteristics of the seismogenic zones in the investigated region. The applicability of that assumption is further demonstrated in figures 3, 4 and 5 that show the frequency-magnitude relationships for different seismogentic zones, where in each case the black line is the best fit for a fixed slope of b=0.96.

The seismo-tectonics of the Cyprus zone and possibly also in the Gulf of Suez are of different character. For both zones there are sufficient data to perform a separate analysis that yield b=1.07 and b=0.98  for the Suez and Cyprus, respectively.  The cumulative frequencies vs. magnitude for those zones are shown in Fig. 5.

 

2. Maximum magnitudes

An important seismicity parameter is the maximum magnitude. Based on previous assessments (see e.g. Arieh, 1967, Arieh and Rabinowitz, 1989, Shapira, 1983, Vered, 1978, Yuceman, 1992, Shapira and Shamir, 1994 and others) we assumed that the maximum magnitude along the DSR is 7.5 with the exception for the Yamouneh fault that may be associated with slightly higher magnitudes (Mmax=7.75). To the faults that are branches of the DSR we assigned maximum magnitudes of 5.5 and 6.0. These estimations are based mainly on the limited seismic history and partially on the length of the mapped fault. These faults, with the exception of the Carmel fault, are currently away from populated areas and for the present time, the uncertainty associated with defining Mmax will not pose a practical difficulty. In the case of the Carmel fault we assumed Mmax=6.5 mainly due to the accumulated length of that fault system and due to its proximity to the population centers. We do not have any record of a strong earthquake (M>6) that has occurred in the past on that fault despite its proximity to the main cities of the Galilee, throughout the history.

Maximum magnitudes associated with the zones that are characterized as zones of background seismicity, follow the seismicity record (see also the references listed above). The estimated maximum magnitudes for the different seismogenic zones are shown in Tables 1,2 and 3.

 

3. Seismicity levels (Return periods)

The black lines in Figs. 3,4 and 5 represent the best fit to the cumulative N(M) function where Mmin=2 and the b-value is fixed to b=0.96.  However, in many cases we may suggest a parallel line (drawn in red on Figs. 3-6) that represents an upper bound from which we may deduct the upper value of “alfa” (i.e. the annual number of events) for M³2.0. The “Maximum observed alfa” coefficients are tabulated in tables 1,2 and 3.

Tables 1 and 2 show the assumed length of each segment of the DSR (Table 1), and that of the branching-off faults (Table 2). Table 3 shows the area of each of the seismogenic zones that are characterized as areas with background seismicity without clearly associated fault segments.

The observed seismicity in the zones constituting the Dead Sea Rift demonstrate systematic rate of activity per unit length (N/km in Table 1). We observe an average activity rate of 0.26 events of M³2.0 per kilometer. This average value is obtained while excluding the Arnona, Aragonese, Arava and the Yamuneh seismogenic zones. The Aragoneze zone that is probably heavily “contaminated” with many aftershocks of the earthquake sequences that occurred during 1983-1996. Yamouneh segment in the north and the Arnona segment in the south are distant from the regional seismic stations and most properly our earthquake catalog for these seismogenic zones is significantly incomplete. The frequency-magnitude relationship for the Arava valley is very different from those of the other segments of the DSR: while for most of the segments of the DSR the b value is within 0.96±0.05, the apparent b-value for the Arava is unacceptably high (close to 2). Furthermore, the seismic history shows no documented evidence of a major earthquake that originated in the Arava. The absent of major earthquakes in the Arava is also supported by the observed low seismicity over the instrumental period despite the fact that the Arava is well monitored by permanent and temporarily installed seismic stations.

Adopting the value of 0.26 events/year/km and considering the length of each zone along the DSR, we obtain an estimate of the annual number of events with M³2.0. As for the Arava Valley, we used the observed upper bound of 25 events per year. The Arava is relatively distant from population centers and even if we are under-estimating its seismicity (and consequently, the seismic hazard), the impact on the population is currently  insignificant. However, the seismicity of the Arava remains a problem to be further investigated.     

 

Table 1 : Estimation of the seismicity level along the Dead Sea Rift

 

The same evaluation process is repeated for the seismogenic zones that constitute faults system that branch from the main DSR. Here again, the rate of activity, N/km, is very similar and thus we may suggest an average value that will be applicable to those zones. The average value (excluding the Malhan fault) is 0.05 events (M³2.0) per kilometer.

When assuming an activity rate of 0.05 events/year/km then, for the 120 km of the postulated Carmel fault we should have 6 events/year with M>2.0. We prefer the “observed” value of 10 events/year that would yield a more conservative estimate of the seismic hazard. This choice is motivated by (a) the proximity of the fault to densely populated areas and (b) previous assessments by Hofstetter et al. (1996) that may suggest a higher rate of high magnitude events.

 

 

 

 

 

 

Table 2 : Estimation of the seismicity level along fault systems that are branching-off the  Dead Sea Rift.

 

 

Table 3 presents the estimated seismicity rate per unit area (squared kilometer) in seismogenic zones that show background seismicity.  The selected background seismicity rate for the Suez, Cyprus and the East Samaria zones are as observed. The seismogenic zones: East Mediterranean-1, East Mediterranean-2 and the Galilee show a similar value of 1.49 events (M³2.0) per squared kilometer. The other low seismicity zones of background seismicity, namely: East Mediterranean-3, Central Israel, North Jordan, Palmyra region and Wadi Sirhan show an average rate of 0.47 events (M³2.0) per squared kilometer. The corresponding “alfa” coefficients, re-scaled to the size of the area, are also shown in Table 3.

 

Table 3 : Estimation of the seismicity level in areas of background seismicity.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Comparison with paleo-seismic and slip rate information.

The seismicity parameters are crosschecked against other information when available. Based on geological interpretations, the slip rate along the Dead Sea Rift is estimated to be in the order of 5-10 mm/year. Equations 3 and 4 relate average slip rates to seismicity parameters.

 

 

 

 

 

 

Table 4 presents the average slip rate inferred from the given (suggested) rate of cumulative activity (see Eq. 9).  Table 5 also shows the postulated alfa value in the case of d=5 mm/year. Evidently there is a huge discrepancy between the two estimations, except for the Aragoneze deep where the 1995, Mw=7.1 earthquake took place. Present seismic activity suggests a slip rate of the order of 1-2 mm/year along the DSR (in agreement with the few geodetic GPS and In-SAR surveys recently conducted along the DSR, see e.g. Baer et al. 2000). It is assumed that the slip rate along the faults that are branches of the DSR is about 10-20 times slower than that of the DS rift (Y. Rotstein – Private communication). Here again, the present day seismic activity suggests a much slower annual slip.

 

 

Table 4: Level of Seismicity and slip Rates

 

 

The slip rate analysis indicates that the estimated seismicity parameters and the associated  earthquake hazard assessment are NOT conservative. On the other end, these are upper bounds to the known seismicity of the analysed regions.

 

It should also be important to compare the evaluated seismicity level with the few available paleo-seismic information (see for example, Marco and Agnon, 1995, Ellenblum et at., 1998, Ken-Tor et al., 1998, Enzel et al., 2000, Zilberman et al., 2000, Ken-Tor et al., 2001). These studies suggest an average repeat time of ~2000 years for M>6 earthquakes in the southern Arava valley and an average return period of 300 years for earthquakes stronger than M=5.5 in the Dead Sea basin.

Table 5 shows estimated return periods for magnitudes 5, 6 and 7 in each seismogenic zone. We may state that the estimations based on seismicity data are  not in conflict with what is currently known from paleo-seismic studies. The most complete seismic history relates to the part of the Dead Sea Rift between southern Lebanon and the Dead Sea basin. In that part of the rift we expect a “return period” of about 75 years and 960 years for earthquake greater that M=6.0 and 7.0, respectively. These estimates match the documented history of earthquake catastrophes in the Holly-Land.

 

Table 5: Estimated return periods of earthquakes in the region.

 

 

 

 

 

References

 

Ellenblum, R., S. Marco, A. Agnon, T. Rockwell and A. Boaz, 1988. Crusader castle torn apart by earthquake at dawn, 20 May 1202. Geology, 26-4:303-306.

 

Enzel, Y., G. Kadan and Y. Eyal, 2000. Holocen earthquakes in the Dead Sea graben from a fan-delta sequence. Quat. Res., 73/74:137-144.

 

Hofstetter, A., T. van Eck and A. Shapira, 1996. Seismic activity along fault branches of the Dead Sea-Jordan transform system: The Carmel-Tirtza fault system. Tectonophysics, 267:317-330.

 

Ken-Tor, R., M. Stein, S. Marco, Y. Enzel and A. Agnon, 1998. Late Holocene earthquakes recorded by lake sediments, Ze’elim plain, Dead Sea, in Israel. Abstr. in Annual Meeting of the Israel Geol. Soc., Mitzpe Ramon. p. 53.

 

Ken-Tor, R., A. Agnon, Y. Enzel, M. Stein, S. Marco and J.F.W. Negendank, 2001. High resolution geological record of historic earthquakes in the Dead Sea basin. J. geophys. Res., 106:2221-2234.

 

Marco, S. and A. Agnon, 1995. Prehistoric earthquake deformation near Masada, Dead Sea graben. Geology, 23-8:695-698.

 

Shamir, G., Y. Bartov, A. Sneh, L. Fleischer, V. Arad and M. Rosensaft, 2001. Preliminary seismic zonation for Israel. GII Rept. No. 550/95/01(1)

 

Shapira A. and Hofstetter A., 1993  Source parameters and scaling relationships of earthquakes in Israel. Tectonophysics, pp 217-226.

 

Wiechert, D. H., 1980. Estimation of the earthquake recurrence parameters for unequal observation periods for different magnitudes. Bull. Seismol. Soc. Am.,  70:1337-1346.

 

Zilberman, E., R. Amit, A. Heimann and N. Porat, 2000. Changes in Holocen paleoseismic activity in the Hula pull-apart basin, Dead Sea Rift, northern Israel. Tectonophysics, 321:237-252.

 

 

 

 

 

Mw

Fig 1. Map of earthquake epicenters showing the area where the historical and instrumental earthquake information is most complete.

	Mw

 

 

 

 

Mw

Fig. 2.  Cumulative frequency-magnitude relationship for the area shown in Fig. 1, with the most complete earthquake information.

 

 

 

 

 

 

 

 

 

 

 

 

                  Fig 3. N(M) relationships for seismogenic zones along the DSR.                                                                         

 

 

        Fig. 4: N(M) relationships for seismogenic zones that are branching off the DSR.

 

 

 

 

 

 

        Fig. 5: N(M) relationships for seismogenic zones with background seismicity.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

        Fig. 6: N(M) relationships for seismogenic zones Suez and Cyprus.