An attempt was made to apply a multi-variable ARIMA model (autoregressive integrated moving average model) to observed non-stationary variations of crustal strains, ground tilts, precipitation and water discharge. The purpose here is to find correlations among these observed values, to make predictions for their future variations, and to detect some possible anomalies that might be ralated to the change of tectonic stress field and the earthquake occurrence.
The model fitting is applied to the records obtained at the Kamitakara Geophysical Observatory. The applicability of this model is carefully tested by the randomness of noise involved in each of the observations, and also by examining independence of each variable after removal of dependent components. The preliminary results suggest that the observed areal strains and ground tilts are somewhat affected by long-period fluctuations of precipitation in a range between 20 and 50 days but not by water discharge in the observation vault, and that the amount of water discharge is subjected to variations of areal strains in addition to the effects of rainfall. It is also suggested by the present analysis that the large amplitude variations of strains and tilts observed during some period in 1978 are within the range of prediction errors and may not be any significant changes due to tectonic origins.

View full abstract

Kinugawa area (N36.0°-36.3°, E139.7°-140.0°), the southwestern part of the Ibaraki prefecture, is one of earthquake clustering regions at depths from 40km to 60km in the Kanto district, central Japan. A seismic moment of an earthquake of August 4, 1974, which is the largest event (M5.8) in the past two decades in the Kinugawa area, is obtained by the least square's method for P wave amplitude data from WWSSN long period records. The seismic moment obtained is 0.77×10²⁵dyne·cm with a focal mechanism of low-angle thrusting to the NNW direction.
There were 2 events of a magnitude larger than 5.4, excluding the largest event, in a period of 15 years from 1966 to 1980. The seismic moments of the events are estimated by the Rayleigh wave amplitudes from the WWSSN long period records of SHK relative to that of the largest event. Thus, the total amount of the seismic moment released in the Kinugawa area (120km²) during the 15 year period is 0.98×10²⁵dyne·cm. This estimation is about 20-30% of an amount that is predicted by a simple scheme on the basis of relative movements of the Philippine Sea and Eurasian plates in the Kinugawa area.

View full abstract

By means of two-dimensional finite element simulations of the stress relaxation in the asthenosphere induced by the occurrence of major underthrust earthquakes at subduction zones, characteristic features of the postseismic time-dependent surface displacements are investigated. In the model of the upper mantle structure involving a subduction zone, the following structural anomalies are taken into account; (1) the presence of a descending elastic slab in a Maxwell viscoelastic asthenosphere beneath a continental elastic plate, and (2) the difference between the thicknesses of oceanic and continental elastic plates.
Both of the structural anomalies of (1) and (2) significantly affect not only the amounts but also the patterns of the displacements. Furthermore whether the coseismic faulting reaches the asthenosphere or not also has a notable effect on the patterns of the displacements. Therefore the displacements for the anomalous structure model are considerably different from those for a simpler layered structure model in amount and pattern. It is confirmed that time-dependent surface movements in subduction zones following megathrust earthquakes must be interpreted by using a realistic model of the upper mantlt structure.

View full abstract

In-situ stresses were measured in four 450m deep boreholes in the seaside district of the Kanto-Tokai area. 25sets of stresses (the maximum and the minimum horizontal compressive stresses) were obtained successfully from water pressure records at the hydraulic fracturings at different depths.
The stresses increase with depth. Linear stress-depth approximations fitted to the data are as follows.
Okabe: S_Hmax=4.3×10^-2h+2.1, S_Hmin=2.8×10^-2h+1.4,
Shuzenji: S_Hmax=2.1×10^-2h+3.6, S_Hmin=1.8×10^-2h+1.8,
Shimoda: S_Hmax=2.9×10^-2h+0.9, S_Hmin=2.0×10^-2h+0.7,
Yokosuka: S_Hmax=2.0×10^-2h+2.7, S_Hmin=1.7×10^-2h+1.8,
where S_Hmax and S_Hmin are the maximum and the minimum horizontal compressive stresses in MPa, and h is the depth in meter. The vertical stress S_v is assumed to be equal to the lithostatic pressure which is calculated from the density. S_Hmax and the maximum shear stresses at Shuzenji and Shimoda are smaller than those at Okabe at the depths. As is well known, in the Izu Penninsula, which includes Shuzenji and Shimoda, crustal movement and shallow earthquakes have been active in recent decades. This suggests that both S_Hmax and the maximum shear stress magnitudes do not simply correspond to current crustal activity. On the other hand, the relative magnitude of S_v, S_Hmax and S_Hmin at the depths for each measurement site derived from the above equations is consistent with that estimated from the type of active fault and earthquake focal mechanisms near the measurement site.
The directions of S_Hmax which were given by the azimuths of the hydraulic fractures were obtained at 11 points in the four boreholes by using impression packers. No change in direction accompanied with the change in the depth of the measurement hole. The average direction of S_Hmax is as follows.
Okabe: N50°W, Shuzenji: N20°E, Shimoda: N35°W, Yokosuka: N5°W.
The above data show that stress direction is not uniformly distributed in this area. However, the direction measured at each measurement site agrees well with that estimated from geological and seismic evidence near the site. Detailed discussion of the stress direction measured and estimated from geological and seismic data in the Kanto-Tokai area will appear in another paper.

View full abstract

A map of the maximum compressive stress direction has been made for the Kanto-Tokai area from the data of in-situ stress measurements, focal mechanisms of earthquakes, and geological evidences. Data compiled, which amounts to 77 points, includes the following: 10 points of hydraulic fracturing stress measurments, 3 points of overcoring stress measurments, 7 points of Quaternary dike trends, 9 points of Quaternary monogenetic volcano alignments (including flank volcanoes), 41 points of active fault movements and 7 points of focal mechanism solutions of shallow earthquakes (depth <20km, M>6.5). The data points which lie close to one another give a similar stress condition: the stress orientation and stress state classified by fault type such as normal, reverse, and strike-slip faults. We have divided the Kanto-Tokai area into seven provinces on the basis of the data compiled after consideration of the plate interation among the Eurasian, the Philippine Sea and the Pacific plates. Stress condition within each province appears almost uniform. The seven stress provinces and their averaged azimuths for the maximum horizontal compressive stresses S_Hmax are as follows: southern part of Nagano Prefecture (NA), N85°W; northern part of the Kanto district (KA), N80°W; Shizuoka district (SZ), N40°W; Nishiizu district (NI), N10°E; Higashiizu district (HI), N40°W; Tanzawa district (TN), N-S; and the Miura-central Boso district (MB), N35°W. The azimuths of S_Hmax in stress provinces NA and KA are both nearly E-W, which is attributed to the westward relative movement of the Pacific plate against the Eurasian plate. The azimuths of N35°-40°W in provinces SZ, HI and MB are considered to be due to the northwestward movement of the Philippine Sea plate relative to the Eurasian plate. The cause for the individual orientation of S_Hmax as N10°E in province NI is interpreted as being a result of the downward bending of the Philippine Sea plate at the Suruga trough. That is, the extension perpendicular to the bending axis is effected on the surface by the downward bending, which result in the parallel orientation of the S_Hmax to the trough. In the case of province TN, the stress orientation in the province is not interpreted in terms of downward bending because the subduction has not been expected in the province. This is one of the subjects for future study. In stress province MB, the direction of S_Hmax agrees well with the horizontal maximum compressive strain accumulated during the period from 1882/1902 to 1973/80. However, the direction does not agree well with that of the strain accumulated during the period from 1924/25 to 1973/80. The difference is one reason for low seismicity in the upper crust in province MB, since current crustal movement does not work efficiently to increase the in-situ differential stress. The difference may be attributed to the long-term after-effect of the 1923 Kanto earthquake.

View full abstract

Late Holocene crustal uplift of Shikine Island, off the Pacific coast of central Japan, has been investigated. Shikine is a small island of about 3km (EW)×2.5km (NS), located on the northern tip of the Izu-Bonin arc, to the south of the Philippine Sea—Eurasian plate boundary of the Suruga and Sagami troughs. So far, historical coseismic uplift of the island, due to the 1498 or 1605 or 1703 great offshore earthquake, has been suspected based on raised fossil barnacles.
In the present study, six localities on the northern and eastern coast of the island, which are situated at the foot of a steep sea cliff, several tens of meters high, have been observed. At Loc. 1 fossil barnacles can be found in the range of 2.6-4.3m amsl (above mean sea level at Okada, Oshima Island). Tetraclitella chinensis (sampling height, 3.2-3.6m amsl), Megabalanus rosa (3.0-3.1m amsl), and Tetraclitella chinensis (2.6m amsl) have been dated at 1980±130(GaK-9718), 1830±130(GaK-9717), and 1350±100(GaK-9720)¹⁴Cy.B.P., respectively, showing a direct relationship between height and date. The ¹⁴C date of 2540±290y.B.P.(GaK-9721) for Saccostrea echinata (2.6m amsl), however, is not in harmony with this tendency. This anomaly is probably due to subsequent reworking of the older shell. At this locality, the upper limit of living Tetraclita squamosa Japonica bed indicates that the present high tide level is 1.0m amsl. At Loc. 2 fossil barnacles and shells can be found in the range of 3.9-4.2m amsl. Here, however, due to a small amount of sample number, only an unreliable ¹⁴C date of 1010±300y.B.P.(GaK-9717) has been obtained for Tetraclita squamosa Japonica (4.2m amsl). At Loc. 3 fossil barnacles can be found in the range of 3.9-6.1m amsl. The ¹⁴C date of the barnacle Tetraclitella chinensis which are scattered throughout 1.1m at section (3.9-5.1m amsl) is 1530±100y.B.P.(GaK-9722). The upper limits of both fossil and living barnacle beds at this locality are somewhat higher than at Loc. 1 or 2, probably due to the high energy of waves in a sea-cave. At Loc. 4, 5 and 6 no fossils are found, but clear raised coastal landforms can be seen. At Loc. 4 there is a raised notch, whose retreat point is 3.2m amsl. At Loc. 5 two steps of raised bench, 4.2m amsl and 2.6m amsl, exist. At Loc. 6 there is a raised bench of 3.7m amsl. Also around Loc. 1 and 2 benches at about 4m amsl can be recognized.
Judging from the above mentioned observations, Loc. 1 offers us the best evidence for the recent vertical movement of Shikine Island. As an evident discontinuity between the fossil bed and the living barnacles exists, it is certain that a sudden uplift took place after the formation of the lowest part of the fossil bed, that is ca. 1400y.B.P. ¹⁴C dates from Loc. 2 and 3 do not conflict with this estimation. The amount of uplift was approximately 3m. Emergence of the most prominent raised benches and sea-caves is probably correlated with that of fossil bed at Loc. 1. Another land uplift might be assumed, taking into consideration the existence of the lower and local bench at Loc. 5, or significantly older ¹⁴C dates from the upper part of fossil bed at Loc. 1. However, there is no enough data to confirm this possibility. Therefore, it is concluded that the distinct crustal uplift of Shikine Island is much older than the historical earthquakes of 1498 or 1605 or 1703.
If this uplift was due to an earthquake, the 684 Hakuho earthquake, a great interplate event of magnitude 8.4 off southwest Japan, might have been responsible for it. However, the possibility remains that an unknown earthquake near the island brought about the uplift. Another possibility is that the uplift might have been related to volcanic activity. All these presumptions, however, are still open to question. Further intensive a

View full abstract

Seasonal fluctuation superposed on elevataion change data was found for the precise levelling along the northwest-southeast trending route of about 25km length between Kakegawa and Hamaoka in the Tokai area, central Japan. The seasonal change amounting 17mm in height difference between the end points exhibits the peak and trough in late January and late July, that is apparently correlated with the change in atmospheric temperature. Correcting the seasonal effect, we obtained a reliable data set of elevation change for the levelling route, along which measurement was repeated thirteen times since 1962.
It has been known that the area is subjected to general subsidence and eastward tilt due to westerly subduction of the Philippine Sea plate from the Suruga Trough. The corrected data, however, revealed more detailed features of the recent crustal movement: (1) A rapid upheaval amounting 2cm at Kakegawa against Hamaoka took place probably in June 1976; and (2) except for the special event, Hamaoka referred to Kakegawa has constantly subsided at a rate of about 4mm a year. The constant rate of subsidence strongly suggests that shallow portion of the continental plate is subject to steady dragging by the subducting Philippine Sea plate. The abrupt crustal deformation in 1976 is interpreted by occurrence of creep dislocation at a deeper portion of the plate interface. Anomalous phenomena similar to the 1976 event may indicate relatively high risk of the pending Tokai earthquake as the deep-seated creep is expected to produce additional shear stress in the hypothetical rupture zone.

View full abstract

A series of experimental studies of long-period microtremors has been continued with the intention of elucidating dynamic characteristics of deep ground soil during an earthquake and the stepwise results have been reported in the previous papers. Through those observations, we pointed out that the temporal variation of microtremors amplitude obviously correlates with that of sea waves and therefore further understanding of microtremors will be advanced by an investigation of sea wave characteristics.
In this paper, using the date of simultaneous observation with sea waves, their relationship is discussed, and the following results are obtained. The temporal variations of both amplitudes of microtremors and sea waves are in clear correspondence. From the comparison of these spectra, it is concluded that the predominant period of microtremors is about a half of that of the sea waves. This fact enables us to insist that the Longuet-Higgins' theory is also applicable to the generation mechanism of the microtremors under consideration. Empirical equations connecting both of the spectral amplitudes of microtremors and of sea waves are newly derived.
Generalizing these results, a revision of the standardized observation scheme proposed in the previous paper was made for the long-period microtremors.

View full abstract

The spatio-temporal variation of seismicity and spectral characteristics of seismic waves are examined for earthquakes occurring offshore of Fukushima Prefecture, of Ibaraki Prefecture and around the Boso Peninsula, eastern Japan, on the basis of JMA earthquake data. From space-time plots of earthquakes with M≥5 and depth≤60km for the period from 1926 to May 1982, it is recognized that before the occurrence of the 1938 Fukushima-Oki large earthquake swarm, seismic activity became gradually high in the south of the 1938 focal area. A noticeable activity occurred in 1935 at the landward margin of the focal area of the 1938 shocks, then considerable quiescence continued till the 1938 shocks. In recent about 20 years, seismicity is admittedly low in all regions investigated. To deduce the characteristics of seismic source spectra during this inactive period, a magnitude M_I based op seismic intensity data is determined for a number of felt earthquakes of M≥5. The difference between M_I and M reported by JMA is used as a measure of the efficiency of the radiation of highfrequency component from the source. The temporal distribution of M_I-M value shows a tendency of increase of the value with time in all the regions. For earthquakes off Fukushima Prefecture, M_I-M has increased systematically since about 1965. In the region off Ibaraki Prefecture, most earthquakes occurring about 10 years prior to July 23, 1982, Ibaraki-Oki earthquake of M=7 reveal considerable large M_I-M. For earthquakes around the Boso Peninsula, M_I-M increase gradually during the last 30 years. These large M_I-M are mainly distributed landward of the focal area of the 1982 Ibaraki-Oki shock and in the northern part off Fukushima Prefecture, while landward and southern parts off Fukushima Prefecture have been seismically quiet of M≥5 shocks since about 1965. It is considered that the predominance of high-frequency earthquakes in recent years is due to the extensive increase of the tectonic stress in the regions investigated. The 1982 Ibaraki-Oki earthquake of M=7 alone does not seem to have released this large tectonic stress, and much attention should be paied to presently inactive seismic regions adjacent to off Ibaraki Prefecture.

View full abstract

A new clock correction system has been developed for the clock calibration of the ocean bottom seismometer (OBS) before its launching and after its recovery. This system has two standard clocks and one portable time calibrator.
One of the standard clocks is synchronized with one of three broadcasted standard time signal waves (2.5, 5.0, 10MHz). Another standard clock uses the oscillation output of a Rb oscillator with a stability about 10^-10 per month. The latter can be corrected by the former standard clock when the stability of the Rb oscillator become worse. The portable clock calibrator is synchronized with one of the two standard clocks. This calibrator is able to read the OBS clock and to measure the time difference between the standard clock and the OBS one, and also to send a trigger signal to the OBS clock for the time correction.

View full abstract

The off-Ibaraki, central Japan earthquake (M=7.0) of July 23, 1982 was followed by the southern Ibaraki earthquake (M=6.0) of February 27, 1983 with a lag time of seven months. Although those two shocks are separated by 200km, a physical correlation between them is suggested since both the earthquakes are the largest events in the respective areas for the last forty years. Investigating past seismicity, it was found that similar phenomenon, which we name the “paired earthquakes”, repeatedly occurred in the Ibaraki region.
The trench side earthquake of the pair is an interplate earthquake associated with the subducting Pacific plate, while the land side one is interpreted as an earthquake representing mutual motion between the Pacific plate and the Philippine Sea plate at a depth of 60-70km beneath Ibaraki Prefecture. The paired earthquake phenomenon is accounted for by accelerated subduction of the Pacific plate which is expected to occur when one of the plate interfaces is ruptured by the preceding event of an earthquake pair.

View full abstract

Theoretical study has been made to investigate seismic directivity effects due to deterministic and stochastic rupture processes on a fault. An earthquake is modeled by a finitely propagating unilateral rupture on a fault plane where fault heterogeneities; fault patches, are randomly distributed. The parameters of the present dynamic stochastic source are (1) seismic moment M₀, (2) fault dimension L and W, (3) fault patch intensity σ, (4) mean fracture time 2πγ^-1 of fault patches, and (5) fluctuation angle α of rupture direction within fault patches. The seismic directivity effect due to the deterministic rupture process is represented by a linear and coherent superposition of fracturing fault segments, which is the Doppler effect. That due to the stochastic rupture process is described by an energy additive superposition of fracturing fault patches. The latter effect, patch multiplexing effect, is prevailling on the source spectra only in a frequency range from corner frequency to patch corner frequency. The fluctuation angle α is an important parameter to control the effect. It has been concluded that the stochastic directivity effect is maximum, when the rupture direction within fault patches is equal likely in every direction (α=π), and minimum when the rupture direction coincides with one for the deterministic rupture propagation (α=0).

View full abstract