H11B-0477
A theoretical model for the wetting of a rough surface
Many applications would benefit from an understanding of the physical mechanism behind fluid movement on rough surfaces, including the movement of water or contaminants within an unsaturated rock fracture. Presented is a theoretical investigation of the effect of surface roughness on fluid spreading. It is known that surface roughness enhances the effects of hydrophobic or hydrophilic behavior, as well as allowing for faster spreading of a hydrophilic fluid. A model is presented based on the classification of the regimes of spreading that occur when fluid encounters a rough surface: microscopic precursor film, mesoscopic invasion of roughness and macroscopic reaction to external forces. A theoretical relationship is developed for the physical mechanisms that drive mesoscopic invasion, which is used to guide a discussion of the implications of the theory on spreading conditions. Development of the analytical equation is based on a balance between capillary forces and frictional resistive forces. Chemical heterogeneity is ignored. The effect of various methods for estimating viscous dissipation is compared to available fluid rise on roughness experiments. Methods that account more accurately for roughness shape better explain the data as they account for more surface friction; the best fit was found with a hydraulic diameter approximation. The theory implies the existence of a critical contact angle that is a function of roughness height to width ratio, below which fluid will spread and above which fluid will recede. The resulting equation predicts the invasion front progressing with a square root of time dependance, consistent with a diffusive process.
H11B-0478
Clast Breakdown on Desert Surfaces by Differential Insolation of Cracks
In the Southwest US, cracks in surface clasts have a preferred orientation independent of rock fabric, rock shape and local conditions. Differential insolation of incipient cracks of random orientations provides an explanation of this preferred orientation through removal of moisture held in the crack. A study of differential insolation of cracks of different orientations at 35 north latitude was undertaken using a numerical radiative transfer code and idealized crack geometry. The amount of energy reaching the bottom of each crack was calculated at five minute intervals over the day for several days of the year to determine hourly, daily, seasonal and annual deposition of energy depending only on crack orientation and depth. By adding to this a formulation in which only crack orientations which effectively shield their interiors and minimize their water loss are able to grow, the observed pattern of cracks is reproduced, including both expressed modes and their deviations from pure North-South and East-West behavior. Given this formulation, the important timescale for water retention is the annual average insolation which is associated with both modes of the aligned cracks while the daily recharge by summer monsoon rains is responsible for the observed deviations of these modes. Thus, both the annual average insolation and the daily pattern of rainfall are recorded in the cracking patterns of surface rocks in the Southwest.
H11B-0479
Linking fault permeability, fluid flow, and earthquake triggering in a hydrothermally active tectonic setting: Numerical Simulations of the hydrodynamics in the Tjörnes Fracture Zone, Iceland.
A good insight into the transient fluid flow evolution within a hydrothermal system is of primary importance for the understanding of several geologic processes, for example the hydrodynamic triggering of earthquakes or the formation of mineral deposits. The strong permeability contrast between different crustal layers as well as the high geothermal gradient of these areas are elements that strongly affect the flow behaviour. In addition, the sudden and transient occurrence of joints, faults and magmatic intrusions are likely to change the hydrothermal flow paths in very short time. The Tjörnes Fracture Zone (TFZ) north of Iceland, is such a hydrothermal area where a high geothermal gradient, magmatic bodies, faults, and the strong contrast between sediments and fractured lava layers govern the large-scale fluid flow. The TFZ offsets the Kolbeinsey Ridge and the Northern Rift Zone. It is characterized by km-scale faults that link sub-seafloor sediments and lava layers with deeper crystalline rocks. These structures focus fluid flow and allow for the mixing between cold seawater and deep hydrothermal fluids. A strong seismic activity is present in the TFZ: earthquakes up to magnitude 7 have been recorded over the past years. Hydrogeochemical changes before, during and after a magnitude 5.8 earthquake suggest that the evolving stress state before the earthquake leads to (remote) permeability variations, which alter the fluid flow paths. This is in agreement with recent numerical fluid flow simulations which demonstrate that fluid flow in magmatic- hydrothermal systems is often convective and very sensitive to small variations in permeability. In order to understand the transient fluid flow behaviour in this complex geological environment, we have conducted numerical simulations of heat and mass transport in two geologically realistic cross-sectional models of the TFZ. The geologic models are discretised using finite element and finite volume methods. They hence have a very high resolution that allows us to investigate the influence of high-aspect ratio structures like faults and thin sedimentary layers on the large-scale fluid flow behaviour in detail. Our numerical simulations account for transient variations of permeability due to mineral precipitation and the abrupt breaking of the rocks caused by hydrofracturing. We also consider different and discontinuous amounts of magmatic fluid production from crystallising intrusions. This allows us to understand how large-scale fluid flow responds to transient permeability variations and how those can trigger earthquakes in magmatic-hydrothermal systems like the Tjörnes Fracture Zone.
H11B-0480
Long-term Alteration of Fractured Glass Blocks in Seawater
Fractured archaeological glass blocks, altered in seawater over 1800 years, are studied because of their morphological analogy with the industrial nuclear glass blocks. They represent a striking example of altered fractured glass on a representative timescale in the prospect of a deep geological repository. Thus, they may contribute to the validation of models devoted to assess the long-term behaviour of high-level waste glass. Here, we present results concerning the alteration mechanisms within the cracks and the influence of seawater transport parameters on these mechanisms. A detailed study of the morphology, mineralogy and chemistry of these blocks allows to see that the alteration thicknesses of internal cracks are thinner than the alteration at the periphery of blocks (1 or 2 orders of magnitude). It suggests that the contribution of internal cracks to global alteration is quite low and that alteration processes depend on the renewal rate of the solution within cracks, so on the aperture and the accessibility to water. We also show that two mechanisms are involved in the glass alteration: i) interdiffusion between glass alkalis and protons of the solution leading to an amorphous gel layer, and ii) glass network dissolution followed by precipitation of less soluble elements of the glass (Si, Al) and some elements of the seawater (Mg) resulting in the formation of a secondary phases layer (smectites) in the middle of the crack. It is of note that the influence of transport parameters causes differences in the contribution of both mechanisms and in their respective kinetics. Indeed, at the periphery of blocks, alteration layers are thick and composed mainly of secondary products, because dissolution is maximal when solution is highly renewed. On the contrary, in cracks, where transport is slow, local equilibrium of saturation (Si) seems to be established quickly, creating favourable conditions to a drop in the dissolution rate. It causes a relative increase of the contribution of interdiffusion. Besides, the sealing due to the precipitation of secondary phases accentuates this influence of transport in cracks. To understand the coupling between chemistry and transport better, mainly the influence of parameters, such as aperture or S/V ratio, on the alteration kinetics, experiments have been carried out. Results show that thinner the fracture, higher pH, which reduces the diffusion coefficient of alkalis, and faster the chemical saturation is reached. These experimental data aim at parametering a model of a fractured glass block alteration.
H11B-0481
Mechanisms of Permeability Enhancement by Seismic Waves at the Pinon Flat Observatory
Seismic waves increase permeability. This was demonstrated by the change in tidal response of the water levels of several wells monitored for more than 20 years at the Piñon Flat Observatory in Southern California. But how is the permeability affected by seismic waves? Do the shear waves permanently mismatch the fractures extending below the observatory? The linear relationship between the amplitude of the shaking and the change in permeability favors this explanation. However, the major fractures of the hydraulic system are horizontal. As they are also only 100m deep, the sigmaxz stress is negligible. Moreover, a fracture mismatch does not explain the recovery observed within the 6 months following the disturbing earthquake. Does the fluid flow induced by the earthquake explain the change in permeability ? In poroelastic media, the seismic waves can induce pressure changes up to 104Pa. With the presence of large heterogeneity, like an open well, large fluid flow may circulate within the fractures. This may induce phenomena like fracture unclogging that temporarily changes the apparent permeability of the fracture network. We model the radial flow within the fracture medium by a finite difference code. If the flow exceeds a threshold value qth, the local permeability is increased by a constant k. We model the progressive unclogging of the medium using the seimic data recorded on the site. The final unclogging front then extends to several meters, which is enough to significantly alter the tidal response of the well. With a threshold of 10-9 m/s and a local permeability enhancement by a factor of 5, we model the changes in permeability observed in Piñon Flat Observatory. The response of the fractured system is thus partially controlled by the properties of the well. This has implication the production enhancement by seismic waves attempted by the oil industry. The control by large heterogeneity may also be an efficient process for inducing fluid flow along active and heteregeneous faults.
H11B-0482
The Effect of Freezing on the Dynamics of Dike Propagation
When magma-filled cracks propagate close to the Earth's surface, host rock temperature is well below the magma solidus. Solidification and substantial increase in magma viscosity can occur, are most pronounced near the propagating tip and can slow or arrest the progress of the dike. Quantitative analysis is required to predict whether a given dike will reach the surface to erupt and the duration of the precursor sequence. This challenging physical problem mixes elasticity, fracture mechanics, heat transfer and fluid flow with strong rheologic gradients due to cooling. We describe the propagation behaviour of such a hydraulic fracture using a laboratory experimental system of a crack fed by a constant flux of paraffin wax from a source reservoir propagating through gelatin below the solidus of the wax. The most novel behaviour is an intermittent regime in which cracks periodically stop advancing due to solidification, then swell at constant length while enhancing the elastic deformation in the surrounding solid before propagation resumes. We present a physical model of this system, based on different balances between driving and resistive forces: the former are elastic stress and liquid buoyancy, the latter are fracture resistance at the tip and viscous resistance. The fracture is represented as a head, behind the propagating tip, connected to the source by a narrow tail. Freezing of liquid close to the tip is assumed to enhance fracture resistance according to a cooling law, and propagation is assumed to occur only when the stress exerted by the liquid is enough to overcome fracture resistance. Our theoretical model reproduces intermittent propagation with precise behaviour depending on the controlling stress balances, and provides a tool to analyse natural systems.
H11B-0483
Numerical Modeling of Multiphase Fluid Flow in Ore-Forming Hydrothermal Systems
Two coexisting fluid phases - a variably saline liquid and a vapor phase - are ubiquitous in ore-forming and other hydrothermal systems. Understanding the dynamics of phase separation and the distinct physical and chemical evolution of the two fluids probably plays a key role in generating different ore deposit types, e.g. porphyry type, high and low sulfidation Cu-Mo-Au deposits. To this end, processes within hydrothermal systems have been studied with a refined numerical model describing fluid flow in transient porous media (CSP~5.0). The model is formulated on a mass, energy and momentum conserving finite-element-finite-volume (FEFV) scheme and is capable of simulating multiphase flow of NaCl-H20 fluids. Fluid properties are computed from an improved equation of state (SOWAT~2.0). It covers conditions with temperatures of up to 1000 degrees~C, pressures of up to 500 MPa, and fluid salinities of 0~to 100%~NaCl. In particular, the new set-up allows for a more accurate description of fluid phase separation during boiling of hydrothermal fluids into a vapor and a brine phase. The geometric flexibility of the FEFV-meshes allows for investigations of a large variety of geological settings, ranging from ore-forming processes in magmatic hydrothermal system to the dynamics of black smokers at mid-ocean ridges. Simulations demonstrated that hydrothermal convection patterns above cooling plutons are primarily controlled by the system-scale permeability structure. In porphyry systems, high fluid pressures develop in a stock rising from the magma chamber which can lead to rock failure and, eventually, an increase in permeability due to hydrofracturing. Comparisons of the thermal evolution as inferred from modeling studies with data from fluid inclusion studies of the Pb-Zn deposits of Madan, Bulgaria are in a strikingly good agreement. This indicates that cross-comparisons of field observations, analytical data and numerical simulations will become a powerful tool towards a more thorough understanding of hydrothermal fluid processes. One such attempt will incorporate geometric data of veins in the Bingham porphyry Cu-Mo-Au deposit into our numerical model. The presentation will introduce the numerical model and show examples and first results of the aforementioned applications.
H11B-0484
The Complex Effects of Simplifying Network Structure: A Modeling Evaluation of Chemical Migration in Fracture-Controlled Aquifers
It is widely recognized that highly permeable features dominate chemical transport in many fracture-controlled aquifers. Recent investigations also suggest that pervasive fracturing and the primary rock porosity are equally important in characterizing the spatial extent and time constants of chemical migration. Relatively little is known about the significance in which features of various attributes influence chemical transport through fracture networks. To address this issue, a suite of discrete fracture network models is constructed with geological and hydrological conditioning from observations at multiple field sites. The resulting models are used to quantify the changes to transport behavior caused by a systematic removal of aquifer structure. The results indicate that the influence of structural features on transport behavior is often significant but highly variable. The correlation of this behavior to the eliminated features is relatively low, suggesting that transport is jointly influenced by the eliminated feature characteristics and their role within the fracture network in altering hydraulic pathways. Moreover, the effects of eliminating features may be ‘enhanced' or ‘restricted', meaning that the removal of aquifer structure may either slow or advance mass migration, mean travel, dispersion, and tailing behavior. This suggests that correcting for the eliminated features does not invariably follow the classic advection-dispersion equation, where for instance dispersion approximates the effect of the eliminated structure as an invariably additive process to the mean flow. Given that the influence of eliminating structure is not easily approximated, the most useful procedure is to employ ensemble averaging from several structurally simplified models that capture the salient aspects of the fracture network connectivity, which significantly reduce but do not fully eliminate the effects of structural simplification.
H11B-0485
Detecting the Presence of a Sub-Porosity in an Open Fluid-Filled Fractures
Fractures and joints in the field often contain debris within the void spaces. These debris originate from many different mechanisms: organic and/or inorganic chemical reactions/mineralization, sediment transport, formation of the fracture, mechanical weathering or combinations of these processes. In many cases, the presence of debris forms a "sub-porosity" within the fracture void space. The "sub-porosity" may partially fill voids thereby reducing the local porosity to lengths scales on the order of sub-microns to tens of microns. It is quite clear that a sub-porosity affects fracture porosity, permeability and storativity. In this study, we investigate how the existence of a sub-porosity affects seismic wave propagation and consequently our ability to probe changes in the subsurface caused by the formation or alteration of a sub-porosity. Laboratory experiments were performed to examine acoustic wave scattering from packings of acrylic beads used to create a sub-porosity within a synthetic fracture. The experimental setup consisted of a synthetic fracture created by the separation of two Lucite cylinders. The aperture of the fracture was controlled using a computer- controlled linear actuator to increment the aperture from contact to 20 mm in increments of 50 microns. The sub- porosity in the fracture was created by using acrylic beads with diameters of 3.0 mm, 5.4 mm, and 9.4 mm. Compressional waves were propagated across the fracture using piezoelectric transducer to send and receive the signal. The ratio of the seismic wavelength to bead diameter ranged from approximately 9 to 0.1 for a broadband frequency range of 0.1 MHz to 1.0 MHz. The seismic response of the fracture was first measured without a sub-porosity for three different conditions: dry, partially saturated and fully saturated. To study the effects of the sub-porosity on wave propagation across a fracture, a single layer of beads was added to the fracture and measured for the same three conditions described above. A time-frequency analysis of the received signals was performed to examine the velocity dispersion of the received signals. The transmission data were compared for a fixed aperture of 10 mm. For a 10 mm aperture filled with water, the velocity dispersion was 550 m/s/MHz because of resonant scattering within the water-layer between the two fracture surfaces. The addition of a single layer of beads reduced the thickness of the water-layer as the bead size increased from 3.0 mm to 5.4 mm to 9.4mm. A low-frequency low-amplitude signal arrives first in samples with a sub-porosity because the low frequency components of the signal scatter the least. From a time-frequency analysis, as the bead size increased the velocity dispersion of the signal decreased from 227 m/s/MHz to 62 m/s/MHz. Two-scattering modes compete when a sub-porosity partially fills an open water-filled fracture: resonance scattering within water-layer and scattering from the sub-porosity (beads). The ability to distinguish open fluid-filled fractures from those partially filled by a sub-porosity depends on the aperture of the fracture, the size of the grains composing the sub-porosity, the relative thickness of the sub-porosity layer to the thickness of the water-layer and the wavelength of the seismic signal. All of these length scales affect both the hydraulic and seismic response of a fracture. Acknowledgments: The authors wish to acknowledge support of this work by the Geosciences Research Program, Office of Basic Energy Sciences US Department of Energy (DEFG02-97ER14785 08).
H11B-0486
Characterization of gas From Seismogenic Depths of the San Andreas Fault at SAFOD
We present data on the molecular composition of drill-mud gas from the lower sedimentary section (1800 - 3987 m) of the SAFOD (San Andreas Fault Observatory at Depth) Main Hole (MH), two sidetracks and from drill core samples thereof. Selected gas samples from drill mud and drill core have been analysed off-line for carbon and hydrogen isotopes; drill mud gas samples also for noble gases. Hydrocarbons, hydrogen, and CO2 are the most abundant non-atmospheric gases at seismogenic depths. The SAF (San Andreas Fault), encountered between ~3100 m and 3450 m borehole depth, is generally low in gas, but is encompassed by two gas-rich zones (2700 - 2900 m and below 3550 m) at the fault margins with enhanced 222Rn activities and distinct gas compositions. Helium isotope composition and the molecular composition of formation gas are distinct on both sides of the fault, implying that the SAF is a barrier for horizontal gas and fluid migration. Furthermore, the overall low contribution of mantle helium (0.4 Ra on the Pacific Plate and 0.9 Ra on the North American Plate) is minimum within the SAF, suggesting that the fault core does not act as a conduit for overpressured fluids from greater depth. Gas influx into the well at depth is related to the lithology and permeability of the drilled strata in general: larger formation gas influx was detected when drilling through organic-rich shale and permeable sandstone. The isotopic composition indicates an organic source of hydrocarbons and CO2 in the entire sedimentary section of the well. Hydrocarbons in sedimentary strata are partly of microbial origin down to ~2500 m borehole depth. The contribution of thermogenic gas increases between ~2500 m and 3100 m. Below 3100 m, carbon isotopes suggest that hydrocarbons almost exclusively derive from thermal degradation of organic matter. Although low in gas, two sections in the fault (3150 - 3200 m and 3310 - 3340 m depth of the MH) are enriched in hydrocarbons with ratios of CH4/(C2H6+C3H8) lower than in the surrounding strata.
H11B-0487
Detectability and significance of the 12h barometric tide in the radon-222 signal, dripwater flow rate, carbon dioxide concentration and air temperature of an underground laboratory.
Radon concentration has been measured since 1995 in the Roselend dead-end tunnel, in the French Alps, together with other geophysical and geochemical parameters. Bursts of radon concentration, reaching 65,600 Bq m-3 and up to several weeks duration, are observed over a background level of ca. 800 Bq m-3. These bursts appear to be related to the bedrock deformation or to the hydrogeological processes associated with the yearly cycle of water level in the nearby artificial Roselend Lake. In order to work out a generation mechanism, for these bursts, we developed tools to characterize the transport properties in the host rocks. Here, we concentrate on the 12h (S2) barometric tide. We first show, using real radon time series integrating synthetic signals, that a modified spectrogram method is more efficient than simple FFT to evidence weak periodic signals in such a context. Then, we apply this method to the radon concentration in the tunnel atmosphere measured by two different sensors: the AlphaGUARDTM sensor based on volumetric detection in an ionizing chamber, and the BarasolTM sensor, based on surface detection. Using the time series recorded by the AlphaGUARDTM, the S2 line, difficult to see with a simple FFT method, emerges clearly with our spectrogram method. This S2 line is not seen using the BarasolTM time series, illustrating the superior sensitivity of the AlphaGUARDTM for this particular purpose. Using a regular spectrogram analysis, we further show that the amplitude of the S2 line in the AlphaGUARDTM data depends on time, and appears particularly strong during the radon bursts. The presence of the S2 line reveals a high sensitivity of radon exhalation flux from the tunnel wall to changes of atmospheric pressure, and thus supports the advective transport mechanism for the radon bursts. A small but clear S2 component is also evidenced using our spectrogram method in a dripwater flow rate time series, representing a flow averaged over a 6 m2 area of the tunnel ceiling, while it is not observed in the flow rate of a more localized dripping. This suggests that some water drippings can also be affected by atmospheric pressure variations. The temporal structures of the S2 component in the flow rate and in the radon concentration, however, are not similar, indicating that water dripping from the ceiling cannot be the dominant source mechanism for the radon bursts. No S2 component is observed in the time series of carbon dioxide, but an interesting pattern is revealed by the S2 component of a temperature profile in the atmosphere. This study illustrates how a refined analysis to extract the S2 component in various geophysical time series can provide interesting clues on the complex processes affecting transport of fluids in unsaturated fractured media under multiple influences.
H11B-0488
Stochastic Analysis of Precipitation/Dissolution and Aperture Alteration in Variable Aperture Fractures Under Gradient-Reaction Conditions
Precipitation and Dissolution reactions within fractures alter fracture apertures, which in turn affects their flow and transport properties. Different types of aperture alteration patterns occur in different flow and reaction regimes. One class of regimes encountered in geological systems is the "gradient reaction" regime, where fluids are essentially in chemical equilibrium with a mineral everywhere, but precipitation-dissolution reactions are driven by solubility gradients that result from variations in temperature or salinity. In many such cases, the solubility gradient is invariant over very long periods of time, and largely unaffected by medium alteration. For instance in a sparsely fractured rock mass, heat transfer is largely conduction-dominated, due to the large heat capacity of the rock, and not significantly modified by fluid flow or the feedback between aperture alteration and fluid flow. Similar behavior has also been postulated during the emplacement of ocean-bed methane hydrates. We present a stochastic analysis to develop equations for the evolution of the mean aperture, aperture variance, spectrum/covariance and effective transmissivity under gradient-reaction conditions. The stochastic analysis consistently predicts (i) a runaway growth of transmissivity in the case of dissolution and (ii) a much slower rate of transmissivity decrease in a variable-aperture fracture than in a parallel-plate fracture. In the case of dissolution, an increase in initial aperture variance leads to a faster rate of transmissivity growth, while in the case of precipitation it leads to a slower rate of transmissivity reduction. Dissolution leads to an enhancement of anisotropy in the aperture correlation structure, with more persistent correlation in the direction of flow. The behavior is opposite in the case of precipitation. The predictions of the stochastic analysis are verified based on high-resolution Monte-Carlo simulations in computer-generated random initial aperture fields. We discuss potential applications of our results to natural and engineered geological processes incuding hypogene karstification, methane hydrates and geothermal systems. We also present preliminary results from ongoing work evaluating the role of convective heat transfer and hydromechanical coupling on aperture alteration in a gradient reaction regime
H11B-0489
He diffusion in SAFOD core samples
Fluid flow in rocks is determined by porosity and fractures in the host matrix. In an effort to quantify fluid flow in the San Andreas Fault (SAF) and since direct fracture fluid sampling of the fault zone was not available, we have adapted a method to extract rare gases from matrix fluids of whole rocks by diffusion. Helium was measured on matrix fluids from the SAFOD main hole at ~3060m (clay-rich shale; American Plate), 3432m (clay-rich siltstone; very near active fault trace), and 3990m (siltstone; Pacific Plate) measured depth in September 2004, July 2005, and August 2005, respectively. Samples were typically collected as 2.54cm diameter subcores drilled into the ends of the cores, or from the core catcher and drillcore fragments within <2hr after core recovery. The samples were placed into ultra high vacuum stainless steel containers, flushed with ultra high purity nitrogen and immediately evacuated. Helium was measured at 4-6 times over a period of up to 1000 days in two sets of samples; one set (collected in 2004) was stored at room temperature, while the second set (collected in 2005) was kept in an oven at 120C. Our measurements indicate a 4He diffusion coefficient of 2-3x10-8 cm2/s at room temperature. Diffusion coefficients measured in the solid phase of sands and clays vary from 1.2x10-18 cm2/s at 21C to 3x10-15 cm2/s at 150C. (Solomon et al, 1996). Using the molecular diffusion coefficient of helium in water (7x10-5 cm2/s; 25C) for a measured matrix porosity (4%) and assumed tortuosity (approx. 5) produces an effective diffusion coefficient of 1x10-8 cm2/s, consistent with our results. Thus, our method effectively isolates the matrix pore fluid 4He from the 4He contained in the rock matrix.
H11B-0490
Groundwater Flow Through a Fractured Rock Aquifer in the Sierra Nevada Foothills of California
The Sierra Nevada foothill areas along the east side of California's Central Valley have experienced a continuous increase in population and land development activities. This has in turn put pressure on the local groundwater supply which is strongly controlled by the complex fracture hydrology of the region. Therefore, understanding how the ground water flows is essential for land use planning and water supply management. Two different approaches were employed in this study. Long-duration pumping tests were conducted to characterize parameters including transmissivity and storativity of the fractured granitic aquifer. δ 18O and δ D isotope ratios were used to trace the history and evolution of the surface and ground waters in the watershed. Results were examined in relation to the distributions and orientations of the fracture systems as observed on the surface by outcrop mapping and lineament studies using satellite and aerial photographs. Because of limited well capacities, constant-head pumping test methods have been found to be more practical than constant-discharge or step-drawdown methods. The results of different pumping tests suggest that both the flow patterns (radial or linear) and the aquifer parameters (transmissivity and storativity) are scale-dependent, and the scale effect is related to the anisotropy controlled by the fracture orientation and connectivity independent of the test methods. To characterize a large area of the fractured aquifer, a pumping test of at least 15 days is required to get a realistic trend line of drawdown versus time. Isotope data collected from a small watershed adjacent to the pumping test area show that the δ 18O and δ D isotope ratios are correlated significantly with the fracture distribution and orientation. Enrichment anomalies of isotopic ratios indicate that the fracture system imposes a strong control on the groundwater flow. Combining the use of pumping tests and isotopic survey can help assess the sustainability of groundwater supply for various locations in the fractured terrain.
H11B-0491
Numerical Modeling of the Fluid Flow through Deformable Fractured Porous Media
The interactions between fluids and deformation in the fractured porous media (e.g. rocks and soils) are critical in geological engineering fields, such as the reservoir engineering, geothermal engineering, the geological disposal of nuclear waste engineering and hydraulic engineering. A new finite element based numerical modeling of the deformation and fluid flow through fractured porous media is presented here with the special attention to the FEM mesh generation of the fractured media. Based on the available rock image data, the rock structure information can be extracted through the converted image data and the interface boundaries between different material/structures can be defined. All the above related information will be described via the specified data during the meshing process and output for the further finite element analysis. Optimized mesh size control is applied to ensure that the fine meshes generated around fractured/boundary zone, while coarse meshes generated in the rock matrixes if nothing special required. In addition, a new pressure-permeability formula is proposed to reflect the relationship between stress and the permeability of deformable fractured porous media. Finally, numerical tests have been performed to investigate the interactions between fluid flow and deformation in fractured porous media and demonstrate the efficiency and usefulness of the proposed algorithm
H11B-0492
HYDRAULIC FRACTURING IN SATURATED COHESIONLESS MATERIALS
Based on the developed experimental techniques, hydraulic fracturing in particulate materials has been directly observed in the laboratory. As a result, we suggested several mechanisms of hydraulic fracturing in particulate materials and determined relevant scaling relationships (e.g., the interplay between elastic and plastic processes). While the ongoing work is likely to change at least some conclusions, it is important that the results reported in this work appear to form the framework for modeling and, perhaps, even for (qualitative) interpretation of field data. The observed fracture geometry and the measured pressure injection curves suggest that hydraulic fracturing occurs in soft sediments in the following sequence: (i) cavity expansion, (ii) fracture front initiation, and (iii) propagation of the developed fracture. Our experiments show that liquid can indeed propagate as a crack-like feature when injected into cohesionless saturated materials. Laboratory observations suggest that at the initial stage, the cavity expansion process ends with fracture initiation. Sometimes, the growing fracture resembles penetration of one movable material into another less movable material, which may be a manifestation of the Taylor-like instability. An important conclusion of our work is that all parts of the cohesionless particulate material (including the tip zone of hydraulic fracture) are likely to be in compression. The compressive stress state is an important characteristic of hydraulic fracturing in particulate materials with low, or no, cohesion (such as were used in our experiments). At present, two kinematic mechanisms of fracture propagation, consistent with the compressive stress regime, can be offered. The first mechanism is based on shear bands propagating ahead of the tip of an open fracture. The second is based on the tensile strain ahead of the fracture tip and reduction of the effective stresses to zero within the leak-off zone. Additionally, an important characteristic feature of fractures in our experiments is the bluntness of the fracture tip, which suggests that plastic deformation at the fracture tip is important. Scaling indicates that fluid pressure does not decrease considerably along the fracture, due to the wide fracture aperture. However, there is a high pressure gradient in the leak-off zone in the direction normal to the fracture. Scaling also suggests the importance of fluid leaf-off in the cavity expansion and, possibly, in the fracture propagation process. First estimates show that large openings at the fracture tip correspond to large fracture energy, an order or two greater than for typical rocks. Unfortunately, it is not currently clear what defines the characteristic dimension at the process (tip) zone, which does not allow devising a comprehensive theoretical model. Without a model, it is not clear how to estimate an in-situ value of the fracture energy (or the corresponding value of the effective fracture toughness), that is, how to "extract" the fracture energy from available data of hydraulic fracturing tests or observations on natural hydraulic fractures (e.g., sand dikes propagated through unconsolidated sediments). However, it may still be possible that the fracture in field conditions is similar to that in conventional cohesive materials. Since what have been observed so far does not contradict to the condition of autonomity at the fracture tip (front), a tip-scale (local) fracture criterion still may be feasible to develop.