Interactive comment on “ Estimation of successive co-seismic vertical offsets using coeval sedimentary events – application to the Sea of Marmara ’ s Central Basin ( North Anatolian Fault ) ” by C

In the deep part of the Sea of Marmara (Turkey), the sedimentation developing upon the North Anatolian Fault is strongly influenced by the associated seismic activity, through gravity reworking (fluidized landslides) and tsunamis. Specific layers (homogenites + turbidites, HmTu), representing individual sedimentary events, have been characterized along three giant piston cores retrieved from the Cinarcik and Central (or Orta) basins. Pre-Holocene, nonmarine sediments, were analyzed, representing the last 12–17 kyr BP (before present). For a 2 kyr long interval, 11 events could be precisely correlated on both sides of the Central Basin's southwestern scarp. For each of them, based on the specific depositional process, the thickness difference between the two sites was considered as a direct estimation of the vertical component of a coeval coseismic offset. The homogenite (upper) component accounts for the major part of the thickness difference (ranging from 36 to 144 cm). These offsets were considered as likely representing dominantly vertical throws, along the transtensional southwestern boundary of the inner, pull-apart Central Basin. In terms of natural hazards, further investigations on this local behavior should rather be directed to tsunami genesis.

Specific point 2: We focused on the Central Basin where we could obtain two giant piston cores (including LGM, Late Glacial, and Holocene periods) on both sides of a major active scarp; which was not the case for the other sub-basins.A first preliminary work indicated that the strong over-thickening in the deepest part (hanging wall) is entirely due to sedimentary events (homogenites-turbidites), being the "homogenite" term responsible for the major part of the difference.A new detailed analysis of these two cores were then planned, providing: 1) the discovery of specific sedimentary events sealing co-seismic scarps in the Lesser Antilles Arc (Beck et al, 2012), 2) the development of Anisotropy of Magnetic Susceptibility along-core profiles (Campos et al., 2013).The study of a long core from the Çinarcik Basin was added (for chronological as well as for paleo-environmental purposes) in order to ensure the regional occurrence of a well-defined "reference" layer (isochron for correlations) at the non marine/marine transition.Both VHR seismic image and cores data indicate that, before this pre-Holocene (non marine) level, the sea-floor stayed flat, due to the complete "erasing" of the fault scarp by reworked sediments depositions (turbidites-homogenites).Specific point 3: During MARMARASCARPS Cruise, directed by R. Armijo and J. Malavielle, shorter Küllenberg type cores, dedicated to the last millennia, were retrieved from the different basins.Beside, a set of 100 short cores (35 cm) were visually picked using VICTOR R.O.V., preserving sediment/water interface.We participated to these two parts of the survey.The processing of the two sets of core were included in L. Drab's and G. Uçarcus' PhD theses.Specific point 4: We separate a mean total sedimentation rate (including "events") from hemipelagic sedimentation rate.The second one is used to estimate the time elapsed C2074 between two successive instantaneous events.Specific point 5: For nowadays marine basins with shallow connections with open sea (e.g.Corinth or Cariaco Gulfs), the problem of reservoir age for 14C is complex and depending on specific site conditions.For the Sea of Marmara, water circulation underwent several changes since 25 kyr BP due to connections changes (Black Sea + Aegean Sea, only Black Sea; see Vidal et al., 2010).Reservoir effects on CO2 are depending: i) on Black Sea and Aegean Sea water characteristics, ii) on their mixing within the Marmara basin, iii) local drainage.In addition, methane seepage may induce the incorporation of dead C into deep water metabolic chains.Thus, the used corrections of reservoir effect, although based on a combination of published values, are questionable.Nevertheless, our final chronological discussion is rather directed to time intervals and frequencies rather than their precise absolute age.Specific point 6: In order to establish event-by-event correlation, a very well-defined layer -usable as an isochron -was necessary.The chosen one belongs to the non marine/marine change level according to previous studies and dating; it is conspicuous, unique in terms of texture, composition, and bioturbation, and could be recognized on the three cores.Specific point 7: I agree with a possible erosion below a density flow, which may eliminate some hemipelagic mud and biases the measurement of time elapsed between two reworking events; especially with a very low sedimentation rate.Grain size and shape, current velocity, and state of compaction of recovered hemipelagites, are the main parameters influencing this possible erosion.Nevertheless, in one of our previously analyzed case (Beck et al. 2007), a 210Pb decay curve and a 137Cs peak were detected below a silty/sandy, 2m-thick, homogenite, thus indicating a negligible erosion with respect to the hemipelagic sedimentation intervals (several centuries).Let's remind that, in coastal macrotidal regimes, few mm-thick clayey-silty layers (from suspended-load settling) are commonly preserved under up to 1.0 m.s-1 tidal currents carrying sandsize particles.There removing may occur but as mud-clasts; in the presently studied C2075 basin, we only observed mud-clasts of hemipelagic laminated sediments in the large pre-Late Glacial event (respectively 8 m-and 5 m-1thick on footwall and hanging wall , Beck et al., 2007).
Specific point 8 and 9: In order to propose an estimation of moment magnitude, we discuss the choice of a dominant displacement (normal vs. lateral).Published tectonic models differ for the general NAF pattern across the whole Sea of Marmara, but they agree with the occurrence of a lozenge-shaped pull apart structure for the Central Basin (Armijo et al., 2005; synthesis in G. Uçarcus's PhD).Thus, a normal component is emphasized, especially for the here-discussed SW limit.This point has been analyzed in detail by Armijo et al. (2005), based on a dense high resolution survey of the sea floor along the submerged trace, in the whole Sea of Marmara.Detailed deep seismic imagery (Becel et al., 2010;Laigle et al., 2008) indicate a strong subsiding component.For the (eastern) Çinarcic basin, Kurt et al ( 2013) defined an oblique slip accommodated along the northern limit.A direct observation of slickensides were done on the western end of the (western) Tekirdag basin on the deeply submerged prolongation of the Ganos fault (Armijo et al, 2005): approximately 15 • .Our attempt to estimate an inclination of the displacement vector remains hypothetical; nevertheless, based on above-mentioned available data, we locally favored a dominantly vertical displacement.In the new version, we test the influence of a 70 • pitch.We do not assess any conclusion concerning earthquake hazards; as previously modeled and discussed by Hébert et al. (2005), if admitting dominant vertical throw, tsunamis genesis represents the main possible risk.Our final estimation -a corollary approach with respect to the sedimentological study -is an additional matter of testing the values of throw (realistic or not).As the sedimentation completely "sealing" the subaqueous scarp (see also Beck et al., 2012) appear composed of well-defined separated unique events, we concluded that each of the later ones correspond to a unique scarp genesis.Armijo, R., et al., 2005.Geochemistry Geophysics Geosystems 6, Q06009, 29, doi:10.1029/2004GC000896.Bécel, A., at al., 2010.Tectonophysics, 490, 1-14.

B) Following Referee 2 (anonymous).
Response to general comments According to chronology, Radiocarbon dating of the three MD core is not dense, and after preliminary measurements (Beck et al., 2007).For our work, only few additional dating were attempted for the Çinarcik Basin.Detailed chronology (Radiocarbon and short-live radionuclids) were conducted by other colleagues on the last millennia (short cores, Drab et al., 2012) and the last centuries (Uçarcus's, PhD).We basically focused on the pre-Holocene part to establish eventby-event correlation in the Central Basin; for that we needed a well-defined correlation level at the Non Marine/Marine transition, independent from any earthquake/landslidetriggered sedimentary event.The chosen one corresponds to these criteria and was also recognized in Cinarcik Basin's long core, indicated a general, widespread, paleoenvironmental change (chemical, mineralogical, and biological modification).We believe that additional attempt for the Radiocarobon dating of the envisaged interval should not bring more chronological precision, because of: error bar, lack of wellconstrained reservoir effect (general and local ï Ą Ďr), and dependence on choice of C-support (debris reworked within turbidite, dispersed organic particles, shells).Conversely, for our purpose, lateral correlations established with different parameters (MS, VHR profiles, and layering) to discuss the differences between deep part and top of the southwestern scarp.Concerning what can be considered as co-seismic sedimentation -which is actually tsunami-generated sedimentation -its characterization has been developed by different teams since several decades and references (although not exhaustive) are mentioned, including our own published contributions.A complete session was dedicated to this topic during the last International Sedimentological Congress (ISC-IAS) held in Geneva (16-22 August 2014).In our draft, the final corollary results are not mentioned in terms of seismic hazards.We do not assess any C2077 conclusion in this sense; as previously modeled and discussed by Hébert et al. (2005), if admitting dominantly vertical throw, tsunamis genesis represents the main potential risk.Our final estimation -a corollary approach with respect to the sedimentological study -is an additional matter of testing the values of throw (realistic or not).Ongoing work concerns the possible genesis of tsunamis by the measured offsets.
Specific point 1: The pre-Holocene period were chosen: 1) because of better-defined tsunami-related deposits; 2) a situation during which, as for other areas of the Sea of Marmara or in other studied cases, active fault traces -with clear vertical componenthave no influence on bottom morphology because of systematic sealing by differential sediment redeposition on depressed areas (including hanging walls).During MAR-MARASCARPS Cruise, directed by R. Armijo and J. Malavielle, shorter Küllenberg type cores, dedicated to the last millennia, were retrieved from the different basins.Beside, a set of 100 short cores (35 cm) were visually picked using VICTOR R.O.V., preserving sediment/water interface.We participated to these two parts of the survey.The processing of these two sets of core were included in L. Drab's and G. Uçarcus' PhD theses.Drab et al. (2012).Polonia et al. ( 2004) and, more recently, McHugh et al. (2013) proposed such inter-and intra-basin correlations, and deduced correlation with historical ruptures which occurred along main strike-slip or oblique slip (not on Central Basin limits).The attribution of sedimentary events to historical ruptures are made through chronological correlations (see also Chapron et al., 1999;Pouderoux et al, 2012;etc.).At the difference our purpose (see Beck at al., 2012) is to directly "connect" an active trace activity to sediments deposited along this trace.Specific point 2: A core information remains punctual, and the Homogenite-Turbidite events are basically caused by reflected tsunami/seiche effects reworking one or several fluidized subaqueous landslides.First, the later ones may occur without seismic shock; second, even if identified as earthquake-induced, the HmTu don't necessarily reveal the location of responsible ruptures.With respect to the relative positions of the two cores, the 29 site is in the center of the basin, and not at the exact foot of the scarp, C2078 but, if following on VHR image the thickest homogenite (pre-LG), it changes appears abruptly close to it, and not progressively (as on the Northeastern side.Although the seafloor morphology displays a unique scarp, the subsurface structure is more complex due to probable clastic injections.Our major argument is the fact that the strong thickness increase of each separated depositional event is systematically related to its position in the deepest part.Specific point 3: I agree with a possible erosion below a density flow, which may eliminate some hemipelagic mud and biases the measurement of time elapsed between two reworking events; especially with a very low sedimentation rate.Grain size and shape, current velocity, and state of compaction of recovered hemipelagites, are the main parameters influencing this possible erosion.Nevertheless, in one of our previously analyzed case (Beck et al. 2007), a 210Pb decay curve and a 137Cs peak were detected below a silty/sandy, 2m-thick, homogenite, thus indicating a negligible erosion with respect to the hemipelagic sedimentation intervals (several centuries).Let's remind that, in coastal macrotidal regimes, few mm-thick clayey-silty layers (from suspended-load settling) are commonly preserved under up to 1.0 m.s-1 tidal currents carrying sandsize particles.There removing may occur but as mud-clasts; in the presently studied basin, we only observed mud-clasts of hemipelagic laminated sediments in the large pre-Late Glacial event (respectively 8 m-and 5 m-1thick on footwall and hanging wall , Beck et al., 2007).Specific point 4: Please, see response to this point in the first paragraph (Response to general comments) Specific point 5: I propose the same response as for Referee 1's Points 8 and 9.In order to propose an estimation of moment magnitude, we discuss the choice of a dominant displacement (normal vs. lateral).Published tectonic models differ for the general NAF pattern across the whole Sea of Marmara, but they agree with the occurrence of a lozenge-shaped pull apart structure for the Central Basin (Armijo et al., 2005; synthesis in G. Uçarcus's PhD).Thus, a normal component is emphasized, especially C2079 for the here-discussed SW limit.Detailed deep imagery (Becel et al., 2010;Laigle et al., 2008) indicate a strong subsiding component.For the (eastern) Çinarcic basin, Kurt et al ( 2013) defined an oblique slip accommodated along the northern limit.A direct observation of slickensides were done on the western end of the (western) Tekirdag basin on the deeply submerged prolongation of the Ganos fault (Armijo et al, 2005): approximately 15 • .Our attempt to estimate an inclination of the displacement vector remains hypothetical; nevertheless, based on above-mentioned available data, we locally favored a dominantly vertical displacement.In the new version, we test the influence of a 70 • pitch.We do not assess any conclusion concerning earthquake hazards; as previously modeled and discussed by Hébert et al. (2005), if admitting dominant vertical throw, tsunamis genesis represents the main possible risk.Our final estimation -a corollary approach with respect to the sedimentological study -is an additional matter of testing the values of throw (realistic or not).As the sedimentation completely "sealing" the subaqueous scarp (see also Beck et al., 2012) appear composed of well-defined separated unique events, we concluded that each of the later ones correspond to a unique scarp genesis.C) Following Interactive comment (Christopher Sorlien).The proposed Mw estimation is a corollary of the main conlusions concerning a very small part of the NAF northern trace; this attempt was developed in order to check if our conclusions about the vertical components of successive offsets were realistic/reasonable or not.We are not discussing the general strike slip-related activity and associated hazards.We do not address any results concerning the whole Sea of Marmara portion of the NAF.Our conclusion in terms of hazards concern the genesis of tsunamis, not regarding any paleo-magnitude.Published tectonic models agree with a pull apart significance of the Central Basin's inner basin.Its SW limit (8 km), which is clearly displayed by detailed morphology/bathymetry (see Armijo et al., 2005), should be submitted to tension and the corresponding throw should be dominantly normal.This has been analyzed and discussed by Armijo et al. (2005), based on a high resolution multibeam survey (see also Uçarcus' PhD, 2010).I agree with the fact that we could use a non vertical pitch C2080 (rather 70 • , which is proposed in the new version).Concerning the approx.15 • -rake (observed onboard R/V Atalante using Victor ROV; Armijo et al., 2005, Fig. 9), we discarded their availability for the small studied portion, for two reasons -and we rewrote this part more clearly -, because: 1) they probably concern a deep submerged prolongation of the Ganos Fault (Armijo et al., 2005) (far from the studied zone) and the vertical component is reverse , 2) it seems difficult to apply this kinematic only along the small analyzed portion.

C2081
Proposed responses for Manuscript NHESSD 2 C1455-C-1458, 2014 (modifications and corrections appearing in new version take all in account) A) Following Referee 1 (L.McNeill).I agree with the fact that the presented work is basically sedimentological with an implication for a co-seismic tectonic activity analysis.Both Title and Abstract are modified this sense.Concerning the seismic stratigraphy/core data links, they have been detailed in a previous published work (Beck et al., 2007).Nevertheless, main linksindicated on Fig. 4 will be underlined.

Specific point 1.
The general geodynamic/tectonic setting is shortened; conversely, the specificity of the Sea of Marmara section of the NAF will be commented, based on the synthesis done by G. Uçarcus (2012, PhD).

Specific point 2:
We focused on the Central Basin where we could obtain two giant piston cores (including LGM, Late Glacial, and Holocene periods) on both sides of a major active scarp; which was not the case for the other sub-basins.A first preliminary work indicated that the strong over-thickening in the deepest part (hanging wall) is entirely due to sedimentary events (homogenitesturbidites), being the "homogenite" term responsible for the major part of the difference.A new detailed analysis of these two cores were then planned, providing: 1) the discovery of specific sedimentary events sealing co-seismic scarps in the Lesser Antilles Arc (Beck et al, 2012), 2) the development of Anisotropy of Magnetic Susceptibility along-core profiles (Campos et al., 2013).The study of a long core from the Çinarcik Basin was added (for chronological as well as for paleo-environmental purposes) in order to ensure the regional occurrence of a welldefined "reference" layer (isochron for correlations) at the non marine/marine transition.Both VHR seismic image and cores data indicate that, before this pre-Holocene (non marine) level, the sea-floor stayed flat, due to the complete "erasing" of the fault scarp by reworked sediments depositions (turbidites-homogenites).

Specific point 3:
During MARMARASCARPS Cruise, directed by R. Armijo and J. Malavielle, shorter Küllenberg type cores, dedicated to the last millennia, were retrieved from the different basins.Beside, a set of 100 short cores (35 cm) were visually picked using VICTOR R.O.V., preserving sediment/water interface.We participated to these two parts of the survey.The processing of the two sets of core were included in L. Drab's and G. Uçarcus' PhD theses.

Specific point 4:
We separate a mean total sedimentation rate (including "events") from hemipelagic sedimentation rate.The second one is used to estimate the time elapsed between two successive instantaneous events.

Specific point 5:
For nowadays marine basins with shallow connections with open sea (e.g.Corinth or Cariaco Gulfs), the problem of reservoir age for 14 C is complex and depending on specific site conditions.For the Sea of Marmara, water circulation underwent several changes since 25 kyr BP due to connections changes (Black Sea + Aegean Sea, only Black Sea; see Vidal et al., 2010).Reservoir effects on CO2 are depending: i) on Black Sea and Aegean Sea water characteristics, ii) on their mixing within the Marmara basin, iii) local drainage.In addition, methane seepage may induce the incorporation of dead C into deep water metabolic chains.