U53A-0015
WDC/National Geophysical Data Center 2004 Indian Ocean Tsunami Data
The World Data Center (WDC) for Solid Earth Geophysics (including tsunamis) is operated by NOAA's National Geophysical Data Center (NGDC). NGDC is one of three environmental data centers within the National Environmental Satellite, Data and Information Service (NESDIS). Operating both World and National Data Centers, WDC/NGDC provides the long-term archive, data management, and access to national and global tsunami data for research and mitigation of tsunami hazards. Archive responsibilities include the global historical tsunami database, the bottom pressure recorder data, and access to event-specific tide-gauge data, as well as other related hazards and bathymetric data and information. The Global Historical Tsunami Database includes data for more than 1,700 events since 2,000 BC and more than 8,500 locations where tsunamis were observed. The tsunami database contains extensive information on the 2004 Indian Ocean tsunami including earthquake and tsunami effects and imagery. The associated tsunami runup database provides information on over 700 locations in 39 different countries, in three different oceans, where the tsunami was observed. The runup data records include information on the location where the wave was observed, arrival time, first motion of the wave, wave periods, distance from the tsunami source, maximum wave height, maximum inundation distance, and effects at the location such as deaths, injuries, and damage. Links are also provided to the tide gauge time- series data where available. All of the WDC/NGDC tsunami and natural hazards databases are stored in a relational database management system. These data are accessible over the Web as tables, reports, interactive maps, and custom CD-ROMs.
http://www.ngdc.noaa.gov/seg/hazard/tsu.shtml
U53A-0016
Marine gravity measurments offshore Sumatra
Gravimetric measurements were carried out offshore Sumatra and Western Java by BGR during several expeditions with R.V. SONNE in 2005/2006 (cruises SO-186 and SO-189) and 1998/1999 (cruises SO-137 and SO-138). Gravity data were collected along profiles with a total length of nearly 33000 kilometers covering the area between Aceh in the north and the Sunda Strait in the south.\\ The oceanic crust is characterized by slightly positive free-air gravity anomalies resulting from the influence of the subduction bulge due to the downgoing lithosphere. Additionally, there are N-S striking gravity anomalies on the oceanic crust, which can be correlated with a series of transform faults.\\ Landward, an about 50 km wide negative anomaly runs from NW to SE, whereby the gravity values increase from about -160 mGal to about -40 mGal. They reflect the Sunda trench with water depths of 6500 m off the Sunda Strait decreasing to 4000 m off Aceh. This suggests that the thickness of the trench fill with sediments increases considerably from South to North. Landward, the gravity anomalies increase and reach a zone of relative maxima from +60 to +160 mGal in the outer arc high. This zone is about 120 km wide, whereby the outer arc high is less pronounced in the South than in the North. North of $2^{\circ}$N the outer arc high moves very close to the trench, reaching its maximum values already in the slope area. This indicates the recent accretion of large volumes of sediments.\\ Further to the NE an elongated negative anomaly follows with values even lower than the minima associated with the trench. It corresponds to forearc basins filled with light sediments of up to 7 km thickness. Around the equator no basin has been developed. The EW striking gravity high in this area shows an obvious geographical correlation between with high free-air gravity associated with the Investigator Fracture Zone on the oceanic plate to the South. A causal link between the two seems likely. Subduction of the fracture zone is a possible cause for both the change in strike of the trench and forearc north of Nias. This would imply an eastward bend of the N-S trending ridge during the subduction process.\\ Isostatic residual anomalies were calculated and a 3D density model of the subduction zone off Sumatra was developed, using BGR multi-channel seismic and refraction seismic as constraints.
http://www.bgr.de
U53A-0017
Contrasting styles of deformation along the Sumatra subduction zone
Along the Sumatra subduction zone we distinguish two styles of deformation and accretion. This study bases on some 9000 km multi-channel seismic profiles acquired by the Federal Institute for Geosciences and Natural Resources (BGR) of Germany during two expeditions in 2006 (cruises RV Sonne 186 and 189) between Aceh in the north and Enggano Island in the south. The first style of deformation is characterized by landward and seaward dipping thrusts. These occur above a gently dipping oceanic crust (about $0.5\deg$ over 250 km) from the eastern flank of the Ninety East Ridge to the deformation front off northern Sumatra. The sediments covering the oceanic crust at the slope toe are between 3.0 and 3.2 s (TWT) thick. The subducting oceanic crust dips at $~3\deg$ under the frontal part of the wedge for at least 60 km. A proto-thrust develops in most cases already some 5 to 10 km seaward of the slope toe. The basally detached trench sediments in most cases are sliding along a seaward dipping thrust plane onto older accreted material forming 3-5 thrust slices. The thrust plane forming the ramp is probably evolving from the set of reverse faults that develop already distal from the deformation front in the sediments covering the oceanic crust. The second deformation style we observe is a conventional imbricate stack of thrust slices. This occurs in the region offshore of Simeulue Island, south of approx $2.5\deg$ N, where the architecture of the fore arc changes significantly. Here, the oceanic crust dips at an angle of $~7\deg$ within the first 50 km beneath the lower slope. Mainly 2.5 s (TWT) sediments fill the deepest part of the trench, but in seaward direction, we observe a sharp decrease in the sedimentary cover to less than 1.0 s (TWT) thickness. Deformation starts at the slope toe with an anticline developing above a landward dipping thrust fault. The steep slope rises continually to the outer arc high in the prolongation of the Islands of Simeulue and Nias. The wedge is built up by a stack of thrust slices bounded by landward dipping reverse faults that rise from a common basal detachment some 0.2 s (TWT) above top oceanic crust reflection. The oceanic crust beneath the trench off the Mentawai Islands (south of $1\deg$ S) is intensively faulted with offsets up to 0.6 s (TWT). These faults are often listric, leading to half grabens or rotated crustal fault blocks. The oceanic crust morphology in the area depicts several ridges and other bathymetric highs. The sedimentary fill of the morphologically expressed trench is about 1.2 to 1.7 s (TWT) at the deformation front. As offshore northern Sumatra, both landward and seaward verging folds developed at the deformation front. The incoming sediment pile is completely involved in thrusting and accretion. The frontal wedge is build up of two to three larger thrust slices. Further upslope a major landward dipping thrust fault, presumably an out of sequence thrust, breaks through the seafloor. Upslope this thrust the slope is merely gently dipping and has a plateau-like morphology.
U53A-0018
Bathymetry Offshore Sumatra � First Comprehensive map of International Data Sets
Knowledge of the bathymetry offshore Sumatra is of great importance for geohazard risk assessment, modelling of tsunami runup heights and development of tsunami early warning systems as well as for the general understanding of plate boundary processes and morphotectonic features. Since the devastating December 26, 2004 Sumatra-Andaman Islands earthquake and tsunami a number of marine expeditions, funded by Canada, France, Germany, India, Indonesia, Japan, United Kingdom and the United States have acquired bathymetric data over the southern part of the earthquake rupture zone but also along strike the whole Sunda trench. Here we present the first compilation of these bathymetric data sets as one bathymetric map. The bathymetric data acquired up to date covers a vast part of the trench, continental slope and in part also of the fore arc basins. The map incorporates the newest data sets from 2005 of the British high-resolution HMS SCOTT survey, the French Marion-Dufresene ''Aftershocks'' and the Japanese Natsushima cruises. While these surveys concentrated on the southern rupture zone of the Dec. 26th, 2004 earthquake, the German RV SONNE SeaCause and Sumatra cruises in 2005 and 2006 mapped the March 28th 2005 rupture area as well as large parts of the central Sunda trench and slope and in part the fore arc basins. Surveys reaching back to 1997 covering parts of the Sunda Strait and offshore southern Sumatra are also incorporated. A nearly complete coverage of the Sunda trench and slope area in the north is achieved. In the south data gaps on the slope still exist. This map compilation is a collaborative international effort initiated and partly funded by InterMARGINS. It is a major contribution to the Indonesian and international science community.
U53A-0019
Tsunami Hazards Along the Chinese Coast from Potential Earthquakes
The recent Indonesian earthquake has awakened great concerns about destructive hazards along the Chinese coast. Scientists have provided a clear record of past tsunamis along East China that clearly indicate the potential for future tsunami damage to China. In this work we will assess from analyzing the probability for tsunami waves to hit the Chinese coast in the next century from large earthquakes coming from the neighboring subducting plate boundaries. This analysis is important because of the sharp increase in coastal population density in China, the intense development of harbors and the exploitation of mineral resources in coastal areas, ranging from Xiamen in the north to Hainan in the south. The probability seismic studies for the South China Sea and adjacent field were based on the relationship of Gutenberg- Richter (G-R) relationship between the number of local earthquake and magnitude. We studied the earthquakes of the global subduction belt. We found the earthquakes of the global subduction belt follow the G-R relationship. The plate boundary model came from P. Bird (2002). According to the historical earthquakes of South China Sea and adjacent field (From NEIC), and the tectonic and focal mechanism (HCMT), the studied field is divided two partitions. The latitude of the first partition is N (12-19 deg.); the second is N(19-23 deg.). There are twelve large earthquakes in the two partitions. Their magnitudes are bigger than 6. The probabilities of earthquakes in the South China Sea are computed by the local G-R relationship. They would determine the seismically-induced tsunami probability. In our study the linear shallow water equation is used for integrating the twelve earthquake induced tsunamis. The numerical scheme for the linear equations is the staggered leap-frog method. The code has been provided by Dr. Fumihiko Imamura, Tokohu University, Japan. We combined the probability of three segments of wave height, 2.0 to 1.0 meter, 1.0 to 0.5 meters, the wave height smaller 0.5 meters, in the tsunami modeling. We forecast that there is 3 percent probability for a ten foot tsunami wave to hit Hong Kong and Macau within the next 100 years.
U53A-0020
Coseismic Deformation of the M9 Sumatra-Andaman Earthquake: Accounting for Tectonic Structure via FEMs
The Great M9 Sumatra-Andaman Earthquake of 2004 ruptured the interface between the overriding forearc wedge and subducting oceanic crust of the Burma and India plates, respectively. The rupture began west of northern Sumatra and propagated 1200 km northward and trench-parallel beneath the Nicobar and Andaman Islands. Prior to the event, frictional strength prevented the crystalline continental crust of the converging Burma plate from driving the compliant forearc wedge over the subducting oceanic slab of the Indian Plate. A suitable model of this coseismic deformational system must honor the spatial variations of material properties that are observed via seismic tomography and reflection, gravity, and bathymetry data. Data gathered from five GPS sites in the Nicobar Islands indicate the earthquake caused about five meters of pure trench-normal convergence. A finite element model (FEM) is constructed to 1) estimate the dislocation distribution for the rupture near the Nicobar Islands based on the GPS data and inverse methods and 2) predict seafloor deformation by loading the FEM with the estimated dislocation distribution. The FEM solves elasto-static governing equations over a problem domain, which includes a material property distribution that simulates the tectonic structure of the system. The analysis is repeated for a FEM that is modified to simulate a homogeneous half-space model. Both FEMs yield similar Green's Functions for displacement in the region that includes GPS sites. Thus, dislocation distributions estimated using inverse methods are also similar for both FEMs. However, prediction misfit of the GPS data from the FEM that honors the tectonic structure is about 20% less than that from the homogeneous half-space model. Although both vertical and horizontal deformation predictions from the two models converge near the GPS sites, the predictions diverge significantly for the seafloor region between the GPS stations and the trench. Even though both model configurations can fit the spatially sparse onshore GPS data, regional scale predictions of seafloor deformation depend heavily on the model configurations. This result has substantial implications for models of coseismic and postseismic deformation, stress-coupling, and tsunami genesis.
U53A-0021
Delayed response of far-field harbors to the 2004 Sumatra tsunami: the role of high-frequency components
During the 2004 Sumatra tsunami, several large vessels (up to 300 m long) broke their moorings in harbors of the Western Indian Ocean (Reunion, Madagascar, Oman). In two instances, these incidents (which resulted in structural damage to port infrastructure in Reunion) took place several hours after the arrival of the waves of maximum amplitude, raising important questions regarding the issuance of an all clear after a tsunami alert. We attribute this timing to the delayed arrival of high-frequency components of the tsunami wave, in the 5 to 12 mHz range, dispersed outside the shallow water approximation, and capable of exciting the resonance of individual harbor basins. In the case of the port of Toamasina, Madagascar, we use small scale nautical charts to conduct a finite-element modeling, documenting a family of eigenmodes with periods of 105 s. At this frequency, the tsunami is expected to travel from Sumatra in non-shallow water at 97 m/s, and thus to arrive around 19:50 (GMT +3), in good agreement with reports from interviewed witnesses. High-frequency components of the 2004 (and other) tsunami(s) have been widely observed, notably on hydrophones and seismometers, and their amplitudes can be quantified in relation to the earthquake source, thus paving the way for the systematic modeling of their effect on distant harbors as part of mitigation efforts and in parallel to the development of inundation maps
U53A-0022
The 2004 Sumatra-Andaman Earthquake: Non-linear Joint Inversion of Geodetic and Tsunami Data
The December 26, 2004 an earthquake of $M_w$=9.2 initiated off the West coast of northern Sumatra and propagated northward up to the Andaman and Nicobar Islands: it was the largest event ever occurred during the last 40 years, since the beginning of the era of modern digital recording. This earthquake generated the most devastating tsunami of the last centuries, causing more than 250,000 fatalities and spreading massive damages along the coast of the whole Indian Ocean. The aim of this work is to infer the characteristic of the earthquake rupture, in terms of slip distribution, rupture speed and rise time along the causative fault, by using different datasets. In particular we use several tide-gauge records distributed along the coast of the Indian Ocean, satellite measured sea level anomalies collected on passing of the tsunami wave at the open sea and GPS measurements of the coseismic static displacement. We divide the fault into a set of subfaults both along strike and dip direction. Tsunami waves are computed by numerically solving the non-linear shallow water equations through a finite- difference technique in a staggered grid. The coseismic GPS offsets are calculated using the semi-analytic global deformation model, which is a spherical, layered, self-gravitating incompressible model which can account for both elastic and viscoelastic responses induced by a point dislocation. The global solution is then computed as the linear superposition of the solutions corresponding to the individual subfaults (Green's functions), for both the case of tsunami waves and GPS displacement. We search for the rupture parameters simultaneously, by means of a two-stages non-linear inversion procedure based on the simulated annealing technique that allows extracting the stable features of the earthquake rupture. We perform several tests, by inverting the data separately and jointly, to validate the method and to estimate the resolution.
U53A-0023
Tectonic and Sedimentary Evolution of the Simeulue Fore-Arc Basin, Northwest Sumatra
During two cruises with the German research vessel SONNE in 2006 approx. 1460 km of MCS data was collected in the area of the Simeulue basin, a NW-SE striking fore-arc basin of the Sumatran subduction zone. Two lines cross industrial wells on the Sumatran shelf which we used for chrono-stratigraphical calibration of the seismic reflection data. We identified three major unconformities in the basin. The lowest unconformity represents the boundary between the Neogene sediments and the acoustic basement. Layers above clearly onlap the uncoformity. A second unconformity is of Middle to Upper Miocene age and corresponds with a relative sea-level lowstand resulting in an erosional surface. The sedimentary column above this unconformity shows significant differences from the northern to the southern basin part. In the north, basinward dipping, high amplitude reflections are intersected by a layer with chaotic seismic reflections in the deeper part of the basin. In the southern basin carbonate build-ups or reef complexes are distinct. The youngest sedimentary sequence is characterized by parallel, continuous high amplitude reflections. At it's base onlap and downlap terminations are widespread and we interpreted the unconformity as the Pliocene � Pleistocene boundary, because a change in deposition from mud and siltstones to sandstones is reported at this boundary. Deformation of the Simeulue Basin is complex. From a correlation of an anticline at the western border of the basin, we interpret a dominant arc parallel strike. The anticline is similar in shape with wrench faults found in fore-arc basins with a highly oblique deformation setting (i.e. Mentawai fault off southern Sumatra). Recent tectonic deformation is manifested by local basin inversions, leading to erosion of the hinge region of a growing anticline with subsequent normal faulting.
U53A-0024
Project SUMATRA: The Fore-arc Basin System of Sumatra
The main scientific objective of the project SUMATRA is to determine or estimate the geological setting and evolution of the Sumatra fore-arc region. RV SONNE cruise SO189 Leg 1 was designed to investigate the architecture, sedimentary thickness, sedimentary evolution and subsidence history of the fore-arc basins Siberut, Nias and Simeulue off Sumatra. During the cruise a total of 4375km of multichannel seismic (MCS), magnetics (M) and gravity (G) data were acquired and additional 990km with M and G alone. Along two lines with a total length of 390km refraction/wide- angle seismic experiments were carried out. 41 MCS lines cover as close grids the three fore-arc basins. Five lines extend nearly orthogonal to the subduction front covering the whole subduction system from the adjacent oceanic plate, the trench and accretionary prism over the Outer Arc High to the fore-arc basins. In the Simeulue Basin it was possible to connect the seismic lines to three industry wells and to correlate the seismic horizons to the results from the wells. The Simeulue Basin is divided into a northern and southern sub- basin. The maximum thickness was determined to be 6s TWT. In the southern sub-basin carbonate build-ups (which were already identified during the SEACAUSE project), bright spots and Bottom Simulating Reflectors (BSRs) are wide spread. The narrowest basin surveyed was the Nias Basin. As the Simeulue Basin the Nias Basin is divided into two sub-basins which are separated by a structural high. Although the basin has a maximum width of only 55km the maximum sediment thickness exceeds 5s TWT. The largest investigated fore- arc basin is the Siberut Basin. It extends over 550km and has a maximum width of 140km between Siberut and Sumatra. The maximum sediment thickness in this basin is 4.8s TWT. The basin geometry is uniform along its axis. At the basins termination on the western side to the Outer Arc High the Mentawai Fault Zone could be traced. In the Siberut Basin BSRs are very wide spread and very good recognizable over the Mentawai Fault Zone. Along the Mentawai Fault and along the eastern rim of the basin the seismic data show strong indications for active venting. As offshore northern Sumatra, both landward and seaward verging folds are developed at the deformation front off Nias and Siberut. For the first time landward verging folds have now been imaged in this domain of the Sunda subduction zone. Two refraction lines were acquired parallel to the subduction front at 2.5N and 1.5S approximately 40-50km seaward of Simeulue and Siberut Island, respectively. The lines were designed to identify the segment boundaries in the subduction system as well as to detect and decipher the subducted aseismic Investigator Ridge. The gravity data set is consists now of over 38,000km (combining the GINCO, SEACAUSE I and II and the SUMATRA data). With this it was possible to compile a map of the free-air gravity from the northern tip of Sumatra (6.5N/95E) to Mid Java (8.5S/110E). Gravity modelling in parallel with refraction seismic data interpretation was carried out along two lines during the cruise. The preliminary results show that the incoming oceanic crust is at 5-6 km unusual thin, both in the south off Nias (5km) and in the north off Simeulue (6km).
U53A-0025
First results from a combined marine and land passive seismic network near Simeulue island
Beginning in October 2005, we have installed a temporary network of 43 ocean bottom and 8 land stations on the forearc of the Sumatran subduction zone on and around the island of Simeulue, with recording times of 3-5 months. The marine work was carried out during Sonne cruise SO186. The temporary network is a component of the SEACAUSE project. Typically 7-8 locatable events are registered per day but this rate increases dramatically after several events with magnitudes larger than 5.5 within the array. All events were relocated by first using a joint inversion for a 1-D velocity model and station statics, and then redetermining locations and location uncertainties by a non-linear location method (oct-tree). Events occur on or near the plate interface, within the downgoing slab and in the overriding plate. The largest concentration of events is found just south of Simeulue island and its extension along strike, near the presumed updip limit of the co-seismic rupture.
U53A-0026
Schlumberger seismic vessel Geco Searcher provides unprecedented images of the Great Andaman Sumatra earthquake megathrust rupture plane
From July 13 to 27, 2006, we carried out a deep seismic reflection survey along two lines on board the WesternGeco seismic vessel Geco Searcher. The vessel was equipped with one 12 km and one 5.5 km Q- Marine streamers. The long streamer was towed at 15 m depth, providing low frequency signal for deep targets, and the short streamer was towed at 7.5 m depth for high-resolution imaging of shallow sediments. The Q- Marine technology, developed by Schlumberger, is the most advanced technology available in seismic industry where individual hydrophones spaced at 3.125 m intervals sample and transmit data continuously back to the vessel. These data are then decimated to the appropriate trace interval, in this case 12.5m, after application of a digital spatial anti-alias filter, providing 960 channels of data. An array of 48 air guns provided a 10,170 cubic inch source with approximately 330 bar-m output. The shot interval was 50 m, providing 120 fold data at 6.25 m CMP intervals. The long streamer would allow us to remove seafloor multiples whereas high fold coverage would be extremely valuable for removing noise. Two deep seismic reflection lines were shot. The first line is 255 km long and runs close to the epicenter of the 26th December event, traversing the subduction front, a narrow accretionary wedge, the Simeulue plateau and the Simeulue forearc basin. The second line is 455 km long and located 255 km farther west: it traverses the whole margin, from the oceanic basin on the Indian plate up to the Andaman Sea, running across the the deformation front, the accretionary wedge, the West-Andaman Fault, the Aceh forearc basin, the submarine Sumatran fault and volcanic arc. Onboard processing of these data shows reflectors down to 18 s two-way travel time (TWTT), i.e. down to about 50-60 km depth. The subducting oceanic crust, including the oceanic Moho, could be seen down to 12 s TWTT. The megathrust that produced the earthquake could be followed from the source region at about 30 km depth to the surface near the subduction front. Extensive processing is under way and should provide unprecedented reflection images of this megathrust and hence insight about the nature of the tsunamigenic Great Andaman-Sumatra earthquake.
U53A-0027
Seismic Wide-angle Experiments Offshore Sumatra Using Ocean Bottom Seismometers
Three wide-angle seismic experiments were performed offshore Sumatra in March 2006 during RV Sonne cruise SO186. The measurements are a part of the Seacause project, which investigates the georisk potential along the active convergence zone between the Eurasian and Indo-Australian plates offshore Indonesia. The first seismic experiment was a 150 nm long profile east of Simeulue Island. Twenty-five ocean bottom stations were deployed along the profile, which was directed from the trench in the south-west to the coast of Sumatra in the north-east. The second and third experiment were 20 nm long seismic profiles south and north-west of Simeulue Island. Seismic tomography is applied to the traveltime data of the first arriving seismic phases to obtain a 2D velocity model. This model is used to compute a depth migrated image from the corresponding multi-channel seismic data.
U53A-0028
Earthquake Trend Around Sumatra Region Indicated by a new LURR Implementation
The LURR (Load/Unload Response Ratio) theory has been proposed for nearly two decades. The current implementation of LURR has shown great promise in intermediate to short term earthquake forecasting in the past practice, but at the same time has also met many difficulties. In this paper a new implementation of LURR method is presented to address one of the problems. The new implementation involves calculation of the maximum faulting orientation (MFO). After compared with the world stress map, the calculated MFO has been found to be in good agreement with the observation from the regional tectonic stress regime. The MFO pattern in Indonesia region has its own feature which may be related to the unique subduction complexity. The MFO in Sumatra in general is different from that in Java region after the 2004 megaevent. This phenomenon may be supported by the evidence of the recent observation that a section in the southern part of the Sumatran Arc is locked. Furthermore, the MFO pattern before the 2004 great earthquake is different from that after the event. Retrospective examination of the Indonesia region by means of this implementation produces significant LURR anomaly not only prior to the 2004 M9.0 Sumatra Andaman Islands earthquake but also before the 2006 M7.7 South Java earthquake. Therefore future great earthquakes like this one have a good chance of being forecasted if the LURR anomaly could be considered precursor.
U53A-0029
The Segmentation of the Subduction Zone Offshore Sumatra: Relations Between Upper and Lower Plate
Modern reflection seismic technologies are very successful in imaging the structure of the upper parts of the earth's crust beneath the ocean in great detail. However, in the case of a subduction zone already at a short distance from the subduction front little information is available about the subducting plate from the seismic images. It is widely accepted that the subducting plate inherits its main structural tectonic features to the overriding plate and to the subduction zone as a whole. A projection of prominent structures like fracture zones from the not yet subducted parts of the oceanic plate into the subduction zone provides a useful approximation of the unknown segmentation of the lower plate. These structures can partly be seen in bathymetric images and gravity data but can most reliably be derived from magnetic seafloor spreading anomalies. Offshore Northern Sumatra, a very thick sedimentary layer hides all but the most prominent morphological features of the igneous oceanic crust in bathymetric images. Where we collected magnetic data during the SeaCause cruises we can confirm the interpretation of Cande et al. (1989). The crustal structure of the Indo- Australian plate east of the Ninetyeast-Ridge is dominated by a series N S striking fracture zones with age jumps of up to 15 m.y. Crustal ages range from 45 to 68 m.y. The rupture areas of the two catastrophic earthquakes from Dec.26, 2004 and March 28, 2005 are separated by a sharp boundary underneath the island of Simeulue. This is the area where one of the fracture zones in the oceanic plate enters the subduction zone. The epicentre of the Dec. 2004 earthquake is located close to the prolongation of the fracture zone. The March, 2005 rupture area terminates to the south in the area where a prominent fracture zone, the Investigator ridge, enters the subduction zone. It seems very likely that these lower plate features play a major role in defining segment boundaries along the Sunda subduction zone. A better knowledge of the subducting plate's tectonic structure might help to define the possible boundaries of future rupture areas.
U53A-0030
Results From Pressure and Seismological Data From Stations Placed Offshore Sumatra During the SEACAUSE & GITEWS Cruises
During the SEACAUSE and GITEWS cruises with RV SONNE in 2005 and 2006, a network of seismological stations was deployed in and around Simeulue Island and on the Indo-Australian oceanic plate close to the trench offshore Sumatra. The marine stations were equipped each with differential pressure gauges and seismometers to record the seismicity. Additionally, absolute pressure sensors were used to record the low frequency tide signals together with higher frequency seismic waves and potential tsunami signals. The absolute pressure sensors are used to recognize tsunamis in the Indian Ocean and were also deployed for the German Indonesian Tsunami Early Warning System (GITEWS) along with two surface buoys. The surface buoys communicate with the ocean bottom instrument, record the surface tidal signals together with meteorological data and transfer data to land stations via satellite. Several earthquake signals were registered with the pressure sensors and correlated with the seismic signals recorded by the seismometers and differential pressure gauges. The earthquake signals arrive instantaneously at the seismometers and pressure sensors because of short distance to the epicenter and the high seismic velocities. A possible tsunami signal would have a time delay, arriving later at the instruments, due to the slower wave propagation of the tsunami (800 km/h) compared to the seismic wave travelling with a velocity of 30000 km/h. The earthquakes recorded with the absolute pressure sensors also show differences in the signal signature, depending on the position of the instrument relative to the earthquake location. Closely spaced neighbouring stations show similar tidal signals. The tidal signals recorded at the positions of the surface tsunami buoys are corrected for the air pressure. A Fast Fourier Transform (FFT) yields the amplitude spectrum of the time series recorded with the absolute pressure sensors. The half-day and diurnal tide signals show the largest peaks. The risk of upcoming megathrust earthquakes is large along the Indonesian archipelago. Therefore the set of instruments described above that were the first instruments to be deployed offshore Sumatra, will help recognize large earthquakes and potential proximate tsunamis as part of the German Indonesian Tsunami Early Warning System.
http://www.gitews.org
U53A-0031
Slip Distributions on the Fault System Responsible for the December 26, 2004 Earthquake Obtained by Inversion of Different Types of Tsunami Data.
The tsunami generated by the December 26, 2004, M=9.3 earthquake, rupturing a very large portion of the Sumatra-Andaman arc, is one of the most devastating events in historical times and probably the best documented event ever. A huge set of data of very different nature were recorded and collected. On the basis of these data, different teams of scientists working in different disciplines have and are still trying to retrieve the detailed spatial and temporal features of the earthquake rupture as well as of the ensuing tsunami. To date, there appears to be a substantial agreement on the gross features of the earthquake source, in particular regarding its magnitude, its North-South extension and its time evolution. But when it comes to finer details, like the detailed slip distribution along the ruptured area, the inversion of different data does not seem to provide a unique solution. In this contribution we infer the slip distribution on the causative fault of the December 26, 2004 earthquake by inverting separately the tsunami elevations registered by three different satellite altimeters, and a set of tsunami run-up heights measured by different field survey teams along the coasts of the countries closest to the source region. The two datasets catch different aspects of the tsunami, and must be inverted by means of different techniques. Three satellites recorded the wave amplitude along three different tracks at different times, and they describe mainly the linear propagation of the tsunami in the open ocean. On the other hand, run-up data are inherently connected to the non-linearity arising in the impact of the tsunami waves on the coasts. We use two different domains and two different approaches to study the two problems. In the case of the satellite- altimeter data the domain embraces a relevant portion of the Indian ocean, while a smaller domain, involving northern Sumatra, the Nicobar and Andaman islands, western Malaysia, Thailand and Myanmar, is used to study the run-up heights. In both cases, we divide the earthquake fault in a given number of sub-faults, and then compute the tsunami generated by each sub-fault by means of two different codes: a finite-difference code implementing the linear shallow-water equations for the ocean-wide domain, and a finite-element code solving the non-linear version of the same equations in the local domain. In the case of the altimeter data, the adopted inversion technique is an implementation of the classical generalized linear least-squares theory. A more sophisticated technique, allowing one to determine also different amplification factors along different coastal segments, is employed to invert run-up data. We discuss the slip distributions obtained separately in the two approaches, and compare the results with those retrieved by modeling seismic and geodetic data.
U53A-0032
A 3D Finite Element Study of Possible Tsunamigenic Earthquakes in South Sumatra.
After the last mega events in the Sumatra region, the likelihood of a large earthquake effecting the west of Sumatra is considerably increased. As already shown by previous studies, the coseismic Coulomb stress increase due to the combination of the December 2004 and the March 2005 events is high under the island of Siberut, and the Sunda trench megathrust has been locked in this area since 1797. Moreover the megathrust under the Mentawai Islands south of Siberut has probably not slipped since 1833. In the region, the possibility of a slip of the order of 10m is likely. An event rupturing the megathrust beneath Siberut might be expected to propagate southward into this area resulting in a great earthquake of magnitude M 8-9 as well as being potentially tsunamigenic. By applying a 3D finite element numerical model we assess the role of several possible rupture models in the deformation of the area, mainly on the vertical seabed displacement. The central aim of our study, together with the analysis of the effects of possible tsunami along the coast of Sumatra, is the investigation of the sensitivity of our model to the complexities of the area, with particular regard to the seabed topography and the lateral rheological heterogeneity. Adopting realistic parameters and a fine bathymetry digitized from nautical charts, we found a relevant tsunami threat for the Sumatra coast with the possibility that in highly populated centres along the coast the water could reach several meters in height.
U53A-0033
Morphotectonics of the Sumatra Margin -- Analysis of new Swath Bathymetry
During several cruises with RV SONNE in 2005 and 2006 (SO186, SO189) systematic multibeam bathymetric surveys offshore Sumatra have been conducted. The bathymetric dataset now covers large parts of the trench and slope between Simeulue Island and the entrance to the Sunda strait. Here, we present a morphotectonic map and structural analysis of this new bathymetric dataset. The incoming oceanic crust, largely draped by a sedimentary cover, depicts a trench parallel conjugate, steeply dipping normal fault set building Horst and Graben structures. Two prominent ridges, the Investigator Ridge and the Wharton Ridge have a relief of 600�-700 m. In the vicinity to these ridges a second normal fault set has developed on the oceanic crust striking parallel to the ridges in N--S direction. These ridges as well as several other topographic highs further south are recently being subducted beneath the Sumatra margin. They deform and offset the lower slope and accretionary prism. The most prominent offset of about 15 km of the deformation front is observed in prolongation of the subducting Investigator Ridge. These ridges account for a general along strike segmentation of the deformation front and accretionary prism and possibly the subduction zone as a whole. The lower slope and accretionary prism are built up by a set of thrust faults and anticlines. In general fold axes strike NW�-SE, parallel to the trench. However, where the ridges subduct beneath the slope, fold axes are rotated and strike NNW�-SSE demonstrating a local rotation of the stress field. The recent activity of this domain is also expressed by an increased number of observed surficial collapse structures and slumps. Additionally, a third orientation of N--S striking lineaments is observed. These lineaments are especially pronounced by transverse channels and furrows feeding piggy back basins further up-slope. Along these lineaments fold axes are also offset, usually in a dextral sense. The structural inventory of the slope and accretionary prism thus depicts a complex interference of tectonic lineaments. This setting is partly due to the oblique convergence along the Sunda subduction zone with associated trench parallel and orthogonal strain partitioning, as well as the recent subduction of large oceanic ridges.
U53A-0034
Slip Distribution of Recent Sunda Trench Earthquakes: Reconciling 3D GPS Inversions with Seismological and Ocean Measurements
Three-dimensional inversion of GPS near- and regional-field observations for the slip distribution of the great 2004 Mw=9.2 Sumatra-Andaman earthquake is performed using a Greens's function approach. To this purpose, a covering of the relevant part of the subduction zone with rectangular patches is generated automatically in a layered earth model. Inversion results in two distinct slip models, depending on the relative weighting in the cost function of moment minimization and smoothing on one hand, and conformance to GPS data on the other hand. Both models fit GPS data very well, but differ in amount of seismic moment released in the southernmost part of the rupture near the epicenter (at 3°N â€â€œ 5°N). This â€Ëœextra' moment makes about 10 per cent of the total moment release. We test these two alternative models against independent observations which include seismological, tide gauge and satellite ocean altimetry data. The slip model with maximum moment release close to the epicenter is more consistent with the seismological tracking of co-seismic energy release by Ishli et al. (2005) and Krueger and Ohrnberger (2005). The same model also shows more consistency with Jason satellite altimetry as well as with tide gauge records, provided the finite velocity of rupture propagation is taken into account. The July 2006 Mw=7.7 Java earthquake excited an unusually large tsunami when compared to the earthquake's magnitude itself. Forward tsunami modeling shows that the finite fault model based on teleseismic broadband data strongly underestimates wave run-ups at Java coast. We employ regional GPS data to derive an alternative fault model of the Java earthquake that better fits observed coastal run-ups References Ishli et al. (2005) Nature, 435, 933-936 Krueger and Ohrnberger (2005) Nature, 435, 937-939
U53A-0035
Towards an Earthquake Monitoring System for Indian Ocean Tsunami Early Warning
The Mw=9.3 Sumatra earthquake of December 26, 2004, generated a tsunami that effected the entire Indian Ocean region and caused approximately 230,000 fatalities. The German human aid program for the Indian Ocean region started immediately after the disaster with substantial funding of 45M Euro for the proposed German Indian Ocean Tsunami Early Warning System (GITEWS). In this presentation we describe the concept of the Earthquake Monitoring System and report on its present status: The major challenge for a Earthquake Monitoring System (ESM) is to deliver information about location, size, source parameters and possibly rupture process as early as possible before the potential tsunami hits the neighboring coastal areas. Tsunamigenic earthquakes are expected to occur in subduction zones close to coast lines. This is particularly true for the Sunda trench off-shore Indonesia, but also in the Macran subduction zone off- shore Iran. Key for an Indian Ocean monitoring system with short warning times is therefore a dense real-time seismic network in Indonesia, supplemented by a substantial number of stations in other countries and territories within and around the Indian Ocean. 40 new broadband and strong motion stations will be installed during the GITEWS project until 2010. The EMS Control Center will be based on an enhanced version of the widely used SeisComP software and the GEOFON earthquake information system prototype presently operated at the GFZ-Potsdam (http://geofon.gfz- potsdam.de/db/eqinfo.php). However, the Control Center software under development at the moment will be more reliable, faster and automatic but with operator supervison. It will use sophisticated visualisation tools, offer the posibility for manual correction and re-calculation, flexible configuration and support for distributed processing. Is large redundancy for algorithms, moduls and hardware assures easy integration into larger multi- sensor, multi-hazard control centers and decision support systems. A first prototype of the EMS Control Center software will be ready in mid 2007.
http://www.gitews.org
U53A-0036
Coastal Effects of the December 26th Tsunami on Ko Phra Thong, Thailand
Ko Phra Thong, located off the west coast of Thailand, is a barrier island with 15.5 km of western facing coastline, fully exposed to the Indian Ocean. The island was impacted by the December 26, 2004 tsunami and was the focus of this post-tsunami study. A series of field surveys were conducted in March 2006, to determine local run-up values and inundation along the coast. Although, we had anticipated similar impact along much of the island's coast, we found that tsunami impact had significant variation along Ko Phra Tong's exposed shoreline. Some areas on the island were completely destroyed, while other areas along the coast had very little visible damage (with the exception of salt inundation that destroyed vegetation). Phra Thong has a variety of terrain ranging from small mountains to open grassy fields as well as tidal inlets with very dense mangrove forests. Our results show that Ko Phra Thong had a wide range of run-up at the coastline ranging from 7 to 20 m above sea level. Tsunami inundation reached up to 2 km inland and most of the mangrove forests were completely destroyed Other impacts documented during the survey include significant coastal erosion, major modification of tidal channels, and the formation of new lagoons. Although this range of run-up could not be seen in numerical models, a recent high resolution numerical modeling study conducted for this area suggests that this range of impact may have been caused by wave focusing from the nearshore bathymetry (Ioualalen et al., 2006). A study done by Kawata et al. (2005) conducted a few months after the tsunami's impact showed run-ups that closely match their model hindcast results. However, the Kawata et al. survey did not cover Ko Phra Thong. The bathymetry and its affects on the numerical model are presented and compared with the wave heights and inundation measurements from this study.
U53A-0037
Post-seismic deformation of Andaman Islands since the 2004 earthquake
On 26 December 2004 the Andaman Islands near Port Blair, on the hanging wall of the Sumatra/Andaman mega thrust, sank 86 cm and shifted 3.1 m WSW. Port Blair since then has continued to move WSW at an exponentially decaying rate that by mid 2006 had attained 25 cm (i.e. 8% of coseismic slip). In contrast the harbour rose 15 cm, an 18% reversal in coseismic subsidence. We have installed ten additional GPS points since the earthquake (http://www.ceri.memphis.edu/people/jpuchkyl/) and these too show post seismic slip, but with spatial variations that require variations in the subsurface process along strike, in addition to downdip post-seismic adjustment processes. The coseismic hinge-line separating uplift from subsidence in the southern Andaman Islands appears to have shifted to the east by 10-20 km. We note that vertical movements in the islands in the past 150 years have been quite complex but are known largely from eyewitness descriptions in many cases, even including the 2004 earthquake. For example, war-time coastal bunkers constructed shortly after the 1941 Andaman earthquake on a raised wave-cut marine terrace had been partly submerged prior to 2004, and then incremented â‰Ë†80 cm lower in the 2004 event. This has led us to install a series of tidal bench marks linked to GPS measurements so that future investigations will be able to quantify these changes better. Pressure gauges linked via leveling to GPS control points define mean sea level at 2-3 points every 5 minutes, while secondary points using hydraulic averaging are surveyed along the coast during their operation. We are currently in the process of modeling these GPS measurements, both as viscoelastic relaxation in response to the coseismic strain change and postseismic slip on the plate boundary interface down-dip of the seismic rupture. Previous studies elsewhere suggest that using both the horizontal and vertical displacements enables the models to distinguish the relative importance of these two processes.
http://www.ceri.memphis.edu/people/jpuchkyl/
U53A-0038
Modeling the Gravity Change Induced by the 2004 Sumatra Andaman Earthquake as Seen by GRACE
We use a modal summation technique for an elasto-gravitational spherical stratified Earth's model in order to compute the co-seismic gravity field induced by the 2004 Sumatra Andaman earthquake. We test several source scenarii and investigate the importance of different elastic structures of the regional lithosphere. The hydrological contributions are also studied from different global models for soil moisture. We compare the predicted gravity signature of this earthquake to the GRACE temporal gravity solutions provided by the Toulouse �Team of Space Geodesy� (CNES/GRGS) in 10 day periods in a spherical harmonic decomposition (up to degree 50). We attempt to detect the post-seismic deformation and induced gravity field in the year following the seismic event. We also compare our modeling to simpler non-gravitating flat models often used in this kind of problems.
U53A-0039
Quantifying hydroacoustic records of the Sumatra tsunami in the 5-12 mHz frequency band
The 2004 Sumatra-Andaman tsunami was recorded by hydrophones of the International Monitoring System at Site HS08 near Diego Garcia, notably in frequency bands extending outside the range of the Shallow Water Approximation. Despite the severe high-pass filtering involved in this instrumentation, we show that the spectral amplitudes recovered around 10 mHz can be successfully explained by generation from the seismic source, in the framework of the normal mode theory of tsunami excitation. At the lower frequencies characteristic of more conventional tsunami waves (T = 800 to 3200 s), the signal is probably present in the hydrophone records, but reliable deconvolution of its spectral amplitude is precluded by the fact that the instrumental filters lowered the tsunami signal to the level of resolution of the instrument digitizer. In the context of distant tsunami warning, hydrophone records could provide useful insight into high-frequency tsunami components, and even at lower, more conventional, frequencies, provided that an unfiltered channel could be recorded routinely.
U53A-0040
Far-field tsunami risk from mega-thrust earthquakes in the Indian Ocean
In the wake of the 2004 Great Sumatran tsunami, we present numerical simulations of a number of scenarios for tsunamis generated by potential mega-earthquakes on the shores of the Indian Ocean. Our sources include (1) a repeat of the 1833 Southern Sumatra earthquake estimated to have reached magnitude 9 and whose probability of reccurence is perceived as enhanced by stress transfer arguments; (2) a major event along the Andaman-Burma plate boundary, recently identified as the locus of a large earthquake in 1762 [Sieh, pers.comm.; also experiencing stress-transfer from the northern end of the 2004 Sumatran rupture]; (3) a magnitude 8.5 earthquake in the Makran, modeled as simultaneously rupturing the faults of the 1851, 1945, and 1765 events. We use the model MOST to calculate far-field evolution of free-surface wave generated by transfering the seafloor displacement directly to the ocean surface. MOST has been repeatedly shown to model the 2004 megatsunami propagation realistically. We also examine the robustness of our simulations with respect to variations in source parameters, such as latitude/longitude, focal mechanism and heterogeneity of slip distribution on the fault plane; we conclude that the far-field evolution is generally insensitive to this level of details, as long as the integrated parameter, i.e., the seismic moment, remains constant. However, the presence of very shallow bathymetry over the source region does substantially reduce the tsunami excitation in the far-field, as illustrated during the 2005 Nias earthquake; for the same reason, the tsunami from the Andaman/Burma source (2) attenuates rapidly as it propagates outside the Bay of Bengal. By contrast, the South Sumatra scenario (1) generates a tsunami with stronger impact than in 2004 on the SW Indian Ocean Islands of the Mascarenes and on Madagascar. The Makran tsunami (3) strongly affects Western India, the Maldives, and the Seychelles, as well as the Kerguelen Islands.
U53A-0041
Geomorphic and seismotectonic segmentation of the Sumatra margin
Upper plate segmentation of the Sumatra trench system is manifested in varying modes of mass transfer as revealed by high-resolution seafloor mapping. The margin segments to the northwest of the Investigator Fracture Zone, which were affected by the 2004 and 2005 megathrust earthquakes, are subject to extensive and complex erosion of the margin wedge. Oversteepening of the lower slope in response to elevated pore pressures and the subduction of pronounced seafloor topography lead to mass wasting processes here. Conversely, neotectonic formation of nascent accretionary thrust folds is limited to the sections of the deformation front southeast of the Investigator Fracture Zone and documents the resumption of frontal sediment accretion in the wake of oceanic relief subduction. The large-scale morphotectonic segmentation of the Sumatra trench system results from subduction of reactivated fracture zones and aseismic ridges of the Wharton Basin and is also reflected in its seismotectonic segmentation as evidenced by the distinct rupture zones of the 2004-2005 earthquake couplet.