Katsuyuki Abe

Tsunami and mechanism of great earthquakes

Abstract The relation between tsunamis and sea-bottom deformations associated with the Kurile Islands earthquake of 1969 and the Tokachi-Oki earthquake of 1968 is studied on the basis of a fairly complete set of seismological and tsunami data. The seismic results are included in the calculation of static crustal deformations. The calculated deformations are compared with the tsunami source area as obtained by the inverse refraction diagram, the first motion of tsunami waves, and the height of the sea-level disturbance at the source. It is found that such deformations as predicted by the seismic results can quantitatively explain the source parameters of tsunamis. These findings strongly favor the idea that tsunamis are generated by tectonic deformations rather than by large submarine landslides and slumps. This conclusion is supported by additional analyses for the 1964 Niigata, 1944 Tonankai, 1933 Sanriku earthquakes. For the 1946 Nankaido earthquake, the source deformation responsible for the tsunami generation is of much greater magnitude than that for seismic waves.

Lithospheric normal faulting beneath the Aleutian trench

Abstract The focal process of the Rat Island earthquake of March 30, 1965, which occurred beneath the Aleutian trench, is studied on the basis of the long-period surface-wave data and the spatial distribution of the aftershocks. The Rat Island earthquake is represented by a normal faulting with some left-lateral strike-slip component. The spatial distribution of the aftershocks shows a remarkable plane-like distribution extending from the surface to a depth of 60 km, and this thin aftershock zone coincides well with one of the orthogonal nodal planes of the double-couple solution. This fact strongly suggests that a shear fracture took place over nearly the entire thickness of the suboceanic lithosphere. It is proposed that a large thrust earthquake of February 4, 1965, which occurred with a spatial and temporal proximity to this normal fault earthquake, lubricated the interface between the oceanic and the continental lithospheres. As a result, a gravitational pull exerted by the denser sinking lithosphere caused a large-scale extensional fracture of the oceanic lithosphere, the event of March 30. The source parameters are determined as follows: fault plane: dip direction N 52° E, dip angle 50°; auxiliary plane: dip direction N 166° W, dip angle 47°; Seismic moment 3.4 × 1027 dyne · cm; average slip dislocation 1.2 m in N 14° E direction; stress drop 33 b; strain drop 0.46 × 10−4; released strain energy (residual strain is assumed to be zero) 7.9 × 1022 erg. Here the fault size and the rigidity are assumed as 50 × 80 km2 and 0.7 × 1012 dyne/cm2, respectively.

Tsunami field survey of the 1992 Nicaragua earthquake

An earthquake with surface magnitude (Ms ) 7.0 occurred 100 km off the Nicaraguan coast on September 2, 1992 (GMT). Despite its moderate size, this earthquake generated a sizable tsunami, which caused extensive damage along the coast of Nicaragua. In late September, about 170 people, mostly children, were listed dead or missing; 500 were listed injured; and over 13,000 were listed homeless, with more than 1500 homes destroyed. Damage was the most significant since the 1983 Japan Sea earthquake tsunami, which killed 100 people in Japan. The Flores (Indonesia) earthquake and tsunami of December 12, 1992, were more destructive than the Nicaragua or Japan Sea events.

Source characteristics of the Nicaraguan Tsunami Earthquake of September 2, 1992

A large, Ms = 7.2, shallow earthquake took place off the Pacific coast of Nicaragua on September 2, 1992, and set off large tsunamis. The ground motion of the main shock was very weak along the whole Nicaraguan coast. A tsunami as high as 10 m was observed at El Transito, whereas a height of about 2 m is empirically expected for an earthquake of Ms 7.2. The tsunami magnitude Mt is estimated to be 7.9 from tide gage data. The Nicaraguan event is a tsunami earthquake which generates unusually large tsunamis for its earthquake magnitude. The source mechanism is studied in detail by using waveforms of body waves and surface waves recorded on global broadband seismographs. The possibility of single force source is ruled out from radiation patterns and the amplitude ratio of Rayleigh and Love waves. The main shock is interpreted as a low-angle thrust fault with strike = 302°, dip = 16° and slip = 87°, the Cocos plate underthrusting beneath the Caribbean plate. The seismic moment from surface wave analysis is 3.0 × 1020 Nm (Mw = 7.6). The source dimension is estimated to be 200 km × 100 km from the aftershock area. The inversion results of body waves suggest bilateral rupture with rupture velocity as low as 1.5 km/s and duration time of about 100 s. The source process time is unusually long, indicating that the associated crustal deformation has a long time constant.

Estimate of the tsunami source of the 1992 Nicaraguan Earthquake from tsunami data

The Nicaraguan earthquake of September 2, 1992 excited large tsunamis which caused significant damage along the Nicaraguan coasts. Some of the inhabitants felt only minimal shock before the tsunami arrived, indicating that excitation of short-period seismic waves was very small. Numerical simulation with the initial tsunami source estimated by the seismic data is carried out to study the characteristics of the tsunami source by comparing the calculated data with the measured data along the Nicaraguan coast. Finding of the comparison shows that the dislocation of the fault estimated from the measured data is 5.6 to 10.0 times larger than that from seismic data. The tsunami source area, which is 200km in length × 100km in width, is used to explain the distribution of measured tsunami heights along the coast and the wave period as shown in the tide record at Corinto. The effect of rise time on tsunami excitation indicates a slow process, which corresponds with the seismic waves. This event falls under the tsunami earthquake category to produces anomalously large tsunamis relative to earthquake magnitude.

Estimate of Tsunami Run-up Heights from Earthquake Magnitudes

A tsunami magnitude scale Mt is unique among the many scales in seismology, because it is based on instrumental tsunami-wave amplitude, its level being adjusted to agree with moment magnitude of an earthquake. Despite the simple definition, the Mt scale reliably measures the physical size of an earthquake as well as the overall potential of a tsunami. By taking account of the definition of Mt and the scaling relation of earthquake fault parameters, a method for estimating tsunami run-up heights from earthquake magnitudes is developed. The application of the method to a number of the height data of previous tsunamis suggests that the uncertainty lies within a factor of 1.5. For practical purposes, the relationships discussed here are summarized into a simple diagram. The present method is practical and robust enough to be used for near-field tsunami warning purposes where rapid evaluation of tsunami heights is required.

The large normal‐faulting Mariana Earthquake of April 5, 1990 in uncoupled subduction zone

A large, Ms = 7.5, shallow earthquake occurred beneath the Mariana trench on April 5, 1990. From the relocated aftershock distribution, the fault area is estimated to be 70 × 40 km2. A tsunami observed on the Japanese islands verifies that the depth of the main shock is shallow. For waveform analysis, we use long-period surface waves and body waves recorded at global networks of GDSN, IRIS, GEOSCOPE and ERIOS. The centroid moment tensor (CMT) solution from surface waves indicates normal faulting on a fault whose strike is parallel to the local axis of the Mariana trench, with the tension axis perpendicular to it. The seismic moment is 1.4 × 1020 Nm (× 1027 dyn.cm) which gives Mw = 7.3. Far-field P and SH waves from 13 stations are used to determine the source time function. Since the sea around the epicentral region is about 5 km deep, body waveforms are contaminated with water reverberations. The inversion results in a source time function with a predominantly single event with a duration of 10 sec, a seismic moment of 2.1 × 1020 Nm, and a focal mechanism given by strike = 198°, dip = 48°, slip = 90°. The short duration indicates a small area of the rupture. The location of the main shock with respect to the aftershock area suggests that the nodal plane dipping to the west is preferred for the fault plane. The local stress drop of the single subevent is estimated to be 150 MPa (1.5 Kbars). The Mariana earthquake is considered to have occurred in an uncoupled region, in response to the gravitational pull caused by the downgoing Pacific plate.

Quantification of tsunamigenic earthquakes by the Mt scale

Abstract A magnitude scale, M t , previously defined by the author, is unique among the many scales in seismology, because it is based on the instrumental amplitude of tsunami waves, its level being adjusted to agree with moment magnitude. Despite the simple definition, the M t scale reliably measures the seismic moment of an earthquake as well as the overall physical size of the source of the tsunami. In reality, the seismic moment estimated from M t for 39 events around Japan is found to be essentially equivalent to that estimated from seismic waves, with an uncertainty factor of two on average. The M t scale is useful for a quantitative discrimination of tsunami earthquakes that generate anomalously large tsunamis for their magnitude. Of 97 tsunamigenic events that occurred around Japan during the period 1894 to 1985, at least eight tsunami earthquakes are identified from the M t vs M s relation.

Source rupture process of the Papua New Guinea earthquake of July 17, 1998 inferred from teleseismic body waves

A large earthquake (Ms 7.1) occurred off northwest coast of Papua New Guinea (PNG), and a massive tsunami attacked villages to cause a devastating damage. In an attempt to ascertain the tsunami source, we investigate the source rupture process using teleseismic data at IRIS network as well as local data at Jayapura, Irian Jaya, station. The source parameters obtained are: (strike, dip, slip) = (301°, 86°, 91°); the seismic moment = 4.3 × 1019 Nm (Mw = 7.0); the duration of main rupture = 19 s; the centroid depth = 20 ± 5 km; the extent of rupture along the fault strike = 40 km; the average dislocation = 1.8 m; the stress drop = 7.3 MPa. The tsunami magnitude Mt determined from tide-gage data at long distance is 7.5, significantly larger than Ms, so that the PNG earthquake is characterized as a tsunami earthquake. Tsunami earthquakes might have been caused by slow rupture, submarine landslide, and high-angle dip-slip. Our teleseisimic analysis precludes the first two candidates and favors the third one as a source of the present earthquake, although it does not necessarily exclude the possibility of an aseismic landslide induced by the main shock or its aftershocks.

Tsunami from the Mariana Earthquake of April 5, 1990: Its abnormal propagation and implications for tsunami potential from outer‐rise earthquakes

The tsunami generated by the Mariana earthquake of April 5, 1990 was observed on the Japanese and Pacific islands as far as Hawaii. The observed tsunami amplitudes are not a simple function of distance from the source but vary with large-scale bathymetry. Numerical computations of tsunami propagation are made for actual bathymetry and the computed amplitudes are compared with the observations. From the comparisons, the seismic moment is estimated to be 1.4 × 1020Nm, very similar to that from seismic waves, and indicates that the seismic waves and tsunami are equally excited. The numerical computations also show that the tsunami propagated toward the Japanese coast through two different paths: one is through the trench system with high velocity and small amplitude, the other along the ridge system with low velocity but large amplitude. The two kinds of tsunami are identified on recently-installed ocean bottom pressure gauges. Since the Mariana earthquake was an outer-rise earthquake in a weakly coupled subduction zone, where the size of outer-rise event is larger than underthrusting events, it is possible that even larger earthquakes occur in this region. The tsunami potential from such events must be considered, including the unusual tsunami propagation.