Karaburun (İzmir) Earthquakes

Journal of Geodynamics

http://dx.doi.org/10.1016/j.jog.2012.05.005

The Dense Micro-Earthquake Activity at the Boundary Between

the Anatolian and South Aegean Microplates

Onur Tan

ABSTRACT

The Aegean coast of Turkey is a transition zone between the Anatolian and the South Aegean microplates. The recent geophysical studies suggest the existence of additional microplates and suggest a modification of the block boundaries; the Karaburun Peninsula is most likely located on one of these boundaries. The earthquake activity in and around the peninsula was very low until the Mw 5.8 Sığacık Bay earthquake of 17 October 2005. The continuing seismicity in the Karaburun Peninsula has been monitored with the help of the dense seismology network. I re-located more than 5000 earthquakes using the double-difference algorithm and obtained fault plane solutions to understand the seismotectonic properties of the region. It is found that the micro-seismic activity has increased towards the NW of the Sığacık Bay. These observations imply the existence of a NW-SE-oriented sinistral strike-slip fault. This fault is possibly a conjugate of the dominant NE-SW-oriented dextral strike-slip faults in this transition region occupying the area between western Anatolia and the mainland of Greece, where a dominantly extensional regime prevails. Moreover, it is clear that the micro-earthquake activity is related to the Uzunkuyu Intrusive located in the middle of an approximately 40 km strike-slip Karaburun Seismic Zone between the Sığacık and Gerence bays.

Keywords

Aegean; Intrusive; İzmir; Karaburun; Micro-earthquakes; Uzunkuyu

 1 INTRODUCTION

The Aegean is one of the most active continental extension areas on the Earth (McKenzie, 1972; Angelier, 1978; Şengör et al., 1984). The driving force of the N-S extension in the Aegean Sea and in western Anatolia is the slab pull force of the Hellenic subduction zone (McKenzie 1978; LePichon and Angelier, 1979) and/or the westward escape of Anatolia (Dewey and Şengör, 1979; Taymaz et al., 1991). The GPS surveys performed in the last two decades enhance our understanding of the kinematics of the region (Billiris et al., 1991; Le Pichon et al., 1995; McClusky et al., 2000; Meade et al., 2002; Reilinger et al., 2010). Block modelling, used with mainly GPS data, shows that the crustal deformation in the Aegean cannot be defined with simplified tectonic models. The recent geophysical studies suggest additional microplates and modify the block boundaries to minimize the residuals between the observed data and the model (Meade et al., 2002; Nyst and Thatcher, 2004).

The tectonic evolution of western Anatolia has been the subject of several recent publications (Hancock and Barka, 1987; Bozkurt, 2001, 2003; Genç et al., 2001; Koçyiğit, 2005; Bozkurt and Rojay, 2005) and thus will not be discussed in this paper. Although large E-W graben systems (i.e. Gediz and Büyük Menderes) with normal faults are dominantly observed in western Turkey, the strike-slip faults that are capable of generating destructive earthquakes are mapped in the area surrounding the Karaburun Peninsula. Figure 1 shows a simplified geological map of the Karaburun Peninsula derived from different studies (Kaya, 1981; Bozkurt, 2001; Emre et al., 2005; Çakmakoğlu and Bilgin, 2006; Uzel and Sözbilir, 2008; Sözbilir et al., 2009; Helvacı et al., 2009). The peninsula contains several Neogene sedimentary and volcanic units (Helvacı et al., 2009). No major faults have been mapped on the main land of the peninsula. The dextral strike-slip Gülbahçe Fault Zone (GFZ) is a boundary between the carbonates and sedimentary units on the isthmus. The Seferihisar Fault Zone (SFZ) is dextral strike-slip deformation zone that is 30 km long. The Orhanlı Fault Zone (OFZ) has several dextral strike-slip segments and extends 45 km (Emre and Barka, 2000; Genç et al., 2001; Uzel and Sözbilir, 2008; Uzel et al., 2012). Çakmakoğlu and Bilgin (2006) published a geological map of the region in studies of MTA (Mineral Research and Exploration Directorate of Turkey) from 1996–1998 and mapped a fault between the Sığacık and Gerence bays. However, there is no additional information about this fault. Ocakoğlu et al. (2005) processed marine seismic reflection data and mapped offshore active strike-slip faults in the south of the Karaburun Peninsula.

Figure 1 also shows the distribution of the earthquakes in the study area (M ≥ 2) reported by Boğaziçi University Kandilli Observatory and Earthquake Research Institute (KOERI) between January 1990 and September 2005. The Sığacık activity from October 2005 (reported by Benetatos et al., 2006; Aktar et al., 2007) and later events are not shown on the map so as to clearly show the background seismicity in the Karaburun Peninsula and surroundings. The earthquake distribution in Figure 1 indicates a scattered seismicity. The two important earthquakes were observed near Seferihisar. The events of 6 November 1992 (Mw = 6.0) and 10 April 2003 (Mw = 5.7) may be related to the SFZ and OFZ, respectively. Altınok et al. (2005) studied historical earthquakes in the region and reported some destructive events (i.e., 1881 and 1949); however, the data for this historical period have large location uncertainties, and the possible link with the recent earthquake activity seems difficult to discern (Aktar et al., 2007). Unfortunately, the national earthquake data catalogues are insufficient to discuss the local fault characteristics in western Anatolia. Hence, the “Multi-Disciplinary Earthquake Researches in High Risk Regions of Turkey Representing Different Tectonic Regimes (TURDEP) Project” has been initiated to observe the activities of the local faults and to improve observation capabilities (İnan et al., 2007).

In this paper, I discuss a seismic zone, the Karaburun Seismic Zone (KSZ), which is not mentioned previously on the Karaburun Peninsula, with a new high-precision earthquake data set (ML ≥ 0.2) obtained from a dense seismology network.

 2 DATA AND METHOD

In the TURDEP Project (İnan et al., 2007), a dense broad-band earthquake station network was established. Hence, 31 online seismic stations have been established in western Anatolia since July 2007 (see Figure 4 inset). The network contains Güralp 3TD digital broad-band seismometers and 40T analogue seismometers with Reftek-130 recorders. To continuously observe the micro-earthquake activity of the region, the waveform data were acquired at a rate of 100 samples per second.

The absolute hypocentre parameters of the earthquakes were computed using the Hypocenter location algorithm described by Lienert and Havskov (1995). I used the P-wave velocity (Vp) model obtained using seismic tomography inversion (Akyol et al., 2006; Table 1) and used 1.75 for the Vp/Vs ratio to calculate shear-wave velocities (Vs). More than 26,383 P- and 23,380 S-arrivals were handpicked to locate the events that occurred between July 2007 and October 2011. During the observation period, approximately 5,160 events were located. The average horizontal and vertical location uncertainty is approximately ±2 and ±3–4 km, respectively. The earthquake focal depth is the most difficult parameter to calculate in location inversion. The changes in the derivative of the depth are minute for earthquake hypocentre depths that are less than half of the distance from the nearest station (Lienert and Havskov, 1995). Based on the network structure, most of the events were first recorded at the stations 10 km away. Therefore, it is difficult to determine the accurate depth of the shallow events (< 5 km) in the study area. The events with large location uncertainty (> 8 km) were not used in this study.

Table 1. P-wave velocity (Vp) model used in this study (Akyol et al., 2006).

To standardize the calculation of the magnitudes from the different instruments and to eliminate operator errors, I preferred to use local magnitudes (ML) rather than duration magnitudes (md) of the micro-earthquakes in the region. To calculate ML, a methodology was introduced into the Seismic Analysis Code (Goldstein et al., 1998). First, the sensor and digitizer responses are removed from the velocity records and are filtered by a band-pass Butterworth filter (0.2 – 40 Hz). Then, each waveform is convolved with the Wood-Anderson seismometer response and integrated to generate the displacement record in units of nanometres. The maximum zero-to-peak horizontal amplitude was selected from the three components recorded at each station, and ML was calculated. The minimum and maximum extreme values (larger than the standard deviation) were removed, and the remaining magnitudes were averaged for that event. The lower ML magnitude limit of the data set is approximately 0.2, and there is no event exceeding ML = 4.8 (Figure 2A). The magnitude completeness of the catalogue is MC = 1.4 (Figure 2B). This small completeness value is the evidence that the catalogue is the best dataset for the study area. Another statistical parameter for earthquakes is the b-value. The characteristics of the event clusters, such as the swarm or normal tectonic activity, can be determined using the b-value. The average b-value is calculated to be 0.9, which is close to a value that is considered “normal”. It can be said that the Karaburun activity is caused by normal tectonic stress changes. The phase readings of the events are performed carefully, and catalogue contamination is prevented from quarry blasts. The hourly event histograms given in Figure 2C clearly show that the catalogue does not contain major seismic activity during the daytime hours. When I split the dataset using a threshold magnitude (MC = 1.4), I found that the micro-earthquake activity during the daytime is less than that of the night-time activity. However, as shown in Figure 2D, the number of events greater than the selected threshold magnitude does not vary significantly with respect to day and night. This is probably the result of the increasing level of the ambient noise with the contribution of human activity (e.g., traffic and industry) during the daytime (Atef et al., 2009).

Several factors, such as network geometry, phase reading quality, and crustal structure uncertainties, control the location errors. I used a double-difference inversion algorithm (hypoDD) developed by Waldhauser and Ellsworth (2000). The algorithm assumes that the hypocentral separation between two earthquakes is small compared to the event-station distance and the scale length of velocity heterogeneity such that the ray paths between the source region and a common station are similar along almost the entire ray path. If so, the difference in travel times for two events observed at one station can be accurately attributed to the spatial offset between the events (Fréchet, 1985; Got et al., 1994; Waldhauser and Ellsworth, 2000). By linking thousands of earthquakes together through a chain of nearby shocks, it is possible to obtain high-resolution hypocentre locations over large distances without the use of station corrections. I used absolute location parameters of the events from the TURDEP catalogue as the starting point of the re-location inversion procedure. The P/S travel time differences between the event pairs were the main inputs of the inversion. Moreover, the high-precision differential travel times from the phase correlation of P- and/or S-wave (cross-correlation, CC) data were also used. The differential travel-time data, with a relative timing precision of approximately tens of msecs, allows for the calculation of the relative location between earthquakes with errors of only a few hundreds of meters.

I used a search radius of 5 km to find neighbouring events and chose the minimum 8 phase observations as a threshold for an event pair. This is the minimum value used to solve unknown parameters of pairs (6 for space and 2 for time). The searched neighbouring events were also used to determine the waveform similarities. The waveform similarities of the events were determined using the coherence algorithm in the MTSPEC package of Prieto et al. (2009). Two waveforms recorded at a common station were considered similar when 70% of the squared coherency values were between 1 and 10 Hz. The examples of the CC data of two events (9 May 2008 10:56 ML = 1.7 and 10 May 2008 02:54 ML = 1.4) that occurred on the northern coast of Sığacık Bay are shown in Figure 3. The P-waveforms recorded at common stations (ZEYE, BEYE and DGBE) during the two events are very similar. The travel-time differences are obtained from the lag of the maximum CC amplitude (Tcc). The average coherency value (C2xy) is used as the weight of the event pairs in the inversion. The high correlation (approximately 80%) of the waveforms and the small time lags (120–130 msec) indicates that these two events might occur on the same fault surface. In the inversion procedure, I used 4 main groups of iterations: the first 5 iterations are for down-weighting (a priori) the cross-correlation data to allow the catalogue data to restore the large-scale picture. The next 5 iterations were used for re-weighting the catalogue data according to the misfit and event separation to remove or down-weight outliers. In the following iterations (11–21), the locations were improved using the catalogue data with separations smaller than 3 km. In the final 10 iterations, I improved the locations of the events by using the waveform cross-correlation data, while the large separation distances were controlled by the catalogue data. Two different damping values were used in this study: 80 for the first and second iteration groups and 60 for the last two groups. Although the damping factor strongly depends on the condition of the system to be solved, I attempted to use a condition number less than approximately 80 in the inversion steps, as described by Waldhauser (2001). This indicated that the re-location solution depends on the data. After the double-difference inversion, half of the events (approximately 2520) in the catalogue were improved. The location uncertainties of the large dataset can be estimated using statistical resampling methods (Waldhauser and Ellsworth, 2000). I used the statistical approach described by Tan et al. (2010). The random numbers that agreed with the absolute location errors (i.e., ±2–3 km) were added to the initial locations of the events in the X, Y and Z directions. This allows for the initial hypocentre location to randomly shift in space and to randomly re-link events in the inversion. The randomly shifted event pairs were re-located with the repeated 1200 well-conditioned inversions. More than 2 million event locations were obtained to include in the statistical distribution. The outliers in the dataset were removed using the interquartile range (IQR) method. More than 90% of the total samples remained after the IQR analysis. The dataset represents a normal distribution, and the 95% confidence interval (±2σ) represents the location variation interval according to the different initial models. The horizontal and vertical location errors were calculated as ±1,000 m and ±2,000 m, respectively.

Table 2. Source parameters of the Karaburun earthquakes. Solution types: FM: First motion polarity,

MT: Regional moment tensor inversion, TI: Teleseismic waveform inversion.

The earthquake fault plane solutions given in Table 2 were determined using the P-wave first motion (FM) polarities using the focmec algorithm (Snoke et al., 1984) and the ISOLA (Sokos and Zahradnik, 2008) local moment tensor (MT) inversion algorithm. If there are enough local seismogram records for full waveform inversion, the faulting parameters from the MT inversion become much more reliable than the FM solutions. Because of the large azimuthal gap problem and/or misinterpretation of polarities, the FM polarities may indicate alternative solutions. The MT inversion may overcome the problems in FM solutions. The events with multiple acceptable solutions indicating different mechanisms or with faulting parameter uncertainties exceeding ±10° were not reported in this study. The best way to determine fault plane solutions of larger (M > 5.5) events is teleseismic waveform inversion. The three moderate earthquakes (M > 5.5, #3-4-5 in Table 2) with reliable waveform data from global seismic stations were analysed with the MT5 (Zwick et al., 1994) teleseismic waveform inversion algorithm. The details of the fault plane solutions are provided in Appendix A.

3 EARTHQUAKE ACTIVITIES AND CHARACTERISTICS

 3.1 The earthquake activity of the Karaburun Nest

The earthquake activity in the central part of the Karaburun Peninsula started in November 2007 (Figure 4, 5). Because the deformation in the area generates a larger number of micro-earthquakes, it may be called an earthquake nest. The total number of events recorded in the observation period is approximately 3,276 in the Karaburun Nest (KN) (Figure 5A). This is 63% of the total events in the study area. The events on the land align along a NW-SE zone. The temporal behaviour of the activity is demonstrated in Figure 5B, and over time, as many as eight clusters can be identified. They localize in two main areas: one is around Uzunkuyu village (U), and the other is in Sığacık Bay (S). The first activity was observed in the sub-cluster U1 (6–14 November 2007, approximately 290 events). This activity began with a small number of micro-earthquakes (ML ≤ 2.0), and the most active day was 8 November, with 115 events. There were eight events with ML > 3.0. The largest event of U1 occurred on 11 November 03:23 (UTC) ML = 4.1. U2 is a small sub-cluster and was activate from 12–25 December. The sub-cluster S1 appeared on the northern coast of Sığacık Bay in May 2008. S1 generated more than 190 events during two weeks and had two events with ML = 3.5. After that, a scattered micro-earthquake sequence occurred between U1 and U3 (south of Uzunkuyu village) in June and July 2008. There was no significant activity in the following year (Figure 5B). Then, a new sub-cluster (U3) started to generate micro-earthquakes on 26 July 2009, and 260 events were recorded during approximately 1.5 months. The most recent large activity with approximately 300 events (ML < 3.0) occurred during May and June 2010 in a different sub-cluster (U4) near the Uzunkuyu village. The determined sub-clusters constitute 40% of the total events that occurred in the Karaburun Peninsula (Figure 5). Although there are two sub-clusters (S2 and S3) in Sığacık Bay, I could not detect a temporal clustering. However, these sub-clusters are observed in the 2005 Sığacık earthquake sequence zone (Aktar et al., 2007). There are 280 and 385 micro-earthquakes in S2 and S3, respectively.

3.1.1 Uzunkuyu Cluster (U1–U4)

The earthquake sub-clusters in the vicinity of the Uzunkuyu village was observed during different time periods (Figure 5). The catalogue is complete down to MC = 1.2 for this area, and the b-value is 0.96±0.04. The Gutenberg-Richter diagrams of each sub-cluster were also analysed to identify the earthquake sequence type. The b-values of the four sub-clusters (U1–U4) are between 0.7 and 1.1 (±0.1). Because the values are close to “normal”, indicating a normal tectonic activity, the sequences cannot be interpreted as swarms.

I re-located approximately 1,200 events using the double-difference inversion in the sub-clusters U1–U4 (Figure 6A). The results reveal that the micro-earthquakes span the narrow zones (i.e., U3 and U4). The fault plane solutions of the larger events (ML > 3.5) indicate a strike-slip motion. The direction of the epicentres is in good agreement with the NW-SE nodal planes. By the classic view, events are spread on the fault surface when examining a fault parallel depth section, and the hypocentres on a strike-slip fault align on a vertical line. However, the depth sections of the events around the Uzunkuyu village produce a strange image in the crust. Figure 6B shows the parallel and normal cross-sections according to the probable NW-SE left-lateral strike-slip fault. The cross-section along the profile A-A’ shows that the hypocentre depths of the U4 events increase from SE to NW. The events occur in a 2 x 2 x 9 km3 volume and are concentrated between the depths of 6 and 8 km. A similar picture is observed on the B-B’ profile, which is parallel to the A-A’ profile. The profile B extends over the sub-clusters U1 (in 3 x 3 x 12 km3) and U3 (in 1 x 1 x 12 km3). The hypocentres in both of the sub-clusters become deeper from SE to NW, as in U4. According to the depth sections, the sub-clusters U1, U3 and U4 dip NNW, NW and NW, respectively. However, it is difficult to identify the dipping orientation of the small sub-cluster U2. The activities are generated in the narrow oblique zones between the depths of 3 and 12 km, as observed in the fault parallel depth sections (A-A’ and B-B’ in Figure 6B). The normal fault depth sections of U3 and U4 clearly demonstrate that the fault plane is vertical (C-C’ and D-D’ in Figure 6B). Also visible on the map is a gap between the sub-clusters located north and south of the Uzunkuyu village. The width of this gap is approximately 1 km. The gap between U1 and U4 can also be clearly seen in the C-C’ section.

 3.1.2 Sığacık Cluster (S1-3)

The micro-earthquake activity of Sığacık Bay first started in 2005 and has continued until the present. The 2005 sequence reported by Benetatos et al. (2006) and Aktar et al. (2007) is shown with the white circles in Figure 7. The clusters following the main three events that occurred in 2005 are not the exact same areas as the present sub-clusters (grey circles). The events of S3 occurred at the NE end of the previous activity in the centre of the bay. The new data show that the activity does not occur only at the centre of the bay; it occurs all along the bay between the northeast and southwest coastlines. The size of this seismic zone is approximately 2 x 15 km2. The new events in the S2 sub-cluster show that the gap mentioned by Aktar et al. (2007) between the two parts of the bay is filled. If the earthquake distribution is examined over a large-scale map as in Figure 7, a single NW-SE seismic zone is evident in Sığacık Bay. The SE-NW conjugate seismic zone starts from the sub-cluster S1. There are also two small isolated earthquake delineations located in the west of the S1 sub-cluster, and they indicate NW-SE left-lateral strike-slip faulting (Aktar et al., 2007).

Figure 8A shows the re-located events with the double-difference inversion algorithm in Sığacık Bay. The fault plane solutions indicate that the sub-clusters S1 and S3 have strike-slip mechanisms. S1 is located at the end of the continuation of the Uzunkuyu activity and has a NW-SE left-lateral character. Because of the SW-NE trend of the events in the bay mentioned above, the sub-cluster S3 developed on a right-lateral fault system. The maximum depth of the events in the clusters is approximately 10 km, and some of the events also occurred between the depths of 10 and 15 km (Figure 8B). The most important point is that the oblique earthquake delineations, as seen in the cross-sections of the Uzunkuyu clusters, are not observed in Sığacık.

3.2 Other clusters

The centre of the Karaburun Peninsula is the most seismically active part of the study area (Figure 4). The small clusters in the north of Seferihisar and on the southwest border of the Gerence Bay are located at both ends of the Karaburun Nest. The dominant earthquake activities in the east of Foça and southwest of Çeşme are the aftershocks of the 5 January 2008 (#12, ML = 4.2) and the 26 March 2010 (#19, ML = 4.8) mainshocks, respectively. The seismicity in İzmir Bay indicates that there are several active faults in the bay (as shown by Ocakoğlu et al., 2005).

Gerence Bay (GB) is roughly rectangular in shape, oriented NW-SE, and enclosed by two small peninsulas (Figure 4). There are approximately 80-recorded events, which bound the bay from the southwest. This small cluster extends in the NW-SE direction and is in good agreement with the shape of the Gerence Bay. Because of the small number of the events and the weak network coverage for this sub-region, most of the events could not be re-located with the HypoDD algorithm, and there is no reliable fault plane solution from recent activities. One of the closest events occurred on 23 December 2011 (#21, ML = 3.9) and indicates NW-SE left-lateral strike-slip faulting. It is most probable that the GB cluster is the NW continuation of the Karaburun Nest. The largest event close to the bay occurred on 6 April 1969 (#2, M = 5.6). McKenzie (1972) has been reporting a NW-SE striking normal fault mechanism for this event. The SW dipping nodal plane that is in agreement with the coastline may be the fault plane, and the event is associated with the NE-SW extension. Event #18 (ML = 4.2) also has a strike-slip mechanism, similar to those of the Karaburun Nest. The largest event (#1, M = 6.8) observed in the study area exhibits a strike-slip mechanism with a normal component, and the NW-SE nodal plane is in good agreement with the activity in the south. The focal depths of the events are between 5 and 15 km, and there are no deeper events.

The cluster observed in southwest Çesme extends in the NW-SE direction. The mainshock (#19, ML = 4.8) and its largest aftershock (#20, ML = 4.2) were followed by approximately 100 events (Figure 4). The fault plane solutions, calculated using the regional moment tensor inversion algorithm, indicate normal faulting with the strike-slip component for both of the events. If the NW-SE nodal planes, which are compatible with the delineation of the event distributions, are assumed to be the fault planes, a southwest dipping normal fault with a left-lateral strike-slip component is responsible for that cluster.

The cluster in the Seferihisar region contains approximately 190 events and extends in the ESE-WNW direction (Figure 4). The two moderate-sized events (#3: 6 November 1992 Mw = 6.0 and #5: 10 April 2003 Mw = 5.7) also occurred in the same region. The teleseismic waveform inversions corresponding to the events indicate right-lateral strike-slip faulting if the SW-NE nodal plane is chosen for the fault plane. In contrast, the recent activity seems to be perpendicular to the main fault traces. There are no additional data to clarify the source of the cluster to the north of Seferihisar town.

During the study, more than 350 events were located in İzmir Bay. The first noticeable property of the earthquakes is that the NW-SE direction of the seismicity, which is parallel to the bay (Figure 4). There are three fault plane solutions at the entrance of the bay. Both events, #4 and #17, present similar NW-SE normal faulting. However, the event (#22) that occurred in the northernmost area of the Karaburun Peninsula indicates strike-slip faulting. The double-difference location analysis of the İzmir Bay earthquakes does not localize the clusters to reveal the fault structure determined from seismic profiles (Aksu et al., 1987; Ocakoğlu et al., 2005) in the crust. Furthermore, there is no large number of correlated events, such as for the Karaburun Nest, and the inversion algorithm discards the uncorrelated events. However, the delineation of the epicentres seems to be in agreement with the normal faults that bound İzmir Bay (Aksu et al., 1987), in contrast to Ocakoğlu et al. (2005). Based on the seismic data, Ocakoğlu et al. (2005) assumed that the main character of the faults in this area is the strike-slip in the N-S direction. There is an apparent conflict between the seismic studies and the micro-earthquake activity reported in the present study.

The single cluster on the land was observed northwest of İzmir Bay. The Menemen earthquake (#12, 5 January 2008, ML = 4.2) exhibits a strike-slip faulting. Its aftershock hypocentres are mainly located in the same area and do not provide any indication regarding which nodal plane is the actual fault plane. This event might have occurred on a NE-SW right-lateral fault similar to the SFZ and OFZ, and the fault is most likely buried under the Quaternary alluvium.

A significant earthquake activity was observed in April-May 2012 in the İzmir Bay. Because the activity started at the end of the review process of this article, I mentioned the properties of these events in Appendix B. This recent earthquake indicates that the oblique faults in the bay are capable of generating significant earthquakes.

4. Discussions and Conclusions

Because of the lack of a sufficient dense seismology network before 2007, the seismological characteristics of the Karaburun Peninsula and its surroundings were not determined. The 2005 earthquake activity mentioned in the previous sections is the first evidence for the seismicity of Sığacık Bay. After that, significant seismic activity began at the end of 2007 that provided the opportunity for observation with our high-density seismology network. I call this most seismically active area a “nest” because the earthquakes occur as different small clusters in an area 20 x 20 km2.

The event clusters of the Karaburun Nest near the Uzunkuyu village exhibit NW dipping characteristics (Figure 6). The micro-earthquake activities occur in thin volumes (approximately 2 x 2 x 10 km3) and do not mimic a strike-slip fault surface. The uncertainties of both absolute and relative event locations do not affect the shape of the event distribution. Based on the correlation between the seismology and previously published local tectonic features, the event focal distribution covers the Miocene-aged cold Uzunkuyu Intrusive that is located somewhat in the middle of the Karaburun Seismic Zone between the Sığacık and Gerence bays. The largest possible area of the Uzunkuyu Intrusive (UI) outcrops compiled from Çakmakoğlu and Bilgin (2006) and Helvacı et al. (2009) is demonstrated with the black polygon in Figure 6A. It is located between the sub-clusters U1 and U4. The left-lateral motion in the zone seems to generate earthquakes around the intrusive and causes dense micro-earthquake activity. The s-shaped gap between U1 and U4 (see Figure 6B, section C-C’) reflects the possible shape of the intrusive in the upper crust. The other two sub-clusters (U2 and U3), located approximately 5 km SE of the Uzunkuyu village may indicate other parts of the intrusive that do not reach the surface. It is not possible to identify the structure of the intrusive(s) in the region with the available earthquake dataset. Future studies (i.e., deep seismic and electromagnetic profiles) may help image the deep structure of the region.

Sığacık Bay has three sub-clusters, and two of them are also discussed by Aktar et al. (2007). The sub-cluster S1 indicates NW-SE left-lateral strike-slip motion, and S2 has similar characteristics. The characteristic of S3 is indicated by an earthquake, which has strike-slip motion with a normal faulting component solution. Aktar et al. (2007) analysed the strong motion waveforms of the main events in 2005, found that the rupture direction of the two events (#7 and #8 in Table 2) is NE and provided the interpretation that this is evidence of a NE-SW fault. Çakır (2010) used InSAR data and reported that the mainshock of the 2005 sequence occurred on a NE-SW fault. The newly observed sub-cluster, which is labelled S3, is located NE of the 2005 aftershocks, and this may indicate that earthquake activity migrates to the NE (Figure 7). Although the NE-SW Sığacık Bay activity is not observed on land, these events may be related to the segments of SFZ mapped by Sözbilir et al. (2009). The right-lateral strike-slip motion observed in 2005 might have led to a decrease in the stress level on its conjugate KSZ and might have triggered high micro-earthquake activity at the northern part of the zone. Furthermore, the 1992 (#3) and 2012 (#24) earthquakes occurred in the OFZ. Zhu et al. (2006) indicate that the 10 April 2003 earthquake (#5) is related to the newly named Beyler Fault. However, I purpose that, considering the location uncertainties, the moderately sized 2003 earthquake occurred in the SFZ. The data given in this study do not show isolated seismic activity that indicates a new fault as mentioned by Zhu et al. (2006). These events exhibit SW-NE right-lateral strike-slip faulting in the cases of the active faults. The SW-NE right-lateral faults may stop the propagation of the SE-NW left-lateral KSZ further to the SE.

The other important observation in this study is the lack of micro-earthquake activity on the three main fault zones (Figure 1). Although the strike-slip SFZ and OFZ are related with the two moderately sized events as mentioned previously, their micro-earthquake activity is very low. Some parts of the zones also never generated micro-earthquakes in the last 4.5 years. Similarly, there is no cluster associated with the GFZ, except in the southern end of the zone.

The October 2005 series of earthquakes in the Gulf of Sığacık and the following Karaburun earthquake activity (2007–2011) that occurred in a NW-SE orientation reveal both dextral and sinistral strike-slip faults in this transition zone (Western Turkey) dominated by the N–S extension.

The recent GPS-based studies indicate a velocity anomaly around the Karaburun Peninsula and its surroundings (Nyst and Thatcher, 2004; Aktuğ and Kılıçoğlu, 2006; Aktuğ et al., 2009; Reilinger et al., 2010) and define a microplate border (Figure 9). Nyst and Thatcher (2004) proposed three segments and showed that the South Aegean moves to approximately to the SE relative to Anatolia. One of the segments (labelled with 9) is located in İzmir Bay. The fault normal and parallel slips are 15 and 9 mm/yr, respectively. The other segment (10) was described by an only 18 mm/yr fault parallel slip. The slips on both of the segments are left-lateral and are in good agreement with the left-lateral fault plane solutions in the Karaburun Nest. The earthquake slip-vectors indicate a SE motion between Sığacık Bay and the northern end of the Karaburun Peninsula. The events occurring in the SFZ, OFZ and Sığacık Bay have SW-oriented slip vectors, and these are associated with a conjugate system. I also selected the SW-oriented slip vectors of the three events that occurred in the north of the peninsula because they are compatible with the boundary segment (8) given by Nyst and Thatcher (2004). The residual vectors between the observed GPS velocities and the results of the block modelling demonstrate that there is an un-modelling deformation in this area (Nyst and Thatcher, 2004; Aktuğ et al., 2009). The left-lateral micro-earthquake activity on the Karaburun Seismic Zone presented in this study can be interpreted as a seismic attribute of the microplate boundary in the region. The Karaburun Seismic Zone may be suggested as a part of the microplate boundary (grey thick line in Figure 9). Thus, a more detailed geodetic study of the region can provide further constraints on the geometry of this microplate boundary.

The results of this study may be summarized as follows:

1.      The dense seismicity of the Karaburun Peninsula indicates a zone between the Sığacık and Gerence Bays: Karaburun Seismic Zone (KSZ). The cluster activities in the zone are not swarms (b-values = 0.7-1.1); the seismicity is caused by normal tectonic activity.

2.      The 2005 Sığacık earthquake sequence might have decreased the normal stress on the left-lateral KSZ, and this might have caused the micro-earthquake activity in the KSZ.

3.      The friction zone around the Uzunkuyu Intrusive that is located in the KSZ probably leads to earthquake sub-clusters in the Karaburun Nest. The high-resolution micro-earthquake data indicate that the intrusive body dips to the NW.

4.      The KSZ is one of the segments of the boundary between the Anatolian and South Aegean microplates. However, this must be tested by block modelling studies. The large number of micro-earthquakes can be interpreted as creeping in the KSZ, but this must be supported by different methods (i.e., geodetic measurements).

Acknowledgements

I appreciate the support I have received from Dr. Semih Ergintav, Dr. Sedat İnan and Zümer Pabuçcu. I thank A. Yörük, M.C. Tapırdamaz, A. Karaaslan, F. Sevim, C. Açıkgöz, A. Cankurtaranlar, A. Tarancıoğlu, Dr. Z. Çakır, Dr. M. Aktar, Dr. E. Papadimitriou, Dr. V. Karakostas and K. Leptokaropoulos. I thank the guest editor (Dr. E. Bozkurt) and two reviewers (Dr. H. Sözbilir; an anonymous) who provided insightful comments that improved the manuscript. This study was fully supported by the “TÜRDEP Project” (TÜBİTAK 105G019) and was partially supported by Turkey-Greece joint project “Seismotectonic Properties of the Eastern Aegean: Implications on the Stress Field Evolution and Seismic Hazard Assessment in a Tectonically Complex Area” (TÜBİTAK 109Y401). The earthquake waveform data are stored on the TRUBA Infrastructure operated by TÜBİTAK ULAKBİM. All figures were drawn using the Generic Mapping Tools (Wessel and Smith, 1998). The research leading to these results has received funding from the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement n° [262229], [EPOS]

References

Aksu, A.E., Piper, D.J.W., Konuk, T., 1987. Late Quaternary tectonic and sedimentary history of outer Izmir and Candarli Bays, Western Turkey. Marine Geology 76, 89-104. doi: 10.1016/0025-3227(87)90019-3

Aktar, M., Karabulut, H., Ozalaybey, S., Childs, D., 2007. A conjugate strike-slip fault system within the extensional tectonics of Western Turkey. Geophysical Journal International 171 (3), 1363-1375.

Altınok, Y., Alpar, B., Özer, N., Gazioğlu, C., 2005. 1881 and 1949 earthquakes atthe Chios-Cesme Strait (Aegean Sea) and their relation to tsunamis. Natural Hazards and Earth System Sciences 5, 717-725.

Aktuğ, B., Kılıçoğlu, A., 2006. Recent crustal deformation of Izmir, Western Anatolia and surrounding regions as deduced from repeated GPS measurements and strain field. Journal of Geodynamics 41, 471–484.

Aktuğ, B., Nocquet, J.M., Cingöz, A., Parsons, B., Erkan, Y., England, P., Lenk, O., Gürdal, M.A., Kilicoglu, A., Akdeniz, H., Tekgül, A., 2009. Deformation of western Turkey from a combination of permanent and campaign GPS data: Limits to block-like behavior. Journal of Geophysical Research 114, B10404, doi:10.1029/2008JB006000.

Akyol, N., Zhu, L., Mitchell, B.J., Sözbilir, H., Kekovalı, K., 2006. Crustal structure and local seismicity in western Anatolia. Geophysical Journal International 166, 1259-1269, doi: 10.1111/j.1365-246X.2006.03053.x

Angelier, J., 1978. Tectonic evolution of the Hellenic Arc since the late Miocene. Tectonophysics 49, 23-36.

Atef, A.H., Liu, K.H., Gao, S.S., 2009. Apparent weekly and daily earthquake periodicities in the Western United States. Bulletin of the Seismological Society of America 99 (4), 2273-2279.

Benetatos, C., Kiratzi, A., Ganas, A., Ziazia, M., Plessa, A., Drakatos, G., 2006. Strike-slip motions in the Gulf of Siğaçik (western Turkey): Properties of the 17 October 2005 earthquake seismic sequence. Tectonophysics 426, 263-279.

Billiris, H., Paradissis, D., Veis, G., England, P., Featherstone, W., Parsons, B., Cross, P., Rands, P., Rayson, M., Sellers, P., Ashkenazi, V., Davison, M., Jackson, J., Ambraseys, N., 1991. Geodetic determination of tectonic deformation in central Greece from 1900 to 1988. Nature 350, 124-129.

Bozkurt, E., 2001. Neotectonics of Turkey - a synthesis. Geodinamica Acta 14, 3-30.

Bozkurt, E., 2003. Origin of NE-trending basins in western Turkey. Geodinamica Acta 16, 61-81.

Bozkurt, E., Rojay, B., 2005. Episodic two-stage Neogene extension and short-term intervening compression in Western Turkey: field evidence from the Kiraz Basin and Bozdağ Horst. Geodinamica Acta 18, 299-316

Çakır, Z, 2010. Türkiye diri fayları boyunca meydana gelen kabul deformasyonun açıklık radar interforemetrisi (InSAR) yöntemiyle incelenmesi [Investigation of the crustal deformation on the active faults in Turkey using InSAR technique]. Project Final Report, TÜBİTAK Project Number 107Y281.

Çakmakoğlu, A., Bilgin, Z.R., 2006. Karaburun Yarımadası’nın Neojen öncesi stratigrafisi [The pre-Neogen stratigraphy of the Karaburun Peninsula]. MTA Dergisi 132, 33-62. 

Dewey, J.F., Sengör, A.M.C., 1979. Aegean and surrounding regions: complex multiple and continuum tectonics in a convergent zone. Geological Society of America Bulletin 90, 84-92.

Emre, Ö., Barka, A., 2000. Active faults between Gediz Graben and Aegean Sea (İzmir Region). BADSEM 2000 Symposium Abstract Book, 24-27 May 2000, İzmir, Turkey.

Emre, Ö., Özalp, S., Doğan, A., Özaksoy, V., Yildirim, C., Göktaş, F., 2005. İzmir yakın çevresinin diri fayları ve deprem potansiyelleri [The active faults of the İzmir region and their earthquake potentials]. Mineral Research and Exploration Institute of Turkey (MTA) Report, No:10754.

Fréchet, J., 1985. Sismogenèse et doublets sismiques. Thèse d’Etat, Université Scientifique et Médicale de Grenoble.

Genç , C., Altunkaynak, Ş., Karacık, Z., Yazman, M., Yılmaz, Y., 2001. The Çubukludağ graben, south of İzmir: tectonic significance in the Neogene geological evolution of the Western Anatolia. Geodinamica Acta 14, 1-12.

Goldstein, P., Dodge, D., Firpo, M., Stan, R., 1998. Electronic seismologist: what’s new in sac2000? Enhanced processing and database access. Seismological Research Letters 69, 202-20.

Got, J.L., Frechet, J., Klein, F.W., 1994. Deep fault plane geometry inferred from multiplet relative relocation beneath the south flank of Kilauea. Journal of Geophysical Research 99, 15375-15386.

Hancock, P.L., Barka, A.A., 1987. Kinematic indicators on active normal faults in western Turkey. Journal of Structural Geology 9, 573-584. 

Helvacı, C., Ersoy, E.Y., Sözbilir, H., Erkül, F., Sümer, Ö., Uzel, B., 2009. Geochemistry and 40Ar/39Ar Geochronology of Miocene volcanic rocks from the Karaburun Peninsula: implications for amphibole-bearing lithospheric mantle source, Western Anatolia. Journal of Volcanology and Geothermal Research 185, 181-202. doi:10.1016/j.jvolgeores.2009.05.016.

İnan, S., Ergintav, S., Saatçılar, R., Tüzel, B., İravul, Y., 2007. Turkey makes major investment in earthquake research. EOS Transaction AGU  88(34), 333-334.

Kaya, O. 1981. Miocene reference section for the coastal parts of West Anatolia. Newsletters on Stratigraphy 10, 164-191.

Koçyiğit, A., 2005. The Denizli graben-horst system and the eastern limit of western Anatolian continental extension: basin fill, structure, deformational mode, throw amount and episodic evolutionary history, SW Turkey. Geodinamica Acta 18, 3-4.

Le Pichon, X., Angelier, J., 1979. The Hellenic Arc and trench system: A key to the neotectonic evolution of the eastern Mediterranean area. Tectonophysics. 60, 1-42.

Le Pichon, X., Chamot-Rooke, N., Lallemant, S., Noomen, R., Veis, G., 1995. Geodetic determination of the kinematics of central Greece with respect to Europe. Journal of Geophysical Research 100, 12,675-12,690.

Lienert, B.R.E., Havskov, J., 1995. A computer program for locating earthquakes both locally and globally. Seismological Research Letters 66, 26-36.

McCaffrey, R., Abers, G., 1988. SYN3: A program for inversion of teleseismic body waveforms on microcomputers. Air Force Geophysics Laboratory, AFGL-TR-88-0099, Hanscomb Air Force Base, MA, U.S.A.

McClusky, S., et al., 2000. Global Positioning System constraints on plate kinematics and dynamics in the eastern Mediterranean and Caucasus. Journal of Geophysical Research 105, 5695-5719.

McKenzie, D.P., 1972. Active tectonics of the Mediterranean region. Geophysical Journal of the Royal Astronomical Society 30, 109-185.

McKenzie, D.P., 1978. Active tectonics of the Alpine Himalayan Belt, the Aegean Sea and surrounding regions. Geophysical Journal of the Royal Astronomical Society 55, 217-252.

Meade, B.J., Hager, B.H., McClusky, S.C., Reilinger, R., Ergintav, S., Lenk, O., Barka, A., Ozener, H., 2002. Estimates of seismic potential in the Marmara Sea region from block models of secular deformation constrained by Global Positioning System measurements. Bulletin of the Seismological Society of America 92, 208-215.

Molnar, P., Lyon-Cáen, H., 1989. Fault plane solutions of earthquakes and active tectonics of the Tibetan Plateau and its margins. Geophysical Journal International 99, 123-153.

Nyst, M., Thatcher, T., 2004. New constraints on the active tectonic deformation of the Aegean. Journal of Geophysical Research 109, B11406. doi:10.1029/2003JB002830.

Ocakoğlu, N., Demirbağ, E., Kuşçu, İ., 2005. Neotectonic structures in İzmir Gulf and surrounding regions (westernTurkey): evidences of strike-slip faulting with compression in the Aegean extensional regime. Marine Geology 219, 155-171.

Prieto, G.A., Parker, R.L., Vernon III, F.L., 2009. A Fortran 90 library for multitaper spectrum analysis. Computer & Geosciences 35, 1701-1710.

Reilinger, R., McClusky, S., Paradissis, D., Ergintav, S., Vernant, P., 2010. Geodetic constraints on the tectonic evolution of the Aegean region and strain accumulation along the Hellenic subduction zone. Tectonophysics 488, 22-30. doi:10.1016/j.tecto.2009.05.027.

Snoke, J.A., Munsey, J.W., Teague, A.G., Bollinger, G.A., 1984. A program for focal mechanism determination by combined use of polarity and SV/P amplitude ratio data. Earthquake Notes 55, 3, 15.

Sokos, E., Zahradnik, J., 2008. ISOLA a Fortran code and a Matlab GUI to perform multiple-point source inversion of seismic data. Computers & Geosciences 34, 967-977.

Sözbilir, H., Sümer, Ö., Uzel, B., Ersoy, Y., Erkül, F., İnci, U., Helvacı, C., Özkaymak, Ç., 2009. 17-20 Ekim 2005 Sığacık Körfezi (İzmir) depremlerinin sismik jeomorfolojisi ve bölgedeki gerilme alanları ile ilişkisi, Batı Anadolu [The seismic geomorphology of the Sığacık Gulf (İzmir) earthquakes of October 17 to 20, 2005 and their relationships with the stress field of their Western Anatolian region]. Geological Bulletin of Turkey 52 (2), 217-238.

Şengör, A.M.C., Satır, M., Akkök, R., 1984. Timing of tectonic events in the Menderes Massif, western Turkey: implications for tectonic evolution and evidence for Pan-African basement in Turkey. Tectonics 3, 693-707.

Tan, O., Tapırdamaz, M.C., Ergintav, S., İnan, S., et al. 2010. Bala (Ankara) earthquakes: Implications for shallow crustal deformation in Central Anatolian Section of the Anatolian Platelet (Turkey). Turkish Journal of Earth Sciences 19, 449-471 doi: 10.3906/yer-0907-1.

Tan, O., Taymaz, T., 2006. Active tectonics of the Caucasus: Earthquake source mechanisms and rupture histories obtained from inversion of teleseismic body waveforms, in: Dilek, Y. and Pavlides, S. (Eds.), Postcollisional Tectonics and Magmatism in the Mediterranean Region and Asia. Geological Society of America Special Paper 409, 531–578.

Taymaz, T., Jackson, J., McKenzie, D., 1991. Active tectonics of the north and central Aegean Sea. Geophysical Journal International 106, 433-490.

Taymaz, T., Jackson, J.A., Westaway, R., 1990. Earthquake Mechanisms in the Hellenic Trench near Crete. Geophysical Journal International 102, 695-731.

Uzel, B., Sözbilir, H., 2008. A First record of strike-slip basin in Western Anatolia and its tectonic implication: The Cumaovası basin as an example. Turkish Journal of Earth Sciences 17, 559-591.

Uzel, B., Sözbilir, H., Özkaymak, Ç., 2012. Neotectonic evolution of an actively growing superimposed basin in western Anatolia: The inner bay of İzmir, Turkey. Turkish Journal of Earth Sciences 21 (4), 439-471. doi:10.3906/yer-0910-11.

Waldhauser F., 2001. hypoDD -- A program to compute double-difference hypocentre locations. USGS Open File Report 01-113.

Waldhauser F., Ellsworth, W.L., 2000. A double-difference earthquake location algorithm: method and application to the northern Hayward fault, California. Bulletin of the Seismological Society of America 90, 1353-1368.

Wessel, P., Smith, W.H.F., 1998. New, improved version of the Generic Mapping Tools released. EOS Transaction AGU 79, 579.

Zhu, L., Akyol, N., Mitchell, B.J., Sözbilir H., 2006. Seismotectonics of western Turkey from high resolutions and moment tensor determinations. Geophysical Research Letters 33 (7), L07316.

Zwick, P., McCaffrey, R., Abers, G., 1994. MT5 Program. IASPEI Software Library, v. 4.

Appendix

Appendix A - Fault Plane Solutions

A.1 Teleseismic Waveform Inversion Solutions

The fault plane solutions of the three events that occurred in the study area are obtained from teleseismic waveform inversion. The source parameters of the 6 November 1992 Doğanbey (Mw=6.0), 24 May 1994 Karaburun (Mw=5.4) and 10 April 2003 Seferihisar (Mw=5.7) earthquakes are summarized in Table 2 and their inversion results are given in Figure A.1-3. The inversion procedure is summarized in the following paragraph.

The digitally recorded seismograms of the three earthquakes were provided from the Incorporated Research Institute for Seismology (IRIS). I inverted the teleseismic body waveforms to obtain the focal mechanism solutions. The approach I followed is that described in previous studies (i.e. Molnar and Lyon-Cáen, 1989; Taymaz et al., 1990; Tan and Taymaz, 2006). I compared the shapes and amplitudes of P- and SH-waveforms recorded by the stations in the teleseismic distance range with synthetic waveforms. To determine source parameters I used the MT5 version of McCaffrey and Abers’s (1988) algorithm (Zwick et al., 1994). The minimum misfit solutions of the earthquakes are shown in Figure A1-3. The solid lines are observed waveforms, and the inversion window is identified by a short vertical bar. Dashed lines indicate synthetic waveforms. The station code is shown to the left of each waveform, together with an upper-case letter that identifies its position on the focal sphere, and a lower case letter indicates the type of instrument (d: deconvolved digital velocity records, d*: digital long period displacement record). The vertical bar beneath the focal spheres shows the scale in microns. Because of the different instrument, two different amplitude scales are given for the 1994 Karaburun earthquake. The source time function (STF) and the time scale used for the waveforms are shown in the middle of the figure. Focal spheres are shown with P (top) and SH (bottom) nodal planes, in lower hemisphere projection. Station positions are indicated by letter, and arranged alphabetically clockwise, starting from north. P and T axes are marked by solid and open circles, respectively. The strike, dip and rake values of the first (generally fault plane) and second nodal planes (NP), centroid depth (h) and seismic moment (Mo) are given at the header.

A.2 P-wave First Motion Polarity Solutions

The fault plane solutions of the five events in the Karaburun Region were determined with P-wave first motion polarities (Figure A.4, Table A.1, Table 2). The take-off angles were calculated according to the same velocity structure used for the location determination. The possible nodal planes which agree with the first motion polarities were searched, running the focmec program (Snoke et al., 1984). No polarity error is allowed in the solutions. The events with multiple acceptable solutions indicating different mechanism or with faulting parameters uncertainties exceeding ±10° were not reported in this study. In Figure A.4, the black and white circles refer to up and down P-wave first motion polarities, respectively. P and T axes are also represented by the solid and open diamonds, respectively.

Table A.1. Number of P-wave polarities used in this study.

A.3 Regional Moment Tensor Solutions

The fault mechanism solutions of the 10 events were determined with the regional moment tensor (RMT) inversion. Only the seismograms that were recorded by stations located between 30 and 200 km were used for the solutions implemented with the ISOLA local moment tensor inversion algorithm (Sokos and Zahradnik, 2008). At least four stations and three components waveforms filtered between the frequency bands of 0.03 Hz - 0.1 Hz were inverted for events. The minimum misfit solutions of the earthquakes are shown in Figure A.5-14. The black and grey lines are un-normalised observed and synthetic waveforms, respectively. The station code is shown to the left of three component waveforms (NS, EW and Z). The percentages of variance reduction (VR) and double couple component (DC) are given below the focal mechanism solution.

Appendix B - 1 May 2012 Foça (İzmir Bay) Earthquake

The most recent activity in the İzmir Bay occurred in April – May 2012 and located in the centre of the bay (Figure B.1, Table B.1). There are six events (M>3.0) before the mainshock with ML = 5.0. The fault plane solutions indicate normal faulting with strike-slip component. The NW-SE nodal planes, which agree with the eastern boundary fault of the bay, are most likely the fault planes. If the fault strike is NW-SE, it is clear that the strike slip component is left-lateral. The previous events (24.05.1994 and 20.07.2008) have also shows similar faulting mechanism (see Figure 4). This left-lateral motion is in good agreement with the Karaburun Deformation Zone. This recent earthquake indicates that the oblique faults in the bay are capable of generating significant earthquakes.

Table B.1. Source parameters of the 1 May 2012 Foça (İzmir Bay) earthquake and its aftershocks.

Solution types: FM: First motion polarity, MT: Regional moment tensor inversion.