The Great Sumatra-Andaman Islands Earthquake and Tsunami of 26 December 2004 II
Presiding: J Park, Yale University; S Bilek, New Mexico Institute of Mining and Technology
U44A-01 15:30h
High Frequency Radiation from the 2004 Great Sumatran Earthquake
The December 26, 2004 Great Sumatran earthquake (Mw=9) is one of the largest earthquakes ever recorded. Here we report on its high frequency (HF) radiation (2 to 4 Hz) that lasts for nearly 500s with a clear azimuthal pattern, 400s in the direction of rupture (azimuth~140°) to 590s in the opposite direction. These estimates can be used to estimate a fault rupture length of 1200 km which is comparable to the length of aftershock distribution . The duration of this earthquake is the longest ever reported, even substantially longer than that of the great Chilean earthquake (340s, Houston and Hiroo, 1986)
U44A-02 15:45h
T waves from the 2004 Sumatra earthquake and its aftershocks
At the time of writing (early February 2005), more than 75 CMT solutions are already available for aftershocks of the great 2004 Sumatra earthquake. The mainshock and most of the aftershocks generated hydroacoustic (T) waves in the Indian Ocean, recorded in particular at the IRIS station at Diego Garcia. We have started to analyze systematically their records, and to compile a database of the parameters GAMMA introduced by Okal et al. [2003, scaling the T-phase energy flux to the seismic moment. Similarly, we compute the duration-amplitude discriminant D introduced by Talandier and Okal [2001]. Both GAMMA and D have been shown to be related to slowness (or snappiness) of the source, and we will explore their possible correlation with epicentral location and focal mechanism since the aftershock database comprises thrust, normal and strike-slip solutions.
U44A-03 16:00h
Ground motion excited by the Tsunami of 26 December 2004: Implications for tsunami warning
Approximately 70 minutes after the great Sumatra-Andaman Islands earthquake of 26 December 2004 and 40 to 130 minutes before the tsunami's arrival, the ground at three island stations in the Indian Ocean started to shake at periods comparable to those of the tsunami waves (~1000 s). The characteristics and timing of these very-long period waves are not those of the normal seismic phases generated by earthquakes, and strongly suggest ground deformation excited by the tsunami waves in a way similar to the generation of seismic waves by the oscillation of a heavy mass pressed against the ground, a technique often used in geophysical exploration. The arrival times of these anomalous waves are consistent with the paths of tsunami-excited seismic waves, along which the tsunami slammed into the east coast of the Indian Ocean and generated seismic waves, which then propagated in the solid earth at a velocity higher than that of the tsunami to the stations in the Indian Ocean. The observation of the same very-long period wave at Hainan, China, at the time expected for the tsunami-excited seismic waves to propagate from the west coast of northern Sumatra, reconfirms our interpretation. We can rule out the possibility that the anomalous long-period waves are the fundamental-mode wave and overtones propagating along the major arcs, because the predicted times are opposite to the observation. Additional studies are required to further understand the coupling between tsunami waves and the solid earth and the manner in which that knowledge might be used to reduce the loss and suffering in future tsunamis.
U44A-04 16:15h
Velocity and Spectral Analysis of Tide Gauge Records of the Indian Ocean Tsunami of December 26, 2004
Digital and analog tide gauge recordings of the Indian Ocean tsunami generated by the great Sumatra earthquake of December 26, 2004 were analyzed to estimate the tsunami's speed of propagation and spectral content. Digital recordings at a sampling rate of 1 sample per minute were obtained from the Australian National Tide Center; the tsunami arrivals were clearly recorded at its stations at Cocos Keeling Island (1702 km from the epicenter) and Hillary's Boat Harbor (4418 km from the epicenter). Spectral analyses of these signals indicate a primary tsunami period of 13.9 minutes at Cocos; this spectral line plus longer period lines up to 50 minutes are seen at Hillarys Harbor. Scans of analog tide gauge records were obtained from the Indian National Institute of Oceanography. These scans were enlarged to a resolution of 0.5 inches per hour and the first breaks of the tsunami arrivals were measured to within 5 minutes. Ten time-distance pairs across the Indian Ocean were used to obtain a tsunami propagation velocity of 657 km/hour. The theoretical value of the tsunami velocity for the average water depth in the Indian Ocean (3.4 km) is 653 km/hour.
U44A-05 16:30h
Down-depth Seismicity Gaps and the Shape of the Seismic Zone along the entire Indonesian arc from Relocated Hypocenters
Using thousands of handpicked P, S, pP, sP, PcP, and ScP phases from digitally recorded seismograms, together with International Seismological Centre reported phases, we obtain improved hypocentral locations for ~2600 earthquakes deeper than 50 km with mb > ~5.0 earthquakes, for the period 1962 to September 1996, along the Indonesian subduction zone. The seismicity distribution is found to be very non-uniform both along the arc and in depth. Gaps in the relocated hypocenters exist along depth in most places of the arc, with its upper edge varying from 100-450 km depth and its lower edge from 350-670 km in different portions of the arc. The relocated hypocenters show that (1) a portion of the Indonesian arc between ~110°E and 123°E longitude, and deeper than ~500 km, is dipping southward at an angle of ~75°, that is, in a direction opposite to the upper part of the north dipping slab, suggesting southward lateral flow in the mantle, relative to the plate motion vector here. (2) East of about 108°E, the seismic zone is wider near 670 km than near the 500 km depth. (3) The seismic zone between 129--131°E in the 100--200 km depth range is the widest along the arc both in strike and downdip. This region, near the highest arc curvature, has the highest seismic activity, and is the only part of the arc with earthquakes continuously occurring from the surface down to below 600 km. (4) The very deep earthquakes under Sulawesi are shown to be part of the west-southwest dipping Seram slab. (5) In the westernmost part of the Banda arc, the slab is under downdip tension in the 50-250 km depth range, while the deepest portion of the slab in this region is under compression. From 128-131°E the slab between 100--200 km depth is under mainly horizontal compression. Our study supports the "two-slab" model for the Banda arc. 3-D computer animations of the subduction zone will be presented. Das, S. (2004) Seismicity Gaps and the Shape of the Seismic Zone in the Banda Sea Region from Relocated Hypocenters,J. Geophys. Res., 109, doi:10.1029/2004JB003192. H.-J. Schöffel and S. Das (1999). Fine details of the Wadati-Benioff zone under Indonesia and its geodynamic implications, J. Geophys. Res., 104, 13101-13114.
U44A-06 16:45h
Structural Setting and Seismicity in the Vicinity of the Great Sumatra-Andaman Islands Earthquake
The earthquake of 26 December 2004 was a thrust event that occurred at the interface between the Burma/Andaman Plate and the Indian Plate. We discuss the seismic and thermal structure of the upper mantle in the vicinity of this earthquake revealed by modeling broad-band surface wave dispersion. The upper mantle model is constructed on a 1 deg by 1 deg grid across a broad region centered on the Sumatra-Andaman Islands earthquake. Seismic and thermal parameterizations are used in the Monte-Carlo inversion resulting in both a shear velocity and temperature model of the upper mantle to a depth of about 250 km. Several issues are discussed: (1) the thermal age of the incoming Indian plate, (2) the ability to resolve slabs subducting at the Andaman, Sunda, and Java trenches, (3) the extent and geometry of the subducting plate, (4) thermal anomalies associated with the Andaman Spreading Zone, and (5) the possibility of high temperature (low velocity) anomalies in the supra-slab mantle wedge that may be associated with slow moment release down-dip of the rupture zone. Slow moment release may be responsible for discrepancies between moment release estimated with the gravest normal modes and with long period surface waves. Finally, historical and recent seismicity is presented in the context of the mantle model. Particular attention is paid to the depth extent of seismicity and the correlation of hypocentral locations with images of the subducted slab.