Tada-nori Goto

In conventional inversion schemes of magnetotelluric data by the linearized least-square method with a smooth constraint, a smoothing parameter, α contributes to all of the electrical resistivity blocks in the smooth constraint term evenly. Therefore, it is difficult to reconstruct a resistivity structure with sharp structural boundaries using the conventional inversion, although subsurface formations including both smooth and sharp structural changes are often targets of magnetotelluric surveys. In this paper, we propose a new two-dimensional inversion algorithm for magnetotelluric data to reconstruct the resistivity structure with both smooth resistivity variations and a sharp boundary. Our inversion scheme needs the location of sharp boundary as a priori information from the other geophysical surveys, such as seismic reflection method. A new hyper-parameter “β” between 0 and 1 is defined for expression of sharpness of the structural boundary. As β becomes 0 to 1, the resistivity variation across the assumed boundary becomes sharp to smooth. Two hyper-parameters are determined using the ABIC-minimizing scheme with a new simple way for searching optimized hyper-parameters efficiently. We applied both conventional and new sharp-boundary inversion to synthetic magnetotelluric data from two models. One model includes high and low resistivity anomalies, and another model consists of layers increasing their resistivity gradually. As a result of the synthetic tests, our sharp-boundary inversion reconstructed the resistivity structure with sharp boundary, if the location of sharp boundary is assigned at the proper position. Even if the sharp boundary is assigned within the smooth structure area wrongly, smooth model is obtained properly. After synthetic tests, we apply this sharp boundary inversion to field data obtained around the seafloor of the Nankai Trough using high-frequency ocean bottom electromagnetometer. The top of Philippine Sea plate, confirmed by a seismic reflection survey, is adopted as a location of sharp boundary in our inversion. As the result of our sharp-boundary inversion, the top of Philippine Sea plate is reconstructed as a sharp structural boundary. We conclude that our sharp-boundary inversion provides more realistic resistivity model compared with the smooth model by the conventional inversion.<br>

The estimation of crustal structure is a key to reveal the tectonic history of the Japanese island arc and to study the future tectonic activity, such as earthquakes, volcanic eruptions, or associated tsunamis. Refraction or reflection wave exploration is one of the methods for deep oceanic crust investigation, but in these methods mainly used is only compressional wave or the vertical component of seismograms. If we use shear wave or horizontal component of seismograms, more precise investigation of crustal structure would become possible. As the first step of such investigations, we focused on the receiver function analysis that estimates the location of subsurface seismic-velocity discontinuities using shear waves from natural earthquake. This study tries to apply receiver function analysis to image subsurface structure using refraction survey data not only from earthquake but from artificial sources on the surface. In the processing of refraction survey data, time migration is first necessary to move each converted point to the true location where refracted wave is generated from the head wave. This is because P-S conversion point, i.e., imaging target, of refracted head wave is different from reflection point of reflection survey data. The f-k migration is then applied to time migrated receiver function for depth imaging. Finally, migrated sections for different source locations or for earthquakes are stacked to produce a single section. Our results demonstrate that the receiver function analysis is applicable to refraction survey data for imaging and results from various seismic methods can be combined after receiver functions are produced and migrated.<br>

Coda-Q reflects subsurface heterogeneity. In the past, from the variation of Q value against time or space, stochastic information on surface medium that seismic waves travel through has been obtained. For example, it was said that Q value shows abnormal value around volcanoes or before earthquakes. However Q value does not indicate any deterministic information on location of scatterers, source, boundary of layers and so on. Coda-Q is influenced by various parameters in the underground because the coda-wave is formed by the summation of scattered waves, so it is difficult to determine factors that influence coda-Q. In this study, we attempt to obtain non-stochastic parameters of the subsurface medium from stochastic parameter Q. We employ a 2-D finite difference method and simulate seismic scattering of a 2D model to see if we can obtain the information on the stress loaded to the model (i.e. the magnitude and the direction of the stress). Our analysis of the coda-wave reveals the proportional relation between the loaded stress and the coda-Q. In addition, the variation of the coda-Q is dependent on the direction of the stress. As a conclusion, we suggest the possibility that we can get the order and the direction of the loading stress if we obtain the time-variant change of coda-Q. It means that we could obtain non-stochastic parameters from the coda-wave that is used as a stochastic measure of the medium.<br>

Natural magnetic fields are attenuated by electrically conductive water. For that reason, marine magnetotelluric surveys have collected data at long periods (1000-100 000 s). The mantle structure has been the main target of seafloor magnetotelluric measurements. To ascertain crustal structure, however, electromagnetic data at shorter periods are important, e. g. in investigations of megathrust earthquake zones, or in natural resource surveys. To investigate of the former, for example, electromagnetic data for periods of less than 1000 s are necessary. Because no suitable ocean bottom electromagnetometer (OBEM) has been available, we have developed a small OBEM and ocean bottom electrometer (OBE) system with a high sample rate, which has an arm-folding mechanism to facilitate assembly and recovering operations. For magnetic observation, we used a fluxgate sensor. Field observations were undertaken to evaluate the field performance of our instruments. All instruments were recovered and their electromagnetic data were obtained. Results of the first experiment show that our system functioned well throughout operations and observations. Results of other field experiments off Tottori support the claim that the electromagnetic data obtained using the new OBEM and OBE system are of sufficient quality for the survey target. These results suggest that this device removes all instrumental obstacles to measurement of electromagnetic fields on the seafloor.

The magnetotelluric response in the transverse magnetic (TM) mode has information of two diagonal components of the resistivity tensor with the cross-strike(ρ_yy) and vertical directions(ρ_zz). Although these two values are regarded as the same single value in the ordinary isotropic inversion, forward MT responses and their sensitivity matrix calculated by anisotropic modeling suggest that TM-mode response is affected by ρ_zz. In this study, a new anisotropic inversion technique that takes into different resistivity values in two directions was proposed. The trade-off between the isotropy and anisotropy is determined objectively by the statistical criterion called ABIC. This method can be applied to both isotropic and anisotropic structures. When applied to the anisotropic structures, we could reconstruct the identical ρ_yy and ρ_zz suggesting anisotropy, while applied to the isotropic structures, we could reconstruct the isotropic structures similar to the ones acquired by ordinary isotropic inversion. The calculations with the synthetic data have showed the effectiveness of the proposed method.<br>

Yatsushiro Sea is one of the high tidal fluctuation areas in Japan, and the water's edge retreats about 200 meters at the low tide. Also there are some submarine fresh groundwater discharge points on the tidal flat of Yatsushiro Sea. VLF-MT survey was carried out to detect some freshwater seepage point on the shore side and the tidal flat of Yatsushiro Sea. In the tidal flat, a lot of observed apparent resistivity and phase values are very low except in the western side. This feature suggests that the deeper part has higher resistivity. The low phase may be caused by fresh water zone. In the western area, observed data show high resistivity, and high phase of more than 35 degrees. Because the basement (tuff breccia) crops out, this characteristic may be caused by this basement. The observed time changes are smaller than detected spatial variation. In addition, two anomalous areas showing high resistivity and low phase are detected in the tidal flat. An anomalous area near the shore side exists on the east side of a well with visible freshwater discharge. Around another 100 m offshore anomalous area, a fresh water discharge was observed by the seepage meter. Therefore, these anomalies are considered to relate to the local fresh water discharges.<br>

Natural magnetic field is attenuated by electrically conductive water. Therefore, the marine magnetotelluric survey has been used at long period range(1000 to 100000 sec), and investigated deep mantle structure. To obtain the crustal structure, however, electromagnetic data at the periods between 1 and 1000s is very important, for example the investigation for mega-thrust earthquake zone. Because there is no suitable ocean bottom electro-magnetometer (OBEM), we developed small sized OBEM and OBE system. On the other hand, controlled-source survey techniques are needed to obtain a shallower detail structure than conventional MT method. The DC resistivity survey is powerful tool for a shallower survey. Our marine DC resistivity sounding system has a 150m long electrode cable, and this cable attached the Deep-tow system. In this article, we describe the mechanism of developed OBE(M) and DC resistivity survey system, and show some results of field tests<br>

The Niigata-Kobe strained belt pointed out by Sagiya et al.(2000), where crustal deformation is concentrated, becomes one of important target areas in “the 2nd new Program of Research and Observation for Earthquake Prediction” in Japan (e.g. Hirata, 2004). A multidisciplinary research for five years on the strained belt using dense GPS and seismological observations and investigation of crustal resistivity structure was started from 2004. In this paper, we will report outline of the MT measurements made in 2004 along a profile across the Atotsugawa fault, which is situated in the Niigata-Kobe strained belt, and also describe preliminary results on the data analysis. Obtained preliminary inversion result shows inhomogeneity in the crustal resistivity structure at depths from 7km to 18km around the Atotsugawa fault.

Electrical resistivities of two mudstone specimens were measured under high confining pressure until about 30MPa and under constant pore pressure at 0.6MPa using a new resistivity measurement system. It is consisting of a pressure vessel of a triaxial compression test apparatus and a resistivity measurement device used for the conventional measurements under atmospheric pressure. The used rock block sample was taken from a mud volcano situated above Kumano Basin of where the depth is about 2, 000m from sea surface. The resistivities of the rock determined by this system have a relatively high reproducibility. In addition, this measurement method employing a rubber jacket and loading some confining pressure can accurately and easily keep the water content of specimen at a constant during the measurement. Thus, the method can more properly prevent the superfluous water existing between the surface of specimen and the jacket than the conventional method by which the specimen was set in a box or laboratory in unjacketed state. As a consequence, it results in that the resistivity may be determined more accurately. The measurement results of the tow rock specimens saturated by sea water used in this study showed that the resistivity increased as the effective confining pressure was elevated. If the influences of pore pressure to the conductive property of clay minerals can be neglected, the fact that there was not a distinct change of gradient of resistivity-pressure curves on loading process suggests that the maximum historical stress of the rock may be larger than 30 MPa, which corresponds a depth of about 2km beneath the seafloor. By comparison between intact and cracked specimens, it was clear that the effect of confining pressure on the resistivity for the cracked specimen was more significant than that of intact specimen due to the closuring of the cracks under pressure loading.

The forearc crust along the Nankai Trough is characterized as an accretionary prism. The unconsolidated sediments on the subducting Philippine Sea plate are underplated to the southwest Japan arc, so that large amount of fluid is supplied to the accretionary prism. Such fluid could be related to various geological phenomena in the accretionary prism such as the accumulation of methane hydrate, expected as new energy resource. However, fluid distribution below the seafloor, in the forearc crust, had not been understood clearly. Marine magnetotelluric soundings around the Nankai Trough and the Kumano Basin were carried out in 2002-2003 to elucidate fluid distribution in the forearc crust. Both time series data of horizontal electric and magnetic fields were obtained at seven sites, and only electric field data were obtained at two sites. Apparent resistivity and impedance phase at each site were calculated from these observed data with removal of spike noises, and a resistivity model below the seafloor was estimated. The model resembles to a seismic reflection section. By interpreting the estimated model, fluid distribution in the accretionary prism and the upper oceanic crust is discussed: high fluid content within the accretionary prism is interpreted, especially along the splay fault. This result implies a pass way of fluid along the splay fault.