T13C-1369 1340h
Character, Provenance, and Recurrence Intervals of Holocene Turbidites in the Kushiro Submarine Canyon, Eastern Hokkaido Forearc, Japan
Eastern Hokkaido forearc, Japan, is located on the Pacific side of the country along the Kuril Trench where large earthquakes and tsunamis have frequently occurred. The earthquake history in this region during the last few centuries is known, based on the historical literature. In order to examine prehistoric earthquakes, geological approach, such as analysis of seismo-turbidites or tusnami deposits, are effective. Two gravity cores (GH03-1033 and GH03-1034) were collected from the bottom of the Kushiro Submarine Canyon, offshore of Kushiro, eastern Hokkaido forearc. The cores contain evidence of several turbidites. Sedimentological, geochemical, and micropaleontological data and high-resolution seismic data have been used to identify character, provenance, and recurrence intervals of the turbidites. Upstream ($<$ 1000 m of the water depth), the Kushiro Submarine Canyon is presumed not to be a pathway for turbidity currents during Holocene, because an accumulation of thick mud lies in the channel. Tephras of Ta-a (AD1739), Ko-c2 (AD1694), Ta-b (AD1667), and Us-b (AD1663) were identified, and used to define the recurrence ages of the turbidite depositions. The provenance of the turbidites in core GH03-1034 collected from the middle part of the canyon was inferred to be the northern slope of the canyon deeper than the shelf edge, on the basis of sand composition and benthic foraminiferal analysis. Strong seismic shocks would have caused repeated turbidity currents into the Kushiro Submarine Canyon. The determined turbidite recurrences (rate of 68 years) are similar to the historical (rate of 79.7 years) and estimated (rate of 77.4 years) recurrence intervals of the earthquakes along the Kuril Trench.
T13C-1370 1340h
A Possibility Of OSL-Dating As An Indictor Of Transportation History Of Turbidites
Turbidite is induced by various geologic events such as earthquake, flood, tsunami, and so on. Especially, seismo-turbidite (e.g. Inouchi et al., 1996) and flood induced turbidite (hyperpycnite: e.g. Mulder et al., 2003) occupy majority of reported turbidites, and are often used as indicators of earthquake and catastrophic river flow, respectively. These researchers revealed such turbidite origins correlating age of turbidites with historical records or taking account of topographic setting, and then attempted to discover inherent descriptive features of seismo-turbidite/hyperpycnite. However, we cannot distinguish one from another at a glance, now. It should be necessary to seek not only usual descriptive clues but also another independent clue for distinguishing seismo-turbidite and hyperpycnite. We attempted to utilize optically stimulated luminescence (OSL) dating method as a new distinguishing method to specify turbidite origins. An OSL signal intensity implies the amount of natural dose accumulated in mineral grains. One can estimate an OSL age dividing the natural accumulated dose by the annual dose. It is important that OSL signal is canceled by sunlight stimulation (bleaching). So, the OSL age generally implies the burial age of sediment after transportation. Hyperpycnal flow should flow directly from river mouth to deep marine environment. Whereas, collapse-induced turbidite, its majority is considered as seismo-turbidite, is formed as a result of collapse of sediments at shallower marine environments. On the assumption that mineral grains are bleached during discharge from land to marine environments through river with a certain probability, some grains of a hyperpycnite should show OSL ages as much as the hyperpycnite depositional age estimated from radiocarbon age of planktonic foraminifers in hemipelagite. On the other hand, OSL age of seismo-turbidite grains must be obviously older than depositional age of the turbidite. Preliminary OSL experiments using KT98-18-P1 and -P2 cores sampled from the off Tokai, central Japan are carried out. Potassium feldspar (K-F) grains with 0.25-0.18 mm in diameter are extracted from five turbidite layers for OSL measurement and the single grain dose distribution of each layer is obtained. Dose distributions of K-F grains sampled from the present flood-induced fluvial deposits are also measured as references. We show detail of the methods and results of the preliminary OSL dating.
T13C-1371 1340h
Pinpoint core sampling at active faults in the Nankai subduction zone by new ROV "NSS"
Piston coring across active faults in shallow water depth, as well as drilling or trenching of fault zone on land, provides us with information about earthquake timing, recurrence of fault activity and fault displacement. In contrast, it is not easy to take samples from deep-sea with pinpoint accuracy. In order to identify the position of the core sampler and approach the target for sampling, a transponder is recently installed on the cable just above the sampler. The pilot vehicle of new ROV "NSS" (Navigable Sampling System) can move by four thrusters, observe seafloor by video cameras, release samplers and measurement tools by command from a mother ship, and real time geophysical or geochemical monitoring through an armored cable. The whole system including winches is removal. The ship with large deck can be equipped with NSS system. Depth capability of the pilot vehicle is 4000m and maximum payload weight is 1.5 Tons. First cruise using NSS was conducted in the Nankai Trough and the Sagami Trough using JAMSTEC R/V "Kaiyo" in 2003. Piston cores were taken from active faults off the Central Japan in the Nankai Trough, and the cold seep sites along the blind fault off Hatsushima, Sagami Bay. NSS using seafloor video camera and thrusters has accuracy of pinpoint sampling within 2m, and allows us to take samples from narrow fault zone or small clam colony. Core samples derived from active fault zones show different sedimentation histories; lack of strata and differences of layer thickness across dip-slip faults and lithological changes with time across strike-slip faults near submarine canyon, and provide information about recurrence time and fault displacement. NSS is one of tools that break through difficulties of pinpoint sampling in deep-sea and will proceed studies of active submarine faults.
http://ofgs.ori.u-tokyo.ac.jp/~ashi/NSS/
T13C-1372 1340h
Twenty-two outsize tsunami deposits at Lake Harutori-ko along the southern Kuril Trench
Along the southern Kuril Trench, which faces the Japanese island of Hokkaido, this fast subduction generated recurrent earthquakes up to magnitude ~8 in two centuries of written history. Eastern Hokkaido's largest well-documented interplate earthquake, the Tokachi-oki earthquake of moment magnitude (Mw) 8.1 in 1952, shook much of northeast Japan and generated tsunami waves 1-4 m high along the Hokkaido coast. Here we use deposits of prehistoric tsunamis in the lake bottom to infer the infrequent occurrence of larger earthquakes. The long stratigraphic record of lacustrine cores at Harutori-ko provides evidence for unusually large tsunamis in the past 9500 years. This lake or lagoon, which contains anoxic bottom water, is separated from the sea by a 5-m-high beach berm that the 1952 tsunami did not cross since 2500-3000 yrBP. Deposits beneath the lake bottom of Harutori-ko consist of diatomaceous laminated mud that alternates with sandy beds 0.1-1.0 m thick. A core 1 km from the modern beach contains 22 such beds. All of these beds grade upward from shell-bearing sand with gravel, through mud-clast breccia and laminated silt and sand, into organic mud. Stratigraphic series of these sand sheets, intercalated with volcanic-ash layers from Komagatake and Tarumai volcano, south western Hokkaido, and radiocarbon dating show that unusually large tsunamis occurred every ca. 430 years on average in the past 9500 years, most recently about 350 years ago in Edo era.
T13C-1373 1340h
Variability Among Tsunami Sources in the 17th-21st Centuries Along the Southern Kuril Trench
Instrumental, historical, and geological records of tsunamis show that successive plate-boundary ruptures differ in size along the southern Kuril trench off eastern Hokkaido. Tsunami source area of the 2003 Tokachi-oki earthquake (M 8.0), the most recent and the best-measured earthquake, is only about 2/3 of that of the predecessor, the 1952 Tokachi-oki earthquake (M 8.2). This difference is apparent from tsunami waveform inversions of the two events. The inversion of the 1952 event, redone with the clock corrections estimated from comparison of the 1952 and 2003 tsunami waveforms, confirms that the 1952 tsunami source extended about 100 km to the east of the 2003 source. The coastal tsunami runup heights were also different; the maximum height in 1952 was recorded by more than 100 km east of that in 2003. An earthquake in 1843 may have resembled the 1952 event, based on tsunami damage distribution recorded in historical documents. For the earthquakes in the 19th century, macroseismic data, i.e., distribution of tsunami heights and seismic intensity, are the only available data to estimate the source areas. As in 1952, the largest tsunami (4-6 m) was estimated around Akkeshi, where the 2003 tsunami was about 2 m. However, historical documents for 1843 are limited around the population center of that time, and no information is available from the Toakchi coast where the tsunami was highest in 2003. Prehistoric tsunami deposits have shown that larger tsunamis occurred in the eastern Hokkaido in an approximately 500 year interval with the last event in the 17th century. In Kiritappu, prehistoric sand sheets extend as much as 3 km inland across a beach-ridge plain. The 1952 tsunami penetrated about 1 km from the coast and the 2003 tsunami did not penetrate at all. Diatom analysis indicates that the sand sheets are of marine origin. Volcanic ash layers aid in the correlation and dating of sand sheets. The youngest of the prehistoric sand sheets shortly predates ash layers from 1663 (Usu volcano) and 1667 (Tarumai). A second, earlier sand sheet postdates a 10th century ash that erupted from Baitoushan on the border of China and North Korea. Three to five sand sheets are commonly found between the 10th century ash and a Tarumai ash about 2500 years old. These deposits are best explained by earthquakes that broke not only the area of the 1952 event but also the adjoining Nemuro-oki segment to the east. The above evidence for variable rupture mode complicates the task of forecasting future earthquakes and tsunamis in eastern Hokkaido. According to a long-term forecast that was issued six months before the 2003 earthquake, probability of an M~8 earthquake was 60 % by 2033. The forecast was correct for the timing but overestimated the earthquake size.
T13C-1374 1340h
Characteristics of deep-sea turbidites by the 1994 Hokkaido-nansei-oki earthquake
To understand the recurrence intervals of earthquakes in a peculiar area, it is very important to describe the phenomena at the sea-bottom occurring during an earthquake event. Based on surface ship and submersible surveys, submarine slope failures were a important phenomena on sea-bottom around the rupture area of the 1994 Hokkaido-nansei-oki earthquake. Earthquake-triggered turbidites were observed in and near the rapture area. The turbidite had sharp basal contact with underlay brown hemipelagic mud, and showed clear fining-upward grading structure from lower fine sand to upper clayey silt and parallel and cross lamination. No surface oxidized layer was observed, indicating very recent deposits. Benthic foraminifera in turbidite sands suggested that the sand was shallow water origin such as shelf to upper slope. Spatial distribution of turbidite thickness and grain size in the Shiribeshi Trough suggested major contribution from the southeastern steep slope with some minor contribution through submarine canyons opening to the northeastern trough floor. Such estimation was supported by grain fabric analysis using anisotropy of magnetic susceptibility. The similar turbidite sand layers were found in a sediment core collected from central Shiribeshi Trough. Radiocarbon age determinations and tephrochronology suggested that the averaged recurrence interval of turbidites was inferred as 1000-1500 years. The recurrence interval is almost the same as those of most active on land faults along the Japan Sea coast indicating higher seismic activity along the plate convergent margin at the eastern margin of Japan Sea.
T13C-1375 1340h
Turbiditic Sequences in Sediment Cores from the Continental Margin of Southern Chile as a Potential Record of Seismic Activity
The tectonically active continental margin of Chile offers an excellent opportunity to study the relationship between local seismic activity and mass movements in the marine sedimentary record. The historical earthquake record of the past 500 years shows that even very strong events of magnitude $>$ 7 (Richter scale) have been fairly common. Seismic events lead to remobilisation of sediments triggering downslope turbidites. Several deeply incised canyons channel most of the sediment into the trench, where the sediment sequence documents a regional record of events. Some intra-slope basins also bear a high potential for the preservation of a turbidite record of local significance. Sediment cores taken from different settings along the continental margin were analyzed for their succession of turbiditic layers. Age-dating - presently in progress - will be based on $\delta$$^{18}$O-; AMS$^{14}$C-, $^{210}$Pb-analyses in all cores. First results are: During the Pleistocene sedimentation rates were high and turbidite layers more frequent due to strong erosion in the hinterland associated with glacial activity and a colder and more humid climate. Even sediment cores in elevated and distant positions of the trench contain turbiditic-layers, although age-datings show that these layers are older than 10ka. An intra-slope basin near the foot of the slope gives a record of the Holocene turbiditic-sequences which contains 6 to 12 strong events. Further detailed age determination of the sequences - presently in progress - and comparison with the historical earthquake record promise an extension of the seismic record into the past beyond the last 500 years.
T13C-1376 1340h
Resolving Normal Fault Growth and Slip Rates Through Time Using Sequence Stratigraphy; Gulf of Corinth, Greece.
The Gulf of Corinth is the most actively extending continental rift zone in Europe. The Aigion Fault, in the western gulf, has 3 main onshore and 3 offshore segments, each of $\sim$1-2km length. It is a young north-dipping fault, and is active concurrently with, and overlaps, the Western Eliki Fault to its south. Detailed swath bathymetry of the eastern fault tip shows that the offshore segments form the southern boundary to a half graben, with multiple synthetic and antithetic splays associated with segment tips. The fault trace appears to propagate beyond the shelf edge. 3D analysis of sequence stratigraphy, using high-resolution marine seismic reflection data, has enabled quantification of fault growth spatially and temporally. Sequence stratigraphic markers relating to the post-lowstand transgressive surface of $\sim$11ka, and the final progradational fan delta topsets of 60-80ka form the boundaries between two distinct packages, enabling quantification of fault slip rates. The maximum observed vertical displacement rate at the fault tip is $\sim$1.2 $\pm$ 0.2mm yr$^{-1}$, averaged over the last 11ka. The displacement rate is seen to increase between 80-11ka and 11ka-present. This may be due to fault propagation and the interaction and overlap between the Aigion and Eliki systems. More detailed analysis will enable determination of displacement rate variations on timescales of thousands of years. Combining these data with onshore geomorphological studies, such as marine terrace uplift, allows correlation of activity across the broader fault system over a similar timescale. This study enables analysis of earthquake behaviour on this fault system over multiple cycles, highlighting evidence for possible clustering of events, and short-term variation in slip rate. It shows that high-resolution seismic imaging taken across submarine fault zones can provide significant additional information that contributes to the understanding of seismological activity and fault growth on normal faults in a continental rift setting.
T13C-1377 1340h
Active Normal Faults on the Forearc Slope of the Eastern Nankai Subduction Zone
Many low angle thrusts run parallel to Nankai Trough on the lower slope of the eastern Nankai subduction zone, whereas normal faults have been developed on the shelf to upper shelf slope (Yamaji et al., 2003). "Central Uplift" composed of a row of 5-10 km long anticlines trending northeast to southwest divides the thrust zone in the lower slope and normal fault zone in the upper slope. The uplift has been inferred to have developed along the Enshu fault which has right-lateral transpressional movement. The normal faults are distributed in the area about 70 km long and 15 km wide on the outer shelf to shelf slope. The faults dip {70-80 $^{\circ}$} north, and have maximum downthrown offsets up to 0.2 second in two-way travel time. On the hanging wall, wedge shaped growth strata fill the depression. Seismic reflection and calcareous nanno fossil age indicate that the normal faults began to be active during 0.91-0.45 Ma (middle Pleistocene). Some faults dislocate the shelf surface that was formed by the sea level fall in the last glacial period, suggesting that the normal faults are still active and have a potential of earthquakes. It is important to study palepseismicity of the normal faults, because they are located near coast. We conducted detailed mapping and sampling from seven sites across one of the normal faults using Navigable Sampling System (NSS) during the R/V Kaiyo (KY03-11) cruise. Seabeam mapping clearly showed a trace of the fault as a linear step of seafloor about 10 m high. NSS has two video cameras which provide seafloor image to operators on line, so we could select a suitable place to take samples. We towed NSS across the fault observing seafloor and collected sediment cores from 5 sites on hanging wall and 2 sites on foot wall. We identified the steep slope on the image, but observed similar seafloor on the both sides of the fault and could not find evidence for cold seepage. Approximately two meters long cores were sampled on the both sides of the fault. One of the core collected from the footwall is composed of semi-consolidated sandy silt overlain by sandy silt. The boundary between the lithologic change indicates the unconformity that is characterized by cobbles encrusted by algae. The semi-consolidated sandy silt contains 0.91-0.45 Ma calcareous nanno fossil, suggesting that the sediments can be correlated to the base of growth strata. On the other hand, the core of hanging wall of the fault composed of alternation of silt and sandy silt. We tried to correlate the sandy silt above the unconformity on the footwall to the sediments on the hanging wall, but could not find key horizon. The average sedimentation rate of the hanging wall is estimated to be about 10 cm/kys, and we infer that slip rate of the fault is larger than the sedimentation rate. At this time, we could not estimate the paleosismicity of the fault, but we will mention further study on cruise of this September using NSS and discuss the potential of NSS for the study of paleoseismicity of offshore active faults.
T13C-1378 1340h
Paleoecological Insights Into Subduction Zone Earthquake Occurrence, North Island, New Zealand.
The boundary between the Pacific and Australian tectonic plates at the Hikurangi Subduction Zone, offshore of New Zealand's east coast, has not ruptured in a large earthquake in historical times. Therefore paleoseismological studies are required to elucidate the nature and history of subduction interface rupture. Multi-proxy reconstructions of paleoenvironments at two coastal sites arcward of the Hikurangi Trough provide evidence for significant Holocene subsidence. Radiocarbon dating, tephrochronology, stratigraphy and analysis of foraminiferal, diatom, spore and pollen assemblages enable assessment of vertical movement of the coastal plain relative to sea-level. Sequences outline a 9500-year history of coastal aggradation and progressive isolation from the sea. However the position of paleo-sea-level indicators almost 6 m below present mean sea level implies that subsidence has taken place since 7200 cal. years BP. A proportion of this subsidence appears to have happened suddenly in association with high-energy marine influx events and rapid infilling of newly created accommodation space. Two inferred coseismic subsidence events at c. 7100 and 5500 cal. years BP can be correlated with previously identified coseismic subsidence and tsunami inundation at a site 10 km away along the coast. Forward elastic dislocation models indicate the observed vertical deformation could occur in Mw 7.9 earthquakes on the subduction interface. However subsidence has not been recovered interseismically at these localities, and flights of raised marine terraces further trenchward suggest a large amount of permanent deformation is occurring along the margin. Reconstruction of earthquake records at additional sites is required to investigate the interplay of rupture on upper plate faults versus the subduction interface.
T13C-1379 1340h
Estimation of Timing and Coseismic Seafloor Displacement along Enshu Fault, off Tokai, Central Japan
Enshu Fault is an active submarine fault, running over 120 km long, from off Hamaoka, Shizuoka, to off Shima Peninsula, Mie, Japan. General geomorphologic feature of the fault is reconstructed based on high resolution bathymetry and seismic profiles, to date the Enshu fault is known to be a transpressional fault to be a right-lateral slip component with a landward uplift component, parallel to the Nankai Trough (Tokuyama et al., 1999). The Enashu fault runs along the escarpment but the several submarine channels dissect the escarpment to cross the fault. The seepage driven, white-colored sea floor mat had been directly observed by submersible Shinkai 2000 (No.1601 cruise), right above the fault. Taking new ROV (NSS), the piston cores to be precisely positioned were obtained from both between the hanging- and foot-wall sides of the fault. The both core samples are consisted of thin-bedded turbidites. The result of bulk magnetic susceptibilities analysis enables to make the core-to-core correlation whereas the correlation based on the sand layer intercalated was not succeeded. In addition, the comparative study of the both cores in AMS fabrics and CAT image help us to identify the internal structure of the fault-displacement induced chaotic deposit to fill the depression in the hanging wall side. 14C age method taking foraminifer test in the cores tells us the accumulation time of the deposit due to the earthquake event. These data sets suggest the time when the most recent earthquake took place, and thickness of the deposit suggest us the amount of sea floor displacement in the time. In this study case, for example, the vertical displacement in the last single event reaches more than 40 cm. We calculated magnitude of the tsunami and way of the propagation as one of basic data of the hazard map.
T13C-1380 1340h
Active Faulting, Surface Deformation and Subduction Earthquakes at Isla Santa Maria, South-Central Chile
The Santa Maria island (37S) is situated above the interplate seismic zone of the Nazca-South America convergent margin. The island is in the boundary zone between distinct seismotectonic sectors of the Chile margin: the Valdivia (46-37S) and Concepcion (38-35S) segments, defined based on the recurrent rupture zones of M$>$8 subduction earthquakes. We used offshore industry reflection lines and local network seismicity to identify active structures and decipher their control on localizing surface deformation during great subduction earthquakes. Their effect on the geomorphology and structural evolution of the island was studied through field work and a high-resolution DEM. 14C ages were used to calculate deformation rates. The island comprises two units: Holocene lowlands formed by a flight of up to 25 emerged strandlines between the present sea level and 18 m elevation, coseismically raised during subduction earthquakes; and an upper surface formed by Tertiary rocks unconformably overlain by 53 to 31-kyr coastal and 31 to 10-kyr-old eolian deposits with paleosol horizons at elevations of 15 to 58 m. The transition from marine to continental depositional environments in the Pleistocene unit indicates complete emergence at 31 kyr BP, when sea-level was 82 m lower than at present, inferring an uplift rate of 3.6$\pm$0.5 m/kyr. The differences between the topography of the base of the near-shore unit, eolian unit, and present-day surface indicate progressive eastward tilting of 0.1%/kyr and synkinematic sedimentation. These surfaces are asymmetric, in the north they are oriented NE and dip SE, while in the south NW/NE. The Holocene strandlines yield a maximum uplift rate of 3 m/kyr, assuming they are preserved since the end of sea-level rise 6 kyr ago. Local network seismicity clusters 7 km NE of the island. Across an E-W transect the earthquakes define a continuous W-dipping zone from 2 km depth to the plate interface at 18 km. Focal mechanisms are compatible with a NNE-striking, steep W-dipping reverse fault. An E-W oriented reflection line across the cluster reveals W-dipping blind reverse faults, fault-propagation folding, and a piggy-back basin documenting eastward propagation of the deformation. This section accounts for $\sim$8% of ongoing shortening since the Pliocene. The structure has a ramp-flat-ramp geometry, probably rooted at the plate interface. Seismic lines SE of the island show N-vergent folding above a WNW-oriented reverse fault. Both the WNW- and NNE-oriented fault/fold systems converge at the center of the island. We therefore conclude that these two convergent systems control the asymmetry of the island. The progressive asymmetric tilting deduced from the base of the Pleistocene sequence, Holocene strandlines, and present-day surface can thus be attributed to lateral propagation of two fold systems that converge at the island. The emergent strandlines and the regional distribution of historic coseismic uplift demonstrate that great subduction earthquakes in this environment trigger crustal-scale reverse faults, which in turn control the spatial evolution of surface deformation. A common assumption, when applying dislocation models to invert surface deformation into slip on the plate interface thrust, is that the crust behaves as an elastic block that deforms without internal failure. Previous inversions yielded slip magnitudes that largely exceed the values expected from plate convergence velocity and earthquake recurrence, which might be explained by the discovery of active faulting below the coast.
T13C-1381 1340h
Reconstruction of Holocene Seismic Events Along Kamchatka's Pacific Coast
Since 1995 we have been conducting detailed neotectonic and paleoseismological field studies of the Holocene at more than 15 localities along the east coast of Kamchatka. Specific methods of these investigations include 1) application of tephra chronology and tephra stratigraphy for dating and correlation of various types of coastal deposits and landforms; 2) study of paleotsunami deposits in order to determine their ages and recurrence rate and to estimate the magnitude of large tsunamis and tsunamigenic earthquakes along the Kurile-Kamchatka subduction zone and north of the latter during Holocene time; 3) analysis of the geologic structure, age, and modern and paleo-topography of marine terraces and beach ridges in order to determine the direction and scale of seismotectonic movements over different time intervals; and 4) identification and dating of inferred seismogenic landslides along the coast. Eastern Kamchatka's short history (since about 1730 A.D.) includes many large and several great earthquakes, although historical coverage of these events is scarce because of both low population density, and also military secrecy. Nevertheless, we can use historical earthquakes and tsunami deposits as benchmarks for pre-historic events. Moreover, our studies have helped elucidate historic cases, including co-seismic deformation, tsunami runup, and tsunami source mechanism. As a result of these studies we have reconstructed key events in coastal evolution during the Holocene, and determined the ages of marine accumulative landforms on various segments of the coast. A combination of all our data enables us to get an idea about spatial and temporal distribution of strong subduction-type earthquakes along the Kuril-Kamchatka trench and 'non-subduction' earthquakes north of the trench terminus.
T13C-1382 1340h
Fresh Submarine Seismic Breaks due to Historical Thrust Earthquakes Offshore Lebanon
The M$>$7 earthquake that struck the shore of Lebanon in 551 AD was associated with a large tsunami that destroyed Beyrouth and other seaports, and thus likely originated offshore. Onshore in the north, the Tripoli thrust, whose 70m-high cumulative scarp cuts the city in half, is responsible for the Plio-Quaternary growth and current uplift of the Turbol-Qualhat anticline. Similarly, the growth of the Aabd\'{e} anticline, and the tilt and emergence of the Ramkine islands group result from slip on the Aabd\'{e} thrust. We have inferred both thrusts to continue at sea. The EM300 multibeam bathymetric data obtained during the SHALIMAR cruise establishes the existence of the active, submarine Mount Lebanon thrust system, which connects the Tripoli and Aabd\'{e} thrusts with the Roum fault. Between Beyrouth and Tripoli the Levantine margin shows its steepest near-shore bathymetric gradient (100 to 1500 m in $<$5 km). The base of this topographic front is bounded by slope breaks characteristic of emergent active faults. Up to 30 km from shore, the seafloor is uplifted and warped by anticlines. Flat-floored turbiditic channels fed by the deep river canyons onshore have been tilted and abandoned, or deviated by folding. Most of the anticlines appear to grow on blind ramps, and show spectacular, active extensional hinge faulting. The strikes of fold axes imply WNW-ESE shortening, consistent with the ongoing rise of Mount Lebanon and uplift of flights of coastal shorelines. The largest submarine shortening faces the steepest relief, where the flexed western limb of the Mt Lebanon mega-anticline tangents the coastline. This is where the highest marine cut terraces are observed. South of Saida, and north of Tripoli offshore the Akkar plain, by contrast, the overall topography of the continental edge is typical of passive margins with little evidence of active uplift. Profiles stretching as far as the Latakia ridge confirm that extensions of the Roum fault do not continue to Cyprus, but act as lateral ramps of the offshore Mt Lebanon fold and thrust belt. Deep-towed SAR side-scan sonar images confirm that the most prominent bathymetric slope-breaks are cumulative, active fault scarps. "Fresh" seismic ruptures are found on 2-3 km-long segments of these scarps, 8 to 18 km from shore between Beyrouth and Batroun. The seismic scarplets face west and have sinuous, stepping traces, consistent with dip-slip faulting. They likely mark the surface trace of past coastal thrust earthquakes, among which the 551 AD event. The multiple vermetid constructed benches seen along this part of the coastline thus likely reflect incremental coseismic uplift ($<=$80 cm). Near bottom studies are in urgent need to date seismic events and better assess seismic hazard along this most populated region of Lebanon.
T13C-1383 1340h
Tsunami Events of the Nankai Earthquakes During the Last 4000 Years
The Nankai earthquakes are interplate earthquakes along the Nankai Trough and one of the most historically documented earthquakes in the world. Nine Nankai earthquakes have been identified from historical documents after the Hakuho-Nankai earthquake at A.D.687 (Usami, 1996; Tuji, 1999). The purpose of this study is to reveal pre-historical occurrences of the Nankai earthquakes by using geological evidence. A small pond named Tadasuga-ike located in the southern Shikoku Island is selected for this study. The last Nankai earthquake of 1946, Tadasuga-ike subsided coseismically and was covered by seawater of tsunami. We collected six 5m long sediment cores from the pond and identified fourteen layers of tsunami sediments. Radiocarbon datings indicated these layers deposited during the time period between 1300 yBP and 3800 yBP. These results suggest the recurrence intervals of the Nankai earthquakes are 200 years in average rate during the last 4000 years.
T13C-1384 1340h
Fault-slip derived paleotsunami compared with historical descriptions: simulation of the 1596 Keicho-Bungo tsunami based on the paleoseismic study of submarine faults in Beppu Bay, Japan
High-resolution shallow-water profiling survey revealed recent activity of submarine fault which was newly found to fill a gap between two subaerial faults, the Asamigawa and the Funai faults running on the southern coast of Beppu Bay. Historical records suggest that this activity on the north-dipping normal fault corresponds to the M7.3 Keicho-Bungo earthquake of September 1, 1596. Our previous profiling survey and correlation study of core samples show that the event caused vertical offset on south-dipping normal faults in the central and northern parts of the bay. Thus it is most likely that the Keicho-Bungo event simultaneously ruptured the faults on the both sides of the Beppu Graben. Furthermore core samples obtained from the both sides of the newly found submarine fault suggest repeated occurrence of this simultaneous rupturing. A comparison of vertical offsets of two prominent ash layers on Beppu Bay faults whose total length exceeds 230km, shows almost linear relationship between the two offsets, indicating characteristic behavior of the fault slip distribution. This enables us to estimate co-seismic fault slip of the Keicho-Bungo earthquake on faults without core samples. On the basis of the estimated fault slip, vertical deformation of the ocean bottom, or the initial height of the tsunami was calculated. A comparison of the simulated tsunami height on the coast and historical description of the 1596 tsunami indicates that the Keicho-Bungo earthquake ruptured not only the Beppu Bay faults, but also the westernmost segment of the Median Tectonic Line active fault zone.
T13C-1385 1340h
Investigating Lateral Variations of Interseismic Strain along the Sumatran Subduction Zone
Paleoseismic, paleogeodetic and GPS data from the Sumatran subduction zone provide an unusual opportunity to understand the physical parameters that control the behavior of a subduction interface. Interseismic strains recorded over the last several decades by coral growth rings and GPS instruments allow us to model subduction zone behavior using a three-dimensional dislocation back-slip model. The kinematic model takes into account the partitioning of slip between the subducting Australian plate and Sunda into dominantly dip-slip motion along the subduction interface and strike-slip faulting along the Great Sumatran fault (GFZ). The sliver between the trench and the GFZ is assumed to behave, in the long term, as a rigid microplate. Given the northward secular motion of the Sumatra forearc, the normal convergence across the subduction zone is 40-46 mm/yr. Our modeling shows that the GPS and coral data are well fit by a simple model that assumes full locking of the subduction interface with lateral variations in the depth of the downdip end of the locked interface. The width of the locked interface is at a minimum value of ~135 km near the Equator and increases to about ~190 km farther south (This corresponds to a variation in depth from ~35 to 50 km). We note that over the past 250 years the subduction interface near the Equator has produced smaller earthquakes ( Mw = 7.7 in 1935 and 7.2 in 1984) than the area farther south, which has produced giant earthquakes in 1797 and 1833 (Mw = 8.4 to 8.7+). This difference in both the seismic behavior and width of the locked fault zone may be related to the structural irregularities associated with subduction of the Investigator Fracture Zone or to lateral variations in the age of the subducting plate and/or the rate of convergence. We investigate these possibilities by modeling the effect of thermal structure on the behavior of the subduction zone.
T13C-1386 1340h
Crustal vertical motions from paleogeodetic data of the Sumatran subduction zone, 1950 to 2003: Steady vs. episodic strain accumulation.
We determined submergence/uplift rates from 42 paleogeodetic sites along the 400-km length of the fore-arc islands. Our analysis is based on data of annual lowest low tide levels, which are accurately tracked by the growth patterns of coral microatolls over the past several decades. The annual oceanographic fluctuation is on the order of a few centimeters, and is statistically removed for determinations of average sea-level changes. The average sea-level changes, then, are corrected for the global eustatic sea-level rise to determine the crustal vertical rates. Records from the Batu islands around the Equator show that the sites from 85-km to 115-km northeast from the Sumatran trench are submerging slowly about 1 mm/yr to near zero, where sites closer to the mainland Sumatra are emerging up to 6 mm/yr. The Batu islands are located above the 1935 rupture patch (Mw 7.7). These data constraint an elastic model suggesting that this part of the subduction interface must be predominated by aseismic slips. The paleogeodetic time-series indicate that the subduction in the equatorial region is steady for the past 30 years. Further south from the Equator, the Sipora, North Pagai and South Pagai Islands, where the 1797 and the 1833 historical events (Mw 8.3 and Mw 8.7) had ruptured the interface beneath the islands; the average rates of submergence are generally higher than those of the Batu islands. The rates vary from about more than 10 mm/yr to 2 mm/yr at the sites 90 km to 120 km off the trench. Most of the paleogeodetic time-series of this region indicates that the submergence may be steady, but interrupted by decadal slip events. The Siberut Island, a 110-km length of the fore-arc ridge, located between the 1935 and the 1833 ruptures submerge at a maximum average rate of about 4.5 mm/yr at the middle of the eastern coast, about 110 km from the trench. The rates generally decrease toward the north and the south-end of the island, but are suddenly reverse to very high rates near the strait of the Siberut and Sipora. Numerous fossil microatolls reveal that the Siberut Island raised more than 1 meter during the pre- historical, early 1600s event. Paleogeodetic records of the Siberut sites indicate that the subduction is episodic. That is the long-term fast submergences were interrupted by years of stable periods. The transitions between these fast and near zero submergence are demarcated by (slip) events. One of these events is a major 1962 aseismic slip, which is well documented at numerous sites from the Equator to about 3.5>| S. We use these rich data of the upper crustal deformations to investigate the behaviors of the subduction zone in a period after the past major earthquakes and before the next one.
T13C-1387 1340h
Dive Surveys for Evaluation of Offshore Active Faults Along the Eastern Margin of Japan Sea
The eastern margin of Japan Sea is an incipient subduction zone between the Amurian and Okhotuk Plates and is characterized by widely distributed active reverse faults. During the 20th century, four major earthquakes (M>7.5) occurred along the margin. The recurrence intervals of the earthqaukes are (probably) longer than several hundreds years, longer than the coverage of historical records. Turbidites have been used as records of paleoseismicity in the margin, and they provided recurrence intervals of earthquakes along several reverse fault zones. Turbidites, however, are not always distributed around reverse faults. In addition, turbidites in a core may be records of earthquakes along different faults around basins. Seafloor disruptions were widely observed by manned submersibles "Shinkai 2000" and "Shinkai 6500" in the source area of the 1993 Hokkaido-Nansei-Oki earthquake (e.g. Takeuchi et al., J. Geophys. Res., 103, 24109-24125, 1998). The dive surveys suggested that earthquakes have been recorded as seafloor disruptions, thus we have been conducting dive surveys to clarify paleoseisimicity of the active faults from visual observation of seafloor. Based on the observation of 15 dives in the margin, we defined two types of slopes which show different disruptions. The first type of slopes is composed of consolidated and highly fractured rocks. The slope is sporadically covered by unstable debris of sand to boulder sizes, and the basin floor on the foot of the slope contains debris layers less than a few centimeters in hemipelagic muddy sediments. We interpreted that the layers are records of ancient earthquakes. The other type of slopes is generally composed of semi-consolidated muddy sediments, and debris is not widely observed on slopes and basin floors. The more widespread disruptions are fissures and small-scale slope failures. We can identify recent ruptures along the fault but hardly determine ages of ruptures on this type of faults. We conclude that observation of seafloor using submersibles is useful to judge offshore faults to be active or not, however, it is not always easy to determine paleoseismicity of faults, particularly for the age. We will discuss possibility and limit of dive surveys of offshore active faults.
T13C-1388 1340h
A Distribution Pattern of Tsunami Deposits in Inter-tidal Zone at Mochirippu, Eastern Hokkaido, Japan
We studied a distribution pattern of tsunami deposits beneath Lake Mochirippu, a lagoon fingering a few km from the coast with narrow entrance on the Pacific coast of eastern Hokkaido. Eastern Hokkaido has been suffered by tsunamis from Kuril subduction zone, and many tsunami deposits from historical and prehistoric tsunamis have been reported from coastal marshes and lagoons. However, detailed distributional pattern of the tsunami deposits has not been established. From over 100 cores in Mochirippu, we identified two prehistoric tsunami deposits. One is from 17th-century giant earthquake, and the other lower one is undated yet. Detailed sedimentological characteristics of the 17th-century tsunami deposit can be summarized as follows; The tsunami deposit can be subdivided into four units named Unit 1 to Unit 4 in ascending order. Unit 1 is composed of poorly sorted medium to fine sand containing many plant fragments and rip-up clasts, and erosionally overlies the peat or mud layers. Unit 2 is composed of well-sorted medium sand, and overlies Unit 1. This unit has coarsest sediments in the four units. Unit 3 gradationally overlies Unit 2, and is composed of fine alternation of plant fragment laminae and fine to very fine sand layers. Unit 4 gradationally overlies Unit 3, and is composed of sandy silt with plant fragment seams. This unit is gradationally covered by mud or peat. This tsunami deposit ranges several cm to 30 cm in thickness. This tsunami deposit extends at least about 2 km from the bay mouth, and shows landward decreases in the thickness.
T13C-1389 1340h
Quaternary Tectonic Tilting Governed by Rupture Segments Controls Surface Morphology and Drainage Evolution along the South-Central Coast of Chile
The Chilean coast represents one of the most active convergent margins in the Pacific rim, where major earthquakes (M>8) have repeatedly ruptured the surface, involving vertical offsets of several meters. Deformation along this coast takes place in large-scale, semi-independent seismotectonic segments with partially overlapping transient boundaries. They are possibly related to reactivated inherited crustal anisotropies; internal seismogenic deformation may be accommodated by structures that have developed during accretionary wedge evolution. Seismotectonic segmentation and the identification of large-scale rupture zones, however, are based on limited seismologic und geodetic observations over short timespans. In order to better define the long-term behavior and deformation rates of these segments and to survey the tectonic impact on the landscape on various temporal and spatial scales, we investigated the south-central coast of Chile (37-38S). There, two highly active, competing seismotectonic compartments influence the coastal and fluvial morphology. A rigorous analysis of the geomorphic features is a key for an assessment of the tectonic evolution during the Quaternary and beyond. We studied the N-S oriented Santa Maria Island (SMI), 20 km off the coast and only ~70km off the trench, in the transition between the two major Valdivia (46-37S) and Concepcion (38-35S) rupture segments. The SMI has been tectonically deformed throughout the Quaternary and comprises two tilt domains with two topographic highs in the north and south that are being tilted eastward. The low-lying and flat eastern part of the island is characterized by a set of emergent Holocene strandlines related to coseismic uplift. We measured detailed surface morphology of these strandlines and E-W traversing ephemeral stream channels with a laser-total station and used these data to calibrate and validate high-resolution, digital imagery. In addition, crucial geomorphic markers were dated by the radiocarbon and optical stimulation methods to better constrain deformation rates. In response to the ongoing deformation, formerly W flowing streams constituting small drainages (< 0.25km2) were inverted and formed closed basins. In contrast, larger streams were reversed or were able to maintain their channels, but formed distinct knickpoints along their longitudinal profiles. In order to reconstruct the Holocene tectonic tilting axis, we connected drainage boundaries of reversed channels and deformation-related knickpoints along more mature rivers. Interestingly, topography clearly indicates that the direction of Pleistocene tectonic tilting was different than that of recent conditions. The Holocene inversion of stream flow associated with continuous uplift may be related to the progressive migration of the tectonic tilting axis in the course of active folding (Melnick et al., this session). The classification of knickpoints and the overall tectonic development also the mainland coast on the Arauco peninsula, during the Quaternary clearly document the surface signature of tectonic segmentation and its spatial evolution through time. The migration of the tilting axes is discussed in relation with active basal accretion and active shortening in the South-Central Chilean forearc.
T13C-1390 1340h
Measurements of Active Tectonic Deformation on the Guerrero Coast, Mexico
The study of tectonic deformation rates using displaced shoreline features is relatively well-established, and has provided much useful information on seismic hazard. Such studies have frequently been complemented by analysis of the coastal sedimentary record, where past marine to terrestrial environmental changes (and vice versa) may be recorded by clear changes in stratigraphy. Studies of this type are particularly valuable for tectonically-active areas where the preservation of former shoreline features is poor, or where long-term subsidence has resulted in their erosion, drowning or burial. The specific objective of this study is to derive rates of tectonic deformation from geomorphic and stratigraphic studies of the Guerrero coastal area, and to examine the feasibility of this stratigraphic approach in the coastal lagoons of the Mexican Pacific coast, in the Guerrero gap. The Guerrero gap coastal area, where a major earthquake is expected to occur, parallels the Cocos plate subduction zone. Here convergence rates vary from 5.2 cm/yr to 5.8 cm/yr. The Guerrero gap has experienced several historical earthquakes, notably the 1911 (7.8 Ms). However, no large magnitude events since the 1911 earthquake and only a few Ms~6 events have occurred near the Guerrero gap edges. It is expected that a major interplate earthquake of estimated magnitude Mw=8.1 to 8.4 has a high probability to occur. Landforms within the Guerrero gap indicate that the coast is subsiding. A series of key indicators such as elongated islands reminiscent of ancient barriers, submerged barriers island, extensive marshy environments, increased depths in the lagoons, and submerged anthropogenic features (shell mounds), among others, suggest active tectonic subsidence of the coast. In contrast, the adjacent northwest area off the Guerrero gap exhibits landforms characteristic of tectonic uplift (marine terraces and uplifted beach ridges), indicating a different seismo-tectonic regime northwest of the gap. Core samples from nine study sites show distinct stratigraphic changes with depth, indicating clear rapid change in depositional environments over time. Preliminary 14C data indicate that the sediment sequences cored are up to ca. 4,626+/- 37 yr BP, and suggest an estimated subsidence rate of about 1mm/yr. Geochemical characterization of the cores collected show the potential for geochemical discrimination of terrestrial-marine and marine-terrestrial environmental changes. Preliminary detailed linking of the geochemistry to the core stratigraphies show downcore variations in sediment sequences that match sedimentary horizons considerably enriched in S, Cl, and Br concentrations, that indicate possible marine inwash events (e.g. tsunami).
T13C-1391 1340h
Structure of Mesozoic Oceanic Crust in the Vicinity of the Cape Verde Islands From Seismic Reflection Profiles
Multichannel seismic reflection profile data have been used to determine the internal structure of Mesozoic oceanic crust in the vicinity of the Cape Verde islands. The data show the oceanic crust to be characterized by both dipping and sub-horizontal reflectors. Several lines of evidence argue against the reflectors being scattering artifacts arising, for example, from rough basement topography. Instead, the reflectors are attributed to tectonic and magmatic processes associated with the accretion of oceanic crust at the mid-Atlantic ridge. The upper crust shows variable reflectivity due to both dipping and sub-horizontal events. We interpret the dipping reflectors, which have been identified on both ridge-normal and ridge-parallel profiles, as sub-surface expressions of normal faults that formed at or near the mid-Atlantic ridge. There is no evidence that the faults are caused by loading of the oceanic crust by either the Cape Verde islands or their associated topographic swell. Some faults, however, can be traced into the overlying sediments suggesting they may have been re-activated since their formation at the ridge. The origin of the sub-horizontal reflectors is not as clear. We believe them to be boundaries of different igneous lithologies, such as that between basalts and gabbros. The lower crust is highly reflective in some areas, whereas in others only a few dipping and sub-horizontal reflectors are observed. Some of the dipping reflectors can be traced into the upper crust suggesting they are also normal faults. Others, however, appear to be confined to the lower crust. The sub-horizontal, discontinuous, reflectors about 2.0-2.5 two-way travel time below the top of oceanic basement are attributed to the Moho.
T13C-1392 1340h
Probabilistic Fault Displacement Hazard Assessment For Pipelines, Mad Dog and Atlantis Field Developments, Deepwater Gulf of Mexico
Seafloor faults having strong geomorphic expression and evidence for late Quaternary activity (i.e. < ~150,000 years) are common geologic features associated with the Sigsbee Escarpment. Waterbottom maps derived from exploration 3D multichannel seismic data provided an early indication that several zones of seafloor faults are in the vicinity of the Mad Dog and Atlantis prospect areas. As part of the site investigation activities for field development BP initiated a study to characterize the potential hazard due to fault displacement. The fault displacement hazard study consists of five components: 1) a site-wide structural geologic characterization of the style and origin of active faulting and fault-related deformation; 2) development of a late Quaternary stratigraphic model to evaluate the history, recency, and rate of fault activity at the site; 3) detailed characterization of faulting within limited study areas designated to capture fault behavior in areas of potential facilities development; 4) a general description of the relationship between Quaternary active faulting and slope failure processes within the field area; and 5) a probabilistic fault displacement hazard analysis (PFDHA) of the potential for fault rupture within the designated study areas that relates annual frequency of recurrence of faulting events to the size of the event. Changes in the style and origin of faulting and deformation of shallow (suprasalt) sediment across the individual field areas primarily is due to differences in the depth, geometry, and movement history of the underlying Sigsbee Salt Nappe. These relationships and the resulting geologic model for structural evolution of the suprasalt section has been used effectively to assess the site-wide geohazards not only for faulting, but also indirectly for slope failure and mass-gravity flows. Hazard from potential seafloor offset at fault crossings is judged to be moderate to low. Fault offsets of the shallowest horizons (less than 15 thousand years old) are typically less than ten meters to several tens of meters. Fault slip rates are on the order of tenths to several tens of meters-per-thousand-years (m/kyr, also millimeters-per-year, or mm/yr), with most values in the range of 2-10 m/kyr. Similarly, the probabilistic annual recurrence of 1-meter events is typically less than 10-3. These studies demonstrate that the presence of potentially active faults does not preclude safe development of seafloor facilities. To evaluate risk associated with potential seafloor faulting, integrated hazard studies can and should be conducted in the early stages of project development, with an underlying intent to understand the causative processes and quantitatively and explicitly evaluate the locations, magnitude and recurrence potential of displacement events.
T13C-1393 1340h
Coastal Deformation Patterns Along the Nicoya Seismic Gap, Pacific Coast, Costa Rica, Central America
The Nicoya Peninsula, Costa Rica provides a unique setting for the study of upper plate deformation along the Middle America convergent margin. Located 60-70 km inboard of the trench axis, this outer fore arc peninsula sits directly above the seismogenic zone. A sequence of emergent marine terraces along the Nicoya coast records the net pattern of late Quaternary uplift associated with the subduction cycle. The last major earthquake centered beneath the peninsula (Mw=7.7, 1950) produced widespread damage and generated 0.5-1.0 m of coseismic uplift along the peninsula's coast. With a large slip deficit since 1950, the Nicoya Peninsula is recognized as a high-potential seismic gap. Ongoing field study of uplifted Quaternary shorelines provides an excellent opportunity to evaluate local deformation patterns and large earthquake repeat times. Recent mapping of late Pleistocene marine terraces along the peninsula's northern coast allows for comparison with those previously mapped along the peninsula's southern tip. The "Iguanazul surface" on the northern coast between Tamarindo and Nosara encompasses at least three separate wave-cut treads that preserve paleo-shorelines at 10-12 m, 18-22 m, and 26-32 m elevation. Preliminary correlations with late Pleistocene sea level high stands at 80-330 ka (OIS 5-9) indicate net uplift rates of 0.1 - 0.3 m/k.y. Holocene beach rock horizons along the active beach yield calibrated radiocarbon ages of 1610 and 760 ybp (Playa Negra and Playa Lagarto), consistent with net Holocene uplift at less than 0.5 m/k.y. In contrast, the "Cobano surface" at the peninsula's southern tip (Cabo Blanco) includes at least five distinct Pleistocene terrace treads separated by well-defined risers at 30-220 m elevation. Holocene uplift rates here range from 1.0 - 6.0 m/k.y. along an emergent Holocene terrace (Cabuya surface). The order-of-magnitude difference in Quaternary uplift rates between the northern and southern Nicoya Peninsula may be linked to sharp contrasts in the roughness, thickness, and dip of the subducting Cocos plate. While the northern peninsula overrides the subducting Cocos plate's "smooth domain", the peninsula's southern tip overrides seamounts of the "rough domain". These contrasting coastal segments may deform independently in response to rupture of different subduction zone asperities (Central Nicoya and Cabo Blanco). Based on historical seismicity data, estimated slip per event, and a rapid convergence rate (9 cm/yr), the recurrence interval for large Nicoya earthquakes is estimated at 50-100 years. While these frequent events may produce meter-scale coseismic uplift along the Nicoya coast, a significant fraction of this is likely recovered during interseismic subsidence. The net result is gradual coastal uplift during the Quaternary at rates ranging from 0.1 to 6.0 m/k.y.
T13C-1394 1340h
Investigation of a 12,000 Year Record of Cascadia Tsunamis and Coastal Subsidence in Oregon
Evidence of great plate-boundary earthquakes and accompanying coseismic subsidence events is revealed at numerous land-sea interfaces along the Pacific Northwest coast. We recently extracted a 27 meter-deep core at a site near the Sixes River on the southern Oregon coast, approximately two kilometers from the Pacific Ocean. A basal date of ~12,300 cal. yr B.P. implies the core has the potential to yield a continuous terrestrial stratigraphic record extending into the late Pleistocene. If correct, this site may provide the first onshore chronology of plate boundary earthquakes along the Cascadia margin on a timescale comparable in length to the turbidite records off shore. However, numerous variables complicate the interpretation of this long sedimentary record, perhaps more so than some of the younger terrestrial records of coastal paleoseismology. These variables include long-term tectonic movement of the landscape on local and regional scales; sedimentation from marine, riverine, and aeolian sources; base-level rise and fall due to tectonic activity and eustatic sea-level changes; and compaction and degradation of the sedimentary record over time. Landscape reconstruction of the coring locality through time, with reference to relative sea level, depositional status, and position of the site relative to the coastal shoreline, aids in the recognition and interpretation of paleoseismic indicators. These indicators include changes in bio-, litho-, and chrono-stratigraphy. This study will potentially expand the toolkit for the analysis of other Late Pleistocene, nearshore terrestrial records on the Cascadia subduction zone.
T13C-1395 1340h
Noncharacteristic Slip on the Northern San Andreas Fault at the Vedanta Marsh, Marin County, CA
Three-dimensional excavations along the 1906 trace of the northern San Andreas fault at the Vedanta marsh paleoseismic site near Olema, CA have yielded new data on the timing and amount of slip during the penultimate earthquake on this fault section. The excavations exposed a 3-m-wide paleochannel that has been offset right-laterally 7.8-8.3 m by coseismic slip during the past two large earthquakes: 1906 and the penultimate earthquake. The paleochannel was eroded into a silty clay marsh deposit and was filled after AD 1400. Both the silty clay layer and the paleochannel deposit are directly overlain by an in situ burn/peat sequence. The penultimate earthquake occurred while the peat was at the ground surface whereas faulting from the 1906 earthquake terminates within an overlying gravel/fill sequence. Preliminary OxCal analyses of radiocarbon dates indicate that the penultimate earthquake occurred in the late 17th to early 18th century. In plan view, two main fault traces were mapped in the excavation. The northwestern portion of the paleochannel is offset across a single fault trace. Just southeast of this portion of the channel the fault splits into two traces. We believe that one of these traces likely slipped only during 1906 and the other trace slipped on during the penultimate earthquake. Unfortunately, the overlying stratigraphic section that could resolve the exact reconstruction of movement on these faults is missing due to the excavation of an artificial drainage ditch at this location in the 1940's. Matching the north margin of the paleochannel to the first exposure of gravel in the zone between the two fault traces gives an offset of 5 m. We have historic records that show the 1906 coseismic slip near the study site was about 5m from field notes of David Starr Jordan (Stanford University Archives) who describes two 16 ft (5m) offsets: one of a tree located about 150m SE of the offset channel and the other of a path to the Shafter barn located about 300m NW. As the locations of these two historical records are so close to the study site, it is reasonable to assume that our excavation site has the same amount of coseismic slip in 1906. Our data indicate that the paleochannel was offset about 2.8 to 3.3 m during the penultimate earthquake which occurred in the late 17th to early 18th century, and that the San Andreas fault at this section is capable of slip in earthquakes smaller than 1906.
T13C-1396 1340h
Coastal Marine Terraces Define Late Quaternary Fault Activity and Deformation Within Northern East Bay Hills, San Francisco Bay Region
Detailed mapping of uplifted marine platforms bordering the Carquinez Strait between Benicia and Pinole, California, provides data on the pattern and rate of late Quaternary deformation across the northern East Bay Hills. Field mapping, interpretation of early 20th-century topographic data, analysis of aerial photography, and compilation of onshore borehole data show the presence of remnants of three platforms, with back-edge elevations of about 4 m, 12 m, and 18 m. Based on U-series dates (Helley et al., 1993) and comparison of platform elevations to published sea-level curves, the 12-m-high and 18-m-high platforms correlate with substage 5e (ca. 120 ka) and stage 9 (ca. 330 ka) sea-level high stands, respectively. West of the Southhampton fault, longitudinal profiles of platform back-edges suggest that the East Bay Hills between Pinole and Vallejo have undergone block uplift at a rate of 0.05 +/- 0.01 m/ka without substantial tilting or warping. With uncertainty of $<$3 m, the 120 ka and 330 ka platforms are at the same elevations across the NW-striking Franklin fault. This west-vergent reverse fault previously was interpreted to have had late Pleistocene activity and to accommodate crustal shortening in the East Bay Hills. Our data indicate an absence of vertical displacement across the Franklin fault within at least the past 120ka and perhaps 330ka. In contrast, the stage 5e and 9 have up-on-the-east vertical displacement and gentle westward tilting across the N-striking Southhampton fault, with a late Pleistocene vertical slip rate of $>$0.02 m/ka. The northerly strike and prominent geomorphic expression of this potentially active fault differs from the Franklin fault. Our mapping of the Southhampton fault suggests that it accommodates dextral shear in the East Bay Hills, and is one of several left-stepping, en echelon N-striking faults (collectively, the "Contra Costa shear zone", CCSZ) in the East Bay Hills. Faults within this zone coincide with geomorphic features suggestive of late Quaternary dextral strike slip and appear to truncate or displace NW-striking reverse faults (e.g., Franklin fault) that do not displace the late Quaternary marine platform sequence. These data support an interpretation that the CCSZ accommodates regional dextral shear, and possibly represents the northern extension of the Calaveras fault. Overall, the marine terraces provide excellent strain gauges from which to evaluate the pattern and rate of late Quaternary deformation throughout the northern East Bay Hills.
T13C-1397 1340h
Seismic reflection profiling of the fault related fold in the eastern margin of the Japan Sea
The fault related fold in the eastern margin of the Japan Sea was investigated to evaluate its activity by geomorphologic and geological surveying. The investigated area is Japan sea off the Tsugaru peninsula where a 16 km long fault related fold had been shown in the previous tectonics map. The submarine geomorphology of the area was mapped by a narrow multi beam survey. In addition roughness and materials of the seafloor were imaged by a side-scan sonar survey. The synthetic picture of these results revealed a geomorphologic character of the continental shelf. A small valley passes through and runs on the axis of anticline from the continental shelf to the continental slope was found. The intensity of reflection on the axis of anticline is high. It seems that the old stratum is exposed. The three dimensional geological structure of a several hundred meters depth below seafloor was clarified by seismic reflection surveys using chirp sonar and water gun. The former target is stratigraphy in shallow part and the latter one is deeper structure. In order to acquire high horizontal resolution and to operate following digital filtering, a short shot-interval and digital recording were applied in the survey. The survey lines run across the investigated area in every 500 meters and along it in every 1000 meters. A thick sedimentary layer of middle Pleistocene was confirmed with reference to the following piston coring and the past exploration drilling in the survey area. The gravity piston cores were sampled at 8 points across the anticline to clear geological age of the acoustic layers. In evaluation of the fault related fold activity, the kink bands around the forelimb were observed on the seismic profiles. The growth pattern of the kink band is identified to hinge migration with a constant dip. As a result of tracing the deformed layer across the dense profiles, two folds were identified in the area, which had been recognized as a long fold in the previous studies. Besides, the upper limit of the deformed layer has the partial variation along the fold.
T13C-1398 1340h
Submerged Shorelines of Pilgrim Banks and the Northern Channel Islands, Southern California Continental Borderland, as Vertical Tectonic Strain Markers
An understanding of the complex tectonic interactions at the boundary between the Transverse and Peninsular Range provinces of southern California is of vital importance to seismic hazard analysis for coastal cities in the region. Tectonic deformation resulting from processes at this boundary is well expressed within the southern California Borderland, with the added benefit of being relatively accessible and undisturbed. In order to illuminate some of the processes occurring at this boundary we have been investigating Late Pleistocene-Holocene submerged low-stand shorelines rimming the Northern Channel Islands and submarine banks within the southern California Continental Borderland, for use as vertical strain markers. Using high-resolution multibeam mapping, direct submersible observations, and available seismic reflection profiles we are able to recognize distinct wave-cut platforms, paleo-seacliffs, and nearshore sedimentary facies characteristic of modern shorelines in the region. Using AMS 14C dating we have successfully dated numerous intertidal invertebrate shells, predominantly Mytilus californianus, from a number of these shorelines. Dates from a -120 m shoreline on Pilgrim Banks, when compared to eustatic sea-level curves, suggest a significant amount of vertical movement during the Holocene. Quantification of the magnitude of this vertical motion is continuing. In addition to vertical motions, the knowledge gained from the deformation of these shorelines by motion of strike-slip faults such as the Santa Cruz Island Fault will aid in constraining earthquake models for these active, seismogenic faults.