2013 UC Ship Funds Proposal: Multi-frequency Imaging of the Sunda Megathrust (MIST) Summary 1. Objective 2. Scientific Rationale

Transcription

1 2013 UC Ship Funds Proposal: Multi-frequency Imaging of the Sunda Megathrust (MIST) Diego Melgar, Emmanuel Garcia, Samer Naif and Robert Petersen Institute of Geophysics and Planetary Physics Scripps Institution of Oceanography Summary We propose a 10 science day cruise to collect acoustical imaging data of the Mentawai patch of the Sunda Megathrust. The Mentawai patch is a seismic gap that has produced large (Mw>8) events in historical times and is the largest unbroken segment of the Sumatran subduction zone. We propose to collect high resolution multibeam bathymetry data across a 400 km long segment of the subduction zone and subbottom CHIRP profiling data on km long trench-perpendicular transects. We expect that the research conducted here can have both short term and long term impacts. It can be used to assess the structural setting of the megathrust by identifying acoustic features that through geomorphic interpretation can be related to the stress conditions of the subduction front. Furthermore shallow stratigraphy of the sediment filled trench and accretionary wedge can be used to infer recent seismic history and behavior. In the long term the cruise data can be used as a pre-event baseline so that when the Mentawai patch produces the next big earthquake, repeat surveys can be conducted to obtain data sets across the same profiles. The value of such repeat surveying and interferometric processing techniques has been demonstrated to constrain the trench-proximal deformation for the Mw9.0 Tohoku-Oki event in a way that land based data cannot. This allows for unparalleled resolution of shallow coseismic slip, which can have a broad impact in understanding the dynamics of the earthquake cycle. We propose a student-led cruise with 4 SIO PhD students functioning as principal investigators. Additionally, we propose that the cruise be conducted as a formal University of California class taught by an SIO investigator and that it include graduate students from the Universities of California Santa Barbara and Santa Cruz in order to fulfill the educational objectives of the UC Ship Funds program. 1. Objective The main objective of the work proposed here is to perform multi-frequency imaging and initiate monitoring of the Mentawai patch of the Sumatra subduction zone offshore Indonesia. We seek to accomplish this by obtaining multibeam and sub-bottom CHIRP profiling data. 2. Scientific Rationale Near Sumatra, the north-south convergence between the Australian Eurasian plates is accommodated mostly by subduction of the ocean plate beneath Sumatra along the Sunda megathrust and horizontal motions along the Sumatran Fault (Fig. 1). After about 170 years of relative quiescence, the Sunda thrust generated a sequence of large earthquakes during the last decade. In 2004, a moment-magnitude (Mw) 9.15

2 earthquake ruptured the plate boundary from Banda-Aceh in northern Sumatra to the Andaman Islands in India (Bilham, 2005; Chlieh et al., 2007; Subarya et al., 2006), generating a tsunami that devastated coastal regions all around the Indian Ocean, taking the lives of more than 200,000 people, and leaving in its wake tremendous economic disruption. The sequence continued southward with the 2005 Mw8.6 Nias earthquake and tsunami, which killed more than 1300 people (Hsu et al., 2006), and the 2007 sequence of Mw7.9 and Mw8.4 ruptures (e.g., Konca et al., 2008) that displaced 30,000 people, creating a risk for disease outbreaks. These great earthquakes also produce a transient change in relative sea level that affects the coastlines of neighboring countries. More recently, the 2010 Mw7.8 Mentawai earthquake ruptured a near-trench patch of the megathrust (Hill et al., 2012; Lay et al., 2011; Newman et al., 2011), generating a large tsunami that devastated the Mentawai Islands, Indonesia, killing more than 500 people. This temporal cluster of large events has increased the likelihood of great earthquakes in adjacent regions. Up to now, the largest unbroken section is the Mentawai segment of the Sunda megathrust, which last ruptured during the 1797 Mw8.8 and the 1833 Mw9.0 earthquakes (Fig. 1). The partial rupture of the Mentawai segment in 2007 and 2009 relieved only a fraction of the slip deficit accumulated since the event that occurred in 1833 (Konca et al., 2008). Currently the greatest and most imminent seismic hazard in the region thus comes from this segment (e.g., Nalbant etal., 2005). However, the variability observed in paleoseismic records from coral microatoll measurements precludes a precise empirical forecast of the next rupture (Sieh et al., 2008).It is of paramount scientific and humanitarian importance to build a better physical understanding of the tectonics at this segment of the plate boundary. The Mentawai patch is an important part of a larger puzzle that can help to elucidate the physics of the earthquake cycle and to shed light on the full seismic potential of the megathrust. Recently Lay et al. (2012) proposed a model for subduction zones with depthvarying frictional properties that could potentially explain the breadth of seismic behaviors typically observed (Figure 2). Persistent frictionally unstable asperities on the megathrust are surrounded by conditionally unstable regions that may or may not contribute to total slip for a given event depending on initial conditions and that will distinguish between large and very large events. At shallower and deeper depths frictionally stable regions behave aseismically or exhibit slow slip and tremor events. Additionally, recent work on the heavily instrumented Parkfield segment of the Central San Andreas fault (Barbot et al., 2012a) has shown that many aspects of the fault evolution, such as the location of microseismicity, the position of hypocenters of large earthquakes and the patterns of surface displacements due to recent earthquakes can be explained in a single simulation. The Sumatran megathrust is incredibly active as evidenced by the many large earthquakes of recent times. This fact coupled with the recent advances in

3 Figure 2: From Lay et al. (2012). Cut-away schematic characterization of the megathrust frictional environment. Regions of unstable frictional sliding are dark regions labeled seismic. Regions of stable or episodic sliding are white regions labeled aseismic. Medium gray areas are conditional stability regions, which displace aseismically except when accelerated by failure of adjacent seismic patches. comprehensive modeling of seismic phenomena make observation of the Mentawai path very appealing as an empirical test of these models. The long term goal of research like this is to ultimately to construct physical models of stress evolution tuned to geophysical and geological observations characterizing the Sunda subduction zone. The model would then allow us to explore what scenarios of future seismicity are possible, based on theoretical knowledge of fault and rock mechanics, and past observations. Several physical models may explain the data equally well while producing different predictions, and this can be addressed by a suite of models that inform us of the possible range of future seismic behavior. By providing quantitative predictions, such as maximum peak ground acceleration, tsunami height, or relative sea-level change, this analysis can form the basis of more informed decisions and better planning for earthquake resilience and survival in the Indian Ocean basin. 3. Scope and Impact of Expected Research Having established why the Mentawai patch of the Sumatra subduction zone is an exciting natural laboratory we describe in detail the expected scope of research. The investigations we plan to conduct hinges on collection of multibeam and sub-bottom profiler data whose scientific impact can be neatly divided into two temporal scales. In the short term we expect to collect high-resolution bathymetry and shallow stratigraphy data for an initial assessment of the prevailing tectonic conditions of the study region. Polonia et al. [2011] demonstrated that CHIRP and high-resolution bathymetry can contribute significantly to understanding the architecture and active deformation of a subduction zone and to better understand seismic hazard by illuminating secondary structures. While their study region (Calabria) has different

4 characteristics than the Sumatra subduction zone, they demonstrated the important constraints that acoustic profiling can provide for structural synthesis. Furthermore, sub-bottom profiling with CHIRP sources can help to assess the recent paleoseismological behavior of important subduction structures (Graindorge et al., 2008; Polonia et al. 2011) or even the contribution of fluid laden sediments to the subduction machinery (Klaucke et al. 2008). To this end and following the Mw9.2 event of 2004, great interest was placed in surveying the Sumatra Subduction zone. Graindorge et al. (2008) conducted multibeam and CHIRP profiling of the northern segment of the Sumatran Subduction zone. From geomorphologic analysis they were able to determine active features on both sides of the trench and map them out with great detail. Singh et al. (2010) employed multibeam data on a small segment of the eastern side of the Mentawai islands to image mass wasting events associated with paleoseismic activity. These studies are important because they provide important controls on the interactions between the downgoing and overriding plates. Furthermore they were able to map structural features on the accretionary wedge and shallow forearc (Figure 2). This is of interest because it remains unclear what role secondary structures on the accretionary wedge and shallow backthrusts play in tsunami generation. For example during the 2011 Mw9.0 Tohoku-oki event it was found that rupture reached the trench (Simons et al., 2011) however an ambiguity remains as to whether the tsunami was generated mostly from slip on the main detachment fault or from inelastic failure on the accretionary wedge structures as well i.e. Ma (2012). This may very well be the case, since for the Tohoku-oki earthquake MacInnes et al. (2013) found that while several different types of coseismic models of fault slip were capable of explaining some of the near and far field tsunami observations, none of them could explain anomalous features such as the large amplitude of the tsunami in north Honshu. Thus we propose to analyze the area around the Mentawai patch, which has not yet been covered with multi-beam and sub-bottom profiler instruments, and perform geomorphic and shallow stratigraphic analysis of the area. Additionally Behrens et al. (2010) found from numerical simulations that fore arc islands, like the Mentawai islands act as natural barriers to near-field tsunami propagation, something that could explain historical observations discussed in Singh et al. (2010) on the high variability of tsunami intensity. If coseismic slip occurs leeward or windward of the islands the ensuing tsunami will have radically different characteristics. Essentially what this means is that the shallower bathymetry between the islands and the trench can exert a first order control on the impending hazard. High-resolution multibeam data can help us better asses it. The proposed research stands to make a large impact on the long term monitoring of the Sumatran subduction zone and in our understanding of the earthquake cycle as well. Despite long and on-going efforts to better document fault slip evolution on the Sunda megathrust and the Sumatran Fault (e.g., Natawidjaja et al., 2007; Prawirodirdjo et al.,2010), it is still notoriously challenging to effectively resolve slip on many locations of the plate boundary. For example, the region near the Sunda trench produces tsunami earthquakes, such as the 1907 Mw7.6 (Kanamori et al., 2010) and the 2010 Mw7.8 Mentawai earthquakes (Bilek et al., 2011; Hill et al., 2012; Lay et al., 2011), that are currently particularly challenging to monitor geodetically in large part because the sources are too far from the current network of available stations on Sumatra and the neighboring islands. Such limitations in observation capabilities result in large uncertainties on the kinematics of deformation and on the models derived from it, no

5 matter how sophisticated. Consider Figure 3A which computes the resolution of a slip inversion using only land based geodetic data using the technique developed in Barbot et al. (2012b), it demonstrates that despite the addition of new geodetic stations to the observation network (Figure 3B), slip close to the trench which is critical to our understanding of rupture mechanics, tsunami generation and hazard assessment will remain unresolved. As such, multibeam data have a very important role to play. Fujiwara et al. (2012) demonstrated via pixel correlation of a repeat survey of the source area of the Tohoku-oki earthquake that multibeam bathymetry can be used to compute displacements of the seafloor (Figure 4). They found seafloor elevation changes as large as 50m, but most importantly that the deformation extended all the way to the trench axis. Thus if indeed the Mentawai patch is the most likely candidate for a large event establishing a pre-event bathymetric baseline is of paramount importance. As evidenced by Figure 3, having information on the amount of deformation in and around the trench will help to develop joint inversions of both ocean and land-based data for a complete picture of the source. In this vein, we propose that repeat surveys of the subbottom profiler data be collected as well, if such interferometric approaches can be used for bathymetric data then research should be conducted into applying them to the shallow stratigraphy imaged by CHIRP. This would go a long way to answering the important question of the contribution of the accretionary wedge in tsunami generation and elucidation of exactly how the shallow part of the megathrust ruptures and contributes to seismic energy budget. For example shallow and deep reflection studies (Singh et al., 2010, 2011) have established that the forearc backthrust is a relevant tectonic feature. However it is perpendicular to the main detachment and thus if it slips coincidentally with the megathrust it is unlikely to be resolved from regional and global seismic data. Bathymetric observations before and after the event could unambiguously determine this. It is unlikely the backthrust system contributes a significant amount to the overall earthquake cycle (Singh et al. 2010) however its shallow and steep dipping nature make it an efficient tsunami generator, therefore understanding its contribution to the subduction mechanism is important. The proposed effort also provides a high-resolution map for selecting suitable locations for possible future seafloor geodetic monitoring that could potentially also provide information about the interseismic period of deformation. Thus the long term effects of the research proposed here can and should be substantive; there is an opportunity not only for education of students but for highimpact science. We posit that with the data collected during this cruise we can begin to attack important science questions and make significant inroads, such as: what are the three-dimensional boundaries of fault segments? Is plate coupling a static property of the thrust, or a feature that evolves with time? Considering the time-dependent evolution of fault slip, is it possible to identify a pattern of slip that leads to nucleation of large earthquakes (Stuart and Tullis, 1996), and assess whether aseismic processes are important elements of earthquake nucleation? Furthermore, when an earthquake strikes, this data set combined with existing geophysical infrastructure and follow up surveys can serve to obtain an image of mainshock, foreshocks and aftershocks with much better resolution, from which we can expect new and exciting discoveries. 2. Data and Instrumentation No extra equipment is needed we expect to utilize the R/V Roger Revelle s EM 120 multibeam and the Knudsen 320B sub-bottom profiler. 3. Cruise Outcomes and Eligibility

6 This project will provide important training for SIO and collaborating graduate students. It will provide hands-on experience with the full life cycle of scientific oceanographic work. From proposal writing through cruise preparation, cruise deployment and collection of data, processing of data, interpretation and paper and report writing. All of the students involved as PI s have a current and continued interest in subduction zones from perspectives that vary from observation to modeling. We all firmly believe that the future of subduction tectonic studies lies in the collection of marine geophysical data. Expected manuscripts to be produced from the cruise can be Figure 4: From Fujiwara et al. (2012) Changes in sea-floor elevation between bathymetric data before and after the 2011 Tohoku-Oki earthquake. (A) Location map with bathymetric survey track shown as yellow line. Coseismic horizontal displacement is estimated over the landward slope indicated by solid portion of yellow line. Cross shows the epicenter. (B) Multibeam bathymetry collected in Red triangles mark the trench axis; the blue triangle marks the landward slope break. Change in sea-floor elevation by subtracting the 1999 bathymetric data from the 2011 data (C), the 2004 data from the 2011 data (D), and the 1999 data from the 2004 data (E). The yellow star marks location of probable submarine landslide. succinctly divided into the short and long term temporal scales discussed above. Immediate work will be for example: Geomorphic characterization of the Mentawai patch of the Sumatra subduction zone from multifrequency acoustic data. Analysis of accretionary wedge structures at the Mentawai patch from acoustic data. And in the longer term manuscripts such as: Surface and sub-bottom deformation observed from repeat acoustic surveys of the 20??

7 Mentawai earthquake. Complete source characterization of the 20?? Mentawai earthquake from joint inversion of land and ocean based geophysical observations. The 20?? Mentawai earthquake. What are the implications for mechanical properties of the subduction megathrust and the earthquake cycle. Furthermore we propose a roster of 4 student PIs and 6 additional graduate students from SIO and 4 additional students from UCSB/UCSC (Section 5) such that the cruise be offered as an official UC class, and we request assistance in funding travel for the participants. Faculty adviser Dave Chadwell (SIO) will function as instructor. With this the 2 highest priority targets of the UC Ship Funds program will be addressed as well as the goal of achieving inter-uc campus collaboration. communication with Earth Science students at SIO, UCSB and UCSC has been met with enthusiasm. We have at least 10 interested students from SIO and 2 from UCSB. Should this proposal be funded we will actively recruit students from our sister institutions to fill these slots. International collaboration is an important facet of the work proposed here as well. The idea for this project nucleated from conversations with colleagues at the Earth Observatory of Singapore (EOS). Scripps alumnus and recently appointed assistant professor at EOS Sylvain Barbot and Emma Hill (also at EOS) and graduate students at EOS and their sister institution the University of Bengkulu are expected to collaborate closely on the project. They have several parallel proposals to increase the density of geophysical observation networks in the region of the Sumatran subduction zone and especially around the Mentawai patch. We expect that EOS will be the regional institution that leads the most significant research efforts into characterization of the earthquake cycle in Indonesia. As such this provides a unique institutional opportunity for the Scripps Institution of Oceanography to obtain a solid foothold and become a major player in a region that promises to produce significant high caliber research in the near future. At a personal level the SIO students involved stand to gain substantially not only from the research and educational experience but from networking with a diverse group of international collaborators. 4. Cruise Plan We have designed a cruise plan comprised of 7 tracks (Figure 5) that contemplates the swath widths determined from the instrument documentation and the water depth. The tracks are approximately 400 km long. Tracks 1-3 have a swath width of 20 km, track 4 of 12 km and tracks 5-7 ok 8 km. The expected time necessary to complete such a survey is 5.5 days plus time to allow for contingencies such as weather and discovery of interesting features, especially at shallower sites. Thus we are requesting 7 days for the multibeam survey. Furthermore we have planned 3 trench perpendicular subbottom profiler (CHIRP) transects. The transects are 100 km long; we are requesting 3 days for that work including transit time between profiles.

8 Figure 1: Tectonic setting. In the last 300 years, patches on the Sunda Megathrust have generated several great to giant earthquakes including the 1797 Mw8.8, the 1833 Mw9.0, the great 2004 Mw9.2 Sumatra Andaman, the 2005 Mw8.6 Nias and the 2007 Mw8.4 ruptures (colored contours). The Sumatran Fault generated at least 17 earthquakes of magnitude greater than 6 since 1850 (stars). The upright triangles indicate the location of the Sumatra GPS Array stations currently in operation.

9 Figure 3: Comparison of the resolution of fault slip with the Sumatra GPS array network in A) the current configuration, and B) with 20 new, optimally located, GPS stations (red crosses). Zero resolution (dark blue) means no sensitivity of the geodetic network, such as near the trench in A). A resolution value of one (white) corresponds to the desirable situation where slip in the area suffers no trade-offs with slip at other locations