NG33B-0172
Downhole seismic monitoring with Virtual Sources
Huge quantities of remaining oil and gas reserves are located in very challenging geological environments covered by salt, basalt or other complex overburdens. Conventional surface seismology struggles to deliver images necessary to economically explore them. Even if those reserves are found by drilling successful production critically depends on our ability to ''see" in real time where fluids are drawn from and how pressure changes throughout the reservoirs. For relatively simple overburdens surface time-lapse (4D) seismic monitoring became industry choice for aerial reservoir surveillance. For complex overburdens, 4D seismic does not have enough resolution and repeatability to answer the questions of reservoir engineers. For instance, often reservoir changes are too small to be detected from surface or these changes occur in such pace that all wells will be placed before we can detect them which greatly reduces the economical impact. Two additional challenges are present in real life that further complicate active monitoring: first, near-surface condition do change between the surveys (water level movement, freezing/thawing, tide variations etc) and second, repeating exact same acquisition geometry at the surface is difficult in practice. Both of these things may lead to false 4D response unrelated to reservoir changes. Virtual Source method (VSM) has been recently proposed as a way to eliminate overburden distortions for imaging and monitoring. VSM acknowledges upfront that our data inversion techniques are unable to unravel the details of the complex overburdens to the extent necessary to remove the distortions caused by them. Therefore VSM advocates placing permanent downhole geophones below that most complex overburden while still exciting signals with a surface sources. For instance, first applications include drilling instrumented wells below complicated near-surface, basalt or salt layer. Of course, in an ideal world we would prefer to have both downhole sources and receivers (e.g. in-situ 4D seismic), but for now VSM may be the most economical alternative. By performing data-driven redatuming with measured Green's functions, these data can be recast into complete downhole dataset with buried Virtual Sources located at each downhole geophone. This step can be effectively thought of as a time reversal and it's remarkable feature is that velocity model between sources and receivers is not required to perform it. We will show various applications of the VSM method to several synthetic and real time-lapse datasets to illustrate the following advantages: 1) ability of VSM to eliminate overburden distortions without knowing velocity model between surface sources and downhole receivers, 2) greater quality of Virtual Sources in strongly scattering environment, 3) beneficial downward only radiation pattern on the Virtual Sources, 4) ability to correct non-repeatability caused by slight changes in acquisition geometry and temporal changes in the near surface, 5) ability to create P-wave Virtual Sources without shear radiation and S-sources without P-waves. Versatility of VSM to handle 1D, 2D and 3D situations and its ability to handle overburdens of any complexity makes it an indispensable tool for the active geophysical monitoring in a challenging geological environments. Although examples presented all come from an oilfield, it is straightforward to envision analogous applications in many other fields ranging from global geophysics to monitoring man-made structures.
NG33B-0173
Active seismic monitoring of changes of the reflection response of a crystalline shear zone due to fluid injection in the crust at the Continental Deep Drilling Site, Germany
In theory and in the laboratory variations of the hydraulic pressure can be detected with seismic methods: A lowering of the hydraulic pressure leads to the closure of micro-cracks within the rock (increase of the differential or effective pressure). Subsequently, the seismic velocities increase. An increase of the hydraulic pressure leads to reverse seismic effects. Consequently, seismic impedance contrasts and associated reflection amplitudes vary in the case of a propagating fluid pressure front in a rock matrix with inhomogeneous permeability - as is the case at shear zones. The largest amplitude changes can be expected with vertical ray inclination on the impedance contrast. Generally, the expected effects are small however (Kaselow, 2004). The practical utilization of active seismics for the detection of pressure changes at large scale in hard rock is currently being studied at the Continental Deep Drilling Site (KTB). The injection of water (200 l/min) in a depth of about 4000 m into the so-called SE2 shear zone in the KTB pilot hole was monitored with active seismics between May 2004 and April 2005. The core of the experiment layout is a fixed 5-arm geophone array consisting of 24 3-component geophones, buried at about 70 cm depth. The source signal is a vertical vibrator sweep of 30 s length with the spectrum 30-120 Hz. The signal is sent into the ground 32 times during each cycle, detected with the array and recorded separately for each geophone channel, without prior correlation with the source signal. This allows maximum post-processing with seismic processing and analysis tools and especially permits the use of array properties to increase the signal-to-Noise ratio. Critical parameters of the experiment are the repeatability of the source signal as well as the stability of the receiver properties. Another pivot is the hydraulic pressure and its distribution built up within the rock matrix. Estimations based on model calculations show that a change of seismic reflections can be detected above a well head pressure of about 15 MPa (Kaselow, 2004). Additionally, the fluid pressure at depth must be distributed within at least the first Fresnel-Zone of the seismic wavefront. After one year of injection only about 12 MPa were reached at the well head of the KTB pilot hole. One consequence is in particular that sensitive seismic signal processing needs to be applied. The standard deviation of the amplitude spectra of the raw data recorded with the array shows values around 36% if the strong direct waves are integrated in the analysis. The weak reflection signals from the target zone show values around 78%. The latter is due to the relatively low level of the wanted signals with respect to the ambient noise level. Band-pass filters and the application of the so-called diversity stack can reduce the errors. Another improvement can be achieved with selective, time-window based amplitude-dependent signal suppression before correlation (Polom, 1999) in an improved manner. References: KASELOW, A. 2004. The Stress Sensitivity Approach: Theory and Application. Dissertation, Freie Universität Berlin. POLOM, U. 1999. Elimination of Noise Caused by Spikes and Bursts in Vibroseis Data. Pure and Applied Geophysics, 156, 319-344.
NG33B-0174
Feasibility experiment for Active Monitoring of Inter-plate Coupling in Tokai region. ---A dense array measurement---
We carried out a long-distance seismic monitoring experiment using ACROSS (Accurately Controlled and Routinely Operated Signal System) for 10 months starting from the end of 2004 in Tokai region central Japan. In this experiment, we attempted to detect reflected phases from the top surface of the subducting Philippine Sea plate and to detect their temporal changes. In the Tokai region, a seismic survey was conducted in 2001. A strong reflected phase was detected and was interpreted as a reflection from the boundary between subducting and overriding plate [Iidaka, 2003]. Yoshida et al. [2004] analyzed the ACROSS signal received by a nationwide seismic network (Hi-Net) and identified several phases which may include the direct waves and reflected phases from the plate boundary. The aim of our research is to confirm the reflected phases and also to detect temporal changes in properties of them. The ACROSS source was continuously operated in Toki City, Gifu prefecture by Tono Geoscience Center. The frequency-modulated signal with frequency band from10 to 20 Hz was precisely repeated with an interval of 50 seconds. As the rotation direction of the source reverses once per hour, we can synthesize linear vibration in any direction. Seismometers were deployed on a survey line between 40 and 70 km distance to the southeast from the source. We also deployed an array consisted by 12 seismometers having 2km aperture at 55 km away from the source on the survey line. We acquired seismograms with the array and stacked to improve S/N ratio. In ACROSS data analysis, we can estimate the errors of the received signals quantitatively in the frequency domain. We stacked the received signals weighted by inverse of the estimated errors in order to reduce the effect of the incidental noises such as earthquakes. We converted the stacked spectral signals into a frequency response by dividing them by the source spectra. Through applying an inverse Fourier transform to the frequency response, we obtained the transfer function in the time domain, in which P, S and some later phases were identified. We applied an array analysis to these transfer functions to identify the later phases which are coherent among the array elements. We calculated the semblance coefficients for the phases arriving after the direct P waves. We searched the combination of the slowness and the incident angle that gives the highest semblance coefficient. As a preliminary result, though we found a large trade-off between the slowness and the incident angle, we also found some coherent phases arriving to the array after direct P phase. Comparing the travel-time with synthetic waveform, we conclude that these phases may include reflected waves from the top of subducting Philippine Sea plate, Moho discontinuity or some other layer boundaries. The data analysis is underway and will be presented in detail in the conference.
NG33B-0175
Preliminary study for active monitoring of the plate boundary using ACROSS: Synthetic and observed seismic records
ACROSS (Accurately-Controlled Routinely-Operated Signal System) has been developed for active monitoring of a dynamic state in the Earth's structure (Kumazawa et al., 2000). Since November 2004, we have conducted an array observation of ACROSS signals in Tokai area, central Japan, to identify any seismic reflection (and hopefully its temporal change) from the lower crust and/or subducting Philippine Sea plate (Kasahara et al., 2004). In this report, we show the recent results and discuss the relevance of several arrivals of wave groups to underground structures using the theoretical travel times and synthetic waveforms. The frequency-modulated ACROSS signals (10-20 Hz) have been continuously transmitted from the sources located in Toki city, central Japan (Kunitomo et al., 2005) and received at 22 temporal seismic stations at the offset distance of 40-75 km from the source. We define the transfer function between a source and a receiver as a nine-element second-order tensor, Hjk, where j and k denote directional components of the observed displacement and the excitation force, and r, t and v represent the radial, transverse and vertical components, respectively. We recognized the significant wave groups within the travel time ranges of 10-18 and of 15-23 seconds at 54-74 km offset distance through stacking the data for about 60 days. Such wave groups also appear on the records of a Hi-Net station at 57.4km by stacking for 30 days (Yoshida et al., 2004). A 2-D velocity structure model was made for our observation area using seismic exploration records across the central Japan (Iidaka et al., 2003). We calculated both travel times by ray tracing method (Fujie et al., 2000; Kubota et al., 2005), and synthetic seismograms by FDM simulation (Larsen and Schultz, 1995). Comparing the observed time series of Hrr and Hzr to the theoretical travel times and synthetic seismograms, we noticed that the wave groups observed at 61-73 km are well corresponding to the theoretical travel times of reflected waves from the bottom of lower crust Moho (PmP) and from the upper boundary of the Philippine Sea plate (PxP). We believe that these results suggest the potentiality for active monitoring of the subtle changes of geophysical properties in the earth's structure using ACROSS signals in future. Acknowledgements: We appreciate the permission to use the seismic waveform data provided by the Research Group for Seismic Expedition in Central Japan.
NG33B-0176
Monitoring Experiment in Tokai, Japan : ACROSS source designed for a phased array
The Tokai district in Japan is the place where a large interplate earthquake is regarded as being impending. Japanese research group for active monitoring are planning to monitor the propagation property of seismic wave in this area using active seismic sources. As a part of the project we are deploying ACROSS vibrators as well as a seismic array in this area. The vibrator we use is originally deployed in another place as a set of four rotational sub-vibrators, which is designed to generate linear acceleration in any direction. As a switching operation method are developed to generate linear acceleration in any direction with a linear combination of the signals of two opposite rotation direction, we use four sub-vibrators as a phased array in the Tokai monitoring experiment. The area we are monitoring has many important aspects for a test site of monitoring experiment. The area is the edge of the source region of the interplate earthquake, so called the Tokai earthquake, where changes are most likely to occur in the preparation stage of earthquakes. A strong reflector at the interface between subducting and overriding plate is detected with explosion experiments. The vibrators and array are located to receive the reflection with good efficiency. Our estimation shows that one-week stacking is enough to detect one percent change in amplitude. A long-term slow slip and sporadic slow slips are observed on the plate interface. The long-term slip has been continuing since 2000, which is revealed by GPS network (GEONET). Sporadic slow slips with durations of several days are detected by tilt and strain meters that are deployed in this area. Low frequency earthquakes or tremors are often observed near the plate interface, whose source is regarded as a movement of fluid. These phenomena can cause the variation of seismic propagation property, especially reflection at the plate interface. In the source site, change in the elastic or inelastic property of the ground around the source is monitored. The experiment mede in Awaji test site shows the importance of monitoring of the source movement. The actual movement of the source can be altered by environmental condition such as temperature, rainfalls or groundwater level. The site we are deploying the source is in the observatory belonging to Nagoya University. It has a horizontal vault of more than 100 m deep in hard rock, where seismometers and strainmeters are already deployed. Groundwater levels are also monitored in deep boreholes. Using these data together with additional seismometers environmental monitoring around the source will be made to calibrate the radiation wave. We start the monitoring experiment in 2006 together with other ACROSS sources being operated by JNC (Japan Nuclear Cycle development Institute) and MRI (Meteorological Research Institute). The nation-wide seismic network (Hi-Net) is also used for monitoring. * ACROSS stands for 'Accurately Controlled Routine Operated Signal System'. An ACROSS vibrator radiates continuous elastic wave of sinusoidal signal for monitoring the earth's crust.
NG33B-0177
Detection of Reflected Waves from Plate Boundary Using ACROSS Source and Seismic Array
ACROSS (Accurately Controlled and Routinely Operated Signal System) is effective in monitoring temporary changes of Earth's interior. A long-term operation experiment near Nojima fault [Ikuta et al.,2004] detected small temporary changes of travel time of P and S waves at tele-seismic events. Toward Tokai monitoring plan to detect the reflected phases from the top of Philippine Sea Plate and monitor its temporal changes, a mid-term continuous experiment was conducted using ACROSS source and a seismic array. The experiment was operated for the period from Dec. 2004 to Sep.2005 in the Tokai area, Pacific side of the central part of Japan. In this region, the expected Tokai earthquake is a serious concern. In addition, slow slip events and low-frequency tremors are observed in this area. A strong reflected phase from the plate boundary was found by the seismic observation using artificial sources [Iidaka et al.,2003]. The purpose of the experiment is to establish a method to detect and monitor the reflection from the plate boundary using ACROSS. The ACROSS source is located in Toki city and operated by Tono Geoscience Center. The ACROSS source continuously transmits precisely-controlled frequency-modulated signals whose frequency band ranges from 10 to 20 Hz with an interval of 50 seconds. We deployed a short-span seismic array at the distance of 55 km from the ACROSS source. The cross-shaped seismic array spanning 2 km consists of 12 seismometers equipped with an offline data logger, amplifier and solarpanel. We stacked the received signal for a month with an interval of 200 seconds in order to improve signal noise ratio. We extracted a series of line spectrum of ACROSS signal. Transfer function can be obtained by dividing spectrum by the source. Applying inverse Fourier transform, we can obtain the transfer function in time-domain. We identified direct P and S phases by comparing with the standard travel time table by JMA. We also found some coherent later phases. By comparing the travel-time of these phases with the travel-time calculated by ray-tracing with a given structure model, these phases may include the reflection waves from the top of Philippine Sea plate, the Mohorovicic discontinuity and some other physical discontinuities. We calculated the semblance of the later phases to estimate the slowness and the incident angle. A later phase with a low slowness can be found after the direct P phase. We also try to evaluate the temporal change of travel time and waveform. The detail results will be presented in the conference.
NG33B-0178
Continuous Observation of Seismic Wave Velocity and Apparent Velocity Using a Precise Seismic Array and the ACROSS Sources
We conducted a continuous monitoring of travel time using a seismometer array and an artificial seismic source, which is known as ACROSS source. The seismometer array was deployed in a vault for crustal deformation observation in the central part of Japan. The ACROSS source, which is located at 2.4 km from the receivers, was operated continuously for one month. The refracted P and S waves, which traveled a velocity boundary between a sediment layer and a basement, and some reflected waves were observed at the array in addition to the direct P and S waves. We investigate a temporal variation of differential travel time and apparent velocity for these phases. We detected a significant temporal variation in the differential travel time. Most of the variation can be correlated to weather variations. We detected a variation that fluctuated in a period of one day with the maximum amplitudes of 0.5 ms for the refracted P and S waves. The variation is well correlated with the change in the atmospheric temperature. Part of this fluctuation can be attributed to the variation of the basement to which the vibration source is attached. The variation can be attributed to and increase of decrease of stiffness of the surrounding ground due to increase or a decrease of a density for dry cracks. We also detected variations with amplitude of about 1.0 ms corresponding to rainfalls for the direct S wave. As little variation is observed in direct P wave, it can be attributed to the change in the density of water-saturated cracks. Direct S wave that traveled in the shallow part of the ground is strongly affected. We have detected no meaningful result for the reflected waves since we cannot gain enough signal-to-Noise ratios for them.
NG33B-0179
Observational Strategy of ACROSS towards the Time-evolving Natures in the Lithosphere
ACROSS (Accurately Controlled, Routinely Operated Signal System) is aiming at the detection of very small changes in physical states in the lithosphere, particularly for the focal region of the anticipated huge earthquakes as demanded socially. Our technical challenge is to device an ideal methodology to enable us to acquire the ideal observation data towards the real understanding of the EarthOs interiors even under the inherent noise and physical limitations. We need light to illuminate the dark EarthOs interiors, eyes to observe them and a brain to interpret the result: The light should be designed well to be really coherent, the eyes with high fidelity should be accurately synchronized to the light transmission and the brain should be smart enough to evolve by itself. In order for the whole system to be robust against noise, we have to devise all that can be done. In addition, a significant demand is imposed onto us; non-destructiveness against our environment. The recent progress of technology makes it possible what was impossible several years, so that we try to find out the ideal way to go. We have spent about 10 years for developmental works, which started a moment before the disastrous Kobe earthquake of 1995. Now we believe that the background theory has been known in addition to some of the basic technology elements, whereas the user-friendly hardware and other auxiliary tools including practical theory and software have not been acquired yet. The examples of the field observation have started to accumulate for demonstration as reported by companion papers. The data acquired by ACROSS in seismology is not seismogram but tensor transfer function (Green function) in frequency domain. The data carry substantially new information with high quality and rigorous estimate of reliability. The availability of ACROSS would change the strategy for underground study in the coming years. We would like to call for your attention and discussion to the next way to go: (1) Combined use of both elastic and electromagnetic waves: (2) Necessity of super-array or array of dense arrays of sensors, the network of ACROSS transmitters, mobile transmitters both on land and on the ocean surface: (3) Protocol of transmitting signals and data format: (4) Software for efficient data processing: (5) New targets of inversion and its methodology to utilize the new type of transfer function data to full extent: (6) Monitoring strategy for prediction of highly non-linear events.
NG33B-0180
Discrimination of the Active Scattering Interfaces
Temporal variations of the scattering sources can be discriminated by the changes in reflectivity and efficiency of energy distributions of transformations of vibration modes (e.g. P-S and/or S-P) at the interfaces among heterogeneous media. The active scattering interface is caused by such temporal variations of material properties of the scattering sources due to various processes associated with crustal deformations and chemical reactions due to the tectonic activities. Occurrence of earthquakes and volcanic eruptions are in some way related to the migrations of super-critical fluid and re-distribution of localized stresses acting on the scattering interfaces. Recently discovered slow slip events and deep non-volcanic tremors are apparently related to the movement of supercritical water supplied by the dehydration of the subducted material, although the mechanisms to generate these phenomena remain uncertain. Intermittent occurrence of these events could suggest that the stress conditions in the crust and upper mantle could be a critical state and triggered by the movements of water vapor. The heterogeneity in the lithosphere can be investigated by monitoring such active scattering interfaces that could be caused by temporal changes of both stress state and heterogeneous distribution of interstitially migrating fluid through surrounding rocks. In order to demonstrate effects of changes in the material properties and/or physical conditions of an embedded low velocity layer, synthetic wave forms are estimated by calculating impulse responses of horizontally layered media with various conditions of scattering interfaces, such as thicknesses, material properties, stress distributions, and so on. Comparisons of changes in waveforms of reflected, transformed, and transmitted elastic waves are made as a function of incident angles, thickness of embedded low velocity layer, and amount of discontinuities of material properties.
NG33B-0181
Monitoring Contamination of the subsurface with Quasi-Static Deformation
Data from a field experiment suggest that we can detect the infiltration of contaminated water (contaminated with 150 ppm of a biosurfactant, Rhamnolipid) into the vadose zone with tilt meters. Three sets of instruments were installed in the vicinity of a 50m x 50m field, which was instrumented and could be irrigated in a controlled manner. Each set consisted of one tiltmeter at the bottom of a 10m cased borehole and one seismometer buried to a depth of about 1m. The instruments were installed in late summer and early fall of 2002. The drift in tilt caused by their installation decayed to near background levels in about one year. The site was irrigated with plain water at 50,000 l/day for 40 days during the fall of 2003. The irrigation was repeated in the fall of 2004 for 50 days, again with 50,000 l/day. This time the irrigation water contained 150ppm of a biosurfactant. This surfactant was chosen to simulate a contaminant through its effect on the modification of surface tension and because it is environmentally benign. It was developed for bioremediation. We expected subtle changes in the relative responses of the instruments as the elastic properties of the vadose zone were altered by the contaminant. This expectation is based on a long series of laboratory measurements, e.g. W. Brunner and H.A. Spetzler 2002. We used natural sources for the excitation of the instruments, solid Earth tides for the tiltmeters and microseisms, i.e. ocean generated noise at about 6s periods, for the seismometers. In the case of the tilt meters we used theoretical site specific tilt and compared that with our measured tilt. We found no significant change in the correlation coefficient between theoretical and measured tilt for the water-only irrigation. The correlation coefficient was generally above 0.9. The correlation coefficient dropped precipitously about ten days after we had started irrigating with the surfactant. It recovered again about ten days after irrigation was stopped. A preliminary analysis of the seismic data shows similar trends, but with much more noise. Using a Born approximation we calculate that a slight perturbation of the complex moduli near the surface does indeed yield changes in tidally-induced tilt near the observed values. Brunner, W. M. and H. A. Spetzler, Contaminant-Induced Mechanical Damping in Partially Saturated Berea Sandstone, Geophys. Res. Lett., Vol.29, No.16, 10.1029/2002GL015455 ,2002