Kenji Satake

Unusually large earthquakes inferred from tsunami deposits along the Kuril trench

The Pacific plate converges with northeastern Eurasia at a rate of 8–9 m per century along the Kamchatka, Kuril and Japan trenches. Along the southern Kuril trench, which faces the Japanese island of Hokkaido, this fast subduction has recurrently generated earthquakes with magnitudes of up to ∼8 over the past two centuries. These historical events, on rupture segments 100–200 km long, have been considered characteristic of Hokkaidos plate-boundary earthquakes. But here we use deposits of prehistoric tsunamis to infer the infrequent occurrence of larger earthquakes generated from longer ruptures. Many of these tsunami deposits form sheets of sand that extend kilometres inland from the deposits of historical tsunamis. Stratigraphic series of extensive sand sheets, intercalated with dated volcanic-ash layers, show that such unusually large tsunamis occurred about every 500 years on average over the past 2,000–7,000 years, most recently ∼350 years ago. Numerical simulations of these tsunamis are best explained by earthquakes that individually rupture multiple segments along the southern Kuril trench. We infer that such multi-segment earthquakes persistently recur among a larger number of single-segment events.

Tsunami generation by horizontal displacement of ocean bottom

Tsunami generation by an earthquake is generally modeled by water surface displacement identical to the vertical deformation of ocean bottom due to faulting. The effect of horizontal deformation is usually neglected. However, when the tsunami source is on a steep slope and the horizontal displacement is large relative to the vertical displacement, the effect becomes significant. We show this for two recent earthquakes which generated much larger tsunamis than expected from seismic waves. In the case of the 1994 June 2 Java, Indonesia, earthquake, the focal mechanism was a very shallow dipping thrust and the source was near a very steep trench slope. In the case of the 1994 Nov. 14 Mindoro, Philippines, earthquake, strike-slip faulting extended from ocean to land perpendicular to the coast line. In both cases, we found that the horizontal motion of slope had an important contribution to the tsunami generation.

Tsunami Source of the 2004 Sumatra–Andaman Earthquake Inferred from Tide Gauge and Satellite Data

Tsunami source of the 2004 Sumatra-Andaman earthquake was esti- mated from a joint inversion of tsunami waveforms recorded on tide gauges and sea surface heights captured by satellite altimetry measurements. The earthquake, the largest in the past 40 years, caused devastating tsunami damage to countries around the Indian Ocean, but the tsunami source, in particular, its northern end, was not well resolved. Although aftershocks and crustal deformation extended from off north- western Sumatra Island through the Nicobar Islands to the Andaman Islands, some seismic-wave analyses indicated a shorter source length, several hundred kilometers. We used tsunami waveforms recorded at 12 tide gauge stations around the source and the sea surface heights measured by three satellites: Jason1, TOPEX/Poseidon, and Envisat. We numerically computed tsunami propagation using realistic bathym- etry; more than 66,000 depth points were digitized from nautical charts and combined with the ETOPO2 data. Inversion of satellite data indicates that the tsunami source extended to the Andaman Islands with a total length of 1,400 km, but such a model produces much larger tsunami waveforms than observed at Indian tide gauge stations. Inversion of tide gauge records and the joint inversion indicate that the tsunami source was about 900 km long. The largest slip, about 13 to 25 m, was located off Sumatra Island and the second largest slip, up to 7 m, near the Nicobar Islands. The slip distribution is similar for different rupture velocities and rise times, with a slow velocity of 1 km/sec and a rise time of 3 min yielding the largest variance reduction.

Linear and nonlinear computations of the 1992 Nicaragua earthquake tsunami

Numerical computations of tsunamis are made for the 1992 Nicaragua earthquake using different governing equations, bottom frictional values and bathymetry data. The results are compared with each other as well as with the observations, both tide gauge records and runup heights. Comparison of the observed and computed tsunami waveforms indicates that the use of detailed bathymetry data with a small grid size is more effective than to include nonlinear terms in tsunami computation. Linear computation overestimates the amplitude for the later phase than the first arrival, particularly when the amplitude becomes large. The computed amplitudes along the coast from nonlinear computation are much smaller than the observed tsunami runup heights; the average ratio, or the amplification factor, is estimated to be 3 in the present case when the grid size of 1 minute is used. The factor however may depend on the grid size for the computation.

The 1964 Prince William Sound earthquake: Joint inversion of tsunami and geodetic data

The 1964 Prince William Sound (Alaska) earthquake, Mw = 9.2, ruptured a large area beneath the continental margin of Alaska from Prince William Sound to Kodiak Island. A joint inversion of tsunami waveforms and geodetic data, consisting of vertical displacements and horizontal vectors, gives a detailed slip distribution. Two areas of high slip correspond to seismologically determined areas of high moment release: the Prince William Sound asperity with average slip of 18 m and the Kodiak asperity with average slip of 10 m. The average slip on the fault is 8.6 m and the seismic moment is estimated as 6.3 × 1022 N m, or over 75% of the seismic moment determined from long-period surface waves.

Depth distribution of coseismic slip along the Nankai Trough, Japan, from joint inversion of geodetic and tsunami data

Coseismic slip distribution on the fault plane, particularly in the downdip direction, associated with large subduction earthquakes can be estimated by joint inversion of geodetic and tsunami data. Two large earthquakes, the 1944 Tonankai earthquake (Mw=8.1) and the 1946 Nankaido earthquake (Mw=8.3), occurred on the Nankai trough, southwestern Japan, where the Philippine Sea plate is subducting beneath the Eurasian plate. The source areas of these events extended over both land and ocean. Coseismic crustal movements on land were measured by leveling, while those in ocean were recorded as tsunami waveforms on tide gauges. The coseismic slip distribution inverted from these data shows that the slip on the shallower part of the fault plane is comparable to that on the deeper parts. This indicates that large coseismic slip can occur beneath an accretionary wedge where current seismicity is low. The result has implications for other subduction zones having a similar tectonic environment such as for the Pacific Northwest region of the United States. Although the possibility of a large earthquake there is still debated, should a large subduction earthquake occur in this region, coseismic slip on the shallow part could be large, and the potential for large tsunamis is high.

Mechanism of the 1992 Nicaragua Tsunami Earthquake

The 1992 Nicaragua earthquake generated larger tsunamis than expected from its surface wave magnitude (Ms 7.2) and is known as a ‘tsunami earthquake’. Seismological studies showed that the duration was very long for its size, about 100 s. Other studies have shown that the seismic moment estimated from tsunamis is an order of magnitude larger than that from seismic waves, even after the long duration is accounted for. Numerical computations of tsunamis from various fault models are made to reconcile this discrepancy. Comparison of calculated waveforms with tide gauge records shows that the fault width is 40 km, much narrower than the aftershock area, and extends only into the upper 10 km of the ocean bottom. Slip amount on the fault is estimated to be 3 m from amplitude comparisons. The fault length is estimated to be 250 km, slightly longer than the aftershock area, from comparison of the tsunami height distribution. The rigidity around the shallow fault may be smaller than that of a standard underthrust fault, and the seismic moment is estimated as 3×1020 Nm, consistent with the seismic observations. A slow rupture on the shallow fault, presumably in the subducted sediments, is responsible to the unusually large tsunami excitation.

Detailed coseismic slip distribution of the 1944 Tonankai earthquake estimated from tsunami waveforms

Coseismic slip distribution on the fault plane of the 1944 Tonankai earthquake is estimated from inversion of tsunami waveforms. Three improvements from a previous study [Satake, 1993] are made. These are: (1) smaller subfaults are used to resolve detailed slip distribution; (2) the sub faults fit better to the plate interface geometry; and (3) finer and more accurate bathymetry data is used. The inversion result shows that a maximum slip of about 3 m occurred on the plate interface off Shima peninsula. The total seismic moment is estimated to be 2.0 × 1021 Nm (Mw 8.2). The result confirms that the 1944 Tonankai earthquake did not rupture the plate interface beneath the Tokai region and supports the existence of the seismic gap in the Tokai region suggested by Ishibashi [1981]. The slip of about 1.5 m on the plate interface beneath Atsumi peninsula, northeast of the large slip area, is necessary to explain the observed tsunami waveforms, although no seismic moment release was estimated from strong motion data by Kikuchi et al. [1999]. This may suggest that the rupture beneath Atsumi peninsula was slow.

Marine incursions of the past 1500 years and evidence of tsunamis at Suijin-numa, a coastal lake facing the Japan Trench

Sandy deposits of marine origin underlie the floor of Suijin-numa, a coastal lake midway along the subduction zone marked by the Japan Trench. The deposits form three units that are interbedded with lacustrine peat and mud above a foundation of marine, probably littoral sand. Unlike the lacustrine deposits, all three sandy units contain marine and brackish diatoms. The middle unit (B) contains, in addition, graded beds suggestive of multiple waves of long wavelength and period. The uppermost unit (C) probably dates to a time in the areas written history when the lake was separated from the sea by a beach-ridge plain at least 0.5 km wide and several metres high. Units A and B postdate AD 540—870, and unit C postdates AD 1030—1640 as judged from radiocarbon dating of leaves and seeds. Unit B pre-dates AD 915 and unit C postdates that year as judged from a tephra within the peat that separates units B and C. The age constraints permit correlation of unit B with a tsunami in AD 869 that reportedly devastated at least 100 km of coast approximately centred on Sendai. Unit C may represent a later catastrophic tsunami in 1611, or perhaps a storm surge that inundated much of Sendai. The lake lacks obvious signs of tsunamis from the regions largest twentieth-century earthquakes, which were centred to the north in 1933 (M 8.1) and directly offshore in 1936 (M 7.5), and 1978 (Mw 7.6).

THE CAPE MENDOCINO, CALIFORNIA, EARTHQUAKES OF APRIL 1992 : SUBDUCTION AT THE TRIPLE JUNCTION

The 25 April 1992 magnitude 7.1 Cape Mendocino thrust earthquake demonstrated that the North America—Gorda plate boundary is seismogenic and illustrated hazards that could result from much larger earthquakes forecast for the Cascadia region. The shock occurred just north of the Mendocino Triple Junction and caused strong ground motion and moderate damage in the immediate area. Rupture initiated onshore at a depth of 10.5 kilometers and propagated up-dip and seaward. Slip on steep faults in the Gorda plate generated two magnitude 6.6 aftershocks on 26 April. The main shock did not produce surface rupture on land but caused coastal uplift and a tsunami. The emerging picture of seismicity and faulting at the triple junction suggests that the region is likely to continue experiencing significant seismicity.