- © 2012 by the Seismological Society of America
The Mw 7.2 earthquake of 23 October 2011 struck an area of Eastern Anatolia with a long historical record and a long earthquake history. The earthquake occurred in a region of rather complex tectonics resulting from the collision of the Arabian and Eurasian continental plates (Tchalenko, 1977; Barka and Reilinger, 1997; McClusky et al., 2000; Sandvol et al., 2003; Armijo et al., 2005). Several studies have recognized the recent and present activity of both the main regional structures and their complex system of conjugate faults (Tchalenko, 1977; Ambraseys, 1989; Berberian, 1997; Bozkurt, 2001; Koçyigit et al., 2001). The area to the east of the Karlıova triple junction and to the north of the Bitlis‐Zagros suture zone is characterized by a north‐south‐compressional tectonic regime, which resulted in a general uplift of the region with the formation of east‐west‐directed thrust and fold belts. A conjugate system of right lateral and left lateral strike‐slip faults paralleling, respectively, the North and East Anatolian faults are also dominant structural elements of the region. The complex compressional structure is also characterized by east‐west‐trending fissures, which resulted in the Plio‐Quaternary activity of the Nemrut, Süphan, and Ağrı volcanoes, as well as in east‐west‐trending basins of compressional origin such as the Lake Van and Muş basins (Şaroğlu and Yılmaz, 1986; Bozkurt, 2001).
The 23 October 2011 earthquake caused heavy damage to Van and several towns and villages around Lake Van, in the districts of Van and Ercis (Fig. 1). It was followed by several aftershocks and another strong event (Mw 5.7) that occurred on 9 November 2011, causing further damage and casualties.
The seismicity of the area is described by some regional parametric earthquake catalogs (Soysal et al., 1981; Shebalin and Tatevossian, 1997; KOERI‐NEMC for the instrumental period) and has been the subject of several studies, such as Tchalenko (1977), Ambraseys (1989, 2001), and Ambraseys and Jackson (1998).
For Eastern Turkey the catalogs by Soysal et al. (1981) and Shebalin and Tatevossian (1997) sometimes provide contradictory information: the number of earthquakes included in the former is smaller than in the latter; Soysal et al. (1981) only covers the time‐window up to 1900 and does not provide magnitudes. For these reasons we select Shebalin and Tatevossian (1997) as a reference (Fig. 2). In the Lake Van area, Shebalin and Tatevossian (1997) report the following earthquakes with a magnitude similar to that of the 2011 event in the time window after 1000: 3 October 1275, located on the western bank of Lake Van (MS 6.8); 14 April 1696, located in the Çaldiran area (MS 6.8); 10 February 1884, located in the Maku area (MS 7.0); 28 April 1903, located in the Malazgirt area (MS 7.0); 24 November 1976, located in the Çaldiran area (MS 7.1).
No earthquake with a magnitude equivalent to the 2011 one is reported by the catalog in the area of the 2011 earthquake; so, apparently, there are no predecessors of this earthquake. However, two questions arise: a) could some events located close to Van have been underestimated or mislocated, as frequently happens with poorly known events? and b) are there gaps in the earthquake history of the Van region?
To answer these questions we analyze the information supplied by studies not accounted for by the Shebalin and Tatevossian (1997) catalog. Tchalenko (1977) was the first to analyze in detail some historical earthquakes in the Van region, although he did not assign intensities. Guidoboni and Comastri (2005) included the Van region in their investigation of earthquakes in the Mediterranean area prior to 1500. Ambraseys and Finkel (1995), and Ambraseys (2009) investigated the available historical sources and proposed new data and interpretation. In particular, the last author, in addition to improving the background information for the many earthquakes in the selected catalog and in other catalogues, supplies information on a few, previously unknown, earthquakes.
SHORT NOTES ON HISTORY AND HISTORICAL SOURCES
The investigated area belongs to the easternmost region of eastern Anatolia, centered on the basin of Lake Van. Its territory has been continuously inhabited since prehistoric times, and Assyrian texts refer to the importance of the Kingdom of Van or Urartu (active between the ninth and sixth century B.C.). Besides the town of Van, we focused on the historical region of Vaspurakan, which was already mentioned in an Armenian “Atlas” (Asxarhac‘oyc‘), probably of the seventh century A.D. (attributed to Ananias of Sirak, seventh century). Vaspurakan included the coastal lands of Lake Van, its territory largely mountainous, extending mostly in a southeast direction along the present‐day border with Iran and Iraq. The region, focused around the town of Van, formed such a cultural and administrative unit that, between 908 and 1021, it became an independent kingdom. After that period of independence, the region fell under the rule of Byzantium, followed by the dominion of the Seljuk Turks, the passages of the Mongols, and the rule of the Persians. Maintaining the same name through the centuries, in 1639 the Vaspurakan region became a stable possession of the Ottoman Empire; in the aftermath of the 1826–1828 War, control of the area passed into the hands of the Russians (Hovannisian, 2000). In all, Vaspurakan was a border region touched by many and different local cultural traditions, with many written languages (Turkish, Armenian, Arabic, and Persian, among others) and types of historical sources.
The region of Vaspurakan had only a few important settlements, such as Van, Ercis, Ahlat, Adilcevaz, Bitlis, most of which were located on or not far from the lake banks. We focused on Van, Ahlat, and Adilcevaz for their history, and because they are the places from which damaging historical earthquakes were reported.
The town of Van had already a long history of local dynasties ruling the area, when the Kingdom of Vaspurakan (908–1021) was established. Van’s importance increased with the creation of the Monastery of Varagavank (Yedi Kilise) which became the richest and most celebrated monastery in the area. Under the Ottomans (from the seventeenth century onwards), Van was first a fortified town close to the border between Ottoman and Iranian territories and then between Ottoman and Russian territories.
Ahlat was an important Armenian town, once ruled by a Seljuk dynasty (1100–1207) and later incorporated in the Ottoman dominion in 1533–1534. The decline of the town began in the sixteenth century, as testified by travelers’ reports (from Lynch, 1901).
There are features in common between Ahlat and Adilcevaz, such as the Kala (Ottoman fortress). Adilcevaz also lost its importance during the sixteenth century, passing its role to towns to the east, such as present‐day Ercis.
Several monasteries, built in the Vaspurakan region as early as the seventh century A.D., and active through the early twentieth century, formed a dense network of potential recording centers of information about historical events. The distribution of the monasteries up to the late seventeenth century can be appreciated in the map (see a detail in Fig. 3) by Eremia Celepi K’eomiwrcean, an important personality of the Armenian community in Istanbul (K’eomiwrcean, 1691). The map, which illustrates the Church of Armenia with its towns, bishops, monasteries and hermitages, indicates their relative importance by means of different insignia. This map is very useful for identifying places whose place‐names changed many times through the centuries.
In summary, most of the earthquake records for the Lake Van region (thirteenth to eighteenth century) come from local historical sources in the Armenian language; these records are contained in manuscripts, and their colophons, which kept alive the local cultural tradition in the several important monasteries of the region. They have been published in the collection “Armenian Short Chronicles,” edited by Hakobyan (1951, 1956).
In the seventeenth century, damaging earthquakes attracted the occasional interest of contemporary chroniclers, such as the Armenian literate Arak’el of Tabriz, the major source of information on the 1646 event. The Ottoman archival documentation (Başbakanlık Arşivi, Prime Ministers’ Archive; Topkapi Saray Archives, Istanbul) investigated so far (Ambraseys and Finkel, 1995; Ambraseys, 2009), and especially the documentation dealing with repairs related to damage caused by earthquakes, is scarcely informative for the remote and badly‐connected areas of Eastern Anatolia considered here.
Sources in Arabic are few for the area, but the history of Mosul by the local chronicler al‐‘Umari (died in 1811) does contribute information on the effects in Iraq of earthquakes that also affected the region of Van. Finally, the periodical press locally produced in this area goes back only to the mid‐nineteenth century, and neither the Turkish nor the western European languages press cover the region of Eastern Anatolia. Starting from the same period, occasional reports from travelers, diplomats, and traders are also available.
In this section we discuss the earthquakes located in the subareas of a) Ahlat/Adilcevaz; b) Malazgirt; c) Çaldiran; d) Van (see Fig. 2).
Earthquakes in the Ahlat/Adilcevaz area (northwestern banks of Lake Van)
We have information on earthquakes felt in this area since the beginning of the thirteenth century.
1208, 1275. According to Ambraseys (2009, p. 337), the earthquake of 14 February 1208 (Ahlat, MS 6.5) hit “the region of Ahlat, near the west shore of Lake Van [...] caused considerable concern, and perhaps great loss of life and triggered a landslide.” This information is taken from Abu Shama (1203–1267), who was living in Damascus. Two other contemporary sources mention the earthquake: i) Ibn al‐Athir (1160–1233) gives the complete date and says he was at Mosul (Iraq) when he felt the event (“not very strong”), and that he learned that the earthquake was felt in other countries as well, but that “the shock was not violent either”; ii) Sibt al‐Jauzi (died 1256) “does not mention the earthquake”, but refers to “the dead bodies he came across on his route”.
The 3 October 1275 earthquake (Ahlat, MS 6.8), according to both Guidoboni and Comastri (2005) and Ambraseys (2009), caused destruction on the northern bank of Lake Van (Ercis, Ahlat). Guidoboni and Comastri add heavy damage (I=8 Mercalli‐Cancani‐Sieberg scale) as far away as Mardin, Diyarbakir, and Silvan (Fig. 4), which would have required a larger magnitude than the one proposed by Shebalin and Tatevossian (1997). Ambraseys (2009) reports that the earthquake was only felt in Diyarbakir and in Anì. This earthquake is the origin of a false event assigned to Romania (Albini, 2011).
1881. Some earthquakes in this area can be associated with the volcanic activity of Nemrut Dag; this is the case with the 30 May 1881 event (MS 6.3), located by Shebalin and Tatevossian (1997), in the southwestern part of the lake. According to Tchalenko (1977) and Ambraseys (2009), this earthquake caused heavy destruction in the villages of Tegut (173 houses were almost completely destroyed), Ahlat (200 houses collapsed out of 600), and Sipradzor, a place not yet identified (13 houses, no loss of life). Ambraseys (2009) adds a later report (Oswald, 1906) that attributes the destruction of the village of Tegut to a mudflow from Nemrut. According to Tchalenko (1977) this earthquake “seems to have been very local and shallow [...] Its importance lies in its connection with the Nimrud Volcano.”
Earthquakes in the Malazgirt area (north of Lake Van)
1891. The 1891 earthquake is located by Shebalin and Tatevossian (1997) in Malazgirt, MS 6.0. Both Tchalenko (1977) and Ambraseys (2009) suggest the date 6 February, instead of 3 April as in Shebalin and Tatevossian (1997). Tchalenko (1977) and Ambraseys (2009) say it caused heavy damage at Adilcevaz (122 houses were “in ruin or dangerously damaged”, 3 people killed) and strong damage at Van (minarets and domes fell, damage to bazar); it damaged Bitlis and Malazgirt also, and produced effects on the natural environment to the north of Adilcevaz. Tchalenko argues that “the information available is insufficient to locate the epicentral region with confidence”. He emphasizes that the most severe destruction was at Adilcevaz, which suggests to us that the location of this event might have been to the south of Malazgirt.
1903. According to Tchalenko’s detailed description, the 28 April 1903 earthquake (north of Malazgirt, MS 7.0) caused heavy damage at Malazgirt (729 people died, all buildings and most of the walls collapsed) close to the presumed epicentral region of the 1891 earthquake, as well as in several villages located south and northeast of it. It was strongly felt at Ahlat.
Earthquakes in the Çaldiran area (northeast of Lake Van)
1696. The earthquake of 14 April 1696 (Çaldiran area, MS 6.8) is described both by Ambraseys and Melville (1982) and Ambraseys (2009). Information concerns many villages in the district of Çaldiran and the upper part of the district of Van (Fig. 5). The Saint Thaddeus monastery (Kara Kilise) was severely damaged as well. The information is too poor to assess whether this earthquake is a twin of the one in 1976 or whether it broke an adjacent segment of the same fault.
1976. The only earthquake in the twentieth century that affected this area is the one of 24 November 1976 (Muradiye‐Çaldiran, MS 7.1). The report by Gülkan et al. (1978) proposes a very detailed damage description and isoseismal lines. As a preliminary attempt, we have plotted the localities mentioned in the report, assigning to them the intensity values of the corresponding isoseismal line (Fig. 6). Only a few data from the far field area are reported; for instance, there is no information about effects at Van and Ahlat. The causative fault of the 1976 earthquake has been identified with the prosecution in Eastern Turkey of the Tabriz fault (e.g., Karakhanian et al., 2004).
Earthquakes in the Van area (southeastern banks of Lake Van)
1111. The first event is reported in the year 1111 (Van, MS 6.5). It is described by the Armenian chronicler Matthew of Edessa (second half of the eleventh century–1144) who gives many particulars, including some that suggest an event similar to an earthquake. What follows is the relevant part of the critical edition, and translation into French, of Matthew’s Chronicle (twelfth century), by M. E. Dulaurier:
“[pp. 274–275] CCV. [...] Cette même année [a.Arm. 559 (22 février 1110–21 février 1111), p. 270], un phénomène terrible eut lieu en Arménie, dans la province de Vasbouragan. Un jour, pendant l’hiver, au milieu des ténèbres de la nuit, un feu éclata du plus haut de la voßte céleste, qui s’entr’ouvrit en lançant des tourbillons de flamme. Ce feu frappa la mer de Vasbouragan (1), dont les flots retentirent de violents mugissements; il atteignit aussi le littoral, et la terre et l’onde, agitées avec violence, tremblèrent. La mer prit une couleur de sang, et la flamme enveloppa toute la surface de l’abîme. A l’aurore on aperçut des masses de poissons morts, accumulées sur le rivage comme des piles de bois. Elles répandirent au loin l’infection. La terre, dans le voisinage, s’entr’ouvrit en crevasses d’une profondeur effrayante.”
“[p. 611] 557 (23 février 1108–20 février 1109)—A la même date, un prodige horrible se manifesta en Arménie, dans la province de Vasbouragan. Un feu éclata du haut des cieux, et tomba dans la mer de Vasbouragan, accompagné d’un terrible fracas du tonnerre. Tous les poissons de cette mer périrent, et la campagne fut empestée de l’odeur fétide qu’ils répandaient: le sol s’entr’ouvrit en divers endroits.”
Apart from supplying a slightly different date, both chroniclers describe the same phenomenon; only Matthew mentions an earthquake among the consequences (...et la terre et l’onde, agitées avec violence, tremblèrent). Ambraseys (2009) concludes that Matthew “clearly sees the earthquake as caused by the meteorite” landing in Lake Van. We agree that this has to be considered a fake earthquake.
1646 (formerly 1648). According to Ambraseys and Finkel (1995), later confirmed by Ambraseys (2009), the only date for the earthquake of 31 March 1648 (south of Van, MS 6.5) in agreement with the many contemporary available records is 28 March Old Style/7 April New Style 1646. Beside many Armenian marginal notes and colophons reporting earthquake effects (Hakobyan 1951, 1956), the account of the contemporary chronicler Arak’el of Tabriz (seventeenth century) supplies a general picture, as well as many details (comments on his possible sources in Sinclair, 1999) on the earthquake effects: “Innumerable houses, merchandise, and edifices were destroyed, and under them were trapped man and animals. The dead were so many that they were taken out of the city” (Hovannisian, 2000, p. 126).
Damage is reported from as far as Adilcevaz, to the northwest, and Baskale, to the southeast; undoubtedly, the most seriously damaged area was the region of Van, including many monasteries and villages with their churches.
Damage in the town of Van was to the walls of the lower citadel, which collapsed in some places and breached in others. Within the walled city, many houses were destroyed. The Armenian church dedicated to St. Sahac was so damaged that it was subsequently demolished and rebuilt. Other Armenian churches suffered structural damage, such as collapse of the roof or the dome. A number of mosques were also destroyed. Many villages around Van suffered a similar level of destruction, both to the buildings and the churches.
Several environmental phenomena are also described by the historical testimonies, from water springs drying up (village of Avan, present‐day Iskele) to landslides, such as the one that destroyed the church in the village of Norghu (present‐day Yolasan), and caused the village to be abandoned. Figure 7 shows the damage distribution, which looks similar to that of the 9 November 2011 event.
1715. This earthquake (MS 6.4) is known to Tchalenko (1977), Ambraseys and Melville (1982), Ambraseys (1989, 2009), and Ambraseys and Finkel (1995) from a record contained in the collection of “Armenian Short Chronicles” (Hakobyan, 1951) as occurring on 8 March (a.Arm 1164) 1715 and having affected the region to the southeast of Van:
“An earthquake happened in Van; a house collapsed and four people died. The dome of Bardoughimeos Araqeal [Church of St. Bartolomeus, in present‐day Albayrak] flew away, the wall of the castle of Hoshab [Hosap/Güzelsu] was shaken; the pavilion of the castle of (Kara) Saray [Karahisar] collapsed; the pavilion of Satmanis [Orenburc] collapsed; the country of Mahmatanay was damaged” (Ambraseys and Finkel, 1995, p. 101).
The damage distribution is shown in Figure 8.
1871. The earthquake of 17 March 1871 (MS 6.8) is located south of Lake Van. Actually, it is the mainshock of a sequence that, according to Ambraseys (2009), was felt in Erzurum, Dogubayazit, Van, Baskale, and Muş. Mushketov and Orlov (1893) confirm: “In the same year, everyday from 5 to 25 March (which style?), many underground shocks in all Armenia and Kurdistan; the first shocks were observed in Bakh‐Kala and Gekkari from where they spread toward Van, Bayazet and Mush, where disappeared toward Euphrates.” Although the earthquake was felt over a large area, there is no information about any damage. The location and magnitude are difficult to assess.
DISCUSSION AND CONCLUSION
In the introduction we stated that no earthquake with a magnitude equivalent to that of the 23 October 2011 earthquake is reported by Shebalin and Tatevossian (1997) for the same area. In the search for the predecessors of the 2011 earthquake, we have tried to answer two initial questions: a) could some events located close to Van have been underestimated and/or mislocated, as frequently happens with poorly known events?, and b) are there gaps in the earthquake history of the Van region?
As for the first question, we have not found earthquakes which could, in principle, be underestimated and/or mislocated by the catalog and that could be re‐located in the area of the October 2011 earthquake with a comparable magnitude. The only large historical earthquake, the one which occurred in 1275, seems to have originated in the Ahlat area. The available information for the 1696 earthquake in the Çaldiran area suggests that it most likely belongs to the well‐known fault system that generated the 1976 event. As for the earthquakes of the Van area, data on the 1646 event (MS 6.5) allowed us to agree with the epicentral location proposed by Shebalin and Tatevossian (1997). The damage distribution of the 1715 earthquake, which covers a large although not very populated area, suggests that the magnitude value provided by Shebalin and Tatevossian (MS 6.1) could be underestimated. Available data do not allow us to properly constrain the epicenter; however, the earthquake could belong to the same fault system as the 1646 event.
As for possible information gaps, Figure 9 shows the earthquake history of the Lake Van region, the same region marked off in Figure 2, as derived from the catalog by Shebalin and Tatevossian (1997); the same subareas of the previous section ([a] Ahlat/Adilcevaz; [b] Malazgirt; [c] Çaldiran; [d] Van) are evidenced. The earthquake history shows gaps between 1275 and 1646, and later between 1715 and 1834. Taking into account that the 1111 earthquake is a fake, before 1275 there are two earthquakes in the Ahlat area, while there are no earthquakes located elsewhere, namely in the Van area before 1646.
It is worthwhile to note that the compilation by Ambraseys (2009) also contributes about fifteen events between the fifteenth and nineteenth century, events that were previously not in the earthquake catalogs. Among them, the one of 1415 caused heavy damage at Ahlat; a few others are reported as slightly damaging at Van. A controversial event, reported as happening on 29 May 1664, might have produced heavy damage in Gerger and Toprakkale, in the Agri region; it could have been felt as far away as Van and Tabriz. However, this event is close in time to the destructive 1641 Tabriz and 1646 Van earthquakes, and the sources may have amalgamated information on those events. For the 1701 March earthquake sequence, Ambraseys (2009) uses the content of the collection of “Armenian Short Chronicles” (Hakobyan, 1951, 1956). On Friday 7 March a.Arm. 1150 (Old Style), corresponding to 18 March 1701 in the Gregorian calendar, started a series of earthquakes, which are essentially reported from Van. After the first shocks, the inhabitants of Van left the town to camp in the open; the strongest earthquake of 15/26 March 1701 caused serious damage to a few houses in Van, but no casualties. The same event caused some houses to collapse in a village to the south of Lake Van, Berdak/Pertek, present‐day Doganlar. Many aftershocks followed, among which was the one on 30 November 1701. At this stage we are unable to determine the parameters of these events, which need to be further investigated.
To further explore the possible information gaps, we compiled the seismic histories of Ahlat + Adilcevaz (Fig. 10a) and Van (Fig. 10b), plotting the earthquake effects in chronological order, including the effects of the above mentioned, newly found earthquakes. We used preliminary estimates of the effects in terms of heavy damage (HD), damage (D), strongly felt (SF), and felt (F). We considered the data for Ahlat and Adilcevaz together, because the two towns are close and because, for each earthquake, we have information from either one or the other town.
Figures 9 and 10 show similar patterns with respect to the information gaps, with the exception of 1415, at Ahlat, a new damaging earthquake, and of 1780, at Van, for which it is recorded that the damaging 1780 earthquake in Tabriz was felt (Ambraseys, 2009). Therefore, the gaps shown in Figures 9 and 10 are probably due to a lack of sources.
In conclusion, nothing can be said about earthquakes in the time‐window before 1646 in the region east of Lake Van. After that year, no prominent candidate earthquake appears as a possible twin of the 23 October 2011 event. The gap between 1715 and 1850, however, must be noted. The damage distribution available for the 1646 (formerly 1648) earthquake suggests a location similar to the recent 9 November 2011, Mw 5.7 event. Although possibly triggered by the 23 October 2011 earthquake, the 9 November earthquake is considered to be independent due to its strike‐slip focal mechanism, which is different from that of the 23 October 2011 earthquake and its main aftershocks. The 9 November 2011 earthquake might be associated with another fault segment, most probably an unmapped segment in the Lake of Van (MTA, 2011).