The present status and the plans of crustal deformation observations, existing projects for earthquake prediction and the results of the GPS observations by the GSI are presented. The future possibilities of GPS for the application to the crustal deformation monitoring are also discussed.

View full abstract

Results by GPS and EDM observations in the case of the 1923 Kanto, Japan, earthquake (M=7.9) are estimated by using horizontal displacements of the first and the second order triangulation stations. It is found that distance and direction of baselines are related to observation value and long distance baselines in the NW-SE direction show larger distance change for this earthquake. Detectabilities by GPS, EDM and continuous strain observations are also compared and discussed. Combination of observations in taking into account their advantage and circumstances is recommended for the earthquake prediction study, though continuous strain observation has about 102 times as better detectability as the others.

View full abstract

We introduce the preliminary results from the observation by fixed-point GPS simultaneous baseline determination network in Kanto-Tokai district, central Japan. Data of initial observation during 33 days (between 97 and 129 day of year, 1988) are analyzed. Satellite constellation was not good over Japan in this period and only one-sided four satellites in the NW direction are tracked. Excess path delays caused by troposphere are calculated using surface meteorological data. Residual zenith delay are also estimated at each site. Satellite orbit parameters are relaxed when more than four sites are tracked. The deviated satellite constellation caused the errors in baseline lengths estimation for NW-SE direction 2-3 times larger than those for the NE-SW direction. The standard deviations are less than 1 ppm for most of NE-SW directed baseline of 70 to 260 km lengths. Estimated zenith delay correction terms vary much more than the variation of the wet term excess pass delay expected in this season. The correction terms for satellite positions range 0 to 200 m, and are also much larger than the accuracy of the ephemerides used. Those improper estimates are probably caused by the errors of the position of the fiducial point of the network and the surface meteorological measure ments, although a part of the deviation could be attributed to the errors in the satellite ephemerides we used.

View full abstract

Changing rates of the baseline lengths since 1984 in global very long baseline inter ferometry (VLBI) networks are distilled into the horizontal velocity of six Japanese, Pacific and European VLBI stations. Four North American stations are fixed as the reference. Obtained velocity vectors of the Pacific and European stations indicate that the Pacific and Eurasian plates have moved without noticeable internal deformation. The Japanese (Kashima) station, however, shows the velocity of about 3 cm/year toward WNW. This is inconsistent with available plate boundary models in and around Japan and suggests substantial intra-plate deformation, i.e. crustal shortening under the compressional stress field caused by the subducting Pacific, plate.

View full abstract

Highly accurate geodetic VLBI measurements has been repeated once a year since 1984 on the same baseline between Kashima and Tsukuba, about 55-km in length. Formal errors of five independent measurements were 6-mm on the average. The formal error of each experiment was studied in detail and found to be the same size as was expected theoretically. The sample standard deviation, however, was much larger than the ex pected error. The large deviation from the theoretical estimation, especially seen in the horizontal components, was investigated from the view of a coseismic surface dis placement due to an earthquake occurred at east off Chiba, in December 1987, the epi center of which was very near to Kashima station. The surface displacement of strikeslip type was calculated with a simple model but found to be too small, one fifth of the observed slip, suggesting that a more complicated model shall be established to explain the observed large horizontal slip, or otherwise, unknown sources of an error should be searched and formulated for a theoretical error estimation.

View full abstract

The range precision of Satellite Laser Ranging (SLR) techniques reaches one centimeter level today. However, even the accuracy of ordinary orbital determination of semi-long arc as five-days-arc remains at ten centimeters level and this high precision of ranging has not been effectively, used. Moreover, if we use longer arc technique to estimate precise baselines, worldwide SLR data should be used. It takes time to prepare worldwide SLR data and to complete the computation. If we estimate precise baseline lengths in a region of a diameter of a few thousands kilometers it is shown that we can use a specific short arc technique named SPORT which uses SLR data obtained simultaneously at two or more nearby SLR stations for successive passes instead of ordinary long arc technique. As an example of application of the SPORT, it is shown that baseline change for Quincy-Monument Peak line of about 884 km in California estimated to be -31 mm/year by using LAGEOS SLR data from August, 1984 to July, 1987. This result coincides with values given by other techniques. The baseline length between Simosato and Titi sima which was measured by using a fixed SLR system at Simosato and a newly developed transportable station by the Hydrographic Department of Japan was also determined as 937, 665.041±0.004 m using the same techni que. We also propose to use this technique for quick detection of symptoms of crustal deformation before a big earthquake.

View full abstract

Horizontal crustal deformation in the Japanese islands which obtained from the geodetic surveys has been discussed and we have obtained some interesting tectonics such as the crustal rifting and spreading in central Kyusyu, the collision between the Kurile outer arc and central Jiokkaido, and the collision between southwest Japan and northeast at the Itoigawa-Shizuoka Tectonic Line.

View full abstract

Crustal movements can be classified into two general modes; one is the anomalous crustal movement which is related with each individual shallow earthquake, and the other is the crustal movement which covers a wide area. The crustal movements which belong to the latter are thought to be caused by irregular motions of a plate, particularly in a subduction zone. An indicator from which those irregular motions of the plate are inferred must be the activities of the subcrustal earthquake which occurs near the plate boundary. The basis for this supposition is the fact that before and after subcrustal earthquakes of over a certain magnitude, there occur abnormal crustal movements. The mode of such abnormal crustal movements is different from that which is related with a single shallow earthquake. Unlike the case of shallow earth quakes, the rate of tilt velocity begins to change just before, or at the same time as the subcrustal earthquake. The change in the tilt-rate lasts for a considerable period of time, with this phenomenon being repeated in a series of subcrustal earthquakes. Judging from the magnitudes and the focal depths of the subcrustal earthquakes, it must be concluded that, the aforesaid modes are not attributable to the crustal move ments caused by a single event, but to an acceleration of the irregular motions of the descending plate, which subsequently brings about crustal movements over a wide area, and simultaneously, subcrustal earthquakes. In this paper some examples observed in the Kui Peninsula along the Nankai Trough are presented although the mode of the crustal movements over a wide area there is different from that of the northern part of Kyusyu-Ryukyu Arc or Southern Peru. The reason for this difference can be attri buted to the differences in subducting angles, thickness of both the continental and oceanic plates, and also to whether the driving force of a plate is "Ridge Push" or "Slab Pull" type .

View full abstract

The temporal variations of tidal strain amplitudes at the Mikawa Crustal Movement Observatory, in central Japan were investigated in detail using the strain data recorded with three silica-tube extensometers. The daily amplitudes of semi-diurnal tide M2 were calculated by the Fourier analysis from many series of 30 days' data which were ob tained by shifting the data successively step by step every day. The standard devia tions of the amplitudes of M2 from each detector usually remain within 5%, but some times the variations clearly exceed standard deviation. Such abnormal variation in tidal strain amplitude appeared four times from January, 1974 to December, 1987. The largest variation of the amplitude was found before the earthquake of March, 1983 (M=5.7, Δ=14 km, h=40 km) which occurred near Lake Hamana. The appearance of these ab normalities seem to be related to the earthquake occurrence around the coastal line. The observed results suggest that the abnormal variations in the tidal strain amplitude are probably reflecting the changes in the crustal properties in the earthquake source regions.

View full abstract

We present a theoretical formula of the gravity change caused by fault motion on a finite rectangular plane buried in a homogeneous half space. We find that the effect of the redistribution on the gravity change is of the same order of magnitude as that of the elevation change. Our result enables the quantitative treatment of the gravity change in terms of crustal deformation. Also important is that the theory leads a way to interpret Bouguer anomaly in the view of the past activities of active faults.

View full abstract

The Median Teconic Line (abbreviated as the MTL), the most important geologic breaks of the basement rocks in southwest Japan, extends nearly 900 km linearly running from ENE to WSW. It's middle part makes active fault system streching about 300 km or more in the central zone of southwest Japan (Fig. 1), and forms a master fault system with predominantly right-lateral component of displacement. Especially, in northern Shikoku basin-facing fault traces continue along the northern foot of the magnificent fault scarp fringing the northern rim of the Shikoku Range (Fig. 2). In Kyushu, southwestward continuation of the MTL obliquely transsects the central part of Kyushu with wide zone. This system extends with the E-W-trending normal faults, which dislocate Quaternary volcanoes in many places and form the graben struc ture at several areas. And also ENE-WSW-trending dextral faults are scatteredly developing with en echelon arrangement (Fig. 1). This zone streches further southwestwards to the sea bottoms of the Okinawa Trough through the Yatsushiro-kai, where back-arc sea spreading is taking place. The distribution of right-slip Quaternary fault in southwest Japan indicates that crustal shortening in a WNW-ESE direction is being accommodated by the MTL active fault system.

View full abstract

Research on regional crustal movements in Japan is reviewed from the points of views of its recent developments and the importance of space-geodetic techniques in it. Some data available for this research are investigated. Horizontal strain data for recent 100 years are digitized and analyzed. The analyses are; (1) filtering of the data for drawing smoothed contour maps of strain changes, (2) pre-estimation of strain rates for a hypothetical geodetic networks, and (3) introduction of Boundary Element Method for interpreting the obtained strain change of Japan by the subducting slab model. Relations of this reasearch with the space geodesy are also discussed. It is stressed that collocation works of different types of space techniques are essentially needed in this research, and that a refined data set (i.e. a standard model) of the regional crustal movements is indispensable in global space geodesy.

View full abstract

For the detection of an anomalous crustal deformation above the subduction zone, it is necessary to closely grasp the physical space-time characteristic lengths observed by the high accurate repetitional geodetic surveys, considering the limit of accuracy of obtained data and referring to various crustal deformation models given by many researchers. Especially, it is necessary to give a definite view on the substantial qualities from the space-time pattern in the secular variation of crustal movements basing on the recent detailed data of crustal structure. In this report, as an example of the investigation, Miura and Boso peninsulas are selected because many geodetic data on them have been accumulated since 1923 Great Kanto Earthquake (M7.9). These peninsulas are confronted with the Philippine sea plate and located under the influence of the Pacific plate and moreover, their tectonic beha viors show the various fluctuations due to these subducting plates. We call them "Time Dependent Peninsula Tectonics (TDPT)" and have been investigating them mainly by geodetic means to detect a uniform deformation in the various fluctuations through the subduction process. The outline of this investigation on the TDPT is as follows;(1) the secular variation related to some periodic change in elevation indicates the pos sibility of about 19-20 years variation. From this result, the long term earthquake prediction along the Japan trench suggested by FUJITA (1978) [1] may become possible in some aspects.(2) the intermediate term variation related to the subducting process indicates the im portance of the plate bending in the E-W direction of the Philippine sea plate due to the subduction of the Pacific plate. Then, recent deep earthquakes near the plate boundary are regarded as a phenomenon in releasing the increasing bending stress of the Philippine sea plate from the Pacific plate.(3) the annual variation related to seasonal change in groundwater or tide indicates the possibility of tidal loading effect. This result, however, is investigated now in detail. Such an investigation on the TDPT through geodetic means including detection of seasonal variation will be a great help in solving the problem on short-range earthquake prediction for practical use in the Kanto-Tokai area, though there are many points which must be solved.

View full abstract

The extension of the measurements of the crustal movements to the sea floor is important to open the window to direct observation of the phenomena which currently occur on the sea floor. This paper reviews the crustal movements measurements which have already been made and examines recent technical developments.

View full abstract

Bouguer anomaly map in the Central Japan delineates two remarkable negative regions. One is around the Central Ranges, Japan, in the Chubu district and the other is around the Lake Biwa in the Kinki district. Low gravity anomaly around the Central Ranges is undoubtedly attributed to the thick continental crust. However sources of the other low Bouguer anomaly around the Lake Biwa have been in question. The Philippine Sea plate is subducting beneath the Eurasian plate along the Nankai-Suruga trough in southwestern Japan and forms a cusp at the junction of the Kinki and Tokai districts. The later phase analysis using subcrustal earthquakes has led to the conclusion that a basaltic oceanic crust with a small velocity (density) structure is subducting with the underlying lithospheric mantle in Western-Central Japan. The upper surface of this oceanic crust is considered to be in direct contact with the bottom of the continental crust beneath the Lake Biwa. We present a new interpretation that the regional negative anomaly around the Lake Biwa can be explained by the contribution of the oceanic crust after correction of light surface sediments and water mass. Our result shows the resultant Bouguer anomalies can be well explained by the threedimensional plate modeling with reasonable assumptions of the thicknesses and density contrasts of the subducting slab and oceanic crust.

View full abstract

Electromagnetic observations of tectonic activities have been intensively carried out in the southern Kanto and Tokai district. Based on the earthquake prediction program, we have 15 proton magnetometer stations in this region. The dense network brought to light that the secular change rate of total intensity differs among some blocks with horizontal extent of a few tens km. This is probably ascribed to regional differences in the crustal stress. Local array stations of total intensity have been established in the Izu Peninsula where a high crustal activity was going on since 1974. Remarkable changes in the total intensity were observed near the east coast of the peninsula where ground uplift and swarm earthquakes took place. Magnetic data suggest that uplift and swarm earthquakes are caused by deep-seated magma and that magma sheets intrude off the Kawanazaki Promontory while a reservoir is formed on the south side of the dyke swarm. Combined observations of geomagnetic total intensity and electrical earth resistivity have been conducted in Izu-Oshima Volcano to monitor the volcanic activity. Distinct precursory changes in electromagnetic data appeared in association with the 1986 eruption. The electromagnetic method is useful to investigate the eruption mechanism.

View full abstract

To derive useful information of regional strain fields from the strain measurements in tunnels, we have investigated meteorological perturbations and reliability of instru ments. With regard to meteorological perturbations, remarkable strain changes of the order of 10^-6 were caused by rainfalls and temperature changes in the 30 m shallow tunnel at Iwakura, Kyoto. However, these disturbances have not been observed in the Amagase tunnel, in which many instruments were installed at positions of 400-500 m from the entrance of the tunnel having a length of 1830 m, and the depth of the observa tion site from the surface is about 140 m. Reliability of instruments has then been examined for several types of extenso meter installed in the Amagase tunnel by comparing the observational results of tidal strains with theoretically predicted values. The result shows that tidal strains observed with the laser extensometer are closest to theoretical ones. Long term strain accumulations observed with laser extensometers in the Amagase tunnel are 3×10^-7/year or so. We thus consider that the regional strain changes of the order of 10^-7 can be detected without meteorological disturbances, if the reliable instruments such as laser extensometers are installed in a deep tunnel having a length of 5001000m.

View full abstract

Some interesting examples and discussions on mechanical correlations between volcanic eruptions and stress (and strain) changes are collected with special reference to big earthquakes in subduction zones. Such a study seems to be useful for understanding the eruptive process and predicting both eruptions and earthquakes .

View full abstract