(cid:0)(cid:2)(cid:2)(cid:3)(cid:4)(cid:5)(cid:3)(cid:6) (cid:7)(cid:8)(cid:9)(cid:10)(cid:11)(cid:2)(cid:12)(cid:13)(cid:4)(cid:5) Sedimentary records of past earthquakes in Boraboy Lake duringthe last ca 600years(NorthAnatolianFault,Turkey) Ulas¸ Avs¸ar, Aure´lia Hubert-Ferrari, Marc De Batist, Sabine Schmidt, NathalieFagel PII: S0031-0182(15)00243-6 DOI: doi: 10.1016/j.palaeo.2015.04.031 Reference: PALAEO7270 Toappearin: Palaeogeography,Palaeoclimatology,Palaeoecology Receiveddate: 4January2015 Reviseddate: 27April2015 Accepteddate: 29April2015 Please cite this article as: Av¸sar, Ula¸s, Hubert-Ferrari, Aur´elia, De Batist, Marc, Schmidt, Sabine, Fagel, Nathalie, Sedimentary records of past earthquakes in Boraboy Lake during the last ca 600 years (North Anatolian Fault, Turkey), Palaeogeography, Palaeoclimatology,Palaeoecology(2015), doi: 10.1016/j.palaeo.2015.04.031 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT Sedimentary records of past earthquakes in Boraboy Lake during the last ca 600 years (North Anatolian Fault, Turkey) T Ulaş Avşar a,b*, Aurélia Hubert-Ferrari c, Marc De Batist a, Sabine Schmidt d, Nathalie Fagel e P a Renard Centre of Marine Geology (RCMG), Ghent University, Krijgslaan 281, 9000 Gent, Belgium I R b Division of Physical Sciences and Engineering, King Abdullah University of Science and Technology C (KAUST), 23955-6900 Thuwal, Kingdom of Saudi Arabia S c Unit of Physical and Quaternary Geography, University of Liège, Sart Tilman, B- 4000 Liège, Belgium U d UMR 5805 EPOC OASU, University of Bordeaux, 33615 Pessac, France N e AGEs, Department of Geology, University of Liège, Sart Tilman, B- 4000 Liège, Belgium A M *Corresponding author: Tel.: +966 2 8080351 , E-mail: [email protected] D Abstract E Multiproxy sedimentological analyseTs along 4.9 m-long sequence of Boraboy Lake, which is located on the P central eastern part of the North Anatolian Fault (NAF), reveal the sedimentary traces of past large earthquakes in the region. The lake has a Erelatively large catchment area (10 km2) compared to its size (0.12 km2), which C renders sedimentation sensitive to heavy rain/storm events. Accordingly, the background sedimentation, which is C composed of faintly laminated reddish/yellowish brown clayey silt, is frequently interrupted by organic-rich A intercalations probably due to heavy rain/storm events transporting terrestrial plant remains from the densely vegetated catchment. In addition to frequent organic-rich intercalations, the background sedimentation is interrupted by four mass-wasting deposits (MWD) of which thickness range between 15 and 50 cm. High- resolution ITRAX XRF data confirms higher homogeneity along the MWDs (E1-E4) compared to the background sedimentation. Based on 137Cs and 210Pb dating and radiocarbon chronology, three MWDs detected xs in Boraboy sequence (E2, E3 and E4) temporally correlate with large historical earthquakes along the NAF; the 1943 Tosya (M=7.6) and/or 1942 Niksar-Erbaa (M=7.1), the 1776 Amasya-Merzifon and the 1668 North s s Anatolian (M=7.9) earthquakes. The youngest MWD in the sequence (E1), which is dated to early 2000s, does s not correlate with any strong earthquake in the region. This MWD was probably a single mass-wasting event due to routine overloading and oversteepening on the delta front formed by the main inlet of the lake. In subaqueous paleoseismology, coevality of multi-location mass-wasting events is used as a criterion to assign a seismic 1 ACCEPTED MANUSCRIPT triggering mechanism, and to rule out mass-wasting events due to routine overloading/oversteepening of subaqueous slopes. Within this context, Boraboy sequence provides a valuable example to discuss sedimentological imprints of single- vs. multi-source MWDs. T P Keywords: Boraboy Lake, lacustrine paleoseismology, radiochronology, seismoturbidite, homogenite, mass- I R wasting deposits C S 1. Introduction U In addition to the most commonly applied technique in paleoseismology, i.e. paleoseismic trenching, geological N fingerprints of past earthquakes can be found by applying subaqueous paleoseismological techniques on marine A and lacustrine sedimentary sequences. The most commonly observed sedimentary traces of past earthquakes are M mass-wasting deposits (MWD). Cyclic loading during earthquakes decreases effective stress in the sediments, which results in instabilities on the steep subaqueous basin slopes. The wasted material can be transported to the D middle parts of the basin by slumping, debris flows and turbidity currents, and archived in the sedimentary E sequence as slump or debris-flow deposits, seismoturbidites and/or homogenites. After the first introduction of T the term “homogenite” by Sturm Pet al. (1995), numerous studies have successfully used homogenites to reveal paleoseismic information (e.g.E, Chapron et al., 1999; Karlin et al., 2004; Monecke et al., 2006; Beck et al., 2007; Schnellmann et al., 2006; CGuyard et al., 2007; Bertrand et al., 2008; Carillo et al., 2008; Beck, 2009; Waldmann et al., 2011; Campos Cet al., 2013; Howarth et al., 2014; McHugh et al., 2014; Moernaut et al., 2007; 2014). Another commonlAy observed consequence of cyclic loading on sediments are in-situ soft-sediment-deformations. Cyclic loading may result in microfaults (e.g., Seilacher, 1969; Monecke et al., 2004; 2006; Beck, 2009; Kagan et al., 2011), microfolds (e.g., Monecke et al., 2004; Kagan et al., 2011), loadcast/pseudonodule/ball-and-pillow structures (e.g., Sims, 1973; Hibsch et al., 1997;Migowski et al., 2004; Monecke et al., 2004; Moretti and Sabato, 2007; Sims, 2012) and liquefaction/fluidization/fluid-escape structures (e.g., Lignier et al., 1998; Chapron et al., 2004; Moernaut et al., 2007; 2009; Beck, 2009; Kagan et al., 2011). Moreover, seismically- induced water oscillations (i.e., seiche) may cause in-situ sediment resuspension and redeposition (e.g., Siegenthaler et al., 1987; Doig, 1990; 1991; 1998a; 1998b), or transportation of material from lake shores (e.g., Avşar et al., 2014b). In addition to subaqueous mass-wasting, earthquakes may trigger landslides in lake catchments, which may enhance sediment flux into lakes and leave sedimentary traces in the sequence (e.g., Leroy et al., 2009; Hubert-Ferrari et al., 2012; Avşar et al., 2014a; Howarth et al., 2012; 2014). In addition to 2 ACCEPTED MANUSCRIPT seismic shaking, surface faulting during earthquakes may also affect sedimentation, especially in shallow water environments (e.g., Mirecki, 1996; Leroy et al., 2009; 2010; Vologina et al., 2010; Bertrand et al., 2011). For example, tectonic subsidence in shallow basins may result in deepening of the lake and inundation of T surrounding land, which may leave traces in the sedimentary sequence. P The sedimentary structures and changes in sedimentation mentioned above are, however, not always of seismic I R origin. For example, lake water level fluctuations or enhanced sediment influx can be simply due to changes in C hydrometeorological conditions, or mass-wasting events can be caused by routine overloading/oversteepening of S subaqueous slopes. Some soft-sediment-deformations can even be induced during coring. In order to rule out U these possibilities, subaqueous paleoseismology basically seeks for coevality of multi-location mass-wasting N events and soft-sediment deformations, and for temporal correlation between observed sedimentary events and A historical seismicity records, if available. In this study, we present observations on mass-wasting deposits M (MWD) within the sedimentary sequence of Boraboy Lake, which is located on the North Anatolian Fault (NAF), Turkey. The region has well-acknowledged historical seismicity records (e.g., Ambraseys, 2009), which D allows evaluating triggering mechanisms of the mass-wasting events. In addition to the seismically-triggered E MWDs, the presence of a very recent non-seismically-triggered MWD in the Boraboy sequence allows T comparisons to be made of sediPmentological imprints of single- versus multi-source MWDs. We will also compare the Boraboy record wEith the record of Göllüköy Lake, which is another lacustrine paleoseismic record located ca 120 km to the Ceast along the NAF (Avşar et al., 2014b) in order to improve our knowledge on the locations of the historiCcal earthquakes in the region during the last 600 years. A 2. The study area Boraboy Lake (40.8038˚ N, 36.1535˚ E, 1062 m a.s.l.) is located at the central eastern part of the NAF, which is a major 1000 km-long dextral strike-slip fault in Turkey (Fig. 1a). It is a small (0.124 km2) landslide-dammed lake with a maximum water depth of 11 m, and it is located at 7 km from the NAF (Fig. 1b and 1c). The catchment of the lake is composed of highly weathered Lower – Middle Jurassic age alternations of tuff, shale, greywacke and sandstone units (Herece and Akay, 2003). The inlet of the lake, which is fed by a densely vegetated and relatively big catchment (10.02 km2), forms a delta at its western shores. The slope angles at the northern and southern margins of the lake basin, as well as at the delta front, reach up to 16˚ to 20˚, which makes them highly susceptible to subaqueous slope failures during an earthquake. In addition, the origin of the lake, i.e. being a landslide-dammed lake, supports the idea that the morphology in and around the lake is highly 3 ACCEPTED MANUSCRIPT susceptible to earthquake-triggered landslides. This segment of the NAF was ruptured by the 1943 Tosya Earthquake (Ms=7.6), which created surface rupture extending 280 km along the NAF. The segment to the east of this one was also ruptured by the 1942 Niksar-Erbaa earthquake (Ms=7.1), of which surface rupture stopped T ca. 40 km to the east of the lake (Barka, 1996). In addition to these instrumentally recorded earthquakes, the P region was also affected by the great 1668 North Anatolian Earthquake (Ms=7.9), of which surface rupture was I detected in a paleoseismic trenching site ca 8 km to the NNW of the lake (ARmbraseys, 2009; Fraser et al., 2009). C 3. Methods S U After collection of bathymetric data by using a sonar equipped Garmin GPS Map 178, one gravity (BR2007-G, N 78 cm) and one piston core (BR2007-01, 470 cm) were taken from the central and the deepest (~11.0 m) part of A the lake in 2007 (Fig. 1c). High-resolution sediment core logging and scanning were done by using a Geotek M multi-sensor core logger (MSCL), an ITRAX μXRF core scanner and a SCOPIX X-ray image-processing system. The magnetic susceptibility profiles were obtained at 2 mm increments by the Geotek MSCL at D the Institute for Applied Geophysics and Geothermal Energy at the RWTH Aachen University. ITRAX μXRF E core scanning was done at the Department of Geology and Geochemistry of Stockholm University, with a T resolution of 2 mm and 12 s expoPsure time with Mo-tube. The radiographic scanning of the BR2007-01 piston core was done on centimeteEr-thick slabs by the SCOPIX X-ray image-processing system at the EPOC laboratories of UniversityC of Bordeaux-I (Migeon et al., 1999). Organic matter content of the sediments was measured at 1 cm incrCements by the method of weight loss on ignition (LOI) at 550˚C for 4 hours (Dean, 1974) at the U. R. ArgileAs, Géochimie et Environnements sédimentaires (AGEs) of University of Liège. The activities of 137Cs and excess 210Pb (210Pb ) were measured on 1 cm-thick sediment slices from 18 levels along BR2007-G. xs The measurements were done by using a low-background, high efficiency well-type gamma spectrometer at the University of Bordeaux (Schmidt et al., 2014). The sediment chronology along the BR2007-01 piston core was established by AMS radiocarbon measurements on terrestrial plant remains. The measurements were done at the AMS facility at the University of Arizona, and the dates were calibrated by using OxCal 4.2.3 software (Bronk Ramsey, 2013) based on the IntCal13 atmospheric curve of Reimer et al. (2013). 4. Background sediments versus event deposits The optical and radiographic images, magnetic susceptibility, organic matter content (LOI ) and ITRAX μXRF 550 scanning results along the gravity (BR2007-G) and piston (BR2007-01) cores of Boraboy Lake are presented in 4 ACCEPTED MANUSCRIPT Fig. 2. The top part of the piston core can be easily correlated with the gravity core, revealing that the top ca. 20 cm of the piston core was lost during coring. The sedimentation in Boraboy Lake is mainly represented by faintly laminated reddish/yellowish brown clayey silt deposition, which is frequently interrupted by organic-rich T intercalations and occasionally by relatively homogenous deposits. P Organic-rich intercalations, which are composed of macrofossils of terrestrial plants (mainly leaves), are I R characterized by relatively darker colors in the optical images, lighter colors in radiographic images, lower C magnetic susceptibility values and higher LOI values (Fig. 2a). Along the ITRAX elemental profiles (Fig. 2b), 550 S organic-rich layers tend to have lower counts-per-second values, probably due to the dilution effect of organic U matter on X-ray fluorescence. Given their high frequency recurrence and terrestrial origin, we interpret these N organic-rich intercalations as the sedimentary traces of storm and/or heavy rain events. A The sequence also contains several relatively homogenous units (E1-E4 in Fig. 2), which are not interrupted by M storm and/or heavy rain events. Most of the proxies do not show any significant variations along these deposits. As it is most clearly seen along E4, they arDe characterized by decreasing upward magnetic susceptibility and optical image grayscale values, and by inEcreasing upward LOI and radiographic image grayscale values (Fig. 550 2a). Among the ITRAX μXRF elemeTntal profiles, Si, K and Ti show decreasing upward trends, where Ca and P Mn do not show any obvious trends (Fig. 2b). A close-up view of E4 is also provided in Fig. 2c. Fig. 3 presents E the plots of standard deviation values of a 2 cm-thick window along the measured proxies, and the average of the C plots. Along the average plot, with a few exceptions, the homogenous units have lower standard deviation values C compared to the background sedimentation. Another observation is that diatoms, which are typically observed in A the background sedimentation, are lacking in these homogenous units (Fig. 4). We interpret the gradual decreasing/increasing upward trends along these deposits as an indication that they were redeposited under the effect of gravitational segregation. For example, the clastic fraction having higher density settles first while the organic fraction is still in suspension, which results in increasing upward LOI values. In addition, the absence 550 of diatoms in these sediments implies instantaneous deposition. Given these observations, we attribute the homogenous deposits (i.e., turbidites/homogenites) along the sequence (E1-E4) to mass-wasting events, which might have been triggered by storm and/or heavy rain events, earthquakes, drastic lake level fluctuations or simply over-steepening of the lake basin slopes over time. We investigated the possible seismic origin of these mass-wasting events by establishing a detailed sediment chronology in order to check their temporal correlation with known historical earthquakes in the region. 5 ACCEPTED MANUSCRIPT 5. Sediment chronology and age of event deposits The 137Cs and 210Pb activity profiles along the BR2007-G core are presented in Fig. 5a. The 210Pb activity at xs xs 3.5 cm depth is significantly lower than the overlying sediments, implying that E1 (3-36 cm) is composed of T reworked sediments. In order to determine the background sedimentation rate, the sediments from 3 to 36 cm P (i.e. E1) and 56 to 62 cm (i.e. storm/heavy rain deposits) depths should be excluded from the sequence, and the I R 137Cs and 210Pb profiles should be re-plotted as a composite sequence (e.g., Arnaud et al., 2002; Schwab et al., xs C 2009). In the composite sequence (Fig. 5b), the main 137Cs peaks in 1986 and 1963 correspond to ca. 19 and 32 S cm depths, respectively. Accordingly, along with assuming the top of the core as 2007, the reference points U along the 137Cs activity profile yield a background sedimentation rate (SR) of 0.70 cm/yr. The exponential fit on N the 210Pb data presented in Fig. 5b has correlation coefficient of 0.89. The data points at 4.5, 6.5 and 24.5 cm xs A depths (blank circles in Fig. 5b) are not included into the exponential fit because of their high discrepancy from M the general trend. The ones at 4.5 and 6.5 cm depths are probably due to a non-distinct storm and/or heavy rain deposits, while the one at 24.5 cm may be related to sediment focusing. Accordingly, the exponential decay of D 210Pb activity results in an SR of 1.04 cm/yr, if “Constant Flux Constant Sedimentation Rate (CFCS)” is xs E assumed. Considering both sedimentation rates from 137Cs and 210Pb results, E1 is dated to early 2000s and it is T xs not related to any strong earthquakPe in the region. E For the older mass-wasting events (E2-E4) along the sequence, the sediment chronology is established by C radiocarbon dating of ten samples of tree leaves at various depths along the piston core (R0-R9 in Fig. 3). The C depths of the samples and the mass-wasting deposits in the bulk and composite sequence, and the obtained A radiocarbon dates are presented in Table 1. During the construction of the composite sequence, the storm and/or heavy rain deposits were assumed to be the part of the background sedimentation due to their high frequency of recurrence, and they were not excluded from the sequence; while the homogenous units (E1-E4) were excluded. In Fig. 6, the calibrated radiocarbon dates, modeled dates of the mass-wasting events and corresponding age- depth models are presented for both bulk (a) and composite (b) sequences. The grey curves along the x axes are the distributions for the single calibrated dates. Although the single calibrated ages have wide ranges of distributions, due to the plateaus in the calibration curve, the sample “R8” at 412 cm depth in the bulk sequence constrains the age as ca. 1460-1660 AD, which results in 0.76-1.17 cm/yr sedimentation rate (SR). In the composite sequence (Fig. 6b), the distribution of “R8” yields an SR of 0.53-0.82 cm/yr, which is slightly lower than the SR calculated by the 137Cs and 210Pb profiles (i.e., 0.70-1.04 cm/yr). This is reasonable because the xs deeper parts of the sequence are expected to have lower sedimentation rates due to compaction. In order to better 6 ACCEPTED MANUSCRIPT constrain the dates of the MWDs, the modeled ages were calculated by OxCal v4.2.3 (Bronk Ramsey, 2013) based on the “P_Sequence” depositional model, which allows random variation from an approximately constant sedimentation rate (Bronk Ramsey, 2008). Accordingly, the marginal posterior distributions for the single T modeled ages with 95% confidence interval are shown as black-filled curves, and the age-depth models are P presented based on the 68% confidence interval of the modeled ages (the grey polygons down-core). As might I be expected, excluding MWDs from the sequence improves the age-depth mRodel, which supports the assumption of instantaneous deposition of E1-E4 (Fig. 6b). C S E1 is dated to early 2000s and it does not temporally correlate with any strong earthquake in the region. Within U the time span covered by the Boraboy sequence in this study, the best-known NAF earthquakes in the region are N the 1943 Tosya Earthquake (Ms=7.6), the 1942 Niksar-Erbaa Earthquake (Ms=7.1) and the 1668 North A Anatolian Earthquake (Ms=7.9). E2 is dated to AD 1925-1820 based on radiocarbon chronology, and to AD M 1957-1932 based on the 137Cs and 210Pb dating (Fig. 6b). Most probably, it was triggered by the 1943 Tosya xs and/or the 1942 Niksar-Erbaa earthquakes. The oldest event (E4) in the sequence, which is dated to AD 1682- D 1622, was probably triggered by the giant North Anatolian Earthquake (Ms=7.9) in AD 1668. This earthquake E was the most catastrophic historical earthquake in Anatolia, of which damage zone extended ca. 650 km along T the NAF from Bolu-Gerede to ErPzincan-Refahiye (Ambraseys, 2009). Probably, that is why E4 is the thickest and the most obvious event aElong the Boraboy sequence recovered in this study. E3, which is the most non- distinct event in the sequenCce, is dated to AD 1761-1674 with 95 % confidence interval. Historical records report an earthquake in 1776 Ccausing damage in the towns around Boraboy Lake. A Based on the constructed chronology, it is obvious that earthquakes trigger subaqueous slope failures in Boraboy Lake. In addition to MWDs, sedimentary traces of increased sediment influx due to earthquake-triggered landslides in the catchment (i.e., catchment response) can be expected in Boraboy sequence, because the lake has a relatively big catchment. According to the compilation made by Avşar et al. (2014a), the median value of the ratio of catchment area to lake area (CA/LA) of 52 lakes in the lacustrine paleoseismology literature is around 14. They report clear traces of catchment response in Yeniçağa Lake, of which CA/LA is 74, as denser siliciclastic-enriched layers in the sequence. The CA/LA ratio of Boraboy Lake is 81, as high as the ratio of Yeniçağa Lake. However, no clear implications of catchment response are observed in Boraboy Lake. Avşar et al. (2014a) illustrate that the sedimentary traces of catchment response are nondistinct in Yeniçağa Lake during the episodes of high biogenic production in the region, i.e. episodes when the catchment is densely vegetated, 7 ACCEPTED MANUSCRIPT which decreases erosion rate. Probably, the traces of catchment response are not clearly observed in Boraboy Lake due to its densely vegetated catchment. 6. Regional paleoseismicity T P In order to better understand the paleoseismicity of the NAF in the region, we compared the earthquake I R sedimentary record Boraboy Lake with the regional historical earthquakes, and also with the lacustrine paleoseismic record of Göllüköy Lake, which is located ca.120 km to thCe ESE of Boraboy Lake (Fig. 7, modified from Avşar et al. (2014b)). The 20th century earthquakes and the 16S68 North Anatolian Earthquake are recorded in both Boraboy and Göllüköy lakes, which are already kUnown as characteristic surface rupturing large earthquakes of the NAF. The earthquakes, which are recorNded only in one of the lakes (i.e., 1825-1827 Erbaa- Tokat, 1776 Amasya-Merzifon-Tokat, and 1543 Tokat-AÇorum), were probably relatively local and smaller ones. None of the remaining historical earthquakes were Mrecorded in the lake sequences, implying that they cannot be large characteristic earthquakes of the NAF. D The record of Boraboy Lake extends back to 1450s, and hence we do not know whether the 1418-1419 Amasya- E Tokat Earthquake, which is clearly observed in Göllüköy record, was recorded in Boraboy Lake or not. This T earthquake is not recorded in any Ppaleoseismological trenches along the NAF (Fraser et al., 2010). Paleoseismic trenching data shows that the NEAF ruptured in that region during the 1668 earthquake and before during a period of high NAF seismicity bCased on historical earthquakes of AD 967 and AD 1035 in the region, and AD 1050 near Gerede, and AD C1043 and AD 1045 near Erzincan (Fraser et al., 2009). Given the wide spacing between paleoseismologicaAl trenches, the 1418-1419 earthquake can still be a M > 6 earthquake rupturing the NAF, but we cannot preclude that it is a large regional earthquake off the NAF like the M=7.2 1916 earthquake (Ambraseys, 1997). 7. Seismically- versus non-seismically-triggered event deposits The Boraboy sequence contains four mass-wasting deposits, three of which (E2-E4) temporally correlate with past earthquakes in the region. Here, it is worth discussing some sedimentological observations, in which E1 differs from the others. First, while the LOI profiles along the E2-E4 show increasing-upward trends, it is 550 almost constant along E1 (Fig. 8). The trends of LOI profiles are also confirmed by the XRF profiles (i.e., Si, 550 K, Ca and Ti), which are indeed affected by organic matter content of sediments. Second, E1 is more homogenous than the other event deposits. Third, E1 has around 2% less organic matter content compared to the 8 ACCEPTED MANUSCRIPT preceding background sediments. However, at their lowermost levels, E2-E4 have almost the same organic matter content as the preceding background sediments, and they get more organic-rich towards their upper levels. Based on these observations, we speculate that E1 was a single mass-wasting event, which took place on T the delta associated with the main inlet of the lake, while the other event deposits were due to multiple coeval P mass-wasting along the slopes of the lake basin. Slight variations of the measured proxies along the E2-E4 I deposits may be attributed to multiple sources of sediments, i.e., multi-loRcation coeval mass-wasting events. Moreover, since there are no significant tributaries on the northern and tChe southern slopes of the lake catchment, the sediments of the delta associated with the main inlet of the laSke can be expected to be more dominant in clastics (less organic-rich) than the sediments deposited on the Unorthern and southern basin slopes. This may be the reason why E1 has less organic matter content than the pNreceding background sediments. A 8. Conclusions M In Boraboy Lake, which is a landslide-dammed lake located at ca. 7 km from the NAF, the sedimentary traces of D the 1943 Tosya and/or 1942 Niksar-Erbaa (Ms=7.6 and 7.1, respectively), the 1776 Amasya-Merzifon and the E 1668 North Anatolian (Ms=7.9) earthquakes are found as mass-wasting deposits (MWD). Being the most T catastrophic earthquake ever known in Anatolia, the 1668 North Anatolian Earthquake resulted in the thickest P MWD within the Boraboy sedimentary sequence recovered in this study. In historical resources, it is sometimes E difficult to evaluate which historical earthquakes were large magnitude regional earthquakes. Combining the C sedimentary records of Boraboy and Göllüköy lakes, which are ca. 120 km apart from each other, allowed C distinguishing between large regional earthquakes and the small local ones. Accordingly, we argue that the A historical 1418-1419 Amasya-Tokat Earthquake was a large earthquake in the region like the M=7.2 1916 earthquake. In addition, the Boraboy sequence allowed a sedimentological comparison to be made between seismically and non-seismically induced MWDs. Basically, seismically-induced MWDs seem more heterogeneous probably due to multi-source sediments transported by coeval multiple mass failures. Acknowledgements This research was funded by the European Commission Marie Curie Excellence Grant Project “Understanding the irregularity of seismic cycles: A case study in Turkey” (MEXT-CT-2005-025617: Seismic Cycles), hosted by the Seismology Section of the Royal Observatory of Belgium. The coring was carried out by collaboration with Istanbul Technical University - Eastern Mediterranean Centre for Oceanography and Limnology (ITU-EMCOL). 9