S43C-1009 1340h
Microearthquakes and Crustal Structures in the Southern Okinawa Trough
Located east of Taiwan, the southern Okinawa Trough (SOT) is a portion of a young continental backarc basin, which is still in the rifting stage. Based on seismicity data, a slab tear was identified along 123.3°E. In order to better understand the nature and role of tectonic features in this region, a passive seismic Ocean Bottom Seismometer (OBS) experiment was conducted from November 19 to December 1, 2003. 15 OBSs were deployed in a 130 x 90 km area including the Ryukyu slab tear and the Cross Backarc Volcanic Trail (CBVT). OBSs with three 4.5 Hz component geophones and an hydrophone recorded about 4000 microearthquakes. Most of them (about 94%) are crustal earthquakes (0-20 km) which occured in the SOT central graben. Just a few earthquakes were recorded from the upper slope of the northern SOT and from the Ryukyu forearc. Most of the earthquake local magnitudes (ML) range from 1 to 2, even if the whole range of magnitudes spans from 0.9 to 4. Three clusters of high microearthquake activity are observed in : (1) the CBVT area (24.7-25°N; 122.5-123°E); (2) the southern central graben area (24.65-24.85°N;123-123.3°E); and (3) the northern central graben area (24.9-25.2°N ;123.2-123.65°E). The southern central graben cluster is also identified by the Japan Meteorological Agency (JMA). The earthquakes determined by the JMA in the northern central graben present larger magnitudes than ours, but are globally located 10-km south of our cluster. The CBVT cluster is not recorded in JMA data probably because the shallow (O-5 km) volcanic activity only involved small earthquakes which are too far from the JMA's network. Based on swath bathymetric and seismic data, most of the epicenters lie along already identified normal faults. Thus, except the volcanic activity, which occurs in the CBVT area, the main factor controlling the SOT tectonic activity is normal faulting. On the northern and southern slopes of the central graben, hypocenters occur along normal fauls which dip toward the south and the north respectively. The deepest earthquakes (15-20 km) are located in the vicinity of the trough axis, within the lower crust. According to the distribution of the seismicity on N-S profiles cutting across the clusters, the dips of normal faults are about 70° and 55° for the northern and southern normal faults respectively. Dips of normal faults are consistent with those calculated from the seismic profiles acquired in this region. Lateron, we will combine the data from the OBSs with those of the 16 onland stations located on Taiwanese and Japanese islands in order to obtain a reliable crustal velocity tomography.
S43C-1010 1340h
Packaged Fault Model for Geometric Segmentation of Active Faults Into Earthquake Source Faults
In Japan, the empirical formula proposed by Matsuda (1975) mainly based on the length of the historical surface fault ruptures and magnitude, is generally applied to estimate the size of future earthquakes from the extent of existing active faults for seismic hazard assessment. Therefore validity of the active fault length and defining individual segment boundaries where propagating ruptures terminate are essential and crucial to the reliability for the accurate assessments. It is, however, not likely for us to clearly identify the behavioral earthquake segments from observation of surface faulting during the historical period, because most of the active faults have longer recurrence intervals than 1000 years in Japan. Besides uncertainties of the datasets obtained mainly from fault trenching studies are quite large for fault grouping/segmentation. This is why new methods or criteria should be applied for active fault grouping/segmentation, and one of the candidates may be geometric criterion of active faults. Matsuda (1990) used _gfive kilometer_h as a critical distance for grouping and separation of neighboring active faults. On the other hand, Nakata and Goto (1998) proposed the geometric criteria such as (1) branching features of active fault traces and (2) characteristic pattern of vertical-slip distribution along the fault traces as tools to predict rupture length of future earthquakes. The branching during the fault rupture propagation is regarded as an effective energy dissipation process and could result in final rupture termination. With respect to the characteristic pattern of vertical-slip distribution, especially with strike-slip components, the up-thrown sides along the faults are, in general, located on the fault blocks in the direction of relative strike-slip. Applying these new geometric criteria to the high-resolution active fault distribution maps, the fault grouping/segmentation could be more practically conducted. We tested this model successfully on the active faults generated the 1943 Tottori earthquake, the Chojagahara-Yoshii fault zone in Chugoku district in southwest Japan, as well as the active fault system in northern Luzon, the Philippines. Thus, we name this conceptual model as _gPackaged Fault Model_h and call the active faults grouped by the model as _gPackaged Faults_h for individual earthquake source faults. Moreover, we come to know that active fault mapping with _gPackaged Fault Model_h in mind enables us to find many new active fault traces (e.g., the Shigenobu fault along the MTL in Japan).
S43C-1011 1340h
Detailed Identification of Large Seismogenic Sources in Calabria, Southern Italy, Based on Recent Shallow Seismicity and Crustal Tomographic Structure vs. Local Geology
Although Calabria is perhaps the most earthquake-prone region of the Italian peninsula and one of the most active areas in the entire Mediterranean, the understanding of the geometry, kinematics and exact location of its main seismogenic sources is still at a rather early stage. We tackled this problem by performing a joint reinterpretation of (i) hypocentral locations and focal mechanisms of the local earthquakes recorded between 1978 and 2003 (about 7,000 events), (ii) the crustal tomographic structure of Vp and Vs seismic velocities derived from the same dataset, (iii) the distribution of potential sources of M5.5+ earthquakes recently hypotesized for the study area, and (iv) selected geologic and geomorphic information. We specifically focused on the portions of the study region affected by significant seismic sequences or swarms during 1978-2003. Then, we analyzed with different methods the hypocentral trends in the separate space-time clusters and performed separate tomographic inversions of earthquake data in the respective sectors. This step was aimed at obtaining the greatest accuracy of the velocity structure in the crustal volumes affected by the sequences. Given the complexity of the study region, during the investigation we encountered different anomalous or unexpected situations, including (a) spots of recent low-magnitude seismicity that do not correspond to any identified potential earthquakes source, and (b) large seismogenic faults that were partially or totally quiescent during the past 25 years. These were interpreted in the frame of the existing geologic and tectonic knowledge and to constrain current seismotectonic hypotheses.
S43C-1012 1340h
Active Faults, Modern Seismicity And Block Structure Of Eurasia
The analysis of on active faults and seismicity shows that the only a northern part of Eurasia should be regarded as an indivisible lithosphere unit. We defined it as the North Eurasian plate (Gatinsky, Rundquist, 2004) unlike the Eurasian plate s.l., which can be used only for paleotectonic reconstructions. The North Eurasian plate is bordered by zones of seismic activity traced along the Gakkel ridge, the Chersky and Stanovoi ranges, the Baikal rift, Altai--Sayany region, northern Tien Shan, Pamir, Hindu Kush and Kopet Dagh, Great Caucasus, northern Anatolia, Rhodopes, Carpathians, eastern and central Alps. Relationships between this plate and Europe west of the Rhine grabens remain ambiguous. The satellite measurements for them seem to be similar (Nocquet, Calais, 2003), but structural and seismic evidences allow suggesting their incipient division. Wide zones between this plate and neighboring ones can be distinguished outside north Eurasia. These zones consist of numerous blocks of various sizes. Block boundaries are mainly characterized by the high seismicity and development of active wrench faults, thrusts or modern rifts. Some of such zones were named earlier as "diffuse plate boundaries" (Stein et al., 2002; Bird et al, 2003). We suggest to name them as "transit zones" because they are situated between large lithosphere plates and as if transfer the stress field of one of them to other. Blocks within the transit zones reveal local divergences in GPS vectors of their displacements in the ITRF system and especially with respect to fixed Eurasia. At the same time data of satellite measurements emphasize the unity of the North Eurasian plate, which moves eastward in absolute coordinates with some clockwise rotation. The stress distribution in inner parts of the continent is being affected by the interaction with different plates and blocks. It can be more effectively illustrated by a triangle of the maximal seismic activity of Eurasia in the central Asia transit zones being under influence Indostan-Eurasia collision. The earthquake distribution, active fault patterns, and satellite measurements clearly indicate the present-day geodynamic instability of Eurasia. Unfortunately, the problem of the depth at which various blocks are displaced so far remains a matter of debate and the relevant information is still controversial. Some blocks have roots in the lithosphere upper mantle (Qaidam, southeast China), but others such as Tarim, Tien Shan are not divided at the lithosphere level. Hot plumes under western and central Europe are ascending from the sublitosphere mantle at the depth of about 400 km (Grosvenor et al, 1995), and European blocks' displacement can be connected with these plumes. At the same time prevailing depths of earthquake hypocenters (20-40 km) and seismic tomography allow to suggest that most of the blocks are shallow-seated and bounded from below by detachment zones localized at the crust base or within the lithosphere. It is displayed by a marked reduction of S-wave velocities at various levels presumably corresponding to the development of anomalous heated mantle at the base of crustal or crustal-mantle blocks. The presented evidence confirms the idea of lithosphere lamination, which was proved in the latest time by investigations in the frame of the project INDEPTH (Zhao et al., 2001; Li et al., 2003 and others). The Program of the RF President supports this work (project NSH-99.2003.5).
http://www.sgm.ru
S43C-1013 1340h
High resolution seismic imaging of an active normal fault in the Agri Valley, Southern Apennines, Italy
The Agri Valley is an intermontane basin located in the Southern Apennine seismic belt (Italy) whose formation in tied to large NW-trending trastensional and extensional faults active since Early Pleistocene. Recent faulting activity in the area is documented by faulted paleosoils and suggested by a M7 earthquake that struck the basin in 1857. On the contrary, present-day background seismicity in the area is extremely low. Despite intense geomorphic investigations, the identification of the source responsible for this historical event and of further large seismogenic faults in the area is still a matter of debate. A new NW trending normal faulting system has been recently recognized based on subtle geomorphic expressions on the ridge bounding the basin westward. Recent faulting activity along this structure is locally documented by a trench. Aimed at yielding new information about the shallow structure of the fault, we conducted a high resolution seismic experiment in a small lacustrine basin, located 4 km south of the trench, in which the presence of the fault is inferred by a linear surface warping but trench excavation is impractical. Both multi-fold wide-angle data and multichannel near vertical reflection data have been collected along a 220-m-long profile in order to obtain an accurate model of the basin combining seismic velocity and reflectivity images. About 3600 first arrival traveltimes picked on 36 wide-angle record sections have been inverted by a non-linear tomographic technique that is specially designed to image complex structures. The tomographic inversion provides a high-resolution velocity model of the basin down to 60 m depth. The model is strongly heterogeneous and displays sharp lateral velocity variations. Seismic reflection processing has been applied to both data sets. Data have been edited for trace quality and first (refracted and direct) arrivals have been muted. A following FK dip filtering on the shot gathers reduced the energy associated to ground roll. The traces were then gathered in CDP and corrected for statics. Note that both statics and the normal moveout correction have been applied using as background velocity the tomographic model. On multi-fold wide-angle data, a further iterative process, which consists of windowing the data using different offset ranges before CDP stacking, allowed to maximize the stack quality. The smooth velocity and stack sections confirm the fault location and delineate more fault splays, which are responsible for sharp lateral velocity changes and reflector truncations. The fault splays are located both beneath the surface warping and along the northern margin of the basin, thus indicating a quite complex faulting process. Moreover, the small amount (about 30 m) of total vertical slip separation imaged for the basin substratum across the fault zone allows explaining the subtle geomorphic expression of the faults.
S43C-1014 1340h
Crustal Seismicity and Recent Faults in Southern Peru
Most seismological studies in southern Peru have been focused on the downgoing slab seismicity in order to constrain the Wadati-Benioff zone. This study deals with the intra-continental seismicity of the southern Peru forearc (17,3$\deg$S - 18,5$\deg$S) in a post-seismic context (Arequipa thrust earthquake, Mw=8.4, 23 June, 2001). It is difficult to identify historical crustal earthquake from available catalogues, however some crustal events teleseismically recorded can be found since 1976; they exhibit normal focal mechanism solutions in the southern Peru volcanic arc and inverse focal mechanism solutions in the Central Depression. Following a notable increase of shallow crustal seismicity located close to the Western Cordillera after the 23 June 2001, a temporary seismic network was deployed between January and March 2003 in order to study the Wadati-Benioff zone and monitoring the crustal seismicity in southern Peru. From the about 1700 events locally recorded by the local network, 300 crustal earthquakes were identified in the Peruvian forearc between Tacna and Moquegua. This crustal seismicity is distributed along a lineament located at depths between 0 and 60 km, dipping at about 45$\deg$ from the Western Cordillera towards the coast, almost perpendicular to the subducting slab; this behaviour was previously observed in northern Chile and in southern Peru, north of the study zone (16$\deg$S). In the Central Depression, seismic activity is not superficial occurring between 25-60 km depth and it is mostly characterized by inverse focal mechanism solutions. Superficial faults situated in the Central Depression and in the Coastal Cordillera can not be associated with the seismic activity observed in this area. However, in the Pre-Cordillera, crustal seismicity occurs at depths between 0-15 km and can be correlated with shallow fault systems recognized by satellite images and on the field. For examples, the Incapuquio fault system which was a transpressive system in Cretaceous time seems seismically active and could be associated with sinestral strike-slips. The Purgatorio fault which has been identified by a pronounced discontinuity of the topography and the presence of terraces of accumulation could be associated with pure component movements. The correlation between seismological data and geomorpho-tectonic observations on field gives new insight in the southern Peru forearc deformation.
S43C-1015 1340h
Active Tectonic in the Southern Central Andes, a Recent Example: the 28 August, 2004 Shallow Mw=6.5 Earthquake
Seismic registration with the permanent local seismic network in central Chile and a temporary seismic network deployed along the Las Le\~{n}as and Pangal river valleys (34$\deg$25'S) between January and May, 2004 permitted to better constrain the abundant shallow intra-continental seismicity previously detected in that region. Although most of the seismicity is randomly distributed in the region, several seisms occur along the trace of the major Chacayes-Yesillo fault system. This fault, recognized between 34$\deg$45' and 34$\deg$30'S, is located at or next to the eastern contact between Mesozoic and Cenozoic deposits in the Principal Cordillera, and is considered to have participated in the development (extension) and tectonic inversion of a widely extended ($>$600 km long) Cenozoic extensional basin along the Principal Cordillera. The apparently associated seismic activity suggests that this structure is still active and participates in the present-day adjustments of the Andean crust. Recorded data show the complexity of the fault system. Several, almost vertical lineaments with depth distributions shallower than 10 km can be observed. Further south, at 35$\deg$S, a Mw=6.5 strike-slip shallow ($<$10 km) earthquake occurred on August 28, 2004, generating moderate damage in the region, reaching a maximum intensity VI MM. The location of this earthquake coincides with the trace of another major fault (El Fierro Fault), that also separates Mesozoic from Cenozoic units (basin bounding fault?). Temporary seismic stations were deployed to follow in detail the evolution of the abundant aftershocks. Analyses show an essentially NS distribution reaching depths lower than 15 km. This behaviour is in agreement with that observed to the north in Las Le\~{n}as-Pangal region. The 2004 shallow earthquake is the second one recorded by local networks in Chile, the previous one occurred in the northern Chile forearc in 2001 (Mw=6.2). The 2004 shallow earthquake is similar to the major intraplate Las Melosas earthquake [Mw=6.9] occurred on September 4, 1958, possibly associated with the Chacayes-Yesillo fault system. The occurrence of the 2004 earthquake offers the possibility to analyze this seismicity from a seismotectonic point of view, in order to understand the present-day crustal adjustments.
S43C-1016 1340h
A seismo-stratigraphic and lithological model for the Upper Cretaceous to Eocene deposits in the Zardob region (Azerbaijan)
A seismo-stratigraphic and lithological model for the Upper Cretaceous to Eoceve deposits in the Zardob region is developed on the basis of the processing and interpretation of seismic and well data. The seismic profiles are obtained using the seismic reflection technique and the common depth point method. The Upper Cretaceous brachyanticline fold in parallel to the Caucasian mountains is revealed in the seismic sections. The analysis of well data show that the Zardob fold is represented by the Maycop deposits. This fold is delineated by the 4400 m isoline with the horizontal dimensions of 4.3 km x 2.0 km and the height of about 100 m. The Zardob fold is marked by a deep subsidence, and the area of the Upper Cretaceous effusives is overlain by carbonates of the same age. The seismic velocities, lithofacial composition, porosity and productivity of the Upper Cretaceous to Eocene sediments are analyzed in this study of the Zardob region.
S43C-1017 1340h
Surface Rupture Segmentation and Slip Distribution of the 14 November 2001 Ms 8.1 Earthquake on the Kunlun Fault, Qinghai-Tibetan Plateau, China
The 14 November 2001 Ms =8.1 earthquake represent the most recent one of a series of large earthquakes along the Kunlun Fault, a highly active left-lateral strike-slip fault that bound the northern Tibet Plateau. Detailed surface rupture mapping, including 291 surficial left-lateral slip measurements and 111 net vertical slip measurements shows that the surface trace was 426 km long, multisegmented and consists of five major segments: the Taiyanghu, Buka Daban, Kusaihu West, Kusaihu East, and Kunlun Pass fault segments with the maximum left-lateral slip magnitudes 3, 5.7, 6, 6.4, and 4.2 m from west to east respectively, averaging 2.7 m. Two primary characteristics of the surface rupture are the en echelon geometry of the principle faults and the predominance of left-lateral strike slip, with local components of normal and reverse displacements caused by local changes in the fault geometry. Transtensional and transpressional structures were observed within releasing and restraining step-over areas respectively. The vertical slip components on all faults are less than 1 m (with as much as of 5.1 m vertical slip), quite variable and show little systematic behavior. Sinistral surficial slip was quite variable along the main trace of the rupture, however there is fairly regular long-wavelength (tens to hundreds of kilometers) behavior to the east of the Buka Daban Feng. Slips as large as 5~6 meters were observed at 5~6 sites that distributed at different surface rupture segments with the asymmetric shape of the slip profiles, which may be reflective of the direction rupture propagated. In additional, field and satellite images evidence indicates that most of the faults that ruptured in 2001 had had late Quaternary displacement. The variations in surficial slip (at both short and long length scales) which is only near-field slip using tape measures, should be considered minimum values and represent real variations in the amount of brittle slip on visible fractures at the surface, and potentially underestimates the actual slip for the earthquake and slip distribution for the overall surface rupture due to the difficulty in recognizing distributed nonbrittle deformation. This calls for caution in discriminating between one or multiple discrete events or estimating the size of past and future earthquakes using displaced deposits in trenches or offset geomorphologic features along strike-slip prehistoric fault ruptures.
S43C-1018 1340h
Seismic Process in Active Faults of the Baikal Rift System
Well known is the role of faults as structural factors controlling ore and non-ore natural deposits and/or other geological formations, fluid permeability rates, certain magmatic phenomena and other geologically ancient and recent phenomena taking place on the Earth's surface. In terms of seismology, faults are considered as concentrators of stresses which generate seismic processes when discharged. The seismic processes take place within areas of dynamic influence of faults (ADIF). Since seismic events controlled by the fault occur at certain time intervals, it is common to use a notion of "tectonic (re-)activation" of the fault or its fragment. Due to considerably different durations of the geological development of large faults and instrumental measurements of seismicity, seismic process modeling is very challenging. However, attempts are undertaken to jointly consider a huge near-fault region of the long-term formation and seismic events that occur chaotically and almost instantaneously in terms of the geological time scale within the given region by analyzing (1) selective reactivation of faults within the real time scale, and (2) regularities of spatial and temporal migrations of seismic events within the limits of ADIFs. Introduced are two new notions: (1) a quantitative index of seismic intensity of a fault, i.e. a number of seismic events n of magnitudes M per a unit of the fault length L (km) for the given time interval t (years), the area of dynamic influence of the fault being given as m (km); (2) an area of dynamic influence of the fault, i.e. a space around the given fault axis wherein alterations of structural and geophysical fields are recorded. The Baikal rift system (BRS) being a well studied object with abundant data on fault tectonics and epicenter fields of earthquakes is taken as a sample to describe spatial and temporal regularities in fault reactivation within the real time scale. Regularities in the seismic process in the BRS are considered within the ADIFs. In a seismic zone, seismicity is predetermined by the behavior of an ensemble of seismically active faults varying in ranks; seismic events within the ADIFs occur in a quasi-oscillating migration pattern. An established regular pattern of the occurrence of seismic events is applicable for medium-scale earthquake prediction in any seismically active region of the world. The research was supported by the Russian Foundation for Basic Research, Grant 04-05-64348.
S43C-1019 1340h
Seismogenic Fault Detection by Different Hypocenter Location Algorithms in the Southern Tyrrhenian Sea, Italy
We investigated the seismicity occurring in the last few decades along the continental margin of the southern Tyrrhenian region. In this portion of the Nubia-Europe contact belt the Tindari fault (TF) is a regional structure capable of up to 6 magnitude earthquakes linking the ongoing contractional and extensional crustal compartments of Western and Eastern Sicily, respectively. According to several investigators, TF represents the northwestward propagation of the Malta escarpment, a normal fault linking Eastern Sicily to Malta island which produced magnitude 7 earthquakes in the last centuries. West of TF in the Tyrrhenian sea the Sisifo fault crosses the compressional compartment and generates seismicity of maximum magnitude over 6. The prevailing off-shore location of these faults has made the data acquisition slow and the definition of the geophysical and geological features of these structures quite problematic. We applied several hypocenter location algorithms to seismometric data collected in the study region by the national and local seismic networks in the last 25 years with the main purpose of improving the accuracy of the local fault detection. Clear improvement in the knowledge of the fault geometry has been obtained applying the Bayesian location method by Presti et al. (BSSA, 2004) to earthquake sequences and swarms recorded between 1978 and 2003. In our investigation of hypocenter locations, we also performed synthetic earthquake simulations to test the significance of the main hypocenter trends found, i.e. we established whether a seismolineament or cluster really reflects fault activity or is a fictitious product of the recording network geometry. The results have been evaluated in the light of the geophysical and geological information available in the literature for the study region.
S43C-1020 1340h
Identification of Quaternary Faults in Southwest Western Australia Using DEM-based Hill Shading
In Australia, the extreme infrequency of large earthquake events means that the historic record of seismicity is poorly suited to the task of assessing seismic hazard. Paleoseismological investigations provide the only viable avenue to obtain constraints on the recurrence intervals of large and damaging earthquakes. However, the prehistoric record is compromised by difficulties related to finding direct evidence for large earthquakes (e.g. fault scarps), which may be subtle or relatively short-lived in the landscape. In recent times, high resolution digital elevation models (DEMs) have emerged as an important tool for defining and mapping of areas of probable elevated earthquake hazard. An examination of selected Shuttle Radar Tomography Mission (SRTM) 3 second DEM tiles and a 10 m resolution Department of Land Administration DEM has resulted in the identification of seven previously unrecognised fault scarps of probably Quaternary age in the southwest of Western Australia (SWWA). This doubles the number of Quaternary scarps known from SWWA, and is an important advance in defining areas prone to large earthquakes. The new features range in length from ~15 km to over 45 km, and from ~1.5 m to 7.5 m in height. As might be expected given the prevailing E-W regional compressive stress direction, the scarps are dominantly north-trending. However, most scarps are also arranged within a broad ESE-trending belt. This belt aligns with oceanic transform faults to the west of Australia relating to the break up with India. Of the fourteen scarps only two have been the subject of detailed palaeoseismic investigation to determine recurrence for large events. Ongoing research seeks to characterise seismicity on these scarps and further explore their large-scale relationship to each other, and to the architecture of the Australian plate. This work has the potential to greatly enhance our understanding of the drivers behind seismicity in intraplate Australia, and hence improve estimates of seismic hazard.
S43C-1021 1340h
The Sources of Destructive Earthquakes Retrieved From Their Regional Intensity Patterns by a new Inversion Technique
On the J.G.R. (2004, Vol. 109, in press) and the B.S.S.A. (2004, Vol. 94, 5, in press) we demonstrated that it is possible to retrieve geometric and kinematic information on the sources of some destructive earthquakes of the past by inverting their regional macroseismic intensity patterns. In fact, in the study cases, the inversion results agree with the seismological instrumental measurements and/or with neotectonic evidence and/or with authoritative tectonic interpretation, independent from our work. The intensity patterns of the following earthquakes were inverted more or less successfully. First, direct validations were feasible in the case of these recent and well-recorded earthquakes: 1987 Whittier Narrows, 1990 Sierra Madre, 1994 Northridge, California; and 1936 Cansiglio, NE Italy, 1990 Santa Lucia, Sicily, Italy; and, less confidently, three small-magnitude events in 1983, 1989 and 2000 in SW Norway. Then, in the case of the earthquakes of the preinstrumental era, which follow, only the compatibility with neotectonic evidence and/or with authoritative seismotectonic interpretation, and the orientation of the maximum horizontal geodynamical compressive stress was demonstrable: we refer to the strongest earthquake that has ever struck the Mediterranean (in the XVII Century, in SE Sicily), and its strong foreshock of Jan. 9, 1693; the 1904 Oslo, Norway, the 1873 Belluno NE Italy, and the 1741 Fabriano, Central Italy, earthquakes. This series of validations, or of plausible results, leads to the hope that more knowledge about pre-instrumental events can be obtained from intensity data: a key towards improving the calculation of seismic hazard, mostly in the Old World. Our technique is able to retrieve 11 geometric and kinematic source parameters that are: the three nucleation coordinates, the fault-plane solution, the seismic Moment, the rupture velocities and rupture lengths along-strike and antistrike, the shear wave velocity in the half-space. We do this by an automatic nonlinear inversion with a Niching Genetic Algorithm (NGA), and by using our simplified KF formula for body-waves that radiate from a linear source (B.S.S.A., Vol. 86, 1019-1027; B.S.S.A., Vol. 89, 1203-1213; B.S.S.A., Vol. 93, 47-60). The NGA inversion finds the minima on the hypersurface of the minimum residuals (calculated-minus-observed intensity at all sites) in the multi-parameter model space. For almost pure dip-slip mechanisms, two minimum-variance source models are found, which resemble the two auxiliary planes of the same theoretical fault-plane solution. In other words, in certain conditions, the problem is bimodal. Some tests suggest that the retrieved fault-plane solutions remain rather stable either in coastal areas or in the presence of asymmetric distributions of the sites surveyed in the field.
S43C-1022 1340h
Seismogenic structures in the Alboran Sea (Westernmost Mediterranean) characterized by marine geophysical data
In spite of the knowledge on the distribution and character of the upper crustal seismicity in the Alboran Sea (Westernmost Mediterranean) and surrounding Alpine mountain chains (Betics and Rif), and the abundance of geological data on the kinematics and tectonic evolution of recent normal faulting, there is a lack of a satisfactory correlation between the observed structures and the shallow seismicity in the region. This is particularly evident for the seismicity at crustal levels through the northern margin of the Alboran Sea, in where the relationship between the different seismic swarms and the crustal structures is far to be well established. Geological and geophysical surveys carried out recently in the Alboran northern margin (called as the Almeria margin) have been devoted to localize major faulting structures nucleated in the brittle, upper crust, characterizing their incidence in the sea-floor morphology, their connection with onshore seismogenic faults, and their timing, time evolution, and three-dimensional geometry. Using detailed swath bathymetry studies, deep-tow backscatter imaging (TOBI, MAK-1) together with high-resolution, single- and multi-channel seismic profiling, we have studied the upper-crustal structure of the northern margin of the Alboran Sea, in an area close to one of the seismic swarms in the offshore region. In the Almeria margin we have detected and mapped seismogenic fault structures, most of them cutting up to the seabed and with an associated fault escarpment, analyze their 3D geometry, the fault-linkage patterns, their possible kinematics and relationships with the shallow earthquakes. Some of these faults control the current physiography of the margin, determine changes in the orientation and slope of active turbidite channels (like the Almeria canyon), and have important implications for the study of submarine seafloor stability in this seismically active area.
S43C-1023 1340h
Two-Step MCMC Seismic Tomography
Damping is commonly used to stabilize the tomography result in conventional linear travel time inversion methods. Concern exists with the subjective damping factor selection and unreliability of solution uncertainty derived from linear inversion. In this study, a nonlinear approach, the Two-Step Markov Chain Monte Carlo method (MCMC) is applied to the travel time tomography problem. In the first step of the MCMC, no damping or smoothing is applied, so as to find the solution and its uncertainty objectively from the observed data. As the estimated uncertainty is not only related to the data error but also the investigation geometry (ray coverage) and the parameterization, the appropriate cell size is found first. The grid is gradually changed from coarser to finer, and the RMS error between the observed data and the predicted data from the mean solution will become smaller, and the estimated uncertainty increases. Once the RMS error won't change significantly, the grid size is considered to be appropriate. The estimated uncertainty (variance) from the first step is used as the damping factor in the second step MCMC when smoothing is applied. In this way, the reliable model information (good ray coverage) extracted from the observed data will be less affected by the smoothing process. This new approach is applied to a complex 2-D data set recorded over a fault zone, including refraction, wide-angle reflection and CMP reflection data. Simultaneous inversion of these data improves the model information and hence the resolution. The MCMC method also provides reasonable estimates of model uncertainty.
S43C-1024 1340h
Studying the Sequence of the April 17, 2003 Delingha Earthquake ($M_L$=6.7) by Regional Moment Tensor Inversion
On April 17, 2003, a strong earthquake of $M_L$ = 6.7 occurred to the northwest of the Delingha City, Qinghai Province, northwest China. The epicenter (37$^{\circ}$33-_N, 96$^{\circ}$27-_E) lies in the Zongwulong Mountain, where the Dachaidan-Zongwulong Mountain fault (DZMF) zone runs through. In this study, we analyzed the focal mechanism of the main shock and several strong immediate aftershocks by regional waveform moment tensor inversion. We collected a comprehensive regional waveform data set that includes the broadband waveforms from the China Digital Seismic Network, the GSN, and a portable broadband seismic network deployed in the near regional distance to the earthquake at the time. One very broadband station equipped with the STS-2 seismometer and 24-bit digital data logger was only 80km from the epicenter, which gave tremendous constraint on the mechanism of the aftershock sequence (Fan and Wallace, 1991; Dreger and Helmberger, 1993). Our results show the dominance of the high angle thrust faulting striking NWW-NW. The distribution of the fault plane solutions reflects the turning and/or branching of the DZMF zone in the area, which has been confirmed by the field geological survey.
S43C-1025 1340h
Geophysical Investigation of the California Wash Fault, Nevada
The California Wash Fault, located in the central Basin and Range province, is a zone of steeply W-dipping normal faults with probable Holocene scarps capable of generating M 6-7 earthquakes (Bidgoli et al., 2003). The fault zone is located ~70 miles northeast of Las Vegas, a highly populated area. If an earthquake were to occur along this fault zone, there are implications for strong ground motions in the Las Vegas basin as well as small towns closer to the fault zone (Bidgoli et al., 2003). The seismic hazard is increased because the local fault-controlled basins, including the Las Vegas basin, contain significant amounts of sand and clay, which have been identified as having increased amplification with strong ground motion. This seismic amplification in basin sediments could potentially cause a great amount of damage to buildings and other structures in and around the Las Vegas metropolitan area (Bidgoli et al., 2003). To better understand the seismic hazards and the fault zone geometry, a seismic reflection survey was conducted during the summer of 2004 along a part of the California Wash fault zone previously mapped and logged to show evidence of Holocene paleoseismicity. The reflection survey maps a shallow portion of the subsurface and images two strands of the fault. Twenty-four seismic recorders were set out along a profile normal to the strike of the fault at four meter intervals and seismic waves were generated by a hammer source. A total survey area of 40 meters by 92 meters was acquired across the fault zone. Seismic Processing Workshop (SPW) was used to processes the reflection data. The interpretation of these data confirms (1) the location of the California Wash fault; (2) that the fault dips steeply, 75 degrees WNW; and (3) that the fault is listric at depth. The vertical displacement along the profiles is 5 m and is consistent with recent magnetic and geologic studies conducted at the same location. An additional step in the analyses of these data will be to determine a velocity model in the survey area and re-process the reflection data with better resolved velocity values. These data will then be integrated with the existing magnetic and geologic interpretations of the area to better understand and characterize the seismic hazard.
S43C-1026 1340h
Integrated Near-Surface Seismic and Geoelectrical Mapping of the Concealed Carlsberg Fault zone, Copenhagen, Denmark
The Carlsberg Fault is located in a NNW-SSE striking fault system in the border zone between the Danish Basin and the Baltic Shield. Recent earthquakes indicate that this area is tectonically active. We locate the concealed Carlsberg Fault zone along a 12 km long trace in the Copenhagen city center by seismic refraction, reflection and fan profiling. We supplement our seismic investigations with multi-electrode geoelectrical profiling. The seismic refraction study shows that the Carlsberg Fault zone is a low-velocity zone and marks a change in seismic velocity structure. A normal-incidence reflection seismic section shows a coincident flower structure. We have recorded seismic signals in a fan geometry from shots detonated both inside the low-velocity fault zone and up to about 500 m away from the fault zone. The seismic energy was recorded on three receiver arrays (1.5-2.4 km long arcs) across the expected location of the 400-700 m wide fault zone at distances of up to 7 km from the shots. Shots detonated inside the fault zone result in: 1) weak and delayed first arrivals on the receivers located inside the fault zone compared to earlier and stronger first arrivals outside the fault zone; 2) strong guided P- and S-waves as well as surface waves inside the fault zone. The fault zone is a shadow zone to shots detonated outside the fault zone. Finite-difference wavefield modeling supports the interpretations of the fan recordings. Our fan recording approach facilitates cost-efficient mapping of fault zones in densely urbanized areas where seismic normal-incidence and refraction profiling are not feasible. The geoelectrical measurements show that the fault zone is characterized by low resistivities (lower than 5 ohmm), indicating that the fault zone is fractured and water-filled. This interpretation is supported by hydrological measurements conducted by others, which show that the Carlsberg Fault zone is highly permeable.
S43C-1027 1340h
Geophysical Imaging of Active Tectonics: A Case Study From the Inter Andean Valley, Ecuador
The Inter Andean Valley is a Pliocene-Quaternary basin filled with volcanic, lacustrine, fluvial and marine sedimentary deposits. A series of faults sometimes collectively referred to as the Delores-Guayaquil Mega Shear (DGM) traverses the length of the Inter Andean Valley posing a seismic hazard to a number of cities including the capitol, Quito. The DGM is a large right-lateral fault system similar in scale and seismicity to the San Andres Fault system which is understudied, especially in the subsurface. A site characterization study utilizing ground penetrating radar (GPR) and near-surface seismic reflection profiling was conducted in two areas of the Inter Andean Valley where geomorphic evidence suggests active faulting. One area, Nono Valley, exhibits extensional characteristics through basin bounding fault and the second area, Saquisili, has a structure consistent with the geometry of a fault propagation fold. Both areas are covered with thinly layered volcanic ash which is clearly seen in outcrop and the GPR profiles. Saquisili, in addition to the ash layers, has a nonuniform layer of pumice near the surface which was revealed in the drilled holes for the seismic source, which helps to account for the quick attenuation of the higher frequencies. The GPR profiles also image abrupt terminations and offset of horizontal layers, often associated with active faulting. We used a 48 channel multichannel seismograph with 30Hz geophones and a 20ft spacing to collect 24 fold common-midpoint profiles using a Betsy Seis-Gun firing 12 gauge blanks. Preliminary seismic data indicates that the frequency content ranged between 20 and 100 Hz with higher frequencies being systematically filtered out with depth. Seismic velocities range between 740 and 2600 m/s, producing a vertical resolution between 2 and 32.5m to a depth of approximately 900m. The GPR data was collected using a GSSI SIR-2 data accusation system with a 100 MHz antenna. The GPR signal penetrated between 120ns and 350ns (9 and 26m assuming a dielectric permittivity of 16). Combination of GPR and near-surface seismic techniques compliment each other by providing varying vertical resolutions and depths of penetrations. This case study provides valuable information that is relevant to future studies utilizing near-surface geophysics to identify and image active structures in the Inter Andean Valley or other geologically similar areas.
S43C-1028 1340h
A Ground Penetrating Radar Experiment along the Warm Springs Valley Fault System, Western Nevada
We conducted a ground penetrating radar (GPR) survey of several fault traces from the Warm Springs Valley fault system in western Nevada. Several trenches have been excavated along the southern part of the fault system, in an ideal setting to experiment with GPR. Our survey was conducted with a pulseEKKO 100 GPR system, using a 1000v transmitter, at an antenna frequency of 200 MHz, with an antenna separation of 0.5 m and a step spacing of 0.1 m. With a wavelength of 0.5 m, this frequency of energy penetrated up to 6 m into the sediments we explored, with the potential to resolve the upper and lower surfaces of stratigraphic units that were 12.5 cm or thicker. We used a limited two-way travel time window of 150-200 ns, to allow imaging of our desired target depth with minimal unwanted reflections (recording instrument, truck, wire fences, etc.). We ran surveys immediately adjacent to two fault trenches (Trenches WSVFS T1 and WSVFS T3) to compare them with established trench logs, and additional surveys over other parts of the fault (WSVFS T2) with a 30 m setback to avoid the noise effects from trenches (trench walls, bottoms, fencing, etc.). The latter survey was run as a series of seven trench-parallel lines, with two perpendicular tie lines to evaluate profile-to-profile features. The results from the approximately 400 meters of radar data collected in this experiment are encouraging. Nearly all mapped faults had fault-type signatures on the radar profiles, and the general stratigraphic features of the upper 1 to 3 m of sediments were imaged fairly well. Major faults were easily recognized as offset reflections, abrupt lateral changes in amplitudes of reflections and overall unit characteristics, and offset groundwater barriers. Numerous secondary faults were also recognizably imaged, although they were more difficult to discern from natural variations in the sediments (textural and structural), and from noise inherent in the GPR method. Liquefied units were found to correspond to chaotic radar reflections. Further research will include experimentation with varying antenna frequencies and orientation, step spacing, and more closely spaced 3-D grid data collection.
S43C-1029 1340h
The Shallow Fault structure of the Incline Village Fault, Lake Tahoe NV, From Offshore Ultra High Resolution CHIRP Profiling and Onshore Paleo-Seismic Trenching.
The shallow fault structure of the Incline Village fault, Lake Tahoe NV, has been mapped using a combination of sub-meter resolution seismic CHIRP profiling, conventional paleo-seismic trenching on-shore, and shallow remotely operated vehicle (ROV) imagery. The pairing of these geophysical and geologic techniques has provided new insights into the fault history of the active Incline Village fault. These data, combined with previous off-shore surveys of the Stateline and West Tahoe faults, will help constrain the extensional history of the Tahoe Basin. Six new CHIRP dip profiles and three strike profiles were collected this past summer along a submerged, late Pleistocene (19.2 +/- 1.8ka) paleo-terrace just offshore of Incline Village, NV, and show a well-defined normal fault zone with >2m of vertical offset across the paleo-terrace. The sediment draped hanging wall shows evidence of at least one prominent colluvial wedge. A 6+ m paleo-seismic trench was also excavated at the Incline Village Elementary School, some 1000 m from the shoreline, which was sited based on a reconnaissance CHIRP profile collected in 2000. Trenching results are complementary with the chirp profiles and show three colluvial wedges with more than 2 meters of vertical offset for each of the events. These preliminary results have strong implications for understanding the evolution of the Walker Lane dextral shear-zone, which defines the western boundary of the Basin and Range.
S43C-1030 1340h
Towards the development of Hyperspectral Images of trench walls. Robotrench: Automatic Data acquisition
Previous studies on imaging spectrometry of paleoseismological excavations (Ragona, et. al, 2003, 2004) showed that low resolution Hyperspectral Imagery of a trench wall, processed with a supervised classification algorithm, provided more stratigraphic information than a high-resolution digital photography of the same exposure. Although the low-resolution images depicted the most important variations, a higher resolution hyperspectral image is necessary to assist in the recognition and documentation of paleoseismic events. Because of the fact that our spectroradiometer can only acquire one pixel at the time, creating a 25 psi image of a 1 x 1 m area of a trench wall will require 40000 individual measurements. To ease this extensive task we designed and built a device that can automatically position the spectroradiometer probe along the x-z plane of a trench wall. This device, informally named Robotrench, has two 7 feet long axes of motion (horizontal and vertical) commanded by a stepper motor controller board and a laptop computer. A platform provides the set up for the spectroradiometer probe and for the calibrated illumination system. A small circuit provided the interface between the Robotrench motion and the spectroradiomenter data collection. At its best, Robotrench ?spectroradiometer symbiotic pair can automatically record 1500-2000 pixels/hour, making the image acquisition process slow but feasible. At the time this abstract submission only a small calibration experiment was completed. This experiment was designed to calibrate the X-Z axes and to test the instrument performance. We measured a 20 x 10 cm brick wall at a 25 psi resolution. Three reference marks were set up on the trench wall as control points for the image registration process. The experiment was conducted at night under artificial light (stabilized 2 x 50 W halogen lamps). The data obtained was processed with the Spectral Angle Mapper algorithm. The image recovered from the data showed an excellent classification of the three types of materials observed on the wall (brick, mortar and green felt) as well as an excellent geometrical restitution. More exhaustive experiment will be conducted on Carrizo Wash paleoseismic site during the fall of 2004.The use of Robotrench will provide an excellent platform to acquire large in situ hyperspectral images that can be used as a proof-of concept to design future devices.