Search

近畿地方において，GEONETデータを解析し均質歪を除去した結果，クリープ性の変位を示す地殻変動境界(クリープ断層)が見出された。このクリープ断層は，明石海峡から琵琶湖北部を通り，多くの活断層を横切り，中部地方北部に達すると推定される。北東―南西走向のクリープ断層の存在は，東西圧縮下にある中部地方から近畿地方にかけての日本列島中央部において，巨大なせん断破壊がゆっくり進行中であることを示すと考えられる。, A GPS array with about 1, 000 stations (GEONET) is operating in Japan. The strain displacement fieldin the Kinki District was obtained from the GPS data analysis. When a homogeneous strain was removedfrom the displacement field, a fine boundary in displacement, which is assumed to be a creep fault, wasfound. The creep fault will extend from the Akashi Strait to the north part of the Chubu District through thenorth Biwa Lake. The existence of the fault will mean that a giant shear fracture is progressing slowly inthe center part of the Japan Island, the Kinki and Chubu Districts, under the E-W compression.

The Hyogo-ken Nanbu earthquake (MJMA7.2), that took place on January 17th, 1995, in Kobe and Awaji area, caused tremendous disasters in the areas. After the earthquake, the Japanese University Consortium for GPS Research (JUNCO) deployed more than 30 GPS receivers around the hypocentral area to find co-seismic and to monitor post-seismic crustal deforma-tions related to the earthquake. Data have been archived first by on-site recordings and later by tele-communcations at the Disaster Prevention Research Institute, Kyoto Universsty. Temporary dense array terminated by the end of March, 1995. Further temporary occupa-tions were conducted in May, 1995, in November, 1995, and in March, 1996, as well as long term continuous monitorings at selected several sites. The monitorings ended in August, 1996. Co-seismic offsets were observed at several sites around the source area. The largest offsets amounted to 45cm at Iwaya site which is about 4 km east of the Nojima fault. These data were used for simultaneous inversion together with strong motion data to clarify slip dis-tribution on the buried rectangular faults. Post-seismic deformations were also found at all of sites. They mostly showed temporal decay and amounted to 2 to 3 cm. The largest one was observed at Iwaya site. The areal distri-bution of post-seismic displacement vectors seems to indicate afterslip on the fault planes, but not the areal visco-elastic readjustments. Relaxation processes at Iwaya and Kawaragi sites were fitted by a logarithmic curve. Application of a theory on the mechanics of afterslip based on the constitutive relationship of a fault surface indicated that the fitting give reasonable es-timate on the frictional rate parameter or the thickness of velocity-strengthning layer of the earth's surface. Assuming that the constitutive parameter is ranging from 0.001 to 0.005 based on the results of rock experiments, thicknesses of velocity-strengthning layer at Iwaya and Kawaragi were estimated to be ranging between 500m and 2 km, which is consistent with other seismological data. Monitoring of fault offsets using real-time kinematic GPS (RTK-GPS) were also conducted at two baselines crossing the Nojima fault and the Arima-Takatsuki tectonic line, respectively. Though significant deformation was not observed, it showed a potential capability of real time monitorings of ground deformations in a few centimeter accuracy.

Movements of the Yamasaki fault in the Kinki District, were monitored by employing precise geodetic measurements once a year during the period from 1975 to 1993.The measurements were performed on a small aperture baseline network across the fault. The Hyogo-ken Nanbu earthquake (M = 7.2) occurred on January 17, 1995 in southern part of Hyogo prefecture. Measurements on the network, which is located at the epicental distance of about 60 kim, were carried out in November 1995.Analyzed data for the period of 20 years from 1975 to 1995 show as follows: No remarkable crustal movement due to the earthquake was detected. On the other hand, the values of average secular strain and tilt rates decreased from 1.4 μstrain/year and 0.95 μrad/year in 1975-1986 to 0.5 and 0.16 in 1987-1995, respectively.

A 3-components small borehole strainmeter was installed at Rokko-Takao station, Kobe, Japan, and observation has been started at July 1993. The station is located at 240 m under the ground surface in the side tunnnel which joins two tunnels for automobiles through the Rokko mountain area (Shin-Kobe and Daini-Shin-Kobe tunnels).Geological features of the Rokko mountain area are mainly characterized by the Rokko granite containing many active faults.The aim of this study is to obtain useful information for crustal strain observation by using the simple borehole strainmeter at the tunnel, the observation in which is disturbed by changes of atmospheric pressure and room temperature and oscillation of the ground due to running automobiles.

GPS/levelling geoid undulation differences were determined at 37 stations in Chugoku, Shikoku and northern Kyushu districts of southwestern Japan.Since the measurement area was widely extended, five reference stations, coordinates of which were accurately determined, were used in GPS measurements.The geoid differences determined by the present study were approximately consistent with the preceding geoids, such as the JHD geoid and the ORI-89 geoid.

Daily measurements by using GPS had been carried out at Kyoto, Kochi, Aso and Tottori sites in Southwest Japan from September 1990 to September 1991. The long term repeatability of the baseline lengths determined was 0.26 ppm, which was comparable to the short term repeatability of them on even as long baseline as about 200 km. The baseline lengths showed annual changes, amplitude of which were about 1 ppm. It seems that such changes were caused by insufficient consideration of tropospheric delay in computation of baseline vector.

Using tunnels and shafts in Ikuno mine, which was closed in 1974, we started in 1987 to observe seismic events and temporal change of 1) electric self-potential, 2) apparent resistivity, 3) underground water level, 4) water seepage from tunnel wall and 5) γ ray intensity. The main purpose of our comprehensive observation is to discuss correlation between the observed data in such a stable circumstance as that without excavation, taking account of the continuous observation of crustal movement which have been carried out before the mine was closed. Since all tunnels and shafts below the underground water table filled up with water now, we expect to catch any behavior of rock body in response to any change of underground water and also any preseismic or postseismic effect. From the observation of about five years, we found that the seismic events (S-P time < 0.3 s; M< 0) still occur even now in the vicinity of mine. From our preliminary analysis, the secular change of electric self-potential seems to correlate to the activity of the seismic event in the mine.

Daily measurements by using GPS have been carried out at Kyoto, Kochi, Aso and Tottori sites in Southwest Japan since September 1990. The aims of the present measurements are to estimate a long term repeatability of the baseline vectors determined, to investigate effects of error factors and finally to improve the accuracy of GPS measurements.Results for the first six months are as follows:(1) The long term repeatabilities of the baseline lengths of about 150 to 500km were 0.2 to 0.4 ppm.(2) The baseline length in summer was determined longer than that in other seasons.(3) There were positive correlations among the changes of the baseline vectors.

In order to investigate the stress state and the strain accumulation in the Earth's crust, andto presume the possibility of earthquake-occurrences, three dimensional absolute stress measure-ment and the stress gradient measurement were carried out at 3 sites in Northwest Kinki District.Respective results obtained by the two different measuring methods, hydrofracturing and stressrelief methods, do not always coincide three-dimensionally with one another due to the condi-tion of the measuring points, however, the azimuth of horizontal principal stress agree well withone another, and most likely reflect the tectonic stress of the region. Moreover, as for the resultsof the horizontal stress measurements by hydrofracturing method at respectiv depths, the scatter-ing is within 7 to 8 in the azimuthal angle. The regional differences observed are due not tothe accuracy of measurement, but probably to the local characteristics of the measuring sites.There is a considerable scattering in the stress values measured at the shallow part near theground surface. As seen in the results measured at Hoden, there appear low stress parts at in-tervals of approximately 20 m. This low values may be caused by regularly existing joints in thelarge rock mass and by low overburden pressure at shallower parts of the upper crust. The stressis constantly released at the large joints or small fracture zone through relative micro-displace-ment of each rock block. The stress depth gradient in this region is relatively lower than theWest Japan standard stress gradient. When the data at Hiraki measured in 1985 and 1989, especially their depth distribution, are compared, it is infererd that the principal stress azimuthand the stress gradient have been changing and increasing. This stress changes may be relatedto the seismic activities that continued since around 1986 near Kameoka in East 40 km awayfrom Hiraki.

For the monitoring of crustal movements and the long-term earthquake prediction, GPSobservation systems were introduced into Japanese universities and institutes in 1988. Thisreport gives the detailed description of the semi-fixed GPS observation system introduced intoD.P.R.I., Kyoto University and the results of some experimental GPS obesrvations. Thepreliminary results, which are derived from only a few observations, show the order of 1 ppmor better repeatability of GPS measurements in 1988, when the constellation of the GPSsatellites over Japan is not so good.

The observation of crustal strain has been carried out at Ikuno Geophysical ObservationStation using 4 components of extensometers since 1977. The observed results at Ikunomake it clear that changes of maximum tilt direction and principal strain axis have occurredsystematically. These phenomena suggest time changes of stress field in the northwesternpart of Kinki District.To confirm these stress changes, in Jan. 1988, measurement of three-dimensional absolutecrustal stress was made using hydro-fracturing methods in three boreholes near the observa-tion station. In Aug. 1988, using the same boreholes, the measurements of absolute stresswere again made repeatedly. Then, stress changes were estimated and further, strainchanges were calculated through the stress-strain relation in the two-dimensional modeling.The obtained data were finally comparative-studied with those data obtained by the exten-someters.Because of the fact that absolute stress values themselves are low, measurements werenot very accurate. On the other hand, the amounts of observed strain changes, due to cavityeffect and inhomogeneity of rock in the observational tunnel, lead to noticeable errors ofanalysis. As a whole, the observed strain changes agree well in characteristics with calcu-lated strain changes from the results of stress measurements within the observational error.But Young's modulus must be taken 200 Gpa, several times of elastic constant (E=26 Gpa)by the uniaxial compression test of in situ rock.Object of this paper, however, is not to show coincidence of both the measured values.The real purpose is to indicate a future prospect which enables Earthquake Prediction Studyto elevate from conventional "qualitative" discussion to "quantitative" one by this compara-tive observation of stress and strain.

Data obtained by precise geodetic measurements on the network across the Yamasaki fault, Southwest Japan, were analyzed for the period of 11 years from 1975 to 1986. In order to monitor the crustal movements across the Yamasaki fault, precise geodetic measurements have been carried out since 1975 annually in November or December on a small aperture geodetic network ("Yasutomi-Usuzuku Baseline Network") across the fault. Baseline measurement, triangulation and trigonometric leveling have been applied here. The accuracy of our measurements are possibly high enough to detect relative movements among the six stations as small as 1mm. Analyzed data show non-uniform strain accumulations with reversed sense between the NE-SW directed principal strains in the northeastern and southwestern part of the network. A secular strain rate of the order of 10^-6 strain per year was estimated based on the strain data from five diagonals of the pentagonal network, and a secular WSW-down tilt rate of about 1×10^-6 rad per year was obtained from height data of all the six stations of the network. Furthermore, episodic strain and tilt changes were superposed on the secular trends. We also pointed out increases in compression rate and deviations of tilt from the secular WSW-down tilt associated with two moderate (M=3.7 and M=5.6) earthquakes having epicentral distances of about 3km from the network, and their focal depths were about 20km. These episodic movements seem to have had begun one/two years before the earthquakes and to have recovered their secular trend after the events. On the other hand, no remarkable crustal movement was detected for the moderate earthquakes having epicentral distances larger than about 15km.

In order to monitor fault movements, precise geodetic measurements have been repeatedlycarried out at the Yasutomi-Usuzuku base line net across the active Yamasaki fault since1975 once a year.The following modes of crustal movements were observed in common before and afterthe two earthquakes of magnitude 3.7 and 5.6, which took place in the vicinity of the netin September 1977 and May 1984 respectively.1. As to horizontal movements, the areal strain of the net begins to show the contractionin one or two years before earthquake occurrence and turns to expand at the time or justafter it.2. As to tilting movements, the ground begins to tilt towards the opposite direction ofthe epicenter in one or two years before earthquake occurrence and amounts .of this tilt arerecovered within one year after.On the other hand, no significant change could be observed as to the earthquakes havingepicentral distances of about 15 km and over.

In order to investigate the behavior of the active Yamasaki fault, a pentagonal base linenet was set up across the fault at Yasutomi-Usuzuku (Hyogo prefecture) in July 1975.Since then, precise geodetic surveys have been carried out every year in November orDecember.The methods employed in the survey are horizontal and vertical angle measurementsusing a theodlite Wild T3, and precise base line measurement using several base line tapes.The angle measurements have been done at night for being freed from fluctuations of targetimages. The accuracy of the present observations is of the same degree as that for first ordertriangulation, consequently relative horizontal and vertical displacements of 1 mm would bedetected.In this report, the outline of the observational method and a part of the results from1975 to 1986 are described.