To investigate microseismicity around the Nagamachi-Rifu fault, the B class active reverse fault with a northeast-southwest strike and a northwest dip, Miyagi Prefecture, northeastern Japan, a temporal seismic observation composed of sixteen high-sensitivity stations was carried out for the total period of one year between 1996 and 1998. In spite of high artificial noise in the urban area of Sendai City, we succeeded in obtaining high-quality seismic data by using a DAT continuous recording system. The total number of the earthquakes recorded is 119, of which the smallest one is magnitude -0.5. The depths of the events range between 3 and 13km. Most of hypocenters are concentrated in the northwestern region of the Nagamachi-Rifu fault, while few earthquakes in the southeastern region regardless of our high sensitivity observation. This indicates that these earthquakes occur only in the hanging wall side of this active fault. Focal mechanisms in this region are not necessarily coincident with that of the Nagamachi-Rifu fault. In particular, focal-plane solutions of strike-slip are predominant along the Kagitori-Okubushi line, which is a flexure with a conjugate strike to the Nagamachi-Rifu fault. The spatial pattern of the horizontal crustal strain also changes its characteristics across this line. These results strongly suggest that not only the Nagamachi-Rifu fault but also the Kagitori-Okubushi line may control the seismogenesis of this area.
A many-body disk model of slip phenomena was investigated by carrying out a model experiment. A two dimensional many-body disk system was used as a model of the boundary layer between slip surfaces. A model experiment with a packing fraction of 0.75 was carried out at a constant rotation rate of 1/80Hz (angular frequency). The time series of torque was measured during the experiment. By analyzing five thousand seconds' time series of torque, 1784 stick-slip events were obtained. We found that the size distribution of stick-slip events obeys a power law, N(s)-s-a, where s is a size of stick-slip events and N(s) is the number of stick-slip events of which size is s. The value of α calculated by the least squares method is 0.93±0.07 (s. d.). The power law size distribution suggests that the many-body disk model of slip phenomena is a self-organized critical phenomenon.
Magnitudes for the 1923 Kanto earthquake and its major aftershocks were determined in JMA (Japan Meteorological Agency) scale. The original definition of the JMA magnitude is a magnitude that is calculated by the Tsuboi's formula from the maximum amplitudes in horizontal components of seismograms obtained by the regional observation network of JMA. The used seismograms were recorded by standard seismographs, which were the displacement type with the natural period of about 5s and damping ratio of about 8. However, the seismometers had not been yet standardized before 1925 and various types had been used whose instrumental responses were quite different from those of the standard seismographs. The purpose of the present study was that the JMA magnitudes of the 1923 Kanto earthquake and its major 3 aftershocks were determined in consideration of the difference of the instrumental responses. Fortunately, unsaturated seismograms by the Imamura's type strong motion seismographs (displacement type) have been preserved at 7 stations of JMA. The natural period and damping ratio of each seismograph have been evaluated from the free oscillation records preserved at each station. The records for the main shock and aftershocks were digitized and corrected in the instrumental responses to calculate the seismograms with the instrumental response of the standard seismograph of JMA. After that, the maximum amplitudes were measured on the corrected records and the magnitude was determined for each earthquake following the definition of the JMA magnitude. The determined JMA magnitude was 8.1±0.2 for the main shock. All the results were consistent within the difference of 0.2 with the customary results, which were determined from the uncorrected amplitude and seismic intensity data. The standard deviations were smaller than 0.2 for all the events, which shows higher reliability of the present results, comparing with the past ones.
The Itoigawa-Shizuoka Tectonic Line (ISTL) in central Japan is a complex 150km-long fault system consisting of north-trending east-dipping reverse, northwest-trending left-lateral strike-slip, and north-trending west-dipping reverse faults. To help to resolve segmentation and to estimate the magnitude of future shocks on the ISTL, we conducted four trench excavations across the Hakushu fault, the Shimotsuburai fault, and the Ichinose fault group in the southern part of the ISTL, where no paleoseismological data for the surface faulting had been available. On the trench walls at the Hakushu fault, we found the evidence for the most recent surface-rupturing event occurring sometime between 6, 650 and 7, 000cal. y. B. P. (BC 4700-BC 5050) with approximately one meter of coseismic slip. On the Shimotsuburai fault, we exposed evidence for three events at two trench sites. The low-angle thrust faults and associated sediments record the most recent and the penultimate events occurring between 1, 370 and 2, 500cal. y. B. P. (AD 580-BC 550), and between 7, 940 and 8, 430cal. y. B. P. (BC 5990-BC 6480), respectively. Vertically offset terrace gravels indicate the dip-slip rate of the Shimotsuburai fault to be about 0.5mm/yr during the past 22, 000 years. We also found evidence for the most recent two surface-faulting events on the scarp of the frontal fault of the Ichinose fault group. The most recent event, which accompanied a coseismic slip of 1.8-2.2m, is inferred to have occurred sometime between 3, 990 and 6, 270cal. y. B. P. (BC 2040-BC 4320). The penultimate event, which appears to have a coseismic slip of 2.1-3.0m, is constrained to have occurred between 9, 520cal. y. B. P. (BC 7570) and 10, 930y. B. P. The recurrence time and slip rate are roughly estimated as 5, 000 years and 0.5mm/yr, respectively. Regarding the long elapsed time since the most recent events, the Hakushu fault and the Ichinose fault group have accumulated enough strain to produce surface-rupturing earthquakes today. Based on such long recurrence times and lower slip rates on the southern ISTL relative to the central ISTL, we suggest that multiple segment ruptures of the central and southern ISTL are unlikely to have occurred during the past 10, 000years. However, partial synchronization of rupture timings for the central and southern ISTL might have occurred around 7, 000y. B. P. though. To evaluate the fault activity in the southern ISTL more precisely and to consider fault interaction with central ISTL, we still need to investigate other fault strands of the southern ISTL, and to gather more paleoseismic evidence.
It is well known that a great earthquake causes stress changes to activate and/or suppress seismicity around the fault zone. Occurrence of a large inland earthquake might be accelerated by stress changes caused by the other preceding earthquakes and/or accumulation of back slip on plate boundaries. In order to verify the hypothesis, we investigated temporal changes in Coulomb Failure Function (ΔCFF) before the occurrence of the recent two major inland earthquakes: the 1995 Hyogoken-Nanbu earthquake (M7.2) and the 1984 Naganoken-Seibu earthquake (M6.8). For the calculation of ΔCFF, we employed a dislocation theory for a semi-infinite homogeneous perfect elastic body proposed by OKADA (1992). First, we found that the value of ΔCFF on the source fault of the 1995 Hyogoken-Nanbu earthquake had increased by about 1.4×10-2 MPa due to the preceding events (1965-1994) and back slip for the 30 years. We can conclude that the Hyogoken-Nanbu earthquake was mainly accelerated by stress accumulation due to interplate coupling. However, since the value is too small, other factors such as block motion bounded by intraplate faults, northward slow slip along a detachment fault, and pre-existing fluid in the source region may also be considered as generation mechanism of the event. Next, we found that the value of ΔCFF on the source fault of the 1984 Naganoken-Seibu earthquake decreased by about 9.0×10-2 MPa due to the preceding events (1965-1984) and back slip for the 20 years. Decrease of the value of ΔCFF for the Naganoken-Seibu earthquake indicates that other factors such as collision due to plate motion and block rotation of microplates must be taken into account in the model to explain generation mechanism of the event. We also attempted to investigate seismic potential in the Hokkaido-Tohoku districts in terms of spatial distribution of ΔCFF for two cases: strike-slip and reverse faults with P-axis in the direction of plate motion of the Pacific plate (N285°E). For the former, correlation between 13 inland earthquakes and calculated increase of ΔCFF is obscure. On the other hand, for the latter, we found fairly good correlation between them. Therefore, there is a possibility that ΔCFF may be useful to evaluate seismic potential in the Hokkaido-Tohoku districts for the case of reverse fault.