Seismic intensity is an important parameter measuring earthquake shaking severity, especially in regions where few strong motion instruments are in operation. This study aims to apply high-precision questionnaire intensity method widely used in Japan to other earthquake countries.
Based on the Modified Mercalli (MM) Intensity Scale, an intensity questionnaire form was prepared and survey was conducted for two California earthquakes and the 1988 Nepal-India border earthquake. First, intensity coefficients are assigned to each item category based on the definition of the scale. For each questionnaire, “average” item intensity is calculated by taking average of intensity coefficients as marked, and also “maximum” item intensity by taking maximum intensity coefficient among responses.
In California, intensities of the USGS isoseismal maps are found larger than the “average” item intensities and is comparable to the “maximum” item intensities, presumably because the USGS guideline of intensity assessment is to choose the maximum damage phenomena.
In order to solve this discrepancy, we introduced fuzzy set theory and expressed intensity coefficients as distribution of likelihood belonging to continuous intensity level. Accumulating membership functions corresponding to selected item categories, the maximum value of distribution suggests the most probable intensity.
New method for questionnaire intensity evaluation is examined for previous data. Questionnaire intensities reasonably correlate with intensities locally reported by the USGS and Nepalese agency. Simple adjustment based on age of buildings is found satisfactory for the case of California but not so for the Nepal-India region.

View full abstract

The Miboro Fault System extends about 65km from NNW to SSE along the upper course of the Shokawa River in the western Hida Mountains, central Japan. The fault system is one of the longest left-lateral strike-slip faults in cenral Japan, and is composed of three fault segments showing a left-handed en echelon arrangement; the Kazura, Shirakawa and Miogo Faults from the north. The Tensho earthquake of 1586 A. D. is inferred to have been generated by the faulting of the Miboro Fault System because severe damage including large-scale slope failures was caused by the earthquake along the Shirakawa and Miogo Faults.
We excavated two trenches across the Shirakawa Fault at Kidani in Shirakawa Village, and three trenches across the Miogo Fault at Terakodo in Shokawa Village, in order to verify the relationship between the latest faulting of the Miboro Fault System and the 1586 Tensho earthquake.
Our excavation has revealed that the latest faulting events on the Shirakawa and Miogo Faults occurred after 2, 500y. B. P., and after 840y. B. P., respectively. The result suggests that the 1586 Tensho earthquake was generated by the latest faulting of the Shirakawa and Miogo Faults, although the participation of the Kazura Fault in this event remains uncertain to be verified.

View full abstract

Based on an epicentral distance dependence of tsunami amplitude observed at tide stations, the tsunami magnitudes were obtained to be m=-2 and m=-1 for the Oshima-Kinkai tsunami of Feb. 20, 1990 and the Tokaido-Oki tsunami of Sept. 24, 1990, respectively. The locality on magnitudes of the tsunamis generated along the Izu-Mariana Island-Arc since 1900 are discussed in relation to earthauake magnitude. According to the statistical relation, the magnitude values (Imamura-Iida scale) of the tsunamis generated near the trench triple junction off Boso Peninsula are one to two grades larger than the average tsunami magnitude. Such tsunamis were mostly caused by low-frequency earthquake with the high-angle dip-slip fault. While, the magnitudes of the tsunamis generated off the west side of Oshima Is. are one grade less than the average value, with caused by earthquakes of the strike-slip type. The magnitudes of a few tsunamis generated on the ridge were larger than the average value. Especially, the 1984 Torishima-Kinkai tsunami had an abnormal magnitude. The regional difference of seismic mechanisms were found by the deviation of tsunami magnitudes.

View full abstract

The focal process of the 1964 Niigata Earthquake was reinvestigated on the basis of hypocentral distribution of its aftershocks. This study indicates that the aftershocks are distributed on a fault plane dipping westward.
Although it has been clear that the fault strike of the mainshock was in N20°E direction, the dip of the fault was not still clear due to a poor resolution of hypocenter of aftershocks. To resolve the difficulty, we reexamined seismological data obtained by the Japan Meteorological Agency (JMA).
Reexamination of seismograms of nearby stations enabled us to supplement more than 1200 new P and S arrivals of aftershocks. We also dentified a number of P and S arrivals from the data which were previously reported as unidentified phases. The Joint Hypocenter Determination method was used to get a more reliable aftershocks distribution. The number of located aftershocks much increased, as about 380 aftershocks are well located by this study.
Aftershocks on the vertical cross section which is normal to the fault strike shows that aftershocks are on a westward dipping plane. The dip of the plane is estimated as 50 degrees which is consistent with the focal mechanisms reported by several studies. Although the dip angle depends on the velocity model used in hypocenter location, westward dipping of aftershocks is valid, independent of several different velocity models. Therefore we estimate that the subduction of the Japan Sea under the north-east Honshu does not occur in the southern part of the eastern margin of the Japan Sea.
The aftershock activity is found to be low around the hypocenter of mainshock which is located near the bottom of aftershock region, suggesting a large strain release around the nucleation point of mainshock. Relative position of forerunning seismic activity which preceded the mainshock by two years seems to be within the shallow part of the aftershock region east of Awashima-island which is located in the western middle of the focal region. The epicentral distribution of aftershocks indicates that aftershock occurrence is scarce around Awashima-island. A similar relation was reported in the case of the 1983 Nihonkai-chubu earthquake between its aftershocks and Kyurokujima-island, which is situated east of the middle of the aftershock region. Few aftershocks occurred in the area around Kyurokujima-island. In spite of the difference in relative location, that is, Awashima is situated west of the aftershock region while Kyurokujima is in the east, this suggests possibilities that crust around the islands cannot sustain enough strain to generate aftershocks or it behave as an earthquake barrier.

View full abstract

For the detection of crustal deformations associated with the convergence of the Philippine Sea plate, interferometric GPS measurements have been conducted in and around Shikoku in Southwest Japan. The first campaign was conducted in March 1990 and the second one in March 1991. During the campaigns, the satellite constellation simultaneously included more than five satellites for three to four hours a day. Under these satellite configurations most of the baseline lengths, which range from 40 to 500km, can be determined with a standard deviation (day-to-day repeatability) of 0.1-0.2ppm or better. Horizontal strains calculated from length changes of the baselines during one year indicate compression of 0.3-0.7ppm in an approximately NW-SE direction. These strain rates and patterns are in agreement with the interseismic strain rates deduced from the geodetic surveys conducted during the past several decades.

View full abstract

A practical method is proposed for measuring the depth of two cracks running extremely close to each other in an elastic half-space. For this purpose a theoretical feature of the diffraction of a P wave at the edge of a slit is used, in a manner similar to its use in a method previously proposed by one of the present authors for estimating the depth of an isolated crack.
A pulse-source is first set sufficiently near to the outside edge of one of the pair of cracks, and then the receiver is moved over the surface on the far side of the cracks, in order to seek the point at which the initial pulse-motion vanishes, its direction on one side of this point being in inverse relation to that on the other. The respective depths of the cracks are regarded as either mutually sufficiently different or almost identical, according respectively as this point is consistently present or absent. In the former case both depths are obtained by means of a relatively easy procedure, essentially the same as that proposed for the case of an isolated crack. In the latter, however, it is difficult to measure both depths, as the measurement procedure must be performed on a different basis—the attenuation of pulse-amplitude with distance must be observed from what is recorded from the receiver, and compared with that derived theoretically. The use of such recorded quantitative measurements is, as is well known, frequently found to result in ambiguity, though adequate results were in fact obtained in our model experiment. It is suggested that the method introduced in this paper constitutes on the whole a significant advance in measurement-technique for the objects proposed.

View full abstract

Groundwater radon anomalies are often interpreted as being related to earthquake occurrences, although their relation has not been clearly established. As a first step in discuss with what certainty the radon anomalies can be related to earthquake occurrences, we have made a statistical study on the significance of the correlation between the groundwater radon anomalies identified under a certain test described below and the earthquake occurrences around the observation site. Continuous observations on groundwater radon concentration coupled with environmental conditions have been made for many years by the Laboratory for Earthquake Chemistry, University of Tokyo. In this study, the radon data at KSM (Kashima, Fukushima, Japan) for a period of 1364 days are used to extract anomalies. The test for anomalous radon data is the absolute difference between the observed values and the values expected from the model we have adopted, whose function is to remove the effects of changes in environmental conditions. We detected 80 radon anomalies during the 1364 days, quite independently of any seismic information. We then selected 75 earthquakes from around KSM, making no reference to the radon data. Comparing those identified radon anomalies with the earthquake occurrences, we found that radon anomalies appeared most frequently of the first and second days after an earthquake. And this tendency is statistically judged to be significant for a 1% significance level. This is our first conclusion. We also found a precursory-like tendency that radon anomalies are likely to appear on 6 and 7 days before an earthquake. To estimate the significance of this precursory-like tendency, we have made a stochastic simulation in which we assume both our first conclusion and the actual occurrence pattern of the earthquakes. And this simulation led us to our second conclusion: the radon anomalies frequently detected on 6 and 7 days before an earthquake cannot be interpreted as precursory for a 5% significance level.

View full abstract

A gravity survey was carried out to estimate fault structure of the Miyakata fault in Ikoma City, Nara Prefecture. The result shows the basement at the western part of the fault is about 200-300 meter lower than that at the eastern part. And a north-south trending basin structure is recognized along the western side of the Miyakata fault. The geological survey for the Ikoma City region reported that the Plio-Pleistocene sediments in this area were deposited about 1Ma ago. After the deposition of these sediments, the Ikoma mountain ranges upheaved. This fact shows the average rate of vertical displacement of the Miyakata fault is about 0.2-0.3mm/yr.

View full abstract

A new empirical formula to determine the earthquake magnitude from the amplitude in the coda-portion of S-wave is obtained. The formula has a form, M_C=alogA+b·logT+c, where A is the rms amplitude in the coda-portion of S-wave at the lapse time, T, from the origin time of an earthquake. The coefficient, a, is around 1.0 and the coefficients, b and c, depend on the observational site condition, in particular, characteristics of the seismometer. The formula is applicable to A more than twice as large as the noise level and to T more than 1.6 times as long as the S-wave travel time. The magnitude derived from the above formula, the coda magnitude, gives a close approximation to the magnitude determined by the Japan Meteorological Agency in the magnitude range from 1.3 to 4.6, when the seismometer of its natural frequency, either 1.0Hz or 4.5Hz, is used.
The coda magnitude has the advantage of covering a wide range of magnitude, because the amplitude in the coda-portion of S-wave can be used even when the maximum amplitude is out of scale. It also has a high reliability, because the coda magnitude is not largely affected by the characteristics of radiation at the source. Present method is very easy to carry out, especially when the digital data are available, as the coda magnitude can be determined by an easy calculation of rms amplitude.

View full abstract