Search

From the volcano-energetics point of view, it is very important to evaluate the velocity, direction and distribution of ejected volcanic blocks. However, in most cases, it is hard to measure directly such quantities during the eruption. Therefore, the volcanic blocks landed on and around the crater have been investigated to reproduce dynamic aspects of eruption after the eruption. In the present paper, ballistic curves of the volcanic blocks are numerically analyzed for various combination of parameters, such as direction of the explosion principal axis of ejection, velocity and direction of the wind, air resistance to the block and dependence of direction of the initial velocity of ejection. The calculations are made on the basis of a generalized ballistic equation of ejected blocks, in which the wind velocity and the air resistance to the block are treated as vector quantities. Calculated results lead to the following characteristic conclusions: (1)The winds above 20m/s along the direction of ejection plays the major role in the ballistic curves, while the lower wind velocity is less effective. (2) Spatial distribution of the landed blocks is mainly affected by the direction of the explosion principal axis. The distribution of the volcanic blocks which was spread to the specific direction is described by inclination of the explosion principal axis. (3) The spatial distribution of the small size blocks depends remarkably on the air resistance.

Seismic waves generated by a nuclear test carried out by North Korea on October 9, 2006 were observed throughout the Japanese islands. Clear arrivals of Pn waves, which traveled below the Moho discontinuity underneath the Japan Sea, were recorded by the Japanese broad-band seismic network, F-net, deployed by National Research Institute for Earth Science and Disaster Prevention (NIED). We analyzed the waveforms of the nuclear event observed at F-net stations as well as some global seismic stations in East Asia. Apparent path-average velocity of Pn waves and their maximum amplitudes are estimated from vertical component seismograms of the F-net stations. We found conspicuous regional variations of apparent path-average velocity and maximum amplitude. For example, fast apparent velocity and larger amplitude are observed in the paths across the Japan basin in the northern Japan Sea, whereas slow velocity as well as relatively smaller amplitudes are found in the path to stations in Kyushu. Such regional variability of Pn waves is likely to make it difficult to estimate source parameters for this small-scale explosive event in the Korean Peninsula, only using the Japanese seismic network. These results suggest the necessity of a precise three-dimensional seismic model of the Japan Sea to utilize Japanese seismic network data for analysis of regional seismic waves that propagate along a variety of paths in the Japan Sea.

The northern part of Hokkaido (Dohoku region) is recognized as the boundary region between the Okhotsk and Amurian plates. In this region, strain concentration is reported based on triangulation and GPS observations. However, there had not been observed any inland earthquakes larger than MJMA 6 in Dohoku region, until the 2004 Rumoi-Nanbu earthquake (MJMA 6.1) occurred. There may be a possibility of occurrence of intraplate large earthquake in the future. The mechanism of the earthquakes observed in the region is strikeslip and thrust type, owing to the east-west compression. In case the thrust type earthquake had been rupturing recurrently, it should be resulted as positive Bouguer gravity anomalies. There exists plenty amounts of gravity observation data in this region, for the purpose of oil and mineral exploration. However, it exists as a hand written paper document. Additionally, these data are basically unreleased. So we performed gravity measurement in the Dohoku region to draw more detailed Bouguer anomaly map. The observed gravity data is listed in this document.

Spatial distribution of volcanic blocks, as well as landing distance from a crater, landing angle and diameter of the blocks has been well investigated on and around the crater in cases of the 1938 Asama and the 1977 Usu eruptions by some researchers. Then, this paper reproduces detailed kinematic aspects of the eruptions so as to interpret the above-mentioned field data, on the basis of a generalized ballistic equation of ejected volcanic blocks proposed by the present authors. Calculations are made for combination of various parameters, such as air resistance to the blocks, direction of the wind, orientation dependence of the initial velocity and so on. The observed wind velocities are referred in the calculation. Comparison between the calculated and the observed results leads to the following conclusion: (1) In the 1938 Asama eruption, the distribution of volcanic blocks shows that the direction of the explosion principal axis is due east (N90°E) and its inclination angle is 30°. The angle dependence of the initial velocity (M = 1.5) is almost the same as that in the Sakurajima volcano. (2) In the 1977 Usu eruption, direction of the explosion principal axis is N170°E and its inclination angle is 20°, and the angle dependence of the initial velocity is large as much as M = 5.0. (3) The air resistance to the blocks is estimated as about Kd = 0.5 in both the eruptions. This value is valid in other volcanic eruptions because the air resistance seems to be same everywhere.

The radiation belts are the region that energetic charged particles are trapped by Earth’s magnetic field. The radiation belts have been observed by a lot of satellites and researched by theoretical approach since the discovery of it. The electron radiation belt is separated into two regions called the outer belt and the inner belt by the slot region in which the flux is very small. It is well known that the outer belt flux vary by several orders of magnitude associated with geomagnetic disturbances (e.g., Baker et al., 1994) and the exact mechanism of such variations has been a puzzle so that the outer belt flux variations have been well researched to explain this in two-decade. On the other hand, the inner belt is more stable and the flux variation in the inner belt is smaller than that of the outer belt so that the inner belt has not been attracted much attention. A radial diffusion model can reproduce an equilibrium structure of the electron radiation belts (Lyons and Thorne, 1973). Radiation belt electrons are supplied from the outer boundary of the radiation belt, and the flux is arranged by the balance of intensity of the diffusion and the loss due to pitch-angle scattering by whistler mode waves in the magnetosphere. The electron flux variations of the radiation belt can be investigated by using a radial diffusion model with time-dependent radial diffusion coefficients and a time-dependent outer boundary condition. Brautigam and Albert (2000) investigated the outer belt electron flux variation by using the time-dependent radial diffusion model. The time-dependent radial diffusion coefficients formulated by Brautigam and Albert (2000) are customarily used in the time-dependent radial diffusion model. They parameterised the amplitude of the electrostatic field variation as a linear function of Kp index, and applied them to the electrostatic coefficient derived by Cornwall (1968). They were also formulated the electromagnetic coefficient as a function of Kp index. Because these coefficients are based on the observation of the outer belt, it is not appropriate to apply them to the slot and the inner belt regions. In fact, extrapolating these diffusion coefficients to the slot and the inner belt and doing numerical simulation, the slot is not formed and the electron flux near the Earth region is extremely large. In this study, in order to find a radial diffusion coefficient that can apply the inner part of the electron radiation belt as well as the outer belt, we performed some simulations by using the time-dependent radial diffusion model assuming that several types of the amplitude and the profile of the electrostatic field variation. The results show that the amplitude of the electrostatic field variation depends on L and is smaller toward the Earth.

Secular and short term strain variations before and after the 2003 Tokachi-Oki earthquake, Mw8.0, recorded at five continuous crustal movement stations located along the Pacific coast are reported. There is no obvious forerunner movement at all stations. However, clear co-seismic and post-seismic strain changes are recorded at all stations. Strain data are digitized by 18 bits with 1 Hz sampling rate so that the complete strain seismogram of Mw8.0 event were recorded at MYR station nearest the source of the event. Strain seismograms recorded at all stations in Hokkaido are also reported.

Cooling of intruded magma, development and decay of hydrothermal systems are investigated by model calculations for various combination of parameters such as hydraulic permeability of the intruded magma and its surroundings. Dependence of the pressure gradient on the hydrothermal systems is also investigated to inspect the effect of the potential flow of the groundwater along the topography. In the case that the horizontal pressure gradient is smaller than buoyancy, the hydrothermal system develops just above the intruded magma, while the greater pressure gradient as well as high permeability of surroundings carries the hydrothermal system to downstream. We applied these examinations to the newly formed Nishiyama geothermal field at the time of the 2000 Usu eruption. Numerical calculations considering the real topography in and around the Nishiyama geothermal field well reproduce the observed temporal variations of geothermal field appeared on the ground surface, assuming the in-situ hydraulic permeability of the intruded magma and the surroundings as l0－12～10－13m2 and 10－l0～10－11m2, respectively.

The energy source of volcanic activity is the heat of magma. Heat discharge rate on the initial stage of Usu 2000 eruptive activity was estimated to be around 1×109 J/sec, that was over one order higher than that on the initial stage of the 1977 activity, and comparable to its peak release rate observed after two years from the beginning of the eruptive activity. Compare to the 1977 activity, much shorter duration of the 2000 activity may be the result of rapid cooling of the intruded magma. Abundant ground water supplied from Lake Toya might remove the heat effectively through active vapor plume rise.

The 1977 eruption of Mt.Usu generated four high-altitude eruption columns which produce BIGⅠ,BIGⅡ,BIGⅢ and BIGⅣ tephra-fall deposits respectively. Expressing the initial eruption velocity in terms of the solid mass eruption rate, the initial temperature, the initial radius and the initial gas fraction from the conservation of solid particles, we apply a steady plume model to the four eruption columns. The initial velocity increases with initial gas mass fraction and initial temperature for the solid mass eruption rate estimated from the volume of deposits, the apparent density of tephra and the eruption duration. For three eruption columns depositing BIGⅠ,BIGⅡ and BIGⅢ, the initial temperature and initial gas mass fraction are estimated at 600K and >0.11. The eruption column depositing BIGⅣ is formed at 800K and >0.15. The estimated initial temperature are low and initial gas mass fraction are high in comparison with general physical properties of dacitic magma. These inadequate values suggest that the growth of eruption columns take the influence of pyroclasts fallout from the column and thermal disequilibrium between pyroclasts and gases.

Ion upflow has been investigated by use of ESR (EISCAT Svalbard Radar), SuperDARN, ACE satellite and CHAMP satellite. The field-aligned ion upflows were observed by ESR, which occurred in night-time between 16 and 17, April 2002, and 19 June 2004. However, ESR didn’t show remarkable increase of electron density, electron temperature and ion temperature, which are generated by auroral particle precipitation or frictional heating. The Solar wind parameters were obtained by ACE satellite. SuperDARN showed variation of convection electric field depending on By component of Interplanetary Magnetic Field. CHAMP satellite observed also variations of neutral density, neutral wind, ion density and electron temperature along the satellite orbits. At the time, ESR was in the polar cap region. Our analyses may suggest that the ion upflow observed by ESR was generated in the upwind region of plasma convection in the high-latitude ionosphere and was moving to ESR station by the convection.

We conducted self-potential (SP) surveys on Usu volcano since July to December 2006. The compiled SP map reveals positive anomalies around Ko-Usu lava dome and at the foot of the volcano, and a negative anomaly on the top of O-Usu lava dome. The SP profile on the summit caldera shows the same pattern of 1985’s. However, the peak-to-peak variation of the SP value is different: the 1985’s of that shows 1000 mV while the 2006’s shows 1400 mV. Topographic effect is clearly shown along the southwestern foot of the volcano, in which coefficient is about -2.5 mV/m. The SP profile corrected of the topographic effect reveals a large and extensive positive anomaly over +600mV and several local positive anomalies over +1000 mV on the summit caldera. The corrected SP also reveals that a positive anomaly at the south part of Ko-Usu lava dome and a negative anomaly at the O-Usu lava dome do not exist. The revealed positive SP anomalies are likely to be affected by an extensive altered layer, located beneath the summit caldera. The largest positive anomaly is shown on the ridge of Usu-Shinzan cryptodome after topographic correction. This anomaly is not regarded to be formed by hydrothermal upwelling, because any indications of fumarolic and geothermal activities have not been observed on the ridge of the volcano. The SP anomaly may be affected by intruded material into the volcano. The SP amplitude in the northwestern part of Nishiyama is very small in spite of rugged topography. An extensive low resistivity layer (< 10 ohm-m) located in the shallow part in the northwestern of Nishiyama probably shields the SP variation.

Many authors have discussed the eruption time sequence and the mechanism of volcanic eruptions by estimating the ejected velocity of the volcanic blocks, with the aid of other factors of eruptive phenomena. Equations of motion for the volcanic blocks have been proposed by considering the factors, such as the air resistance, the influence of the wind, the inclination of the explosion principal axis and so on. However, the equation of motion and its solution has not been considered that the air resistance power was vector quantity for the flight of volcanic blocks. In this paper a new equation of motion for the ejected volcanic blocks is proposed by considering the air resistance power as vector quantity.

The present authors pointed out, in the last paper, a mistake in the equation of motion for volcanic blocks. Reliable assumption of the air resistance to the blocks is required for accurate estimation of initial velocity of volcanic blocks. However, little attention has been given to the point. This paper organizes the problems about the estimation of the air resistance and proposes a new method of the estimation. Also this paper organizes the effect of factors on the distribution of volcanic blocks. Model calculations revealed the effect of the atmospheric air density cannot be ignored. For the effect of explosion principal axis on the distribution of volcanic blocks, this paper proposes an extended equation which volcanic blocks eject to all the directions around the crater.

We performed temporary dense observation at Usu volcano in summer 2006, aiming to reveal micro-earthquake activity under the summit crater during dormant period of the volcanic activity. During about 80 days of temporary observation, more than 330 volcanic earthquakes were observed. Of these earthquakes, well-recorded 142 events are selected for relocation analysis assuming the 3D-velocity structure model which is a result of the explosion seismic experiment at Usu volcano in 2001. Following the result of relocation, we estimated focal mechanism of 7 earthquakes whose magnitudes are more than 1.0. The temporal observation and analyses gave us three new insights about seismic activity; 1) Continuous activity of volcanic earthquake is seen at the shallow part under the southern crater floor, 2) Hypocentral regions are distributed along the SW end of the U-shaped fault which was built in the 1977-1982 eruption associated with the uplift of Usu-Shinzan crypt dome, 3) Strike direction of the 7 major earthquakes are almost parallel to the U-shaped fault, and dominant mechanism is dip-slip type which suggests subsidence of summit domes relative to the southern crater floor.

The radiation belts are the region that energetic charged particles are trapped by Earth’s magnetic field. It is well known that the energetic particles flux are varied during geomagnetic disturbances. Many researchers have studied about the flux variation of radiation belt, but the mechanism of the variation has not been understood in detail. A radial diffusion model can reproduce a basic structure of the radiation belts. Radiation belt particles are supplied from the outer boundary of the radiation belt, and the flux is arranged by the balance of intensity of the diffusion and the loss due to pitch-angle scattering. An intensity and a distribution of the electrostatic field and the whistler waves which determine the magnitude of the diffusion and the loss of radiation belt particles are uncertain. Cornwall (1968) derived the radial diffusion coeffcient due to substorm convection electrostatic field. Brautigam and Albert (2000) derived the root mean square of electrostatic field which is a linear function of Kp index from observations. Lyons et al. (1972) calculated the pitch-angle diffusion within the plasmasphere and the lifetime of the particles. Solving the radial diffusion equations using these functions, the slot region is not formed and the flux near the Earth region is too large. Diffusion coeffcient is reduced by 1/100 at L < Lpp, the slot is formed and too much flux injection from outer belt is suppressed. The change of radial diffusion coeffcient at the plasmapause location may affect the structure of radiation belts.

The thermal activity of Iwodake volcano is characterized by the predominant volcanic gas ejection at the summit crater and the heat discharge from high ground temperature area which is distributed widely from the summit crater to hillside of the mountain. The volcanic gas at the summit crater is of magmatic origin and its temperature is observed to be 880°C in maximum. The total amount of the volcanic gas discharged from the summit area is estimated to be 150kg /sec from the SO2 measurement and the chemical composition of the volcanic gas. The heat discharge rate form the high ground temperature area at the hillside, which is distributed around the fumaroles is estimated to be 50MW from the surface temperature measurement. These thermal activities of Iwodake volcano is thought to be continued for more than 800 years. A plausible model to explain the long degassing of large amount of volcanic gas is magma circulation in a conduit which extended from the deep-seated magma reservoir. The continuous active degassing probably causes the hydrothermal system within the volcano because the volcanic gas ascending the conduit from the top of the magma to the surface is diffused to the surrounding formation. The development of such hydrothermal system is studied using the mathematical simulation. The result shows that the overall thermal activity of Iwodake volcano such as the volcanic gas ejection at the summit crater, widely distributed ground temperature anomaly at hillside and hot springs along the coast can be caused by the degassing activity. The important factors in order to induce the wide-ranging hydrothermal system are permeability of the volcanic edifice and the depth of the degassing. The simulation indicate that the permeability of 10-13m2 and the degassing at sea level are suitable condition for the Iwodake thermal activity.

Many researchers have studied the thermal state of the crust and the upper mantle and its tectonic implications in Hokkaido, Japan. Terrestrial heat flow data have revealed the following characteristic features: (1) High heat flow (84〜125 mW/m2) is observed in the backarc side of the volcanic front, while low heat flow (< 42 mW/m2) is observed in the forearc side, showing the typical thermal aspects of the island arc. (2) The central axis zone of Hokkaido (the Kamuikotan and Hidaka metamorphic belts) shows relatively low heat flow (42〜80 mW/m2). The low heat flow zone extends for Sakhalin. (3) Extremely low (25〜34 mW/m2) and high (125 〜210 mW/m2) heat flow is distributed locally on the Ishikari-Tomakomai low land and its neighbor on the west, respectively. Based on the heat flow data and other geophysical data, such as the seismic velocity structure, aeromagnetic data, and so on, the Curie point depth, the solidus temperature depth, temperature at the Moho, and the mantle heat flow have been estimated. The estimated thermal structure implies the temperature of the lower crust comes to the solidus temperature beneath the backarc side of Hokkaido. The result is well supported by the electrical resistivity soundings: the low resistive layer (10~40 ohm-m) develops in the lower crust and is ascribed to the effect of partial melting of the crustal rocks.

We investigated geomagnetic changes due to dome-like deformation by using a truncated circular cone model. In the case of the normal magnetization, magnetic total field decreases in the northern side of the uplift center, while it increases in the southern side. We found that the original topography prior to the mound formation does not affect the pattern of this magnetic change. We then apply this calculation to the mound-forming activity in the 2000 eruption of Usu Volcano to infer the magnetic change on the foot of the mound. The expected range of magnetic total field change at the station U01 is -80 to -800 nT which is in the opposite sense to the observed actual change (+70 nT). We examined some alternative mechanisms to compensate this decrease due to the deformation. None of them, however, can be a promising candidate. The most reasonable explanation so far for this contradiction is the reversed magnetization of the Tertiary lava that may comprise the mound.

An earthquake (MJMA 6.1), occurred in December 14, 2004, at southern area of Rumoi sub-prefecture, northern Hokkaido, Japan. The earthquake intensity was assumed to 6 in Obira town. In Hokkaido, this is the first inland earthquake larger than M6 since Teshikaga earthquake (MJMA 6.5, 1967). Detailed aftershock distribution shows that the aftershocks occur on an eastward dipping plane of 10 km square, depth of which 5 km. The mainshock assumed to have occurred near the southern end of this plane. Despite the sparseness of existing gravity stations over the focal region, it was found that the aftershocks occur just above the gravitational rise of about 20 mGals. So the gravity survey was operated in Nov. 2005 to Aug. 2006, to make a detailed Bouguer anomaly map and to explicate the subsurface structure of the focal region. The measurement was done with CG-3 type gravimeter. Measurement positions were carefully decided by operating differential GPS observation. The accuracy of altitude decision is within 1 meter in all of the measurement points. Besides, data provided by Japan Petroleum Exploration Co. Ltd. was very helpful. The appropriate assumed density was detected to 2300 kg/m3, using density log of deep borehole near the focal region, and the correlations of gravity correction terms. The Bouguer anomaly map by current dataset also shows good coincidence with aftershock distribution.

我々は異常地殻変動検出手法の開発に取り組んでいる。時定数の短い変動を捉えるためには，キネマティック測位を応用した手法を用いて，電離層や対流圏遅延などの誤差要因を除去しなければならない。これらを移動平均によって±1cmの精度まで取り除くことができた。さらに長基線に展開するため，新たに空間フィルター法を適用し共通誤差成分を取り除くことによって，基線の長さに関係なく±1cmの精度で議論できるようになった。また衛星軌道暦の違いによる測位精度への影響はほとんどないので，準リアルタイムモニタリングシステムの構築が可能になった。, We are taking with the development of the technique which detects the anomalous crustal deformation by using GPS now. The static positioning method which used a daily solution was the main stream with the positioning which used GPS till now. However, the short-term change can’t be caught in this method. So, we think that the kinematic GPS positioning method being used for the moving object was applied. But, we must remove various error factors which exist in this technique. As for the short baseline which went until last year, we could remove the influences of the ionospheric and tropospheric delay, and the positioning accuracy is +/- 1 cm. We extend a baseline more this year. We succeed in removing the error factors from the sky by using a spatial filter, and we can get the same result as the short baseline. And we must remove the error to result from the satellite orbit to monitor it in real time. As that result, a large difference wasn’t seen in the positioning accuracy which satellite orbit announced from IGS was used.

マグマ貫入後の熱水流動のシミュレーションを行い，熱水系の消長によって地表での地磁気がどのように変化するかを数値的に評価した。シミュレーションで得られる温度分布から火山体内部の磁化分布を見積もり，地表での地磁気変化を計算するポストプロセッサーを開発し，マグマの貫入の深さおよび浸透率が地磁気変化に与える影響を調べた。その結果，より浅くマグマが貫入する方が，地磁気変化が大きく，地磁気変化が観測される時間が短いこと，また，浸透率が小さいほうが，地磁気変化が大きく，地磁気変化が観測される時間が長いことがわかった。, We have developed a postprocessor to calculate the geomagnetic field variations caused by hydrothermal activities in the volcanic region. They use distributions of temperature inside the volcano, which are results of a numerical simulation for hydrothermal activity after magma intrusion. The most dominant mechanism of volcanomagnetic effects was considered here: thermomagnetic effect. We examined the influences of host rock permeability and depth of magma intrusion on thermomagnetic effect. Results showed that the shallower depth of intrusion causes larger total intensity anomaly and earlier decline of it and that larger anomalies appear at later times in cases of lower permeability host rocks.

We estimated the predominant period of surface waves from the dispersion curves of the Love and the Rayleigh fundamental mode, calculated from 1-D verlocity structure models extracted from the Osaka basin model.The estimated predominant period is from 1 to 10 s for the Love waves and 1 to 6 s for the Rayleigh waves in the Osaka plain. The linear relationships between the predominant period of surface waves and the bedrock depth are confirmed. Predominant period is proportionality to the bedrock depth. Coefficient proportionality approximately equals 5 s/Km for Love waves and 2 s/Km for Rayleigh waves. We compared the estimated predominant periods with the observed ones at strong ground motion observation stations maintained by CEORKA (Committee of Earthquake Observation and Research in the Kansai Area)for the main shock(M_J7.4)and aftershock(M_J6.5) of the 2004 off the Kii peninsula earthquake. Except for some CEORKA sites, where the depth of the bedrock is deeper than 1.5 Km, the estimated predominant period of surface waves is consistent with the observed one. This result is the first approximation,effective for the predominant period of surface waves in the Osaka plain. However,there is a possibility of a variation of the predominant period for the surface waves coming from a different direction.

Wide-band magnetotelliric and gravity surveys were conducted at the Teshikaga area, located in eastern Hokkaido, Japan, where intra-plate earthquakes had teken place. We inferred shallower (

A high dense gravity measurement composed of 323 sites was carried out in Amami-Oshima Island,Kagoshima,in order to reveal the detailed gravity anomaly distribution. The obtained distribution shows that Amami-Oshima Island has the positive anomalies ranging from +32～+66 mgal with a trend from NE to SW parallel to Ryukyu trench. The trend-reduced gravity anomaly distribution also indicates the northern and southern areas in the island with negative gravity anomalies.

Series of Mogi's single source model display characteristic family curves on the logarithmic graph of vertical displacement versus horizontal distance. The family curves on log-log plots are simply characterized by the flat part at short distances and constant decay at farther distances. The specific corner distances, which are defined by the bending corner of the curves, are mainly controlled by the source depth. The edifice scale deformation associated with the 2000 eruption of Mt. Usu is far from that predicted by Mogi's single source. The vertical displacement shows much more sharp decay with distance around 5-6 Km and uplift-to-subsidence reversal at much farther distances on the logarithmic graph. This steep decay and a farther reversal can be well explained by a simplified twin spherical source model, with shallow inflation and deep deflation. It provides better fitness to the data, when the deflation volume at 10 Km is assumed to be nearly equivalent to the inflation volume at 2 Km. Similar result is also evident for the horizontal model. The combined twin source model of shallow inflation and deep deflationi is also applied to the available leveling data for the histolical 3 eruptions of Mt. Usu in 1910, 1943-45 and 1977-82. Edifice-scale deformations accompanied by 4 dome-building eruptions in 20th century can be also well explained by shallow inflation and deep deflation source. Strong similarity between the 2000 and the 1910 eruptions is especially interesting considering the smilarity of mutual surface eruptives activities and crypto-dome formations.

We have developed rotation seismometer using four translation seismic sensors (L22d), which have the same amplitude and phase characteristics. A negative input impedance amplifier and band-pass filter have eliminated mechanical deviations of individual sensor and adjusted to have suitable characteristics. Temporal earthquake observations have executed using translation and rotation sensors, at three places, which have different surface geology. They were Mitsuishi (MUJ), Obihiro (OBI), and Usu volcano (SWZ) in Hokkaido. We demonstrate some seismograms of rotation and translation components. Between maximum amplitudes of rotation (Ar) and translation (At), an emprical formula; Log Ar=p Log At (1) has been found, and p takesdifferent value for each observing site. We have obtained p value to be 1.39 for MUJ, and 1.29 for OBI, from local earthquakes seismograms. At SWZ, we obtained 0.84 from volcanic swarm type seismograms, which contains some low frequency type.

We examined 2-D heterogeneous structure in the eastern Hokkaido region, using the S-wave coda amplification factor (CAF) with seismic storong-motion date of Kyoshin Net(K-NET). CAF is defined as the spectral ratio of coda waves among different stations after corrections for source, site, and overall propagation.(i.e,cada-Q) effects. Confirming that code-Q values are nearly independent of both source and station, the site effect at each station was esetimated by the coda-normalization method. The source effect was removed by taking the spectral ratio of S-wave coda among stations for each earthquake. We evaluated the CAF value of each source-station pair for three components,using 95 seismograms for five local earthquakes recorded at 21 stations of K-NET. A map of CAF values at the frequency band of 16 Hz shows a systematic variation, implying that heterogeneities are concentrated at the central area (Shibeccha) in the eastern Hokkaido region, manifested locally as a large degree of scattering.

We have investigated the source process and the aftershock distribution of a large earthquake with M=7.1 occurred on 29 November 2004 off the coast of Kushiro (or Shiranuka) in the eastern Hokkaido,Japan. The source parameters are summarized as follows:the total seismic moment M_0=3.4×10^{19}Nm,thatis,M_w=7.0;(strike,dip,ship)=(238°,33°,117°);depth of the initial break point=48 km;and source duration time =10sec. The aftershocks were relocated using the three-dimensional velocity structures of P-and S-waves,and we found that the aftershocks concentrated on a plane with the area of 30×15㎞＾｛2｝dipping 22°toward the landside whose orientation agrees closely with that of the northwest-dipping nodal plane of the focal mechanism. The stress drop was estimated to be 89 bars. The aftershock area does not overlap with the asperity ruptured by the main shock,but outlined the asperity. These observations strongly suggest that the main shock itself is a very usual interplate event at depth of around 50 km in the Japan subduction zone. However,taking the tectonic circumstance into account,thisi is an outstanding event since the focal area was located between the five-years-lasting seismic quiescence area and the asperity ruptured by the 1973 Nemuro- Oki earthquake.

We performed gravity surveys and compiled pre-existing gravity data to obtain gravity anomalies of the Akan Caldera, eastern Hokkaido, and a new gravity map around the caldera was produced. In particular, we conducted gravity surveys on the frozen lake surface in winter to reveal a fine gravity structure. A new Bouguer anomaly map shows that the Akan Caldera is characterized by three closed mimima and takes its minimum value(～70mgal)in the central area of the caldera. It is consistent with our geologic studies suggesting that the Akan Caldera is multiple caldera. Characteristic low anomalies in the Akan Caldera gradually increase towards topographic rim outside and reach their maximum around the rims. Post-caldera volcanoes are distributed not along topographic caldera rim but along a periphery of the low-anomaly region, which is situated inside the topographic rim.

A lot of strong ground motion data from the 2003 Tokachi-oki earthquake (Mw8.3) were obtained by a dense strong motion network including K-NET, KiK-net and so on. We examine the long-period ground motions with a period of about 20 seconds based on these data. First, we estimate propagation directions and phase velocities of long-period Rayleigh waves at 64 seismic array sites in Hokkaido arranged from the dense network; the radius of the array is about 15 Km. The semblance analysis is applied to band-pass filtered, verticalcomponent waveforms at each array; the central frequencies are 0.03 and 0.05 Hz. The propagation directions of these long-period seismic waves basically radiate from the source region of the 2003 Tokachi-oki earthquake. However, the propagation directions seem to turn from the northwest to the west across the Hidaka Mountains, extending from the north to the south in the central Hokkaido. The phase velocities are 3 to 4 Km/sec; the phase velocity of the 0.03 Hz central frequency wave is slightly higher than that of the 0.05 Hz central frequency wave across one array. Next we examine ground velocity waveforms and spatial distribution of peak ground velocities (PGV) based on all data in the eastern Japan from the 2003 Tokachi-oki earthquake. Long-period waves with a period of about 20 seconds predominate on the ground velocity waveforms; at some sites shorter-period seismic waves are overlaid on the longperiod waves. The PGV distribution shows peculiar features; PGV in Hokkaido, the northern part of the epicenter, are larger than those in Tohoku, the south-western part of the epicenter, at the comparable distance. The PGV attenuation relation also shows peculiar features; the decay rate changes around a distance of about 200 Km and the large scatter with one order exists at a distance of about 300 Km. We interpret these features by comparing the observed waveforms with the synthetic ones calculated by using the CMT solution and 1-D velocity structure. The direct S-waves contribute PGV at distances less than about 200 Km, while the long-period surface waves (Rayleigh waves) contribute PGV at distances larger than about 200 Km. The source radiation pattern of long-period surface waves generates the large scatter of PGV at a distance of about 300 Km. This study concludes that the source radiation patterns of long-period S and surface waves mainly contribute the spatial distribution of PGV for large earthquakes such as the 2003 Tokachi-oki earthquake.

Geometry of the upper boundary of the subducting Pacific plate or slab was estimated in the Hokkaido region, Japan, using the ScSp phase: the converted phase to P wave at the boundary from the S wave reflected at the core-mantle boundary and propagating nearly vertically (i.e., ScS phase). Taking the advantage of a dense seismic network named “Hi-net” recently deployed across the Japanese islands, we applied several seismic array analyses to the recorded waveform data for a large nearby deep earthquake, in order to enhance weak ScSp signals in record. At first, we set up five blocks for the region in the subducting direction of the plate. After aligning the travel time of the ScS phase and stacking seismograms among stations in a same sub-block perpendicular to the plate subduction, we searched for the optimal plate model (i.e., two-dimensional geometry of the upper boundary of the plate) for each block. The model was parameterized with six depths, and the seismograms were stacked based on the travel time of ScSp as time lag at each sab-block, so that the optimal model would yield the maximum amplitude of ScSp after stacking. The searches were done, using ray tracings of the ScSp phase with a reference velocity model and the non-linear inversion scheme called Neighbourhood Algorithm. THe optimal model of each block was combined each other by cubic spline interpolation, in order to construct three-dimensional geometry of the upper boundary of the plate. We then performed the frequency-wavenumber (f-k）spectral analysis to refine the above result. Assuming each station as a reference point, we made each beam output with adjacent 7 stations as a function of wavenumber(kx,ky) and frequency. The peak of its power spectrum was considered as the ScSp signal, estimating the wavenumber vector, that is, the azimuth of arrival and slowness,so that we can estimate the position and depth of the corresponding S-to-P conversion. In the frequency range of 0.5 to 1.5 Hz, we could estimate the conversion points for 21 stations, and refined the geometry of the upper boundary already obtained by the above non-linear stacking approach. The final plate model was compared with the distribution of intraplate earthquakes in the Pacific plate. This comparison clearly reveals that the upper seismic zone merges with the lower from 150 to 200 Km in depth, by systematically deviating away from the upper boundary of the plate.

An exposed cross section of the tilted Tottabetsu plutonic complex allows direct observation of its 2-D cross-sectional shape and pretilting vertical variation in lithology and small-scale internal structures. Furthermore, the cross-sectional exposure allows directevaluation of pretilting vertical density variations in both the pluton and the country rocks, which serve as a strong constraint in gravity modeling that complements information on the‘missing’original horizontal dimension of this tilted pluton. The pluton is rectangular in cross-sectional shape, measuring ～10-Km thick by ～8-Km wide. It is stratified with the uppermost thin granitic unit (～1-Km thick) and the underlying thick gabbro-diorite units (～9-Km thick) that preserve a stratigraphic record of numerous botter replenishments in the form of alternation of originally horizontal mafic sheets and cumulate layers. The paleohorizontals (i.e. floor of the magma chamber at any one time) inferred from the sheets, pipes and cumulate layering and foliation dip steeply eastward, being roughly perpendicular to the original side walls and subparalell to the original roof of the pluton. Thus, in terms of the 2-D cross-sectional exposure, the pluton apparently had a vertically-elongated cylindrical form with vertical side walls and a flat roof before tilting. Both the pluton and the country rocks show systematic density increase with paleodepth, but density contrast of the pluton with the country rocks varies between each unit. The gabbro-diorite units show positive density contrasts (Δp=0-0.2g/㎝＾{3}) with the tonalitic and metasedimentary country rocks, while the granitic unit shows negative density contrasts with the country rocks (Δp=～0.1ｇ/㎝{3}). The new Bouguer anomaly map shows distinct correlations between the bedrock geology and local gravity field. As expected from the positive density contrasts of the gabbros and diorites relative to country rocks, a local gravity high is seen associated with the exposure of the gabbro-diorite units. At the same time, a weak gravity low is seen associated with the exposure of granite, concordant with the negative density contrasts of the granites relative to the country rocks. The gravity model shows that the subsurface contact between the pluton and the country rock gently dips westward, being roughly perpendicular to the steeply east-dipping paleohorizontals inferred from the surface geology. This strongly supports that the hidden pretilting side wall of the pluton was also nearly vertical. At first sight, the cylindrical shape with vertical side walls and a flat roof could be interpreted as being the result of emplacement at dilational sites along major faults and shear zones. However, such an emplacement mechanism invoking space creation by horizontal displacement of country rocks contradicts gradual vertical stacking of the originally horizontal mafic sheets that represent recurrent hotter injections. We suggest that space for the successive magma batches injected into the Tottabetsu magma chamber was created by vertical displacement of fault-bounded blocks of roof or floor country rocks with little horizontal displacement.

The great Sumatra-Andaman earthquake of December 26, 2004 is studied by using extensometer (strain-meter) as very-broard-band seismograph. We analyze the strain seismograms obtained by 100 m extensometers installed at Matsushiro Seismological Observatory. Japan Meteorological Agency. We compare the observed strain seismograms with thouse calculated by normal mode theory and obtain fairly good agreement between the observed and calculated seismograms for a five-sourcse model with a moment magnitude of 9.3 and a source duration of about 600 sec. We find a peculiar strain change right after the passage of Rayleigh wave in the NS component of the strain seismograms. This strain change cannot be explained by the summation of normal modes and double-couple type sources.

We point out that sometimes a STA/LTA trigger algorithm does not work appropriately for strong-motion observations. The STA/LTA trigger algorithm evaluates the ratio of short-time average(STA) to long-time average(LTA) and is adopted to recognize an earthquake for recording by an event recorder. We observed strong-motion records in the northern part of the Ishikari plain during the 2003 Tokachi-oki earthquake and its largest aftershock. Peak ground accelerations of the records are about 1 m/s^{2 } and peak ground velocities, about 0.3 m/s. However,two of the stations which adopted the STA/LTA trigger algorithm failed in recording during the main shock and the largest aftershock. We evaluate STA/LTA ratios by using the records obtained at other stations in order to investigate factors in not triggering these events. Weak long-period initial motion of P-wave, which was observed during the 2003 Tokachi-oki earthquake and its largest aftershock, make the STA increase slowly. Since the LTA also increases during the slow STA increase, the STA/LTA ratios do not increase enough to trigger an even if peak ground motion is large. We note that the STA/LTA trigger algorithm alone is not perfect for strong-motion observations,especially at an urban site.

We examine source characteristics of intraslab earthqakes based on the eleven source models estimated by the empirical Green’ｓfunction method. This method provides strong motion generation areas that reproduce near-source ground motions in a broadband period range(0.1～10sec). The storong motion generation areas roughly coincide with the asperity areas where stresses are largely released. Fault parameters of the asperity source model are outer fault parameters (entire source area and total seismic moment) and inner fault prameters (area of asperities,stress drop on each asperity and short-period level of the S-wave acceleration source spectrum). We find self-similar scaling of total seismic moment to the asperity areas. The number of asperities is one for Mw6 events,two to three for Mw7 events and five for Mw8 events. We also find self-similar scaling of total seismic moment to the short-period level of the S-wave acceleration source spectrum that is directly related to the asperity parameters (area of asperities and stress drop on each asperity ).Further we find that the aftershock area overestimates the entire source area.This means that estimation of the entire source area is somewhat difficult for intraslab earthquakes. Based on our results we propose a procedure for characterizing the source model for future intraslab earthquakes. Step 1:setting the total seismic moment (Mo), Step 2:setting the asperity areas (Sa) from the empirical relationship of Sa-Mo, Step 3: setting the short-period level (A) from the empirical relationship of A-Mo, Step 4: setting the number of asperities, Step 5: setting the stress drop (Δσ）on each asperity from the theoretical relation of A, Sa and Δσ, Step 6: setting the entire source area (S) from the theoretical relation of Mo, Sa, Δσ and S. The theoretical relations are derived from the multiple-asperity model.

The 1975 Hokkiado-Oki earthquake (Ms7.0) was characterized as a tsunami earthquake which had anomalous long source process time and spectrum of seismic waves. Predominant period of seismic waves were situated in several tens to several hundreds seconds. This abnormal long-period seismic wave generated very weak intensity distribution compared to the seismic moment (Mo) and epicenter distance. Aftershock distribution and focal mechanismw indicated this event occurrde on very shallow plate boundary close to the trench axis, in where seismicity has been very low. If next tsunami earthquake will occur near Hokkaido, JMA will be likely to underestimate the M, because JMA seismographs are able to record the waves only less than several tens of seconds. This may cause error in judgment for urgent tsunami warning because of smaller M, than real Mo. Therefore, a reliable quick Mo estimation method is proposed by using near-and far-field strain seismograms and far-field broadband seismograms especially for tsunami earthquake warning. In this method Mo is estimated from source time function τ which is detected from the duration time of near-field huge pulses or far-field S-wave pulses for strainmeter and of P-wave pulse for broadband seismographs. Strain seismographs near the fault are able to fully record whole period strain change without falling off and saturation because of its flat response from several Hz to DC component. Trial Mo estimation from the observed near-field pulse associated with the 2003 Tokachi-oki earthquake (Mw8.0) shows in good agreement with Mo calculated by teleseismic waveform inversion. This implies that realtime Mo estimation will be done from only watching the growth process of strain seismographs near seismic fault.

Nagoya City, one of the largest cities in Japan,is located on the Nobi basin where the future occurrence of Tokai and /or Tonankai earthquake disasters is feared. The prediction of strong ground motions in this region is a pressing concern in order to mitigate these disasters. In this study,I constructed a subsurface velocity structure model for an area of 60 Km×40 Km in the Nobi basin on the basis of geophysical exploration data. Basement structure of the Nobi basin has the following features. Large vertical offsets of 2.0 to 2.5 Km exist between the mountain area and the sediment site at the western boundary of the basin by the Yoro fault. The basement beneath the plain gradually becomes shallower toward the northeast direction,and the outcrop of the basement rock is observed at approximately 40 Km from the Yoro fault. I assume four sedimentary layers on basement. This model explained the observed PS-P times at most sites. Next I carried out 3D finite-difference waveform modeling for two events. The maximum amplitudes and waveforms of S-wave were reproduced well at most sites by the simulation. The generation and propagation of the Rayleigh wave were also reproduced by the simulation. These results indicate the validity of the proposed 3D velocity model.However,some differences were seen in the propagation character of the Rayleigh wave. These are discrepancies between observed and simulated group velocities of the Rayleigh wave. These differences suggest that proposed model have to be modified in detail.

The great Sumatra earchquake on December 26, 2004, which devastated the coastal areas of the Indian Ocean by the subsequent tsunami, generated tremendous numbers of surface-wave trains which traveled around the globe for many times. We analyse vertical-component seismograms of the event recorded by the F-net broadband seismic network in Japan, which has been deployed and maintained by NIED (National Research Institute for Earth Science and Disaster Prevention, Japan). By applying a band-pass filter between 3 and 5 mHz, we extract the long-period Airy phase of the fundamental-mode Rayleigh wave. The long period records display a magnificent series of multi-orbit Rayleigh wave trains that have circumnavigated the Earth for more than 6 times; i.e., we are able to observe clear signals up to R13-R14 phases. Although much later arrival phases than R14 become less visible mainly due to lateral heterogeneities in the Earth as well as dispersions of surface waves, we are still able to detect some smaal signals up to R18.

第四紀火山・大山（鳥取県）において，伏角計（dip circle）を用いた地磁気伏角の簡易測量を行なった。山体の帯磁による標高の増加に伴う伏角変化が確認できた他，山体中心部の溶岩円頂丘・弥山の麓で，伏角が減少することが分かった。過去に行なわれた全磁力測定結果（宮腰他，1985）では，この伏角が減少する付近で全磁力異常値の正負が反転しており，二つの成分の空間分布に対応が見られた。これらの地磁気異常は地下の構造（山体下に潜在する小規模カルデラ？）に起因する可能性がある。, Daisen volcano, a vast Quaternary composite volcano is situated in western Tottori prefecture, southwest Japan. I carried out measurements of the geomagnetic inclination on the volcano, to be more exact, Misen lava dome. It is suggested that a small-scale caldera submerges under the lava dome (Tsukui, 1984), and the total intensity data of the geomagnetic field also implied the existence (Miyakoshi et al., 1985). Simple dip circle was used for measurements of the inclination. It is very handy and light, and in need of no electric supply. I tried whether the resultant geomagnetic anomaly of the caldera can be found out with the equipment.