Cent. Eur. J. Geosci. • 5(3) • 2013 • 423-434 DOI: 10.2478/s13533-012-0141-8

Central European Journal of Geosciences

Time-dependent analysis of events and structural impacts on intraplate crustal seismicity of the Van (Mw 7.1, 23 October 2011), E-Anatolia

Research Article

Mustafa Toker1∗

1 Yuzuncuyıl University, Department of Geophysical Engineering, Van/Turkey

Received 2013 May 19; accepted 2013 July 31

Abstract: The Van earthquake (MW 7:1, 23 October 2011) in E-Anatolia is typical representative of intraplate . Its thrust focal character and aftershock seismicity pattern indicate the most prominent type of compound earth- quakes due to its multifractal dynamic complexity and uneven compressional nature, ever seen all over Turkey. Seismicity pattern of appears to be invariably complex in its overall characteristics of aligned cluster- ing events. The population and distribution of the aftershock events clearly exhibit spatial variability, clustering- declustering and intermittency, consistent with multifractal scaling. The sequential growth of events during time scale shows multifractal behavior of seismicity in the focal zone. The results indicate that the extensive hetero- geneity and time-dependent strength are considered to generate distinct aftershock events. These factors have structural impacts on intraplate seismicity, suggesting multifractal and unstable nature of the Van event. Multifrac- tal seismicity is controlled by complex evolution of crustal-scale faulting, mechanical heterogeneity and seismic deformation anisotropy. Overall seismicity pattern of aftershocks provides the mechanism for strain softening process to explain the principal thrusting event in the Van earthquake. Strain localization with weakening controls the seismic characterization of Van earthquake and contributes to explain the anomalous occurrence of aftershocks and intraplate nature of the Van earthquake. Keywords: the van earthquake• intraplate • aftershock seismicity • multifractal behavior • strain softening © Versita sp. z o.o.

1. Introduction volcanism [2,3]. Thrust faulting process [4] and aftershock seismicity pattern of the Van earthquake indicate the most prominent type of compound-complex earthquakes [1] due to its multifractal dynamic complexity and uneven com- The Van earthquake (M 7.1, 23 October 2011) occurred w pressional nature, ever seen all over Turkey. in E-Turkey, typically "intraplate" (plate boundary related) The 23 October 2011 Van earthquake (Mw, 7.1) was fol- earthquake [1]](Fig.1b). Its is located nearby lowed by the 9 November 2011 Edremit (Mw, 5.6) earth- "volcanic intraplate environment", surrounded by active quake (Fig.1b), reported by Kandilli Observatory and Earthquake Research Institute (KOERI). Hypocentral and ∗E-mail: [email protected] source parameters of these earthquakes estimated by dif-

423 Time-dependent analysis of aftershock events and structural impacts on intraplate crustal seismicity of the Van earthquake

The 2011 Van earthquake is unique for its unusual focal depth (below 5 km) with a peak slip of about 5.5 m, vertical displacement of 60 cm, the total seismic moment (4.6x1019 Nm) [4] and volcanic intraplate location. The main rupture zone striking N700E and its main cluster was confined to a northeast -southwest area of 70 kmCE 20 km in map view seen in Fig. 2, with about 80% primary slip con- centrated at depths of 5-15 km. Several strong events including magnitudes (Mw≥5.0) occurred along the main- shock area, between Lake Van and Lake Erçek (see LE in Fig. 2). This area appears almost ellipsoidal, also includ- ing the causative fault. It is apparent that the 2011 Van and Edremit events seen in Fig. 1b are caused by the prevailing regional N-S compression [3]. Similarly, two intermediate earthquakes with oblique thrust type having moment magnitudes of 5.4 on 1988 June 25 and 5.3 on 2000 November 15 occurred in the south-eastern part of Lake Van (Fig. 1a). Aftershock data (Fig. 2) reveals that intraplate nature of the Van earthquake can’t be explained with linearly uni- form, elastic fracture mechanics, because the time delays between the observed individual clusters in aftershock seismicity are too long to result from elastic processes. It is clear that the Van earthquake dynamically loads the surrounding volcanic region. Such a compound earth-

Figure 1. a) recorded events (Mw 3.0-7.3 for the last 111 years) in quake can result from viscoelastic relaxation in the im- Lake Van and its vicinity and focal mechanisms of some mediate postseismic period. This results in nonlinear and earthquakes recently occurred (KAN, USGS, EMSC) [43], heterogeneous redistribution of loads, such as volcanic- b) Epicentral distribution and focal mechanism solutions of the October 2011 Van and the November 2011 Edremit magmatic coupling within the accretionary crustal blocks. earthquakes and the main aftershocks (Mw¿5.0) by differ- Seismicity pattern of Van earthquake, as well seen in most ent institutions (EMSC), (KAN: B.U. Kandilli Observatory and Earthquake Research Institute (KOERI); EMSC: Eu- compound earthquakes [1], often has delay times of hours, ropean Middle East Center; AZUR: Nice Uni- days and months (Fig. 1). In the case of delay times versity, GeoAzur Laboratory, France; GFZ: Geoforshung Zentrum, Potsdam, Germany; ERD: Disaster Manage- of months, seismic coupling may be important, as shown ment and Emergency Presidency, Ankara Turkey; HARV: by the earthquake waveforms recorded in Van and high Harvard CMT; INGV: Insituto Nazionale di Geofisicae Vul- seismic b-values [6–8]. The point is that Van earthquake canologia, Italy; USGS: United States Geological Survey [43]. as a composite system has input parameters more than one and strongly characterized by composite sequences of event instabilities. These are well recognized in time- dependent distribution of aftershock seismicity (Fig. 2). ferent organizations and seismological institutes are sum- marized in map view given in Fig. 1b. Aftershocks seen in Fig. 2 associated with the 2011 Van The Van event is the largest thrust earthquake known to still continue to occur till today and provide have occurred in Van area and Turkey, since at least the 6500 events during October 2011-August 2012. This gives 1976 Çaldıran-Muradiye event of Ms, 7.3 (Fig. 1a) [4, 5]. a reliable dataset of aftershocks that enabled this study The 1976 event caused the uplifting (about 16 cm) of the to carry out a detailed investigation of events in the fo- northern shore of Lake Van and was followed by several cal region of the 2011 Van earthquake. Aftershock pat- Ms, 5.0 aftershocks in November 1976 and January 1977. terns used in this study contain all of event features for The 9 November Mw 5.6, 2011 Edremit earthquake (5-7 a given time scale (Fig. 2). These patterns can be exam- km depth) occurred offshore, in the south of Van along the ined for evidence of precursors and characteristic features north dipping, Edremit fault, a normal oblique-strike-slip in overall seismicity [1, 9–11]. Aftershocks are the most fault (Fig. 1b). The epicentre locations and the source ubiquitous, observed to follow almost all shallow tectonic mechanism solutions of these earthquakes indicate that earthquakes of any significant size [1]. They have the most they occurred on different faults. well-defined characteristics of any of the earthquake se-

424 M. Toker

Figure 2. Topographic map of Lake Van area and its vicinity showing major provinces, structural features and epicenter distribution of all events (5304 events) between 2011.10.23 and 2012.03.28 (157 days) extracted from KOERI catalogs. An inset with Turkey shows KOERI seismic network area in E-Anatolia and defines two stations (VANB and VANT) located in Van (VANB) and Edremit (VANT) areas. The colored and dashed line shows Muž Suture (MS) zone, separating Bitlis Pötürge Massive in S from Eastern Anatolia Acretionary Complex in N, (NV: Nemrut Volcano; SV: Suphan Volcano; PVD: Parasitic Volcanic Domes; CC: Collapsed Cones; ÇSZ: Çarpanak Spur Zone; LE: Lake Erçek; F: Fault). Variation of the magnitude, depth and frequency of occurrence (foo) of 5516 events, as a function of time (177 days) during the period from 2011.10.23 to 2012.04.17 within the focal zone. Note that the figure locations used are shown as lines including their time intervals in the foo.

quences. In particular, the decay of aftershock sequences unstability of event propagation in aftershock seismicity. follows the Omori law [12]. This brings an idea that is intraplate seismicity of accre- tions. This kind of seismicity may give a possibility of The long-time anomalous occurrence of the aftershock the extensive investigation of post-collisional [3] rheology events in Van region is typically related to very heteroge- of accretionary complex of E-Anatolia and also the Van neous crustal activity, dynamic thrust faulting and strain earthquake. However, in this study, I briefly aim to show incompatibility problems [13, 14]. Crustal heterogeneity is some structural impacts on intraplate crustal seismicity of characterized by major irregularity in the event rate and

425 Time-dependent analysis of aftershock events and structural impacts on intraplate crustal seismicity of the Van earthquake

the Van earthquake by analyzing of available aftershock (35 s) of tremor-like signals was recorded on 13 and 14 data and to expect clear ideas to comment on this principal November 2000, a few days before the Van earthquake of thrust event. 15 November 2000, Mw 5.3. This event occurred at the southern shore of Lake Van (Fig. 1a) [6]. The spectral analysis of earthquake data from Lake Van provided three 2. Seismicity of Lake Van basin types of seismic events, hybrid event, long period event and tremor, nearby the mainshock area of the October 23 Lake Van basin indicates seismotectonic paradigm and ex- Van earthquake [6]. treme intraplate seismic deformation for crust-forming pro- cess in Turkic-type orogens [3]. The lake is subjected to various seismic activities as given by historical and in- 3. Aftershock data and method strumental records (Fig. 1)[4, 5, 15–19]. Recorded events with magnitudes (3.0≤M≤7.3) for time interval (1900- Earthquake data is provided by KOERI catalogs, pre- 2011) are also given in Fig. 1a. These events include pared by the National Earthquake Monitoring Center the earthquakes with magnitudes M ≥5.0 [4], reflecting of Turkey that has a gradual increase of the num- results of the extensive compilation from the studies given ber of stations operating in E-Anatolia since 1970 (see above. Fig. 2). Aftershock data used is recorded at VANB ◦ ◦ In Lake Van basin, from 1970 to the 23 October 2011 Van (38.595 N, 43.388 E) broadband station of KOERI seismic mainshock, the events with magnitudes 4.0≥M<5.0 are network (Fig. 2). The aftershock activity including recent ◦ ◦ reported [20, 15, 4] (Fig. 1a). The epicentre distribution large event in Van area, E-Anatolia (38.10 N-39.20 N, ◦ ◦ of these events, their source parameters and focal mecha- 41.90 E-44.03 E) was monitored by 19 stations; 14 short- nisms (M ≤5.0) are extensively detailed and documented periods, 5 broadband for the time period be- [4]. The past strong events (M≤5.0) in the instrumental tween October, 2011 and August, 2012 (Fig. 2). Due to period are reported to be a sequence that started in June the single-active broadband station located at the Van city 1900 and lasted several months [4, 21–23]. These events (VANB), localization errors of aftershocks are considered are reported as to be felt strongly in the Van [4], regarding to be about ± 4-8 km, as reported by KOERI. the reported macroseismic effects of the previous studies To show structural nature of the Van earthquake and its [21] (Fig. 1a). aftershocks in E-Anatolia, I used earthquake data set from The 1903 Malazgirt (MS 7.0) is one of the largest earth- KOERI [32] and seismic reflection profiles collected in quakes that occurred in the study area in the instrumen- Lake Van (ICDP-2004, PaleoVan project) [33–35]. I il- tal period [4, 24, 25]. This earthquake had been strongly lustrate the aftershock clustering patterns and event rates felt across Van and Malazgirt area [26] The 1913 Tatvan in the focal zones of the Van (Mw 7.1) and Edremit (Mw event was instrumentally located 27 at about 14 km south 5.6) earthquakes [4]. I used the depths and of Tatvan area (Fig. 1a) and the 1915 (MS 5.6) Nemrut seismic density patterns collected from the seismograms event occurred NE of the Nemrut volcano (Fig. 1a) [4, 20]. of 6500 aftershocks. I present here a summary of time- The 1941 Erciž (MS 5.9-6.0) earthquake caused damage dependent aftershock observations that have been made in the N-NW of Lake Van (Fig. 1a) [4, 24, 25]. The most in the clustering earthquake areas of Lake Van. This fo- destructive event located along the E-shore of the lake is cuses on characteristic sub-sequences for the origin and the 1945 Van earthquake sequence as the largest shock generation of the long-time occurrence of the repeated occurred on 20 November (MS 5.5-5.8) [21, 22, 24, 25, events. It is observed that aftershocks spatially extend 4]. The 1972 Gevaž and 1976 Van earthquakes (MS 5.0) over the focal zone and its vicinity larger than the entire appear to be felt strongly at Van, Edremit and Gevaž [25] rupture zone of the earthquake (Fig. 2). and their focal solutions are also reported [4, 28].

The 1976 Çaldıran (MS 7.3) earthquake is the largest one 3.1. The scaling process and seismicity pa- that occurred in the area (Fig. 1a). This large event rameters ruptured two-segmented fault zone about 40 km N-NW of Lake Van with a 55 km-long [4, 24, 29]. The 1976 Anomalous effects of aftershocks recorded that persist Çaldıran event was followed by several MS 5.0 aftershocks for periods of a few months after the Van earthquake in November 1976 and January 1977 [4]. After the 1988 are briefly reviewed with selected examples in this study Van earthquake in the lake [30] (Fig. 1a), the last record- (Figs. 2-7). The scaling process of aftershock pattern as able event occurred in the region on the November 15, a function of time includes a composite data set. This the 2000 Van earthquake [4, 31] (Fig. 1a). A small part data is composed of a wide range of punctual events with

426 M. Toker

Figure 3. Variation of the magnitude, depth and frequency of occurrence (foo) of 321 events, as a function of time (1,25 days or 31 hours) during the period from 2011.10.23 to 2011.10.24 within the focal zone (see the foo in Fig. 1 for time interval of this figure). The Van earthquake (Mw 7.2) and the following aftershocks (Mw 5.7 and 4.8) are also shown by plot of magnitude variation, recorded in 23.10.2011. Topographic map shows epicenter distribution of 321 events, trending northeast-southwest into the lake basin.

available information between them. Seismicity parame- of events (foo) (Fig. 4). These intervals are well enough ters used are compiled to show variation and correlative to illustrate dynamic fluctuation and propagation of clus- relation in magnitude, focal depth, the frequency of oc- tering events and strong local interactions for short time currence (foo) and the clustering events in the focal area period. The 21th, 38th and 75th days for long-range ac- (Fig. 2). These parameters well illustrate time-dependent tivity are particularly taken into account to show low reso- evolution and propagation of aftershock events. This can lution pattern of considerable and irregular drop of events exhibit an overview of seismic property on a wide range (foo), firstly below about 100 (Fig. 5), continuous sharp of scales in and around the mainshock. drop below 50 (Fig. 6) and slight drop below 30 (Fig. 7). The main approach of this research deals with detection The continual drop of events remains almost stable below of clusters (C) and their temporal variation by examining 50 for the rest of time (Fig. 2). Hence, these key inter- large groups of aftershocks statistically for achieving high- vals appear to be enough to illustrate long-range dynamic quality results with various key intervals (see Fig. 2 for fluctuation, propagation of clustering events, and partic- intervals) to show the event clusters that shift to local ularly the event migration and/or shifting to a series of clusters. local clusters for long time period. The first three days with 683 events for short-range ac- In sequential analysis of aftershock data, the behavior of tivity are considered to show high resolution pattern of a sequence of aftershocks is well classified by detecting the first day maximum (Fig. 3) and initial sudden drop the event variations of the interevent times (∆t), given

427 Time-dependent analysis of aftershock events and structural impacts on intraplate crustal seismicity of the Van earthquake

Figure 4. Variation of the magnitude, depth and frequency of occurrence (foo) of 683 events, as a function of time (3 days) during the period from 2011.10.23 to 2011.10.26 within the focal zone (see the foo in Fig. 1 for time interval of this figure). Clusters are shown by C. Topographic map shows epicenter distribution of 683 events, trending northeast-southwest and clustering in the lake basin. Note that uneven peaks (Clusters, C) at the foo of events form very distinct series.

Figure 5. Variation of the magnitude, depth and frequency of occurrence (foo) of 2553 events, as a function of time (21 days) during the period from 2011.10.23 to 2011.11.13 within the focal zone (see the foo in Fig. 1 for time interval of this figure). Clusters are shown by C. Topographic map shows epicenter distribution of 2553 events, trending northeast and clustering in northeastern end of the lake. Note that the last 49 events occurred within 6 hours resulted in a prominent cluster (C) at the foo. The of these 49 events are numbered from 1 to 49, shown in the map by curved line (see arrow).

428 M. Toker

by average values of the interevent times (Cv). If Cv is (Fig. 3) and 683 events occurred in the first 3 days (Fig. larger than 1, the event distribution is referred to as clus- 4). Frequency of occurrence (foo) is clearly irregular and tered in time, known as the temporal behavior of cluster continually shifting to local clusters (C) during time scale (C) [36]. In this study, it is considered that analysis of (Figs. 3 and 4). Configurational pattern of aftershock the event time intervals is more appropriate for detecting events seen in Figs. 3 and 4 is systematically continued and quantifying temporal clustering. Detailed characteri- and well observed in different time windows. zations of the temporal statistics of clusters are given by Fig. 5 illustrates magnitude, depth and the foo variation specifying the distribution of the time intervals between of 2553 events for 21 days, showing strong and chaotic all events with magnitude bigger than magnitude cutoff pattern of aligned clusters (C). The last 49 events occurred (Mc). The variations in minimum magnitude for the en- within 6 hours resulted in the cluster seen in Fig. 5. tire KOERI catalog roughly range between 1.4 and 2.3, Irregular variation in magnitude, depth and foo is also simply giving Mc ≈ 2.5 for this study [37]. This value distinct for 38 days (Fig. 6). The event rate in the clusters represents the threshold of events over which a temporal considerably lowers below 20 events in the foo. As seen cluster is defined in analysis of aftershocks. Thus, clusters in Fig. 7, total 4147 events for 75 days are recorded of repeating events with peaked statistics are interpreted during time scale, showing irregular peaks at events and to understand fault-related heterogeneity. Previous stud- local clusters (C). This shows that the foo continually shifts ies also found that small asperities generated clusters of to a series of local clusters. It is clear that aftershocks highly repeating small earthquakes with peaked statistics with magnitudes (Mw ≥ 4.0) frequently occurred for the centered on events (M ≈ 1.0) [36, 38]. first 3 days (Figs. 3 and 4) and that they continued to In the interpretation of aftershock data, I used an ap- occur for 21 days (Fig. 5). This produced the second proach of the range of size scales (ROSS). Since, it is the large earthquake (Mw 5.6, the 9 November 2011, Edremit key physical parameter in the effects of heterogeneities earthquake, southern part of the mainshock area) (see the (fault instabilities and asperities) on earthquake dynam- interval between 16th days and 18th days in Fig. 5). ics [36, 39]. The ROSS also plays the significant role of The events with magnitude, larger than Mw 4.0, are also a tuning parameter [40, 41]. It is reported that extrap- distinctly recorded for 38 days (Fig. 6) and 75 days (Fig. olations of statistics based on low-magnitude seismicity 7). to behavior of large events are valid only for disordered It is observed from data that the foo shifts to local clus- systems with a wide ROSS [36, 40, 41]. Model real- ters (Fig. 3). This shifting produces a distinct seismicity izations with heterogeneities are also characterized by a pattern, characterized by propagation and fluctuation of wide ROSS, representing strong geometrical disorder and clustering events for a given time scale (Fig. 4). This disordered immature fault zones [36, 39]. Such model re- seismicity pattern appears to be densely covered by a se- alizations can produce frequency-size statistics following ries of clusters by sudden increase in the event rate (foo) the Gutenberg-Richter relation over the entire temporal and magnitude (Fig. 5). The event rate (foo) suddenly statistical range of events and clustered (or random) tem- decreases on 25th days and remains unchanged (below poral distribution of large aftershock events (see Fig. 2). 20) (Fig. 6). In Fig. 7, irregular pattern of clusters with Moreover, small events can follow in all cases power law peak at about 20 events, from 25th days to 75th days, is frequency-size distribution and are also clustered in time also observed. This pattern is characterized by the events [36, 39] (Figs. 3 and 4). In this study, the effective anal- with magnitudes (Mw ≤ 5.0). ysis and interpretation of aftershocks with a wide ROSS approach well provides a bridge between strongly dis- ordered individual fault zones and broad regions with a 5. Interpretation and discussion diverse population of faults and/or asperities [36, 39]. 5.1. Anomalous occurrence and distribution 4. Time-dependent seismicity anal- of aftershock seismicity ysis of aftershock events Aftershock pattern seen in Fig. 2 indicates the longer sequence of events. This exhibits two extremes of seis- Time period characteristics (magnitude-depth-frequency mic temporal pattern, typically the mainshock-aftershock of occurrence) and seismic properties of aftershock se- pattern [42] with many smaller events and sub-sequences quences of the Van earthquake are given in Figs. 2-7. of similar magnitude, within a small focal area (Fig. 2). Figs. 3 and 4 show magnitude and depth configuration This pattern reveals sequential duration, evolution and of the repeated 321 events occurred in the first 31 hours propagation of the aligned events and includes subse-

429 Time-dependent analysis of aftershock events and structural impacts on intraplate crustal seismicity of the Van earthquake

Figure 6. Variation of the magnitude, depth and frequency of occurrence (foo) of events, as a function of time (38 days) during the period from 2011.10.23 to 2011.11.30 within the focal zone (see the foo in Fig. 1 for time interval of this figure). Topographic map shows epicenter distribution of events, clustering in eastern and northeastern ends of the lake. Note that the last event (5.0) occurred, together with smaller-sized events (2.0≤Mw≤3.0), nearby Van city, shown in the map by curved line.

Figure 7. Variation of the magnitude, depth and frequency of occurrence (foo) of 4147 events, as a function of time (75 days) during the period from 2011.10.23 to 2012.01.06 within the focal zone (see the foo in Fig. 1 for time interval of this figure). Clusters are shown by C. Topographic map shows epicenter distribution of 4147 events, clustering in the further east of the lake. Note that the last event (4.3) occurred, together with the last 10 events (2.0≤Mw≤3.0), nearby Lake Erçek (see LE in Fig. 1 for the location), shown in the map by curved line. A series of these last events on 75th days produced a prominent cluster (C) at the foo.

quences such as swarm-like response (120th-125th days) short-range characteristics, with a rapid development and and typical earthquake series (128th-183th days). Fig. 2 decay, tendency for enhanced, repeated seismicity after shows that the aftershocks densely take place at shallow the large-medium events (Mw≥4.0), and a very high b- depth between 5 and 10 km. The minimum magnitudes value (>2.0). are mostly 1.9-2.0. Aftershock seismicity in Fig. 2 also displays an aftershock decay pattern and dense clusters, Anomalous occurrence and sequential pattern of after- with most of the activity within the first few days (see and shock events are also prominently characterized by in- examine Figs. 5-7 for time scale). However, detailed pat- dividual and peculiar clusters of events. The 49 events tern of aftershock seismicity seen in Figs. 3-7 has varying seen between 20th and 21th days (Fig. 5) occurred in 6 hours only and caused a distinct cluster with a prominent

430 M. Toker

group of events (M ≈ 4.0). The occurrence order of these heterogeneity of the dip-slip faults (thrusting) and mul- 49 events is curve-linearly aligned and concentrated in tifractal pattern of the accretionary crust, or slip insta- the epicentral area. This implies strong structural het- bility of the failed elements. These are probably driven erogeneity of the ruptured zone. It is to note that this by complex evolution of intraplate crustal faults, mechan- cluster is distinctly observed within time interval of irreg- ical heterogeneity and seismic deformation anisotropy, as ular and discontinuous drop of events below about 100 discussed below. (Fig.5), indicating chaotic picture of events. Similarly, distinct clusters with two significant events (5.0 and 4.3) are also seen in Figs.6 and 7. These clusters are inter- 6. Structural impacts on intraplate preted as indicating fault-related heterogeneities and/or crustal seismicity small asperities that generate clusters of highly repeating small events with peaked statistics centered on events (M All the events are clustered and centered in the epicentral ≈ 3.0). area of the mainshock where the causative thrust faulting The aftershock data used for the all days of the events prominently resulted in an uplifting structure (ÇSZ-F-LE make it easy to get a reliable indication of changes or line, together with NE-delta in Fig. 2). This was accom- variations in aftershock seismicity. Hence it seems from panied by rapid subsidence in central Tatvan basin. the plot of magnitude with time (Fig. 2) that it has a Seismicity pattern of aftershocks appears to be invariably decreasing average magnitude with time. It appears that complex in its overall characteristics of aligned events and there is also a tendency for the foo of per event to decrease clusters (Figs.2-7). This pattern is interpreted as the ge- with time. The resulting distribution has a shape similar to ometrical irregularity of many faults and extreme hetero- that found for the clustering of natural seismicity [1]. Thus, geneity in the various physical parameters in the rupture a progressively organized pattern of events develops even process [1, 12]. This may be resulted from both in irregu- when the initial heterogeneity is set to be random [12]. larity in the propagation of the rupture and in the distri- Irregular variation of aftershock events in magnitude and bution of stress release within. It is obvious from the data the foo in Fig. 2 suggests time-dependent long-range in- that the heterogeneity may be considered to generate dis- teractions of events in the focal zone. The shifting clusters tinct aftershock events and various subevents recognized result in the prominent and variable sequential pattern in this study. It is reported by previous studies that envi- of events, suggesting dynamic fluctuation, propagation of ronmental effects on time-dependent strength are impor- clustering events and strong local interactions. These in- tant for the generation of aftershocks [1]. These effects teractions are probably resulted from dynamic event in- can result in variable friction as a function of time and stabilities and faulting complexity. Aftershock events are also allow subcritical crack growth to occur, causing the interacted by triggering of seismicity with local nearest event instability. Thus, it is considered that these effects neighbor interactions, implying that event instability dy- significantly control an aftershock picture of post-seismic namically takes place in aftershock pattern. The event behavior and seismic triggering in the focal area. instability indicates that aftershocks appear to have a con- Overall seismicity pattern of aftershocks and their anoma- tinuing trigger mechanism, either by a strain transfer, or lous distribution may provide a mechanism for strain soft- by a slow irregular slip event, in the mainshock area. As ening process [1] to explain the principal thrusting event well seen by Figs. 3 and 4, the events drop and/or in- in the Van earthquake. This suggests that the causative crease is continually redistributed to its nearest neighbors thrust faulting may have become much more of a zone of during a time scale, suggesting complex and multifractal upper crustal weakness progressively during its deforma- seismic nature of the Van earthquake (see Fig. 2). tional history. Strain is further localized by thrusting in Briefly, this study of aftershock seismicity shows time- the focal zone for a given time scale. As a result, extreme dependent distribution of events, event instabilities and strain localization and fault weakening control the seismic their local interactions. The population and distribution characterization of earthquake occurred and contribute to of the aftershock events clearly exhibit spatial variability, explain the anomalous occurrence of aftershocks and in- clustering-declustering and intermittency, consistent with traplate nature of the Van earthquake where the rupture multifractal scaling [1, 12]. It is found that the distribution process has high strength, moving slowly. This process of clusters of events observed is power-law in both space shows strong deformation anisotropy in the accretionary and time. The sequential growth of aftershock events dur- complex sites, forming a typical fractal set of seismicity ing time scale is considered as multifractal behavior of [12]. seismicity in and around the focal zone. It is suggested From the combined interpretation of structural cases men- that multifractal seismicity can be dominated by strength tioned above and aftershock seismicity given here, it is

431 Time-dependent analysis of aftershock events and structural impacts on intraplate crustal seismicity of the Van earthquake

to say that the slip event in the causative thrusting of ity contributes to a better understanding of the possible Van earthquake seems to have been much more irreg- physical processes and records of many small events. The ular [1]. This reveals the much greater irregularity of October 23, 2011 Van earthquake combined with an ex- thrust fault topography in the slip-normal than the slip- tensive examination of aftershocks provided new structural parallel direction [1] in the focal zone. It is also to note impacts on intraplate crustal seismicity and different in- that the causative thrusting event may have been seg- sights into seismic coupling between the principal thrust- mentary with many smaller, secondary faults. Such a ing event and its aftershocks. Finally, I suggest that after- structural case may result in a spectrum in the dynamic shocks resulted from an appear to rupture segmentation characteristics in the upper crustal be unique data sets. These may serve to retrieve detailed block [1]. This assumes that the spatial heterogeneity and information on the upper crustal structure of the focal zone dynamic complexity of the causative thrusting and many and related strain patterns. secondary faults can easily introduce many degrees of freedom. Therefore, considerable irregularity of aligned aftershock events, propagation of event instability and a 8. Acknowledgements large number of events for a longer time scale should be expected in multifractal aftershock seismicity of intraplate I would like to thank the anonymous reviewers for Van earthquake. This is best characterized by a power- their careful reviews and comments that improved the law size distribution of clusters observed and also by the manuscript. I sincerely thank Kandilli Observatory and causative thrust faulting and its possible length distribu- Earthquake Research Institute (KOERI) of Boğaziçi Uni- tion that obeys a power law and constitutes a fractal set. versity (Turkey) for providing the earthquake catalogue and G. Berkan Ecevitoğlu for useful discussions and com- ments about the aftershock data set. I am grateful to Alper 7. Conclusions Çabuk of Anadolu University (Turkey), Research Institute of Satellite and Space Sciences for his help in provid- This study shows that there is a great amount of variation ing the seismological laboratory and related instruments in the aftershock seismicity pattern of the Van earthquake. to work. This research was logistically supported by Re- There are a large number of small events in a small focal search Institute of Satellite and Space Sciences, Anadolu area, followed by the renewed clusters of aligned events. University (Eskižehir, Turkey). It is possible to say that the anomalous pattern of after- shock seismicity may be a precursor of volcanic or tectonic events (or both) as a result of intraplate crustal seismicity. References Seismicity parameters compiled of the data sampling pro- cedure exerts a strong influence on the inferred fractal [1] Scholz C.H., The mechanics of earthquakes and fault- seismicity of Van earthquake. Hence, aftershock events ing. Cambridge University Press, 1990 used in this study are better interpreted as extensive het- [2] Şengör A. M. C., Özeren S., Genç T., Zor E., East erogeneity in seismic deformation. This has structural im- Anatolian high plateau as a mantle-supported, north- pacts on intraplate seismicity, rather than the linear and south shortened domal structure. Geophys Res Lett., uniform fracture mechanics. This suggests multifractal and 2003, 30 (24):8045 doi: 10.1029/2003GL017858 unstable nature of the Van earthquake. This is supported [3] Şengör A.M.C., Özeren M.S., Keskin M., Sakınç M., by inhomogeneous strain patterns and highly anisotropic Özbakır A.D., ve Kayan I., Eastern Turkish high nature of accretionary crust where tectonic instability of plateau as a small Turkic-type orogen: Implica- crustal blocks and thermal inequilibrium are frequently tions for post-collisional crust-forming processes in observed. The reverse dip-slip character of Van earth- Turkic-type orogens, In Earth Science Rev., 2008 doi: quake, local strain softening mechanism and related strain 10.1016/j.earscirev.2008.05.002 localization and fault weakening process well explain the [4] Utkucu M., 23 October 2011 Van, Eastern Anatolia, intraplate nature of the Van earthquake. The crustal het- earthquake (Mw 7.1) and seismotectonics of Lake Van erogeneity and deformation anisotropy, associated with Area. Journal of Seismology, 2013, (in press). the surrounding volcanoes contribute significant pertur- [5] Utkucu M., Implications for the water-level-change bation to intraplate strain regime. These findings have triggered moderate (M≥4.0) earthquakes in Lake Van played a key role in the formation of the continued af- basin, Eastern Turkey. Journal of Seismology, 2006, tershock seismicity since the occurrence of the 2011 Van 10:105-117 DOI: 10.1007/s10950-005-9002-y earthquake (Mw 7.1). This study of aftershock seismic- [6] Horasan G., and Boztepe-Güney A., Observation and

432 M. Toker

analysis of low-frequency crustal earthquakes in Lake Persian and Ottoman domination. In Armenian Van and its vicinity, eastern Turkey. Journal of Seis- Van/Vaspurakan, edited by Richard G. Hovannisan, mology., 2006, doi: 10.1007/s10950-006-9022-2 UCLA Armenian History and Culture Series, Historic [7] Toker M., Krastel S., Demirel-Schlueter F., Demirbağ Armenian Cities and Provinces, 1, Mazda Publishers, E., İmren C., Volcano-Seismicity of Lake Van (Eastern Inc., California, 2000, 117-131 Turkey): A comparative analysis of seismic reflection [18] Mangaltepe İ., Seyyahların gözüyle Van (XIX yüzyıl). and three component velocity seismogram data and Kitabevi Yayğnevi, İstanbul, 2010 (in Turkish) new insights into volcanic lake seismicity. Proceed- [19] Daniel J.E, Khazai B., Wenzel F., Vervaeck A., ings of International Earthquake Symposium, 22-26 The CATDAT damaging earthquakes database. Nat. October, in Kocaeli, Turkey, 2007, 103-109 Hazards Earth Syst., 2011, 11, 2235-2251, DOI: [8] Toker M., Krastel S., Demirel-Schlueter F., Demirbağ 10.5194/nhess-11-2011 E., ve Şengör A.M.C., Structural Impacts on Crustal [20] Ayhan E., Alsan E., Sancaklı N., Üçer S.B., Türkiye Seismicity of Lake Van Basin (Eastern Turkey): Seis- dolayları deprem kataloğu 1881-1981. Boğaziçi mological Constraints for Crustal Deformation. Pro- Üniversitesi Kandilli Rasathanesi, İstanbul, Turkey, ceedings of 6th Congress and Technical Exhibition of 126, 1981 (in Turkish) Balkan Geophysical Society (C11, BGS 2011), 2011, [21] Lahn E., A note about earthquakes in Van area (July- Budapest, Hungary November 1945). Maden Tetkik ve Arama ve En- [9] Fedotov S.A., Regularities in the distribution of strong stitüsü 1(35), Ankara, Turkey, 1946 (in Turkish and earthquakes in Kamchatka, the Kuriles, and north- French) eastern Japan. Akad. Nauk USSR Inst. Fiz. Zeml., [22] Taşman C.E., Varto and Van earthquakes. Publ. Miner Trudy, 1965, 36, 66-95 Res. Explor. Ins. Turk., 1946, 36, 287-291 (in Turkish) [10] Mogi K., Dilatancy of rocks under general triaxial [23] Öcal N., Türkiye’nin sismisitesi ve zelzele coğrafyası, stress states with special reference to earthquake 1850-1960 yılları için deprem kataloğu. Kandilli precursors. J. Phys. Earth., 1977, 25, 203-217 Rasathanesi Yayınları, İstanbul, 1968 (in Turkish) [11] Mogi K., . Academic Press, [24] Ambraseys N.N., Engineering seismology. Earthq. Tokyo, 1985 Eng. Struc. 1988, 17, 1-105 [12] Main I., Statistical Physics, Seismogenesis and Seis- [25] Eyidoğan H., Güçlü U., Utku Z., Değirmenci E., mic Hazard. Reviews of , 1996, 34-4, 433- Macroseismic guide for large Turkey earthquakes 462. DOI: 8755-1209/96/96RG-02808 1900-1988. Istanbul Technical University, İstanbul, [13] Toker M., Ecevitoğlu G.B., Van Gölü Havzasının Tek- 198, 1991 (in Turkish) tonik ve Sismolojik Özellikleri (6000 artçı deprem [26] Pınar N., Lahn E., Türkiye depremleri izahlı kataloğu. ve 180 günlük sismolojik ve sismotektonik modellen- Bayındırlık Bak. Yapı ve İmar İç Reis, 6-36, 153, 1952 mesi). Anadolu University Press, Satellite and Space (in Turkish) Sciences Research Institute, Eskişehir, Turkey, 2012a [27] Alsan E., Tezuçan L., Bath M., An earthquake cata- (in Turkish) logue for Turkey for the interval 1913-1970. Kandilli [14] Toker M., Ecevitoğlu G.B., Shallow seismicity of the Obs. İstanbul, Turkey, and Seism. Inst. Upsala, Swe- Van earthquake (ML, 7.2, 23 October 2011, East- den, Report No. 7-75, 166, 1975 (in Turkish) ern Anatolia): monitoring and analysis of the seismic [28] Suetsugu D., Source Mechanism Practise. Japan In- data. Proceedings of National Conference with In- ternational Cooperation Agency, 1995 ternational Participation "Geosciences 2012" by Bul- [29] Toksöz M.N., Nabelek J., Arpat E., Source properties garian Geological Society (BGS), Sofia, Bulgaria, of the 1976 earthquake in eastern Turkey: a compar- 2012b, 125-126 ison of field data and teleseismic results. Tectono- [15] Kalafat D., Güneş Y., Kara M., Deniz P., Keko- physics, 1978, 49, 199-205 valı K., Kuleli S.H., Gülen L., Yılmazer M., Özel [30] Kalafat D., 1964-1994 yılları arasında Türkiye ve N., A revised and extended earthquake catalogue yakın çevresinde etkili olmuş depremlerin makrosis- for Turkey since 1900(M≥4.0). Boğaziçi University, mik gözlemleri. Deprem Araştırma Bülteni, 1995, 73, Kandilli Rasathanesi ve Deprem Araştırma Enstitüsü, 60-97 (in Turkish) Bebek, İstanbul, 2007 (in Turkish) [31] Pınar A., Honkura Y., Kuge K., Matsushima M., Sez- [16] Ambraseys N., Earthquakes in the Mediterranean gin N., Yılmazer M., Öğütçü Z., Source mechanism and Middle East: a multidisciplinary study of seis- of the 2000 November 15 Lake Van earthquake (Mw micity up to 1900. Cambridge University Press, 2009 = 5.6) in eastern Turkey and its seismotectonic im- [17] Kouymijan D., Van under Mongol, Turkmen, plications. Geophys. J. Int., 2007, 170, 749-763, doi:

433 Time-dependent analysis of aftershock events and structural impacts on intraplate crustal seismicity of the Van earthquake

10.1111/j.1365-246X.2007.03445.x Res. Let. (SRL), 2013 (submitted to publication) [32] KOERI Boğaziçi University, Kandilli Observatory [38] Nadeau R., and McEvilly T.V., Fault slip rates at and Earthquake Research Institute, 2011-2012, depth from recurrence intervals of repeating mi- http://www.koeri.boun.edu.tr/ croearthquakes. Science, 1999, 285, 718-721, doi: [33] Toker M., Tectonic and Magmatic Structure of Lake 10.1126/science.285.5428.718 Van Basin and its Structural Evolution, Eastern Ana- [39] Ben-Zion Y., Stress, slip and earthquakes in models tolia Accretionary Complex (EAAC), E-Turkey. Phd. of complex single-fault systems incorporating brittle Thesis, Istanbul Technical University, Istanbul, 2011 and creep deformations. J. Geophys. Res., 1996, 101, (in English) 5677-5706, doi: 10.1029/95JB03534 [34] Toker M., Şengör A.M.C., Van Gölü havzasının temel [40] Hillers G., Ben-Zion Y., and Mai P.M., Seismicity on yapısal unsurları, tektonik ve sedimanter evrimi, a fault controlled by rate- and state dependent fric- Doğu Türkiye, İstanbul Teknik Üniversitesi Dergisi/D, tion with spatial variations of the critical slip dis- Mühendislik, 2011, 10, 119-130 (in Turkish) tance. J. Geophys. Res., 2006, 111, B01403, doi: [35] Cukur D., Krastel S., Demirel-Schlüter F., Demirbağ 10.1029/2005JB003859 E., Imren C., Nissen F., Toker M., and PaleoVan- [41] Hillers G., Mai P.M., Ben-Zion Y., and Ampuero J.- Working Group., Sedimentary evolution of Lake Van P., Statistical properties of seismicity along fault (Eastern Turkey) reconstructed from high-resolution zones at different evolutionary stages. Geophys. seismic investigations. Int. J. Earth Sci. (Geol Rund- J. Int., 2007, 169, 515-533, doi: 10.1111/j.1365- sch)., 2012, 6-10, DOI: 10.1007/s00531-012-0816-x 246X.2006.03275.x [36] Ben-Zion Y., Collective behavior of earthquakes [42] Mogi K., Some discussions on aftershocks, and faults: continuum-discrete transitions, progres- and earthquake swarms-the fracture of a semi-infinite sive evolutionary changes, and different dynamic body caused by an inner stress origin and its relation regimes. Rev. Geophys., 2008, 46, RG4006, doi: to the earthquake phenomena. 3. Bull. Earthq. Res. 10.1029/2008RG000260 Inst. Tokyo Univ., 1963, 41, 615-658 [37] Toker M., Time and depth-dependent variation in [43] JMO., Van (Tabanlı-Edremit) depremleri raporu. Je- seismic b-value of the Gutenberg-Richter relation and oloji Mühendisleri Odası Raporu, Ankara, 2011, 52 implications for stress state of the October 23th, 2011 (in Turkish). Van earthquake (Mw 7.2), Eastern Anatolia. Seism.

434