NS32A-01
Interpretation of the seismic response of a fracture as a function of the scale.
The seismic response of a fracture is affected by intrinsic length scales associated with the fracture as well as lengths scales associated with the seismic measurement. Fracture geometry for instance, has many length scales (e.g., radii of areas of contact, apertures, correlation lengths). All these specific length scales can be compared to the wavelength of the seismic probe, as well as the length scale over which the seismic probe samples. In this study, we focus on the effect of the scale of observation on monitoring alteration of a fracture caused by reactive flow. The investigation was performed on fractured carbonate rock (150 mm in diameter and 76 mm in height) and on a standard sample made of acrylic with approximately the same dimensions. To control the scale of observation, an acoustic lens system was used to produce pseudo-collimated acoustic beams with diameters of 15 mm, 30 mm and 60 mm from water-coupled spherically-focused 1 MHz piezoelectric transducers. Seismic measurements (as a function of beam diameter) and volumetric flow rates were made on the fractured sample prior to and after reactive flow with an HCl solution. In each case, the different beam diameters were used to probe the same region of the fracture. Thus, the number of signals obtained at the 15mm, 30 mm and 60 mm scales were 16, 4 and 1, respectively. Statistical analysis was used to compare measurements made at the 60 mm scale with averaged values from the 15 mm and 30 mm scale, and to compare measurements made before and after reactive flow. In addition, high resolution two-dimensional maps over an 80 mm x 80 mm region of the fracture were obtained through diffraction-limited acoustic mapping without the use of the lenses. The statistical analysis showed that prior to reactive flow multiple measurements on the local scale (15 mm and 30 mm) can be averaged to obtain the same seismic response of a fracture as on the global scale (60 mm). However, after reactive flow, data obtained with the 15 mm probe was the only scale that exhibited a statistically significant change in the mean and cross-correlation coefficients, i.e., averaging the local scale measurements did not result in the same interpretation of fracture properties as those based on the 30 mm and 60 mm. This corresponds to the alteration of the spatial correlation length of the fracture geometry caused by the reactive flow. A two-point correlation analysis of the high resolution seismic maps showed that the spatial correlation length of the fracture was roughly isotropic prior to reactive flow but became anisotropic after reactive flow. In one direction, the correlation length increased from 11 mm to 25 mm. In the measurements prior to reactive flow, the correlation length is smaller than the 15 mm and for that reason the data are very similar on all scales. But after reactive flow, the correlation length increased to 25 mm, which is greater than smallest probe. For this fracture, changes in the flow rates from chemically etching the fracture can only be detected/interpreted from multiple local scale measurements, (measurements made below the scale of the spatial correlations within the fracture). Acknowledgments: LJPN wishes to acknowledge the Geosciences Research Program, Office of Basic Energy Sciences US Department of Energy and the University Faculty Scholar program at Purdue University
NS32A-02
Influence of Fluids on Compressional to Shear Wave Conversion in Reservoir Rocks
Converted shear waves (P- to S-) are clearly seen at large offsets in conventional seismic reflection profiling using 3-component seismometers. However, at normal or near normal incidence it is unusual. One such observation where the horizontal component signal was more pronounced than the corresponding vertical component prompted this inquiry. Possible explanations are scattering and anisotropy. A different possibility considered here is the conversion due to lateral motion of fluids in the reservoirs. This lateral motion of fluids induces shearing forces on the matrix due to a combination of viscous drag and pressure differences. Compressional waves were induced in a slab of aluminum bonded to a Berea sandstone block and reflected waves, both compressional and shear, were recorded for varying offsets. In the first run the rock was dry. In a repeat run the rock was wetted with water. In one of these experiments significant differences between the observed shear wave amplitude and that predicted by theory were observed. Another experiment investigated the influence of fluids on the transmitted wave. In this experiment compressional and shear motion in two orthogonal directions were induced on one face of the core sample and for each set-up all three motions were recorded on the opposite face. The data suggests that the converted shear is enhanced by the presence of fluids, in this instance water. These results are preliminary and further work is in the planning stages.
<a href='http://puma.kvcc.edu/kbalachandran'>http://puma.kvcc.edu/kbalachandran</a>
NS32A-03
Geologic Observations of Asymmetric Rock Damage Across Large Strike-Slip Faults and Ruptures along Bimaterial Interfaces
Analytical and numerical results indicate that mode II ruptures along a bimaterial interface tend to propagate for ranges of conditions in the direction of motion of the slower velocity medium. In such cases, the tensional quadrant of the radiated seismic field is persistently on the medium with faster velocity. This is expected to lead over geological time to asymmetric pattern of rock damage across the interface. The theoretical results indicate further that significant generation of damage is limited to the upper 3 or so km of the crust. To test these predictions, we mapped systemically the symmetry properties of rock damage across the principal slip surface (PSS) at sites along the San Jacinto fault (SJF) and the San Andreas Fault (SAF) in southern California. The mapping is done at scales of cm to meters (fault-core), meters to 10s of meters (fault-zone) and 10s to 100s of meters (damage-zone). In the SJF, cm-scale fracture density within the fault core is systematically higher on the NE side of the principal slip zone at 3 exposures of the fault near Anza. Inversion of fault zone trapped waves south of Anza shows that a 100 m wide low velocity zone is shifted to the NE side of the fault. Seismic imaging of the velocity structure in the Anza area shows that the NE more damaged side of the fault has faster seismic velocities at seismogenic depth. On the SAF in two sites near Littlerock, the fault core scale damage is concentrated on the NE side of the PSS. In Palmdale, a 60 m wide bedrock fault- zone shows considerably more SAF-related damage NE of the PSS. Mapping of pulverized basement rocks along 140 km stretch of the SAF in the Mojave shows that they occupy about a 100 m wide tabular zone parallel to the fault, and that 70 percent of those rocks were found on the NE side of the fault. Several lines of evidence suggest that the mapped rocks were pulverized at shallow depth. Seismic imaging in this area indicates that the NE side has higher seismic velocity. The results are thus compatible with the theoretical expectations for rupture along a bimaterial interface. We note that the theoretical properties of such ruptures can be relevant for various other important issues, including the "heat flow paradox"½, short rise time of earthquake slip, effective constitutive laws, and expected ground motion.
NS32A-04
Measured total permittivity of a silt loam soil with low surface electrostatic charge.
Dielectric dispersion and signal attenuation strongly affects the performance of ground--penetrating radar. Both phenomena are determined by soil total permittivity, which, for a given water content, below 100 MHz is thought to be controlled by the quantity of bound water associated with the clay--size mineral fraction. Total permittivity of an unfractionated Fairbanks silt loam soil and selected particle--size fractions were determined from 10 MHz to 6 GHz by time--domain spectroscopy. The reported mineralogy of the soil is largely quartz and albite, with smaller amounts of montmorillonite, kaolinite, and hematite. The reported relative proportions of the soil particle--size fractions were sand, 7 %; coarse silt, 49 %; fine silt, 35 %; and clay, 9 %. The soil had a reported cation exchange capacity of 16 cmol (+)$\bullet$kg${}^{-1}$ and a measured specific surface of 54.7 m${}^{2}$$\bullet$g${}^{-1}$. The volumetric water contents of the samples were between 0.25 and 0.36. The particle--size fractions studied were $>$ 20 $\mu$m, 10--20 $\mu$m, 2--10 $\mu$m, and $<$ 2 $\mu$m. The measured dielectric dispersion at 10 MHz for these fractions were 22.2, 18.5, 52.3, and 59.3, respectively. The measured dielectric dispersion at 10 MHz for the whole soil was 28.6, which could be explained solely as the contributions from free water and bound water associated with the coarse and fine silt--size fractions, without invoking the contribution of bound water associated with the clay--size fraction. This work was supported by the US Army AT24 project entitled, "Effect of clay particle size on the dielectric permittivity of soils" and Strategic Environmental Research and Development Program, Military Munitions Project UX 1440.
NS32A-05
Multi Asperity Micromechanical Model of Rock Joints With Application to Wave Propagation
The authors have developed a kinematically driven micromechanical methodology for rock joints that utilizes, (1) a directional distribution function of asperity contact orientations as an additional measure of surface roughness, and (2) an iterative procedure to obtain the asperity contact forces at each load increment recognizing that the asperity contact force distribution is not always known a priori [1]. The derived model is utilized here to understand the effect of surface roughness and asperity shear resistance on the initial normal and shear stiffness behavior of joints. We find that the initial normal stiffness is nonlinear, and is related to the normal stress and the initial joint closure. Interestingly, the closure behavior varies from an exponential to a power law behavior depending upon the initial closure. Power laws, hyperbolic and exponential functions with asymptotic maximum closure have been utilized in the past to model closure behavior. However, the functional form of the closure behavior critically depends upon the initial closure which may be a result of the joint mismatch. We also find that the initial shear stiffness critically depends upon the asperity contact orientation and the asperity friction. Also, the joint initial shear stiffness varies nonlinearly with roughness. We also observe that for very smooth joints, the ratio of initial shear and normal stiffnesses becomes proportional to the asperity contact stiffness and independent of asperity friction. The calculated initial normal and shear stiffness are used to investigate plane wave propagation behavior through joints utilizing the well established imperfectly bonded interface model. We find that the amplitudes of the reflected and transmitted waves as well as the group time delay of the wave-packets are significantly influenced by the joint normal stress and initial closure conditions. Since the derived model explicitly incorporates the asperity heights and radii of curvature, scale-dependent overall joint stiffnesses are obtained. The model may thus be used to elucidate the frequency dependency of wave transmission in rock joints. References: 1. Misra, A. Mechanistic model for contact between rough surfaces, J. Eng. Mech. 1997, 123(5):475-484.
NS32A-06
Quarter Wave Resonances Between Targets and Stratigraphic Interfaces Within GPR Profiles of Shallow Frozen Lakes
Resonances commonly appear within GPR reflection profiles. They can be caused by the target itself or by reverberation between a target and a stratigraphic interface that defines a high contrast in permittivity. The most common interface is the ground surface. I discuss resonances that occur at depth beneath shallow, frozen lakes located on an artillery and bombing range in interior Alaska. They originate near ice bottom and freeze/thaw interface horizons and are similar to those caused by shallow utilities. The transmitted pulse spectra were centered near 300 MHz. I interpret the resonances to have originated with reflections from metal targets because automatic gain control and low-pass filtering appear to differentiate them from natural clutter. The time delays between their onset and that of the overlying interface reflections correspond with spectral peaks well below the transmitted pulse bandwidth and predicted by in situ quarter-wave resonance theory. Consequently, they are not resonances of the targets themselves. Although the time delays cannot resolve both layer thickness and permittivity, reasonable values of permittivity give layer overburdens above the targets on the order of 10 cm, which is consistent with the persistence of the resonances. A simple, one-dimensional model shows that strong dielectric contrasts across the overlying interfaces are needed to make these frequencies significant. The appearance of other spectral peaks may be natural target resonances, so that the two phenomena may be complementary in the presence of strongly reflecting interfaces. The presence of shoreline craters, the unlikelihood that these lakes have ever been inspected, and the infrequent occurrence of the resonances amongst natural clutter lead to the conclusion that the targets may be unexploded ordnance (UXO). Late winter is an ideal time to survey for such objects in these lakes because the ice affords exact and repeatable positioning for the antennas and because thermal interfaces between strong permittivity contrasts are then well developed. The pulse bandwidth appears to be ideally suited for detection and discrimination of these resonances by filtering.