MR31C-0517
Difference in Gas Emission Pattern under Different Compression Rates
\hspace*{5mm}In order to explain precursory changes observed in radon, hydrogen and methane concentrations in groundwater or soil-gas immediately before earthquake occurrences, experimental works of gas emission prior to and associate with rock fracture events have been carried out since 1970' s. A pioneering work by Giardini et al. (1976) showed that H2, CH4, H2O, N2, CO, O2 and CO2 were emitted from granite and gneiss under uniaxial compression to the rock fracture. In this study, we particularly focused on methane, investigating the characteristic features of methane emission with varying the compression rate and the distribution of methane of emitted gas in the granite. \hspace*{5mm}In last AGU fall meeting, we reported methane emission behavior under two different compression rates. In case of loading rate=11.1 kN/m2sec along the long-axis, the concentration of methane in the vacuum chamber began to increase from the background level when the pressure exceeded roughly half of the failure pressure (28.39±2.25 MN/m2). Then, it was increasing in an accelerative manner toward the maximum at the failure point of the sample. In contrast, under the 2-fold compression rate (22.2kN/m2sec), the methane concentration increased very slowly or negligibly before the failure and suddenly increased at the time of the failure. From this result, relationship of loading rate and temporal variations of the methane concentration expects to indicate the relationship between the concentration of emitted methane and both state of the rock and crack generation. We carried out a continuous loading experiment for detailed discussion on gas emission before and after the rock fracture. Sample used in this study were cylindrical Inada granite (d=50 mm, l=100 mm) were placed in a vacuum chamber equipped with a uniaxial compression device and were compressed continuously with different loading rate. The chemical composition of emitted gases in the chamber during a course of uniaxial compression were analyzed by a quadrupole mass spectrometer every 7 seconds (Hal-201, HIDEN, UK). \hspace*{5mm}We reported the result of temporal variations of the methane concentration in wide compression rate.
MR31C-0518
Stress Transfer During Pressure Solution Compression of Neighboring Axi-Symmetric Asperities Pressed Against a Flat Semi-Infinite Solid
In a previous work, we developed a numerical model of compression by pressure solution (PS) of a single axi- symmetric asperity pressed against a flat semi-infinite solid. The dissolution rate along the contact was determined by (1) computing the normal stress distribution from the present shape of the asperity, and (2) solving the diffusion equation inside the fluid saturated solid-solid interface, including local dissolution source terms corresponding to the stress field previously determined. The change in shape of the asperity during an infinitesimal time interval can then be calculated and the entire procedure repeated as many times as desired. Our results showed that, as the contact flattens and grows during PS, the initial elastic deformation is partially relaxed and the stress transferred from the contact centre to the edge. Our current goal is to demonstrate that, among a population of asperities, stress can also be transferred from one contact to another and that the overall compaction rate can be significantly affected by this process. For this purpose we extended our previous numerical model to simulate PS of two spherical asperities simultaneously pressed against a flat semi-infinite solid. We considered two cases: a) identical asperities with different initial (unstressed) separations with respect to the flat solid, b) asperities with different radii of curvature and identical initial separation. In both cases, stress was transferred from the most strained asperity to the least, and the overall PS displacement rate was decreased. We conclude that elastic stress transfer should be taken into account in determining the overall compaction rate, particularly for granular aggregates that are characterized by very broad distributions of both asperity separation and radius of curvature.
MR31C-0519
Equivalent Elastic Models
We theoretically find the aspect ratios of the inclusions required to match the elastic-wave velocity of rock as predicted by the Differential Effective Medium theory (DEM) with those predicted by classical empirical rock physics equations of Wyllie et al. (1956) and Raymer et al. (1980). The aspect ratio range thus established is remarkably narrow (between 0.1 and 0.2) and stable in the porosity range between zero and 0.4. This conclusion is valid for pure mineralogies (calcite and quartz) as well as for mixed sand/shale mineralogies. It is not only applicable to the P-wave velocity but also matches the S-wave velocity as predicted by such empirical equations as of Castagna et al. (1993) and Krief et al. (1990). The goal of this exercise is to find the range of inputs into a fundamental micromechanical model (DEM) that is relevant to real rock data. Once the stability of such input is established, it can be perturbed to predict the elastic-wave velocity versus porosity and pore geometry beyond past experimental evidence which may be especially essential for carbonate rock where the shapes and structures of the pores may sufficiently vary in accordance with depositional and diagenetic history. For example, using aspect ratio 0.1 with DEM allows us to accurately match two carbonate laboratory datasets in a wide porosity range.
MR31C-0520
Laboratory measurements and modeling of seismic attenuation in saturated limestones
There is growing interest in using seismic attenuation for reservoir characterization. We measured limestone samples in the laboratory to analyze attenuation and elastic moduli as a function of frequency from 3~Hz to ultrasonic. We observe that there is a positive correlation between bulk modulus dispersion and permeability. We also measure different modes of attenuation and find that the attenuation in our fully-saturated samples is dominated by bulk compressibility losses rather than shear or viscous losses. We observe that attenuation significantly increases when brine replaces a light hydrocarbon. In addition, the shear modulus shows sensitivity to brine saturation. These results suggest that changes in seismic wave attenuation can be used as an attribute in surface seismic exploration. This might be of particular utility in the areas of time-lapse monitoring of reservoirs for either enhance production or CO2 sequestration. Finally, we use the Kramers-Kronig relations to model attenuation from the measured moduli as a function of frequency.
MR31C-0521
Fault strength drop due to phase transitions in the pore fluid
We conducted tri-axial friction experiments for quartz gouge at hydrothermal conditions to examine the effect of phase transition of water. We observed a decrease of ~ 0.05 in its coefficient of friction during the transition between liquid and vapor. This can be interpreted as a local pore pressure increase in the gouge-filled layer during the phase transition, which is caused by the surface tension of bubbles created between the two different phases. The local pore pressure increase deduces effective normal stress in the gauge layer, which makes a decrease of the frictional strength. The transient friction drop on a fault can play an important role in triggering an earthquake not only in hydrothermal areas but also in typical seismically active areas in the crust where water often contains CO2 of various concentrations because the CO2 density in the binary system of H2O-CO2 controls the pressure-temperature condition of liquid/vapor phase equilibrium.
MR31C-0522
Deformation Behavior of Rocks by Micro Focus X-Ray CT under Various Confining and Pore Pressures
To visualize in detail the manner of deformation in rocks under various confining and pore pressures, we used a micro focus X-ray CT system to obtain three dimensional image with high resolution of 5 micron. A new pressure vessel was developed to simultaneously supply both confining pressure and pore pressure to rock specimen. Berea sandstone and Noto diatomecious mudstone specimen of 10mm diameter and 20mm length were deformed hydrostatically. The diameters at every 15 degree angle were measured under the various conditions of confining and pore pressures. The average diameter decreased monotonically with increasing effective confining pressure, following the law of effective confining pressure. Sensitive orientation around the specimen diameter during pressurization was recognized, deformation anisotropy under pressurization then emerged and the fluctuation of diameter reached a maximum value of 8 percent in the case of Noto diatomecious mudstone. In this paper, we demonstrated deformation measurements by means of micro focus X-ray CT. This measurement method is applicable for deformation measurements of small and irregular rock specimen under various confining and pore pressures.
MR31C-0523
Measurement of contact angles of aqueous solutions on some rock forming minerals
Wetting properties of fluids on earth's materials are controlling fluid flows and dynamics of the geological systems. Although the wetting behavior of industrial materials have been widely examined often by contact angle measurements, contact angles of rock-forming materials have not been commonly measured. Therefore, we have been measuring contact angles of some representative rock-forming minerals. The surfaces of solid samples were polished successively by emery papers then by grinding powders (alumina: up to \sharp3000: grain size about 5 micrometers). Water droplet from a micro-syringe needle are placed on solid surfaces by moving up the sample stage. Images of water drops on the solid surfaces are captured from the horizontal direction with a CCD camera. Contact angles can be determined from the height and the length of the images by assuming them to be parts of circles. Over 60 measurements of contact angles of pure water on (101) and (011) faces plates cut from a natural quartz single crystal were repeated. The average contact angles of pure water on (101) and (011) faces of quartz were 48 ± 5 degrees and 52 ± 3 degrees, respectively. Contact angles of pure water on a natural calcite single crystal was also measured in the same way to be 37 ± 8 degrees. Contact angles of various aqueous solutions such as NaCl and NaHCO3 on these minerals will also be measured in order to evaluate wetting properties of natural rock-water systems.
MR31C-0524
Effects of solution composition on water permeability in a Fontainebleau sandstone
Permeability of aqueous fluids is an important transport property in geological systems. Extensive work has been carried out to investigate the controlling factor of permeability such as porosity and pore geometry. Although natural water has a wide variation in chemical composition, pure water has generally been used for the permeability measurement and the effect of solution composition remains to be understood. In this study, we performed the permeability measurement using aqueous solution with different composition (pure water, NaCl and NaHCO3) to evaluate the effect of solution composition on water permeability. A Fontainebleau sandstone, composed of ~100% quartz, was used as the sample. We will report the preliminary results of these experiments.
MR31C-0525
Acoustic Properties of Carbonate Rocks and Their Relation with Porosity and Mineral Composition
Carbonates are complex rocks characterized by a wide range of facies, texture, micro-structure, and rock fabrics. Understanding how this complexity affects the acoustic properties of carbonates is a key issue for interpreting and predicting changes in seismic images and acoustic log. Questions arise from the study of the porosity versus velocity relation for carbonate rocks which often consider the large scatter around the main velocity-porosity trend predominantly related to the porosity. We started a comprehensive laboratory study on carbonate rocks to understand how mineral composition, together with porosity, controls seismic wave propagation. The samples were collected capturing a wide range of porosities (from 1-52 percent) and different depositional environments in order to represent at best pore fabric and mineralogical heterogeneity in carbonates. Results of the hydraulic, transport, and acoustic properties of the collected samples were compared with those reported in the literature. The main results of this research show that a quite heterogeneous mineral composition of the samples (calcite, dolomite, and anhydrite), other than the pore type, controls the elastic behavior of carbonate rocks, and thus, the velocity-porosity trend. In particular, the samples showing the biggest departure from the general velocity-porosity trend show a non-negligible amount of anhydrite. Compared to calcite, anhydrite 1) causes rock softening and in turn, a decrease of P-wave velocity, because of the lower bulk modulus (k =56 GPa); 2) is characterized by a finer grain size (silt-size) which may create two elastic domains separated by a critical porosity approximately 30 percent.
MR31C-0526
Analysis of Rock Properties for fining upward sandstone and sandstone-carbonate mixed systems in the Resistivity - Velocity domain.
P-wave velocity and electrical resistivity are standard measurements generally acquired when a well is drilled, and far more recently, also when seismic and controlled source electromagnetic (CSEM) surveys are performed jointly. Examining how these two geophysical measurements relate to each other is, therefore, important for reservoir characterization. We review bounds and several effective medium models for P- and S-wave velocities, and resistivity, and compare them with well log data for water-saturated unconsolidated sandstones and a mixed sandstone- carbonate system. Using these data allows us to focus in studying only the effect of porosity and lithology in velocity and resistivity. We also plot these data, bounds and estimates in the resistivity-velocity domain, creating two pair of "crossbounds", following Wempe (2000). Crossbounds are the result of combining the bounds of P-wave velocity and the bounds of normalized resistivity. The velocity Hashin- Shtrikman (HS) bounds can be modified to include the critical porosity. For resistivity, there are equivalent HS bounds, plus a modified upper bound defined by Wempe (2000). The inner crossbounds include: a lower and upper crossbounds. The lower crossbound is the HS bound for velocity versus the lower HS bound for resistivity. The upper crossbound is the upper HS bound for velocity versus the upper empirical bound for resistivity. The outer crossbounds are obtained by crossplotting the upper velocity bound and the lower resistivity bound; and the lower velocity bound and the upper resistivity bound. The data for a fining-upward sandstone tend to plot within the inner crossbounds in the normalized resistivity - velocity domain, suggesting that both elastic and electrical properties are changing in a similar manner, mostly due to the change in porosity and sorting. In the case of the mixed sandstone-carbonate system, we found that carbonate-rich sediments can be identified in the resistivity-velocity domain, since they are characterized by higher velocity and resistivity values. Another observation was that the inner crossbounds for S- wave velocity and resistivity are broader, therefore analyzing and modeling the data in the normalized resistivity-S- wave velocity space using specific models, such as Archie and the "soft sand" model, may be more useful than doing it in the normalized resistivity - P-wave velocity domain.
MR31C-0527
Dispersion of Bulk Modulus of Partially-Saturated Porous Media
The elastic properties of partially-saturated porous media lie somewhere between a volume average of drained and undrained properties in the high frequency limit, and drained behavior in the low-frequency limit. The frequency dependence is related to the distribution of cluster sizes of the wetting fluid and the permeability. Johnson (2001) developed an approximation to Biot theory to predict the frequency dependence of the dynamic bulk modulus, which requires a solution to a Poisson equation of an auxiliary function over the saturated regions. To apply this theory, a coupled percolation-continuum evaporative drying model was used to produce simulated tomographic images of water distributions in a porous medium. The auxiliary function was then calculated using a finite difference method in order to calculate the full frequency spectrum of the bulk modulus of the patchy- saturated medium. The elastic moduli were then compared to those calculated from velocity data measured ultrasonically as a function of saturation for Massillon, Spirit River, and Schuler-Cotton Valley sandstone samples. Predictions from a heuristic model for determining elastic moduli, one that utilizes a Voigt volume average of the elastic properties along with the ad hoc rule that water-filled cells on the air-water interface behave in a drained manner, have already been shown to mimic the experimental data. The parameters found using the heuristic model were compared to those from the frequency-dependent bulk modulus calculations since the specific surface area arises naturally as a key geometrical parameter in both.
MR31C-0528
Development of Techniques for Subsidence Estimation Considering the Bulking Factor of Rock Mass
Since 1990s, the surface subsidence in an abandoned mine area has been dealt as a serious public problem in Korea. In order to minimize the damage from the subsidence, many researchers have continued their researches about the mechanism of subsidence, the restoration of subsidence area, the compensation for a public welfare, etc. The great part of researches on the subsidence has been performed by numerical study. In these cases, the physical properties of rock masses and the dimension of gangway were used for major input data. The volumetric expansion ratio of rock mass according to the collapse of roof in gangway has never been used for major input data in numerical studies. When the bulking factor of rock mass is overlooked in numerical study, the mined-out area will be considered as an empty space rather than a filled space which is piled up high with the broken rocks in the gangway. Therefore the new estimation techniques for subsidence considering the bulking factor of rock mass would be reapplied to the studied area to check whether the amount of subsidence has been overestimated or not. In this study, most popular types of rock mass in Korea have been chosen for obtaining their bulking factors. Gneiss, granite, shale and limestone have been crushed in laboratory. After the particle size distribution, each particles were piled up and the bulking factors were identified with every steps; naturally stacked step, well interlocked step, and particle squeezed step in which the rock mass behaviors like a new intact rock. With the bulking factor of rock mass, the subsidence area will be re-estimated and their results would be used for generation of geo-hazard map in Korea. Key Words: subsidence, bulking factor, geo-hazard map.
MR31C-0529
Seismic Response of Carbonate Cemented Sandstones
This study focuses on how carbonate cementation precipitated at the key sequence stratigraphic surfaces impact the seismic impedance. Our goals are two-fold: (1) to identify the sedimentological variations within carbonate- cemented sandstones and (2) to quantify their effects on P-impedance. To accomplish this goal, we identify the relationship between carbonate cementation and key stratigraphic surfaces, such as, the incision surfaces and the flooding surfaces. Next, we use effective medium models to quantify the impact of sediment parameters on P- impedance. We find that the carbonate cemented sandstones are extremely heterogeneous in nature, even within a depth interval of 60 meter in our study area offshore Equatorial Guinea, West Africa. Their grain-size, sorting, mineralogy, clay-content, amount of cement and degree of leaching vary considerably. We identify two distinct clusters of data in the P-impedance vs. porosity plane. The carbonate cemented sandstones from the base of incision are usually associated with lower shaliness, lower porosity and higher P-impedance. On the contrary, data from the top of flooding surfaces exhibit higher shaliness, higher porosity and lower P-impedance. The contact cement model fails to predict the trend shown by the later cluster of data. The predictions using the constant cement model with 1% constant carbonate cement, and the modified stiffsand model with 15% critical porosity agree reasonably well with the data. Furthermore, we find that the modified differential effective media model with 40% percolation porosity, and Berryman's self consistent model with 20% percolation porosity fit P- impedance vs. porosity trend of the carbonated cemented sandstones. In conclusion, the carbonate cements are different than the siliciclastic cements in terms of sedimentological parameters, and the commonly used rock physics model for quartz cemented sandstones are not always suitable to predict P-impedance vs. porosity trends for the carbonate cemented sandstones. We recommend testing the predictions of rock physics models against data, classified by key stratigraphic surfaces.
MR31C-0530
Elastic wave velocities and Poissonfs ratios of amphibolite up to 900 ° C at 1.0 GPa: Effect of dehydration melting on Poissonfs ratio of mid- to lower crustal rock
Vp and Vs measurements at high pressures and high temperatures (up to 900 ° C at 1.0 GPa) were carried out with piston-cylinder apparatus having a 34 mm inner diameter at high-pressure laboratory at Yokohama National University. The amphibolite was collected from Central Graben South, Mariana Trough by JAMSTEC cruise KR02-01. This rock is fine-grained homogeneous rock (grain size < 0.2 mm) without discernable oriented fabric. It consists mainly of hornblende (48.2 vol.%), plagioclase (44.4 vol.%), and substantial amount of magnetite (7.4 vol.%). Vp and Vs were measured using the pulse reflection method for a cylindrical rock sample having 5.7 mm diameter and ~5.0-5.5 mm length enclosed in a welded Pt capsule. Both Vp and Vs linearly decrease while Poisson's ratio increases from room temperature to ~500 ° C. The temperature derivative of Vp is -3.3 × 10-4km s-1 ° C -1and Vs is -2.6 × 10-4 km s-1 ° C -1 below 500 ° C. The temperature derivatives of Vp and Vs show a remarked change above ~500 ° C. The relatively lower Vp and Vs and higher Poisson's ratio at higher temperatures above 500 ° C are attributed to dehydration melting of amphibole-bearing assemblages. We identified substantial amounts of glass (4.1 wt.% at 700 ° C, 12.2 wt.% at 800 ° C and 20.3 wt.% at 900 ° C) in the quenched run products. The glass occurs as isolated pockets and/or pools among grain boundaries. Modal proportion of the glass linearly correlates with the increment of Poisson's ratio. The observed increments in Poisson's ratio with temperature are 0.02 at 700 ° C, 0.12 at 800 ° C, and 0.16 at 900 ° C, relative to the value at 500 ° C.