Palaeoseismology of the North Anatolian Fault near the Marmara Sea: implications for fault segmentation and seismic hazard

Abstract

We conducted palaeoseismic studies along the North Anatolian fault both east and west of the Marmara Sea to evaluate its recent surface rupture history in relation to the well-documented historical record of earthquakes in the region, and to assess the hazard of this major fault to the city of Istanbul, one of the largest cities in the Middle East. Across the 1912 rupture of the Ganos strand of the North Anatolian fault west of the Marmara Sea, we excavated 26 trenches to resolve slip and constrain the earthquake history on a channel-fan complex that crosses the fault at a high angle. A distinctive, well-sorted fine sand channel that served as a marker unit was exposed in 21 trenches totaling over 300 m in length. Isopach mapping shows that the sand is channelized north of the fault, and flowed as an overflow fan complex across a broad fault scarp to the south. Realignment of the feeder channel thalweg to the fan apex required about 9+1 m of reconstruction. Study of the rupture history in several exposures demonstrates that this displacement occurred as two large events. Analysis of radiocarbon dates places the age of the sand channel as post AD 1655, so we attribute the two surface ruptures to the large regional earthquakes of 1766 and 1912. If each was similar in size, then about 4 -5 m of slip can be attributed to each event, consistent with that reported for 1912 farther east. We also found evidence for two additional surface ruptures after about AD 900, which probably correspond to the large regional earthquakes of 1063 and 1344 (or 1354). These observations suggest fairly periodic occurrence of large earthquakes (RI ¼ c. 283+113 years) for the past millennium, and a rate of c. 16 mm/a if all events experienced similar slip.We excavated six trenches at two sites along the 1999 Izmit rupture to study the past earthquake history along that segment of the North Anatolian fault. One site, located in the township of Köseköy east of Izmit, revealed evidence for three surface ruptures (including 1999) during the past 400 years. The other trench was sited in an Ottoman canal that was excavated (but never completed) in 1591. There is evidence for three large surface rupturing events in the upper 2 m of alluvial fill within the canal at that site, located only a few kilometres from the Köseköy site.

Recent behavior of the North Anatolian Fault: Insights from an integrated paleoseismological data set

“…Rockwell et al [2009]did not use an OxCal model for this site, we ran a model using all of the dates and found a very poor fit, which we attributed to the reworking of sample 6. By excluding this sample our model is consistent with the appropriately vague interpretation of historical earthquakes described by Rockwell et al [2009].…”

“…These data were used along with additional data by Rockwell et al [2009] to constrain the timing of two paleoearthquakes and clear evidence of the 1999 İzmit earthquake. Using the fraction modern carbon presented by Rockwell et al [2009], we determine the unrounded conventional radiocarbon ages that are included in the paleoearthquake model.…”

“…These data were used along with additional data by Rockwell et al [2009] to constrain the timing of two paleoearthquakes and clear evidence of the 1999 İzmit earthquake. Using the fraction modern carbon presented by Rockwell et al [2009], we determine the unrounded conventional radiocarbon ages that are included in the paleoearthquake model. Rockwell et al [2009]did not use an OxCal model for this site, we ran a model using all of the dates and found a very poor fit, which we attributed to the reworking of sample 6.…”

“…Therefore, the results from existing paleoseismic investigations have been determined using slightly different methodologies, making it dubious to draw comparisons between them. With a long historical earthquake record [e.g., Ambraseys and Jackson, 1998], a relative abundance of paleoseismic data [e.g., Hartleb et al, 2006;Kozacı, 2008;Pantosti et al, 2008;Rockwell et al, 2009], and relatively simple fault geometry (i.e., linear without major interacting faults) along the eastern two thirds [e.g., Barka, 1992], the North Anatolian Fault (NAF) is an ideal natural laboratory to study long-term (thousands of years) patterns of seismicity.…”

Kinematic analysis and palaeoseismology of the Edremit Fault Zone: evidence for past earthquakes in the southern branch of the North Anatolian Fault Zone, Biga Peninsula, NW Turkey

“…When we considered these earthquakes' locations, formation timings and magnitudes, it seems that the large earthquakes, which eventuated on the EFZ and were quite close to the YGFZ, triggered each other. The palaeoseismological studies carried out in the northern branch of the NAFZ suggested that the estimated recurrence interval is 150-300 years for surfaceruptured large earthquakes (Hartleb, Dolan, Kozaci, Akyuz, & Seitz, 2006;Ikeda et al, 1991;Kozacı, Dolan, & Finkel, 2009;Kozacı, Dolan, Yönlü, & Hartleb, 2011;Özaksoy et al, 2010;Rockwell, Barka, Dawson, Akyuz, & Thorup, 2001;Rockwell et al, 2009), while palaeoseismological studies on the southern branch are very limited, except for Yoshioka and Kuşçu (1994), Kürçer et al (2008), Belindir (2008) and Özalp et al (2013); however, Özalp et al (2013) reported that the return period of the greater earthquakes that occurred on the southern branch was not regular, but these fault segments have the potential to produce earthquakes on the northern branch. Additionally, Belindir (2008) reported an earthquake recurrence interval on the YGFZ that ranged between 253 and 925 years.…”