North Anatolian Fault Field Trip Report
Report on the field trip along the North Anatolian Fault Zone (August 1-15, 1997) organized by Aykut BARKA.
In Turkey, "Deprem" and "Fay" are the names for earthquake and fault respectively, and these two terms are well known by the Anatolian shepherds and countrymen. Indeed, they can show you the fault traces crossing mountains and agricultural fields without any kind of hesitation, which is not the case in many parts of Mediterranean regions where earthquake faults do not clearly emerge at the surface.
The North Anatolian Fault extends in an east-west direction along ~ 1500 km, showing several segments associated with historical large seismic events, and in particular with the well known westward migrating earthquake shocks of Ms 7.9 1939, Ms 7.1 1942, Ms 7.6 1943, and Ms 7.4 1944. The earthquake sequence gives an unrivalled opportunity to examine a well exposed and very active fault system and therefore, Aykut Barka - presently Professor at the Earth Science Department of the Institute of Technical University (ITU) of Istanbul - had the idea to organize a field trip with an international group of geologists and seismologists (see the list below). Beside the last large earthquake of 1992 (Ms 6.8), Aykut includes the Ms 6.9 1949, Ms 6.9 1951, Ms 7.0 1957, Ms 6.8 1966 and Ms 7.1 1967 seismic events that occurred along the fault. The idea of Aykut was indeed brilliant and due to an impeccable organization, the fifteen days were dedicated to examining outstanding examples of successive coseismic faulting along the magnificent Anatolian landscapes. The objective was to bring together in the field earthquake geologists with different backgrounds and provide a stimulating forum in which to discuss previous theories and observations (coseismic faulting and displacements, fault segmentation, slip rates from field measurements and GPS data, recurrence time of large seismic events from trench data, NAFZ peculiar earthquake sequential occurrence and seismicity migration from the Coulomb-friction model, etc...).
Over successive days, from Erzincan region to Istanbul and Marmara Sea, the minibus and "Coccinella - Volkswagen" convoy crossed the different regions, mountains, cities, villages from early in the morning to late evening. Every morning, an hour of explanations, questions and discussions prepared the group for that day,s field trip. The colorful landscapes and extraordinary faults were illustrated into a rich guidebook also prepared by Serdar Akyuz (ITU, Istanbul) and containing, general itinerary maps, detailed geologic-tectonic maps and photos with explanations, 1/250 000 topographic maps together with a heavy volume of main papers.
On the first day, almost as a "side-dish", a workshop on "Recent developments in active fault studies" was held at the ITU. Diverse kinds of earthquake faults were described from all over the world including Australia, Japan, California, Israel and Middle-East, Europe and Turkey. This was an occasion for the participants to get to know each other before spending the next 14 days together ..., and Aykut revealed the travel-expeditionnary conditions : no withdrawal was allowed !!
Coseismis rupture traces and fault segments associated with each major earthquake of the seismic sequence along the North Anatolian Fault zone are described from east to west in the following sections (many local names are given such as any reader can follow the trip on a general geographic map):
1992 Erzincan earthquake fault. On the second day the wake up call was at 4 a.m. ... heading to Erzincan by air. Because of the size of the 13 March 1992 earthquake (Ms 6.8) in the Erzincan region, and the fact that considerable damage was still visible (650 victimes and about 1 billion $ losses), we spent 3 days along this fault section and the related basin. East of Erzincan basin, two large earthquakes with surface breaks occurred in 1949 (Ms 6.9) and in 1966 (Ms 6.8) along two adjacent segments, each 20 to 30 km long. The junction with the 1939 and 1992 coseismic faulting corresponds to a fault segment showing a gap of large seismic events in the historical catalogue. Trench investigations from the work of Koji Okumura (Geol Survey of Japan) and Steve Wesnousky & Aykut Barka (University of Nevada and ITU Istanbul, respectively) show the probable occurrence of a large earhquake in 1784, even if trench analyses led to different interpretations (Koji thought that visible displacements reflect several smaller sized earthquakes). Along the NAFZ north of Mutu, fault-sections made by river meander-entrenchments display high-angle thrust ruptures within colluvial deposits indicating the occurrence of successive coseismic movements in the recent past. Geologic units belonging to the Anatolian ophiolitic melange complex, mainly represented by serpentinites, are juxtaposed with alluvial river deposits, while sag-ponds develop to the ESE along the fault. In the basin at Eksisu, we visited a mineral water factory and a borehole with geyser-like resurgence which results from the hydrothermal activity along the fault. About 10 km of surface cracks appeared during the 1992 earthquake at this site, which also reveals an alignment of basaltic outcrops along the fault.
Aykut outlined different rupture scenarios related to the seismic gap and his most pessimistic prognostic is that east of Erzincan, the 75 km long fault segment, which includes the 1949 and 1966 fault sections, could break along the entire length during an Ms 7.1 earthquake.
The 1939 coseismic fault section. Three days (from 3rd to 5th August 1997) were necessary to visit the main tectonic features of the 360 km long fault and related coseismic movements associated with the Ms 7.9 1939 large earthquake. At Kuyumcu, general and detailed views of the fault reavealed displaced stream drainage, aligned old tree (with about 2.5 m of right-lateral movement), and a transpressive fault-section along road-cuts. Near Bahik, strongly folded probably late Pleistocene deposits were exposed along the fault plane, which itself displays low-angle slickensides with right-lateral frictional sense of movements. The hospitality and extreme kindness of local inhabitant and farmers were clearly evident when we were invited for a fruit-break, time. WNW of Bahik, a panorama of displaced stream beds and associated fan crests exhibits 2 to 3 m of right-lateral movement likely to be due to the last large seismic event. i.e. the 1939 earthquake. At the village of Mihar, the fault crosses a large valley and dissects young alluvial deposits and our "measurement-team" estimated 6 m and 12 m of right-lateral displacements across channels. According to Ambraseys (1970), who provided a detailed map of 1939 earthquake surface effects, the epicentral location might be near Mihar, which seems to be confirmed by the large displacements observed along this fault zone. Immediatly to the west, we climbed the Kara Dag, a 2832 m high "black mountain" and discovered a superb fault lineament picked out by a marshy green grass narrow strip running across a semi-desertic high plateau. Here again, sag-ponds, en echelon fault geometry and displaced river channels illustrate successive right-lateral movements on the fault with a probable coseismic 1939 displacement of about 6 m. After a long, dusty and winding road, we reached the Cukucimen locality where the fault, forming a large pull-apart basin, laterally displaced a presumable former road by about 8 m. At Tumekar village, the 1939 earthquake severely damaged the old Mosque and, according to local witnesses, the rupture crossed nearby agricultural fields where displaced aligned trees showed about 6 m of right-lateral movement. In continuation to the west and along several kilometers, the fault is pointed out by small pull-apart basins and fold structures corresponding to pressures ridges mapped in detail by Kocyigit (1985). Near Baro village, the southern side of a beautiful little lake is formed by two small hills separated by the fault, showing about 500 m of displacement. The magnificent fault bordering the lake was the setting for an unforgetable "pique-nique" and lunch-time discussion. We arrived at Sushehri basin in the afternoon and visited a trench of Koji Okumura (Hiroshima University) dug across a sag-pond formed along the fault which experienced 5 to 7 m displacement during the 1939 earthquake. At Ortakoy site, a field boundary is deformed by the fault forming a scarp with an average 6 m of dextral displacement. The city of Resadiye is close to the western end of the 1939 surface fault that displaces streets (3.7 m according to an old photo taken right after the earthquake), and forms a small pull-apart basin. Oblique to the fault (45°), fractures and cracks affect a prominent (several meters high) travertine construction resulting from the hydrothermal activity associated with hot water springs and outdoor "hammam" (Turkish bath) facilities. It is of interest to mention that from Susehri to Resadiye and until Niksar (about 150 km), the Kelkit river follows the fault trace. Away from the fault, the city of Tokat, which is famous for its hand-printed scarf and table cloths and the antique market in the old han, was our next stop.
1942 earthquake fault. (August 6th , 1997) One of the largest pull-apart basin along the fault is the Niksar and Erbaa depression. On the northern side of the basin, the Ms 7.1, 1942 earthquake occurred on the main fault, showing about 60 km of surface ruptures with a maximum of 1.8 m of right lateral offset, with the epicenter near Erbaa. Sag-ponds mark the fault trace and we noted and discussed the fact that the southern block of the fault is often uplifted. At least two main uplifted river terraces exist along strike which may result from successive vertical movements on the fault. West of Buzkoy village, the fault is always parallel to the Kelkit river. Near Tepekisla village we visited an old graveyard where 0.30 to 0.40 m offset gravestones resulted from the earthquake, as reported by Blumenthal (1943). An impressive natural mega-trench along the Destek river, near Alparslan village, reveals the fault in section and shows a 20 m wide fault zone with a folded southern block cut by secondary small normal faults (bending moment normal structures). Amasya, its gorges of the Yisilirmak river and impressive Hittite caves were a salutary stop to recover from the long trips of previous days ...
1943 seismic event and related 260 km long fault segment. (August 7th , 1997). The historical catalogue mentions a seismic crisis in August 17, 1668, studied by Ambraseys who interpreted it as a single large event with significant surface breaks from Kelkit valley to Bolu. Further to the west, between Sepetli and Destek villages, displaced river channels described by Allen (1969) show consistent dextral offsets along the fault. 400 m west of Destek, the total offset of an old, dry river channel, measured from an outstanding displaced crest, reaches 20 m and even 60 m. It is reported that during the 1943 earthquake the maximum offset was about 2.5 m. Indeed, we,ve found a local witness - a retired policeman who was 13 years old at the time of the earthquake - who decribed us the damage at his locality. Fruit gardens were displaced by 2.5 m to 1.5 m at two parallel sites along the fault, which probably corresponds to a splaying of the main fault into two branches. The lunch-break near Ladik lake was the opportunity to recall the different offsets, discuss possible slip rates, but also to admire the landscape. At the Arslantac site, the fault is visible in a quarry which exposes a 100 m wide fault zone and a succession of faulted paleosols. To the west, an impressive succession of sag-ponds partly covered by small alluvial fan deposits within a total fault length of 500 m was the site of a discussion about the best potential trenching sites for an extensive paleoseismological investigation along the fault. Displaced fences were also measured near this site to show an average 1.6 m offset, and along the same trend, a deflected wall was interpreted as apossible trace of the 1943 earthquake displacement. Near Yenice, and after a long walk, we crossed the meandering Tersahan river and arrived in front of an outstanding natural fault exposure with a 10 m wide shear zone, comprising oriented gravels and a kink-band with associated thrust structures, affecting Plio-Pleistocene formation. Blumenthal (1945) observed 1.0 m offset during the 1943 earthquake at this site. On our way to Corum we crossed the Suluova basin, another branch of the NAFZ. The 8 of August, 1997, was a break day and in this country with full of historical sites from the Hittites to Mustapha-Kemal Attaturk, we could not refrain from sacrificing a day for a cultural visit. We visited the Alacahoyuk (Hittites archeological sites northeast of Ankara) and Hattusas old city (Hittite capital) and its magnificent temple (1275 - 1250 BC), and end up in a carpet shop with the guide Attila!. That night, we stopped at Tosya.
Between Amasya and Tosya and to the west, the 1943 fault segment displays a curved trend. After a long climb in the morning of August 9th, 1997, we reached the Berçin village (north of Tosya). Here, the fault zone exhibits displaced river-channels with abandoned streams, and the relationships between the different drainage patterns show that at least two recent successive coseismic movements took place at this site, the last one most likely to have been the 1943 event. The gorgeous wood-house village which was built in 1918 was totally destroyed after the 1943 seismic event but has been reconstructed. The northern fault block is uplifted along this section, and in the Ilgaz-Eksik area west of Berçin, we measured several right-lateral offsets of channels of 3.0 m, 5.0 to 6.0 m, and 2.3 to 3.8 m along the edges of streams. A few km to the west, and after a winding road, we went to the pull-apart Kirkpinar basin where our co-traveller Sugai described his two trenches and the minimum two pre-1943 earthquakes at this site (Sugai, in preparation). The next stop was in Safranbolu, a remarkably beautiful and picturesque city, a remnant of the magnificent Ottoman time in this part of Anatolia.
1944 earthquake and related fault segment. The Ismet Pacha village creeping site, (August 10th, 1997) was first discovered by Ambraseys (1970) who identified some sections of the 1944 coseismic movements along the 150 km long active fault, and reported 0.24 m of right lateral progressive offset of a wall near the railway station in 1968. From 1969 to 1978, Altun (1982) measured 1 cm of annual displacement. This creep movement and offset are also regularly measured by means of a strainmeter that showed 0.38 m in 1993 (according to Wesnousky and Barka). We measured a maximum of 0.41 m which makes a total of 0.16 m in 25 years (= 0.8 cm/yr.) ; at this site between 1980 and 1985 the strainmeter gave a slip rate of 1.0 cm/yr. (Barka, personnal communication). Near this site along the fault zone, one can also notice displaced iron steps, thrusted concrete masonary and displaced drainage ducts. In 1944, both the wall and railroad tracks were affected by the fault and showed about 1.5 to 2.0 m of lateral displacements. East of Gerede town, remarkable pressure ridges (identical to those related to thrust faults) with clear en echelon patterns and about 0.7 m of vertical offset correspond to the still fresh traces of the 1944 fault breaks. The 1944 coseismic offset is estimated to be 3.0 to 4.0 m at this site, while in comparison displaced river-channels show a minimum of 4.5 m and a maximum of 27.0 m right lateral offsets. At another locality east of Gerede and along the fault zone, called Allen,s site (by Aykut), a stream with meanders is displaced by about 10.5 m (Allen, 1969). 12 km east of Gerede, Koji Okumura opened a trench where several events, including a 1668 and the 1944 earthquakes, were recognized, and estimated to indicate a slip rate of 1.7 cm/yr. (Okumura et al., 1994). The fault traverses the city of Gerede and extends to the west and southwest until the Burnuk huge quarry showing a large and outstanding view of the fault zone with folded units and successive displacements of young alluvial deposits and paleosols. West of Gerede, sag-ponds and elongated pressure ridges trace the fault trend. Finally, we also discussed the overlap of the 1944 fault segment and coseismic surface faulting with the Ms 6.9 1951 (easternmost fault section, 1.0 m offset) and the Ms 7.0 1957 earthquakes (westernmost fault section with 1.5 m offset). The night,s stop was at Bolu.
Overlaping fault sections and the 1957 - 1967 seismic events. (August 11th, 1997). West of Bolu, the NAFZ extends along 110 km toward the WSW with an overlap faulting section near the Abant Lake, which separates the 1944 and the 1957 fault breaks on one side, from the 1957 and 1967 fault breaks on the other side. We were not able to see much because of bad weather conditions and heavy fog obscuring the landscapes. However, by the afternoon we headed to Iznik and near Mekece, where the fault displays a nice morphology of active structures with about 150 m of topographic relief and curved set of river channels consistent with right lateral movement. According to Yoshihiko and Kuscu (1995), it is likely that east of the Geyve basin a large event occurred at about 200 BC while the last earthquake at Iznik took place at 128 AD. We also visited two trench-sites along this fault segment (Ikeda,s and Barka,s trenches) but both trenches were unsuccessful in exposing the fault. A discussion followed about how much can be reliable a morphological signature of faulting with respect to a real estimation of the fault activity. That night,s stop was at Iznik. The following day (August 12th, 1997), a half-day was devoted to the visit of Iznik Museum (of prehistory and archeology) and a walk through the city famous for its "Mosaic" design. The Iznik Lake is considered as a pull-apart structure along the fault which is here characterized by left-stepping faults. Uplifted terraces are clearly visible along the southern side of the lake, and the fault scarp extends toward the Marmara Sea and Gemlik town.
The northern strand of the NAFZ along the Sapanca-Izmit bay area was joined in the afternoon where normal faults affect marine terraces of interglacial stages e5 or e9 (Aykut personnal communication). The Sapanca and Izmit Pull-apart basins were the site of two main old projects: 1) a canal excavated to connect the Sapanca lake toward the Marmara Sea (a project started during Roman times and also Ottoman times when the work was aborted in 1591), and 2) a bridge joining the Hersek delta with the northern side of the bay. These two projects provided a large amount of geological data on earthquake faults but trenching are necessary to investigate the occurrence of earthquakes in the past and their relationship to the archeological remains. We came back to the ITU of Istanbul late in the evening.
1912 earthquake fault and the seismic hazard of Marmara sea . (August 13th, 1997) This earthquake (Ms 7.4) affected the northwestern coast of the Marmara Sea along a 80 km long fault segment that extends from the Gulf of Saros to the city of Gazikoy. According to a press report photograph, surface faulting seems to have occurred in this area and Ambraseys estimated 3.0 m of lateral displacement. The fault scarp is clearly visible on aerial photograph (1:20 000) forming sag-ponds along strike. We visited a trench opened by Tom Rockwell and Aykut Barka across a 0.30 to 0.40 m vertical scarp and had the opportunity of in situ explanations of a trench-exposure showing a vertical fault offset of deposits dated between 1900 BC and 887 AD. Three, and possibly four, earthquakes were recognized in this trench, the penultimate event being between 428 AD and 778 AD. While visiting the Gelibolu peninsula (the site of a famous First World War battle), a large debate followed about active faults along the Marmara Sea, their seismotectonic characteristics and the implications for seismic hazard assessment in Istanbul region. Aykut mentioned the tsunami-induced 1509 earthquake that affected Istanbul, could not hide his strong concern and discussed his model of successive pull-apart basins forming the Marmara Sea. The need for a complete inventory of active faults in the Marmara Sea region and detailed paleoseismological investigations were stressed. A very late evening travel to Istanbul ended the trip.
The aim of this report is to give, as much as possible, details of the field trip. However, fifteen days were definitely not enough to make a thorough visit to the North Anatolian Fault, but we were able to see much of the fault zone and imagine the effect of a large earthquake produced along this fault. Among the numerous discussions we had in the field, a recurrent one was the comparison with the San Andreas Fault and the seismic hazard implications with the straightforward question: is an earthquake sequence similar to the 1939-1944 seismic events likely in the future? Indeed, the 1939 earthquake sequence provided a large amount of data through the various descriptions of surface faulting. The strike-slip faulting behavior is, however, not yet well understood and much of the fault segments need to be studied. Although paleoseismological programmes were recently launched we are still in the prologue of a long story of the NAFZ.
Post scriptum: A complete field trip description would have included the various oriental "cuisine", the friendly and nice atmosphere, and the evening backgammon contest between Ariel, Tom and Arzu; to fill this gap, Lloyd Cluff prepared a book of selected photos such as each participants could visualize the good time we shared in August 1997. Tachakurlek Lloyd
References.
Allen, C. R., 1969, Active faulting in northern Turkey, Div. Geol. Sci., Calif., 32 pp, Caltech, Pasadena.
Ambraseys, N., 1970, Some characteristic features of the north Anatolian fault zone, Tectonophysics, 9, 143-165.
Aytun, A., 1982, Creep measurements in the Ismetpasa region of the north Anatolian fault zone, Progress in Earthquake Prediction, eds A. M. Isikara and A. Vogel, Friedr. Vieweg and John, F. R. G. 2, 279-294. Barka, A., & K. Kandisky-Cade, 1988, Strike slip fault geometry in Turkey and its influence on earthquake activity, Tectonics, 7, 663-684. Blumenthal, 1945, Ladik deprem hatti, Bull. Mineral Res. and Explor. Ints. Of Turkey, 1/33, 153-174.
Kocyigit, A., 1985, The Karayazi fault (in Turkish), Bull. Geol. Soc. Turk, 28, 67-71.
Okumura, K., Yoshioka T., and Kuscu, I., 1994, Surface faulting on the north Anatolian fault in these two Millenia, Proceeding of the workshop on Paleoseismology, 18-22 Sept 1994, Marshal, California, U.S. Geological Survey open file report 94-568, 143-144. Yoshioka, T., and Kuscu, I., 1994, Late Holocene faulting events on the Iznik-Mekece fault in the western part of the north Anatolian fault zone, Turkey, Bull. Geol. Surv. Japan, 45, 677-685.
List of participants
Serdar Akyuz (ITU, Istanbul), Rivka Amit (Geol. Survey of Israel), Ryo Anma (ERI, Tokyo), Yoav Avni (Geol. Survey of Israel), Yasuo Awata (Geol. Survey of Japan), Hidir Aygul (ITU, Istanbul), Aykut Barka (ITU, Istanbul), Ann Blythe (USC, Los Angeles), Lloyd Cluff (Calif. Seismic Safety Commission, San Francisco), James Dolan (USC, Los Angeles), Ariel Hieman (Geol. Survey of Israel), Sonya Lenz (Dept. of Industries and Energy, Australia), Shmulik Marco (Geol. Survey of Israel), Ari Matmon (Geol. Survey of Israel), Kevin McCue (Dept. of Industries and Energy, Australia), Mustapha Meghraoui (CNR, Italy), Tom Rockwell (UC San Diego), Yakayasu Saka (Waseda Univ., Japan), Arzu Sekirden (Bogazici Univ., Istanbul), Toshihiko Sugai (Geol. Survey of Japan).