U51A-0001
Slip distributions of the southern Sumatra earthquake doublet of September 12, 2007 estimated from teleseismic inversion
The great earthquake of 26 December, 2004, ruptured the plate boundary offshore the Andaman and northern Sumatra Islands. This earthquake and the following tsunami had a devastating impact throughout the Indian Ocean. The Nias earthquake occurred on 28 March, 2005, rupturing the next plate boundary to the south of the 2004 Sumatra earthquake. Recently, a pair of magnitude 8.4 and 7.9 interplate earthquakes occurred on 12 September, 2007, in a region south of the 2005 Nias earthquake. We investigated the slip distributions of the 2007 Sumatra sequential earthquakes by inverting teleseimic body and surface waves. The slip model of the magnitude 8.4 earthquake showed a large slip zone of up to 7m about 100km northwest of the epicenter. The moment magnitude was estimated to be 8.47 which was very similar to that obtained by U.S. Geological Survey. In the magnitude 7.9 earthquake model, a significant slip patch of about 3m was imaged at the northern end of the slip region of the magnitude 8.4 earthquake. The magnitude 7.9 slip area did not reach the southern end of the plate boundary ruptured by the 2005 Nias earthquake. This indicates that there may be a seismic gap of about 120km between the 2005 Nias and the 2007 sequence. The total slip zone of these models was consistent with the 1833 rupture zone estimated by using coral data, which extended from approximately Eggano Island to the southern portion of Siberut Island. However, the amount of slip of the pair of the 2007 earthquakes was slightly smaller than that of the 1833 model.
U51A-0002
Preliminary Numerical Simulations of the September 12, 2007 Southern Sumatra Tsunami: Forward Modeling and Comparison With Publicly Available Tsunami Data
An earthquake of magnitude Mw=8.4 (USGS source) was registered on September 12, 2007 offshore the southern coasts of Sumatra, Indonesia. The epicenter was located about 130 km offshore the city of Bengkulu. The earthquake generated a tsunami which, according to some preliminary surveys posted on the Internet by the Indonesian BMG, produced significant damage especially in a region north of Bengkulu, with maximum run-up of 3.6 m in Muko-Muko. The tsunami was also recorded by a number of "tsunameter" stations (including DART) all over the Indian Ocean, including some stations along the coast of Sumatra itself, like Padang. The aim of the present study is to perform some forward modeling of the tsunami and to try to put some constraints on the position and geometry of the causative fault. We basically follow a trial-and-error procedure by adopting some initial fault models, all sharing the same magnitude (provided by the Harvard Moment Tensor Solution), but being different as regards the position and the geometry, including possibly some slight variations of the focal parameters with respect to the solution provided by the Harvard CMT. For each fault choice, we simulate the ensuing tsunami and compare the obtained results with the available experimental data, and in particular with the available tide-gauge records in the Indian Ocean and the run-up measurements that will possibly be collected in the close future. As for the models adopted, we formulate the simple hypothesis that the tsunami initial condition coincides with the coseismic vertical displacement component of the ocean floor, which in turn is computed by means of the classical elastic half-space approach. The propagation of the tsunami is simulated through the numerical finite- differences code UBO-TSUFD, developed by the Tsunami Research Team at the University of Bologna (Italy), which solves the linear Navier-Stokes equations in the shallow-water approximation and in spherical coordinates, and is especially suited to study the tsunami propagation features in the open ocean. To study the tsunami impact close to the source region, and in particular the distribution of observed run-ups, we use the finite-element code UBO-TSUFE, developed by the same research team.
U51A-0003
Rupture process of the September 12, 2007 Southern Sumatra earthquake from tsunami waveform inversion
The aim of this work is to infer the slip distribution and mean rupture velocity along the rupture zone of the 12 September 2007 Southern Sumatra, Indonesia from available tide-gauge records of the tsunami. We select waveforms from 12 stations, distributed along the west coast of Sumatra and in the whole Indian Ocean (11 GLOSS stations and 1 DART buoy). We assume the fault plane and the slip direction to be consistent with both the geometry of the subducting plate and the early focal mechanism solutions. Then we subdivide the fault plane into several subfaults (both along strike and down dip) and compute the corresponding Green's functions by numerical solution of the shallow water equations through a finite difference method. The slip distribution and rupture velocity are determined simultaneously by means of a simulated annealing technique. We compare the recorded and synthetic waveforms in the time domain, using a cost function that is a trade-off between the L1 and L2 norms. Preliminary synthetic checkerboard tests, using the station coverage and the sampling interval of the available data, indicate that the main features of the rupture process may be robustly inverted.
U51A-0004
Imaging of the 2007 Pagai Earthquake with Back-Projection of the Hi-net Data
The location and geometry of the 2007 September 12th Pagai earthquake in Sumatra are well suited for application of the back-projection method using the Japanese High-Sensitivity Seismograph Network (Hi-net) array. This technique allows imaging of the relative strength of high-frequency radiated energy as a function of space and time. The 2007 Pagai earthquake appears to have ruptured unilaterally at a nearly constant speed of about 2.9~km/s toward the northwest from the epicentre. The event lasts for about 100~seconds, and the estimated total rupture area is consistent with that expected for magnitude~8.4. The inferred rupture area also agrees well with the aftershock distribution except for a cluster of events around Sipora island. This sequence starts about a day after the magnitude~8.4 event, and includes an event with magnitude~7.0. A heightened level of seismicity in this region was also observed after the 2005 Simeulue-Nias earthquake. The locations of these events and the gap of seismicity between this cluster of events and the rupture area of the 2007 earthquake suggest that they may be two separate sequences. The rupture of the magnitude~8.4 Pagai earthquake starts at the northwest edge of the slip region of the 2000 magnitude~7.9 Enganno earthquake, and terminates near the southern boundary of the slip region of the 1797 earthquake. The 2007 event broke an area that had significant slip during the 1833 earthquake, but did not cover its entire extent. It is also in good agreement with locations of strong seismic coupling along Sumatra that have been obtained from paleoseismic and geodetic data. The series of large earthquakes in Enganno (2000, M~7.9), Sumatra-Andaman (2004, M~9.3), Simeulue-Nias (2005, M~8.6), and Pagai (2007, M~7.9) cover most of the area along the Sunda trench from -7°S to 15°N. However, there are two gaps, both in ruptured area and seismicity. One covers a region between the Pagai and Sipora islands which is thought to have broken twice during the 1797 and 1833 earthquakes. Another is an area northwest of Sipora island, with the last giant earthquake occurring in 1797.
U51A-0005
The 12 September 2007, Mw=7.9 Pagai-Sipora Earthquake, an Impulsive Rupture on the Central Sumatra Megathrust.
On 12 September 2007, at 23:40 UT, a large Mw=7.9 earthquake occurred offshore central Sumatra, 12 hours after a Mw=8.4 rupture that initiated about 200 km to the southeast. We analyze this event using teleseismic P and SH waveforms, together with GPS co-seismic displacements measured at cGPS stations from the the SuGAr network. The SuGAr network includes 5 receivers on the outer arc islands of Pagai and Sipora, in addition to stations on Sumatra mainland therefore providing an excellent coverage of the rupture area. Moreover, the high sampling rate of the stations (1 to 15 s sampling for most stations), allows to properly identify the respective contributions of the Mw=8.4 and major nearby aftershocks. Both seismic waveforms and the geodetic data indicate a moment release of about 1.1e21 N.m due to the rupture of two distinct asperities about 130 km apart along the subduction interface. The seismic waveforms require the first subevent to be extremely impulsive, with a short rise time of a few seconds at most, and a highly peaked slip distribution close to the epicenter. The geodetic and seismic data were inverted jointly to derive some kinematic model of the source that is found to reconcile well the two datasets. The first subevent is located near the eastern border of South Pagai Island with a maximum slip of 5.5 m released in less than a few seconds. The slip distribution due to that event abuts with the slip distribution due to the Mw=8.4 event. The second patch is located 130 km to the northwest, along the north- eastern coast of Sipora Island, with a maximum slip of 3.5 m. The rupture velocity is estimated to be about 2 km/s. Both the 2007 Mw=8.4 and Mw=7.9 events occurred within the estimated rupture area of the Mw>8.6 historical earthquake of 1833, a well defined patch that had remained locked during the interseismic period. A same portion of the megathrust can rupture in different ways. http://tectonics.caltech.edu/slip_history/
U51A-0006
Effect of The Heterogeneous Structure Near The Source on The Teleseismic Body Waveforms From The Sunda Trench Earthquakes Analyzed by a 2.5D Finite-Difference Method
Teleseismic body waveform analysis is one of the major methods in the source process study of large earthquakes. It is also possible to relocate small earthquakes by using teleseismic waveforms to provide improved distribution of earthquakes (e.g., Okamoto 1994). For shallow subduction zone earthquakes, it is quite important to consider the effects of the laterally heterogeneous structure near the source on the teleseismic body waveforms: the thick ocean (water) layer with dipping ocean bottom and the thick sediments with low seismic wave velocity often cause very large effects which are not reproduced by flat-layered model structure (e.g., Wiens 1989; Yoshida 1992; Okamoto 1993). In this paper, we study how the lateral heterogeneity in the Sunda trench affect the teleseismic waveforms and discuss the effect on the study of the source process of the recent large earthquakes on 12 September 2007. We preliminary compared the observed waveforms and 1D synthetics computed for a flat-layered model structure from a shallow event (2007/09/14 06:01 Mw6.2) that is one of the aftershocks closest to the trench axis. Relatively large later phases were observed at stations located close to one of the nodal lines (i.e., stations at about 180 to 280 degrees in azimuth): these phases were not reproduced by a 1D model structure near the source assumed based on the laterally varying crustal model obtained by a detailed seismic experiment (Kopp et al. 2001). This suggests that the lateral heterogeneity near the source is the likely cause of the later phases. The structural effect appears to be minor on the waveforms at stations far from the nodal lines. We will present analysis using teleseismic waveform modeling with a 2.5D finite-difference method (Okamoto 1994; Takenaka and Kennett 1996) which incorporates the lateral heterogeneity in the Sunda trench.
U51A-0007
The Comparison of Earthquake Rupture Processes along Sumatra-Andaman Subduction Zone Revealed by CNDSN
Following the occurrence of September 12, 2007 earthquake, more and more seismologists keep eye on the Sumatra-Andaman subduction zone because a sequence of great earthquakes in this particular region have occurred during the past four years. Back-projection method by Ishii et al. (2005) is very successful in imaging the progression of the transient rupture front of the 2004 great Sumatra earthquake. This paper we present our results based on the newly built China National Digital Seismic Network (CNDSN) by using the Back-projection method. CNDSN consists of 48 real-time transmission broadband seismic stations at 500-600 km spacing. Despite of being sparse, it can play an equivalently important role as other dense regional networks (e.g., Hi-Net in Japan and GRSN in Germany) in deciphering earthquake rupture processes with a specifically designed imaging condition. Our results showed that: (1) those earthquakes had extra long duration of P wave indicating long rupture length; (2) among all the source parameters, the moment magnitude or its equivalent, seismic moment, source mechanism, focal depth as well as the earthquake rupture process, the last one plays a key role in the tsunami generation process. We carried out a comparison study on the characteristics of the rupture process of the tsunamigenic earthquakes occurred from 2004 to 2007. This research fits for the urgent need of studies on the rupture process for prevention and mitigation of tsunami disaster.
U51A-0008
Post-Seismic surface deformation studies between India and Burma and the Indian plate kinematics by GPS-Geodesy due to recurrent earthquakes off the coast of Sumatra.
The recurring earthquakes off the Sumatra coast even after the devastating earthquake of December 26, 2004, the studies on the influence of these earthquakes on the deformation between India and Burma and the Indian plate kinematics assume a greater urgency. The earlier studies that have been independently carried out using GPS –Geodesy and in a global network solution requires reevaluation. After the December 26, 2004 earthquake, the global network that was formed earlier consisting of 15 IGS GPS stations in and around India encompassing all the plates surrounding India, and having HYDE and IISC in the central India as the core stations are being reassessed now after the September 12 2007 earthquake. Earlier 18 days of GPS data including before and after the earthquake were used in the global network analysis using Bernese software version 4.2 in ITRF-2000 Reference Frame. The baseline lengths from Hyderabad to other chosen sites and the rate of changes were also estimated. Our results show a significant shift of 7.3 mm for HYDE and 7.9 mm for IISC respectively, the stations, which are of primary importance as far as Indian plate kinematic studies are concerned. These two sites which are located in the southern hemisphere of India and around the Northern Indian Ocean show an increasing trend of convergence towards NE and specifically towards the Burmese plate that's corroborated by Vigney C. et al., Nature, July 2005. All these already estimated results are being reprocessed in ITRF-2005 Reference Frame by including a larger data set from 1995 to till date. Our earlier estimation also showed that the stations in Eurasia remain almost constant with time and in essence the whole Indian peninsula has moved by about 7.3mm towards east longitude and Burmese plate, which are also being reevaluated.
U51A-0009
Status of GPS Based Investigations on the Recent Sequence of Earthquakes on the Sumatran Trench
Within only a few years, almost all the length of the Sumatran trench accommodating the Indo- Australian/Sundaland convergence ruptured. After the giant earthquake of December,26 2004 and the march, 28 Nias earthquake, the last sequence of earthquakes along the Mentawai Islands (12 september) broke the remaining gap. According to the distribution of aftershocks, it is not completely clear whether the last sequence of earthquakes reproduced the 1833 rupture (Sieh, 2006), i.e. leaving an untouched section of 100-200 km roughly in front of the Siberut Island and city of the Padang (capital of west Sumatra). Immediately after 09/12, we re- occupied with GPS a network of 30 benchmarks we had installed in the Padang area and surveyed initially in June 2007. From these measurements, we expect to be able to constrain the rupture of the earthquake, and especially to quantify the co-seismic slip near the northern end of the rupture. We also installed 4 cGPS stations along the Sumatran coast around Padang to monitor post-seismic deformation in this area, and help determine if this area slipped or is still accumulating strain for a future earthquake. Along with GPS results, we will show the role played by Coulomb stress, shear stress and normal stress increase, as well as pore pressure modification, on past and future earthquake triggering; on the trench itself and on the Great Sumatra Fault, the continental structure that accommodates the strike-slip component of the oblique plate convergence behind the trench.
U51A-0010
Stress, triggered earthquakes, and modelled tsunamis on the Sunda Trench
The current earthquake sequence off the coast of Sumatra illustrates both the potential, and the difficulties, of using stress based models to assess the likely location of future large earthquakes. Following the 26 December 2004 event, we showed that stress had increased on the Sumatra Fault as well as further south along the subduction zone, and published concerns that both structures had the potential to experience large triggered earthquakes. The 28 March 2005 event occurred in the area of concern on the Sunda Trench although the Coulomb stress changes at the hypocenter were only on the order to 0.1 bar. The 2005 earthquake, in turn, increased stress along a large extent of the Sumatra Fault as well as further south on the subduction zone. The latter is of particular concern as paleogeodetic data show that the last large event under the northern Mentawai Islands was in 1797, while to the south the trench last ruptured in 1833. In order to assess the tsunami hazard in the region we modelled 100 possible earthquakes, and their tsunamigenic potential, along the subduction zone; all models had the common feature that they involved slip under Siberut Island in the northern Mentawais. On 12 September 2007, an M=8.4 earthquake occurred along the Sunda Trench. The hypocenter was south of the Mentawai islands and its location had not experienced a significant stress increase from the 2005 event. The earthquake re-ruptured a portion of the trench that failed in 1833 but did not propagate as far north as Siberut Island. Here we present an updated view of the Coulomb stresses along the Sunda Trench and discuss the potential for a further large event under Siberut Island; this area has now been loaded from both the north and the south. Additionally, we show results for modelled tsunamis resulting from possible future earthquakes in this region.
U51A-0011
Modeling Stress Changes Following the 2004-2005 Sumatra Earthquake Sequence: Exploring Triggering of the 2007 Rupture and its Relationship to the 1797 and 1833 Events
The September 2007 Sumatra M8.4 earthquake initiated in the southern section of the 1833 rupture segment. Twelve hours later a deeper M7.9 aftershock ruptured further to the north. Their occurrence, close in time and space to the 2004 M9.2 and 2005 M8.7 Sumatra earthquakes, suggest the possibility of these being triggered events. Earlier studies of vertical motion, derived from coral growth histories, suggest that interseismic strain accumulated along the 1833 segment has approached levels relieved in the historic earthquake. Along the 1797 segment the accumulated interseismic strain appears to have exceeded previously relieved levels. Following the 2004 and 2005 events, investigations of coseismic and viscoelastic deformation show that the 1797 segment had higher Coulomb failure stress changes than the 1833 segment. Therefore, it is puzzling that the 1797 and northern 1833 rupture zones did not recur before the southern 1833 segment. We model additional postseismic processes, including afterslip and poroelastic rebound, in an attempt to find scenarios that can explain why the southern 1833 segment was triggered prior to the 1797 segment.
U51A-0012
The Sumatra September 2007 Sequence of Large Earthquakes and Tsunami Warning
The Sumatra sequence of large earthquakes in September 2007 started on 12 September with three earthquakes of magnitudes 8.4, 7.9 and 7.1 occurring within 15 hours along a stretch of the Sumatra coastline about 500 km long. The M 8.4 event generated a small tsunami of about 1 m at the nearby city of Padang. The damage was more from the earthquake shaking than from tsunamis. The three earthquakes acted as three main shocks ruptured three parts of the Sunda Arc. Each has its own aftershocks. The three aftershock areas add to be about equivalent to the size of the aftershock area of the March 28, 2005 Simeuleu-Nias earthquake of M 8.7 to the north of this area. Before this sequence and after the December 2004 M 9.2 Sumatra earthquake, there were also the March 28, 2005 M 8.7 Simeuleu-Nias earthquake and the July 17, 2006 M 7.7 Java earthquake. Based on the high seismicity of large earthquakes, many believe that more "big" ones are likely. This study attempts to analyze the seismicity and characters of each of the large earthquakes using seismic arrays. Results will be presented.
U51A-0013
Social Uptake of Scientific Understanding of Seismic Hazard in Sumatra and Cascadia
The importance of science within hazard mitigation cannot be underestimated. Robust mitigation polices rely strongly on a sound understanding of the science underlying potential natural disasters and the transference of that knowledge from the scientific community to the general public via governments and policy makers. We aim to investigate how and why the public's knowledge, perceptions, response, adjustments and values towards science have changed throughout two decades of research conducted in areas along and adjacent to the Sumatran and Cascadia subduction zones. We will focus on two countries subject to the same potential hazard, but which encompass starkly contrasting political, economic, social and environmental settings. The transfer of scientific knowledge into the public/ social arena is a complex process, the success of which is reflected in a community's ability to withstand large scale devastating events. Although no one could have foreseen the magnitude of the 2004 Boxing Day tsunami, the social devastation generated underscored the stark absence of mitigation measures in the nations most heavily affected. It furthermore emphasized the need for the design and implementation of disaster preparedness measures. Survey of existing literature has already established timelines for major events and public policy changes in the case study areas. Clear evidence exists of the link between scientific knowledge and its subsequent translation into public policy, particularly in the Cascadia context. The initiation of the National Tsunami Hazard Mitigation Program following the Cape Mendocino earthquake in 1992 embodies this link. Despite a series of environmental disasters with recorded widespread fatalities dating back to the mid 1900s and a heightened impetus for scientific research into tsunami/ earthquake hazard following the 2004 Boxing Day tsunami, the translation of science into the public realm is not widely obvious in the Sumatran context. This research aims to further investigate how the enhanced understanding of earthquake and tsunami hazards is being used to direct hazard mitigation strategies and enables direct comparison with the scientific and public policy developments in Cascadia.
U51A-0014
Modeling Indian Ocean Tsunami Propagation and Inundation with TsunamiClaw---Adaptive Finite Volume Methods.
TsunamiClaw is a freely available software package that we have been developing to model free surface depth- averaged flows such as the shallow water equations for tsunami modeling. The underlying numerical methods are based on adaptive finite volume methods for wave propagation that we extended to tsunami modeling prior to the December 2004 Indian Ocean Tsunami. The adaptive Cartesian gridding allows global-scale teletsunamis and local inundation to be modeled in single and efficient computations by using multiple evolving grids to track waves. Though adaptive mesh refinement presents some new difficulties, we believe that it is a powerful solution to the problem of disparate spatial and temporal scales exhibited by tsunamis, and it allows large as well as detailed simulations on small personal computers. \par The initialization of the tsunami is generated by dynamically moving the seafloor bottom during the computation according to spatial temporal models of fault displacement for a given seismic event. \par We have modeled Indian Ocean tsunamis for multiple fault models of different Sumatran seismic events. Various simulations will be shown demonstrating tsunami propagation and inundation with different fault models. Additionally some case studies of local inundation will be shown. We hope that this will provide comparison with other numerical models as well as field studies.
U51A-0015
Modeling Ruptures and Tsunamis That May Follow Event of September 12
The seismic series of the 12-13 September near the Padang/Bengkulu region in Western Sumatra, Indonesia, emerged to, but did not rupture the three locked patches at the subduction plate interface as recently derived from geodetic and paleogeodetic studies (Chlieh et al., 2007), likely increasing probability of the next giant earthquake off Padang. Future earthquake, when rupturing all of the three locked patches, may reach a magnitude of about M=9 and repeat giant historical event of 1797 and partly that of 1833 followed by severe tsunamis. We have modeled possible earthquake and tsunami scenarios assuming future ruptures coinciding with each of the presently locked patches. Our modeling is performed in a framework of the currently developed German- Indonesian Tsunami Early Warning System (GITEWS). Resulting tsunami wave height in Padang is very sensitive to the location of the ruptured patch. The shallow patch under the Siberut island and the deep patch just off Padang, both result in M=8.4 earthquakes. However, the latter rupture would result in considerably shorter arrival time and in more than 5 times higher run-up in Padang. We show that GPS stations at islands and at Padang can be used to estimate rupture parameters and to predict tsunami heights just after the earthquake. Our modeling also demonstrates that historical records about the 1797 tsunami are more consistent with the simultaneous rupturing of both the shallow and deep patches off Padang.
U51A-0016
Thermal Models of the Sumatra Subduction Zone: Implications for the Megathrust Seismogenic Zone
We have developed 2-D finite element models for the thermal regime across the Sumatra subduction zone and temperatures on the subduction thrust fault. The main objective is to examine thermal controls of the downdip extent of great earthquake rupture. The rupture updip and downdip limits are important both for tsunami generation and earthquake shaking. However, another important application is the dehydration and upward fluid expulsion from the downgoing slab with increasing temperature and pressure. Two recent well-studied M9 great earthquakes occurred in 2004 and 2005, the Sumatra-Andaman and the Nias earthquakes, respectively. Finite element thermal models on three 2-D profiles are developed, allowing for smoothly varying subduction dip, thermal properties of the rock and sediment units, frictional heating, the thermal regime of the incoming plate, etc. The common updip thermal limit to seismic behaviour of 100-150°C occurs directly at the trench. The common downdip limit of 350-450°C occurs globally at depths of 30 to 50 km, 200 to 220 km inland from the trench. An alternate downdip seismogenic limit, the intersection of the subduction thrust with the aseismic forearc mantle, is poorly constrained but is estimated to be at a depth of ~30 km. For a profile 50 km north of Sumatra 350 and 450°C on the thrust are well below the forearc mantle intersection and the latter is inferred to apply. For two profiles through Sumatra, 350°C occurs on the thrust near the forearc Moho with the transition zone to 450°C extending to greater depths. Either the thermal or forearc Moho limits therefore may apply for this region. The maximum downdip limit of rupture in the two great earthquakes is not well determined, but most estimates are at a depth of ~30 km, which supports the idea that the intersection of the downgoing plate with the aseismic forearc Moho is governing the downdip limit of the seismogenic zone in northern Sumatra, ~ 160-200 km landward from the trench. The updip limits of rupture are predicted to be at the trench in the two northern profiles and about 15 to 20 km inland from the trench in the southernmost profile. These values fit well with results of tsunami models.
U51A-0017
Likely Human Losses in Future Earthquakes in Central Myanmar, Beyond the Northern end of the M9.3 Sumatra Rupture of 2004
We estimate that the city of Rangoon and adjacent provinces (Rangoon, Rakhine, Ayeryarwady, Bago) represent an earthquake risk similar in severity to that of Istanbul and the Marmara Sea region. After the M9.3 Sumatra earthquake of December 2004 that ruptured to a point north of the Andaman Islands, the likelihood of additional ruptures in the direction of Myanmar and within Myanmar is increased. This assumption is especially plausible since M8.2 and M7.9 earthquakes in September 2007 extended the 2005 ruptures to the south. Given the dense population of the aforementioned provinces, and the fact that historically earthquakes of M7.5 class have occurred there (in 1858, 1895 and three in 1930), it would not be surprising, if similar sized earthquakes would occur in the coming decades. Considering that we predicted the extent of human losses in the M7.6 Kashmir earthquake of October 2005 approximately correctly six month before it occurred, it seems reasonable to attempt to estimate losses in future large to great earthquakes in central Myanmar and along its coast of the Bay of Bengal. We have calculated the expected number of fatalities for two classes of events: (1) M8 ruptures offshore (between the Andaman Islands and the Myanmar coast, and along Myanmar's coast of the Bay of Bengal. (2) M7.5 repeats of the historic earthquakes that occurred in the aforementioned years. These calculations are only order of magnitude estimates because all necessary input parameters are poorly known. The population numbers, the condition of the building stock, the regional attenuation law, the local site amplification and of course the parameters of future earthquakes can only be estimated within wide ranges. For this reason, we give minimum and maximum estimates, both within approximate error limits. We conclude that the M8 earthquakes located offshore are expected to be less harmful than the M7.5 events on land: For M8 events offshore, the minimum number of fatalities is estimated as 700 ± 200 and the maximum is estimated as 13,000 ± 6,000. For repeats of the historic M7.5 or similar earthquakes, the minimum is 4,000 ± 2,000 and the maximum is 63,000 ± 27,000. An exception is a repeat of the M7.5 earthquake of 1895 beneath the capital Rangoon that is estimated to have a population of about 4.7 million. In the case of a repeat of the 1895 event, a minimum of 100,000 and a maximum of 1 106 fatalities would have to be expected. The number of injured can in all cases be assumed to equal about double the number of fatalities. Although it is not very likely that the 1895 event would be repeated in the same location, it is clear that any medium to large earthquake in the vicinity of Rangoon (at a distance similar to the M7.2 earthquake of May 1930) could cause a major disaster with more than 10,000 fatalities. In spite of the uncertainties in these estimates, it is clear that the capital of Myanmar, and the provinces surrounding it, will likely experience major earthquake disasters in the future and the probability that these could occur during the next decades is increased. We conclude that major efforts of mitigation, using earthquake engineering techniques, and preparation for seismological early-warning capabilities should be undertaken in and near Rangoon, as well as in other cities with more than 100,000 inhabitants (e.g., Phatein, Bago and Henzada).
U51A-0018
Is A Great Earthquake Imminent Or Not?
For the earthquake prone Indonesia, the question about whether there will be a big earthquake is probably insignificant. Instead, a more pertinent question should be when it will happen. In the aftermath of the 2004 M9 Sumatra-Andaman Sea Earthquake, the M8.4 earthquake that occurred on September 12th, 2007 has certainly brought some panic. Because this earthquake occurred within a so-called gap where some great historical earthquakes occurred several hundreds of years ago, a very natural question is raised: Is a bigger earthquake treading the heel of this event? The author is trying to answer this question using the LURR method. By examining the temporal and spatial pattern of LURR evolution in Sumatra region, significant LURR anomalies appeared not only prior to the 2004 mainshock, but also prior to the 2006 M7.7 South Java earthquake. Moreover, all the earthquakes larger than M7.0 after the 2006 Java earthquake (2006.7.17 M7. 7) occurred in the predicted zone in the PAGEOPH paper which submitted and accepted before the recent earthquakes occurred (also see the follwing web site). Based on the currently availble data, we find that a bigger earthquake is unlikely to occur in the near future. Although the LURR temporal pattern shows high anomaly before the mainshock of the M8.4 earthquake, the LURR value has dropped since the mainshock, indicating the stability in this region is being restored. On the other hand, the spatial pattern shows the area around the 2004 M9 earthquake has higher LURR value than other areas, indicating that the northern part of Sumatra is still the most unstable. Because of insufficient data in this region, the result and the conclusion are preliminary and subject to further revision when more data is available. http://www.esscc.uq.edu.au/~xing/tidal-deformation.htm
U51A-0019
Numerical Modeling and Forecasting of Strong Sumatra Earthquakes
ESyS-Crustal, a finite element based computational model and software has been developed and applied to simulate the complex nonlinear interacting fault systems with the goal to accurately predict earthquakes and tsunami generation. With the available tectonic setting and GPS data around the Sumatra region, the simulation results using the developed software have clearly indicated that the shallow part of the subduction zone in the Sumatra region between latitude 6S and 2N has been locked for a long time, and remained locked even after the Northern part of the zone underwent a major slip event resulting into the infamous Boxing Day tsunami. Two strong earthquakes that occurred in the distant past in this region (between 6S and 1S) in 1797 (M8.2) and 1833 (M9.0) respectively are indicative of the high potential for very large destructive earthquakes to occur in this region with relatively long periods of quiescence in between. The results have been presented in the 5th ACES International Workshop in 2006 before the recent 2007 Sumatra earthquakes occurred which exactly fell into the predicted zone (see the following web site for ACES2006 and detailed presentation file through workshop agenda). The preliminary simulation results obtained so far have shown that there seem to be a few obvious events around the previously locked zone before it is totally ruptured, but apparently no indication of a giant earthquake similar to the 2004 M9 event in the near future which is believed to happen by several earthquake scientists. Further detailed simulations will be carried out and presented in the meeting. http://www.scec.org/aces2006/