H23G-1688
Analyzing pumped-well impeller logs to ascertain vertical hydraulic conductivity variations
Horizontal variations in the hydraulic conductivity of aquifers are generally well characterized through simple pump test analyses. However, vertical variations are often poorly understood and misrepresented in the regional models used by regulatory bodies and water companies. Understanding these is key for predicting flow paths and hence the behavior of contaminants in the aquifer that might present a risk to public drinking water supplies. Traditionally, packer tests were used to characterize these variations, but they can be time consuming and costly to perform. However, other techniques have been developed which can quantify these variations, including impeller logging. This study aims to present new, more rigorous methods of analyzing impeller flow log data. Impeller logs were taken under pumped conditions in open wells in a chalk aquifer located in N. England. Theoretically, hydraulic conductivity can be obtained from the gradient in flow rate with depth. However, data are typically noisy due to turbulent flow and hole diameter variations with depth; so directly converting the flow rate gradient to hydraulic conductivity leads to rapid non-physical variation and negative hydraulic conductivity values. Correcting for hole diameter variations using caliper logs proved difficult due to phenomena such as jetting, whereby when the water enters a widening, it does not instantly slow down. In order to obtain more realistic hydraulic conductivity profiles, we firstly tried a data smoothing algorithm, but this approach distorted the data and still gave an unacceptable noise level. Instead, a layered modeling approach has been developed. A hydraulic conductivity profile consisting of a discrete number of uniform layers is constructed, and layer thicknesses and hydraulic conductivities are varied until a satisfactory fit to the observed flow log is achieved. Results from field sites on the confined Chalk aquifer of East Yorkshire in the United Kingdom showed good correlation to packer test analysis. The absence of significant ambient flows at this test site made the final analysis relatively simple. By testing boreholes across the aquifer a pattern of hydraulic conductivity variation with depth can be established, and compared to the proposed geological and climatic reasons for the variations' existence.
H23G-1689
Development of Pneumatic Inverse Model for Estimating Permeability and Porosity of Unsaturated Soils
The distributions of permeability and porosity are key factors that control airflow and gas phase transport in unsaturated soils. This study develops a pneumatic inverse model to estimate the distributions of permeability (K) and porosity (n) in unsaturated soils. Based on the concept of tomography surveys, the developed model accounts for compressibility and density of air and estimates the hydrgeologic parameters using air pressure measurements from sequential cross-hole pneumatic pumping or injection tests. Two synthetic examples and a laboratory sand column test were used to validate and test our inverse model in estimating the distributions of permeability and porosity. Results of the comparison show that our pneumatic inverse model can exactly reconstruct the property (i.e., permeability and porosity) fields if the well-defined conditions are met. With relatively small number of available measurements, the proposed model can closely capture the patterns and the magnitude of estimated properties in unsaturated soils. The validation with one-dimensional column test showed that the proposed inverse model can clearly map the patterns of different soils in the sand column. However, the estimated values of permeability and porosity show slightly difference from the values measured from the laboratory test.
H23G-1690
Can the Continuous Time Random Walk method be used to improve the upscaling of heterogeneous permeability fields for reservoir simulations?
Upscaling a fine-scaled geological model to a coarse, but computationally efficient, flow model is a standard process in reservoir engineering. However, the upscaled permeability and porosity fields do not preserve the original heterogeneity any longer. There are also many different methods for upscaling permeability and porosity. Multi-phase flow simulations in an upscaled hydrocarbon reservoir are hence less accurate and are likely to yield wrong predictions of oil recovery and water cut. The aim of this study was to improve permeability and porosity upscaling of a turbidite reservoir analogue by applying computationally fast conservative tracer transport simulations to each upscaled realisation and analysing the resulting breakthrough curves using the Continuous Time Random Walk method. Realisations that show anomalous transport behaviour that is comparable to the original, fine-scale, base case are likely to preserve, at least on the integral scale, the heterogeneity the best. They hence should provide oil production rates and water cuts that are similar to the base case. We have confirmed this assumption by carrying out two-phase flow simulations for the base case and selected upscaled permeability and porosity fields. Upscaled permeability and porosity fields that show an anomalous transport behaviour that is comparable to the base case yield production rates and water cuts that are similar to those of the fine-scale base case. Conservative tracer transport simulations and Continuous Time Random Walk analysis is hence a very powerful yet computationally efficient tool to improve the quality of upscaling and accuracy of reservoir simulations by allowing us to discard those permeability and porosity fields that will lead to oil production rates and water cuts that are significantly different from the expected fine-scale behaviour.
H23G-1691
Probability Density Functions of Hydraulic Head and Velocity in Three-Dimensional Heterogeneous Porous Media
In this study, we assess probability density functions of hydraulic heads and specific discharges in three- dimensional bounded heterogeneous porous media by Monte Carlo simulation (MC). We discuss their empirical shapes and demonstrate that the intuitive use of obvious information on boundedness leads to parametric distribution functions which fit surprisingly well. On the basis of statistical moments of hydraulic heads and velocities up to fourth order, we discuss the spatial dependence of the empirical distributions and their dependence on the variance of log-conductivity. Comparison of the first and second central moment to the results from classical numerical first-order second-moment (FOSM) analysis reveals that FOSM predicts these moments surprisingly close for hydraulic heads. Based on this fact, we demonstrate that fitting the chosen parametric distributions for hydraulic heads to FOSM moments is promising for the sake of estimating exceedance probabilities. Our Monte Carlo scenarios vary in variance of log-conductivity (0.125 to 5.0), in the type of multivariate dependence, in correlation scale and types of boundary conditions. Our study illustrates that, in contrast to the common assumption, FOSM is a reasonable choice for evaluating multivariate and univariate moments for heads, if used in conjunction with additional information on distribution shapes. In absence of utilizable additional information, we demonstrate that second-moment methods are mostly inadequate for assessing distributions accurately. Significant deviations from Gaussian distributions occurred for discharge components even at a variance of log-conductivity as low as 0.125, and we found that the distributions of transverse discharge components are extremely fat-tailed. The observed non-Gaussianity questions the results of approximate approaches in solute flux and dispersion studies where velocity fields are assumed to be multi- Gaussian and then directly represented by or generated from their covariances. The main implication is to apply more accurate schemes such as exact non-local methods, extensive MC or higher-order stochastic Galerkin approaches, and to include higher-order moments, at least if no additional assumptions on the shape of distributions are available or justifiable.
H23G-1692
Water Flow and Solute Transport in Heterogeneous Soils: A new Multicompartment Sampler and a Theoretical Toolkit for Data Analysis
Water flow and solute transport in soils are invariably affected by heterogeneity and often by preferential flow, both typically occurring within 1 square meter. Paradoxically, we need to understand flow and transport at this small scale to quantify them at the field and regional scales. This paradox arises from the geometry of soils: the scale in the direction of the flow is orders of magnitude smaller than the scales perpendicular to it. We present a coherent package of experimental and theoretical tools to observe and analyze small-scale variations (within 0.1- 1 square meter) of water and solute fluxes. Multicompartment samplers can measure small-scale water and solute movement in space and time, particularly in temperate climates. The latest generation of samplers allows repeated extraction of percolate samples in situ under controlled suction to minimize disturbance of the unsaturated flow field. After discussing the general principle of such samplers, a method will be presented to estimate the required total sampling area of a sampler from the degree of flow convergence in a soil. In recent years, we improved our ability to analyze the data produced by multicompartment samplers. The spatial solute distribution curve as the spatial equivalent of the travel time distribution was parameterized and physically interpreted. Both distributions were unified in the leaching surface, which has tremendous potential for detailed interpretation and model evaluation. Multicompartment samplers can also help identify the nature of the solute transport process. Recently, we expanded the theory of solute dilution to make it applicable to multicompartment sampler data. We will demonstrate how dilution theory can be used to determine the predominance of a convective-dispersive or a stochastic-convective transport regime during a tracer experiment.
H23G-1693
A field assessment of high resolution aquifer characterization based on hydraulic tomography with cross-hole slug tests
Slug tests have traditionally been utilized as a means to determine the hydraulic conductivity of an aquifer at a relatively small scale. When performed in a cross-hole mode using a tomographical measurement array, however, slug tests can be used to gain information about the distribution of hydraulic properties between two or more wells. Despite the common perception that a slug test only affects a small volume of the aquifer in the vicinity of the test well, response data with a reasonable signal to noise ratio can be collected at distances of over several hundred times the radius of the screen of the test well. In this work, we use a hydraulic travel time approach to evaluate the cross-hole slug tests. The approach, which is based on the inversion of travel times, is analogous to seismic ray tomography. The key element of this procedure is a travel time integral relating the square root of the peak travel time to the inverse square root of the hydraulic diffusivity for a Dirac point source. The potential of hydraulic tomography using cross-hole slug tests was investigated at a well-characterized sand and gravel aquifer located in the Leine River valley near Goettingen, Germany. The data base for the inversion consists of 200 cross-hole slug tests performed between five wells in which the positions of the sources (injection ports) and the receivers (observation ports), isolated with double packer systems, were varied between tests. The delay in the travel time of the pressure pulse due to wellbore storage and inertial effects was quantified and corrected using semi-analytical solutions based on the assumption of a homogeneous aquifer. The inversion results demonstrate the effectiveness of the proposed approach, as we were able to reconstruct four 2- D diffusivity tomograms with a resolution of 0.15×0.15 m2. The reconstructions were verified using single-well slug tests, grain-size statistics and geophysical well logs. The results of this study show that the combination of tomographical travel time inversion techniques and analytical evaluation approaches is particularly well suited for providing information about the structure and spatial variability of hydraulic properties of an aquifer, which are prerequisites for prediction of contaminant transport.
H23G-1694
A Stochastically Based Method for Multiphase Flow Modeling in Heterogeneous Media - Experimental Observations and Model Design
The results of two collaborative studies of multiphase flow that involve new modeling and experimentation are summarized. The goal was to explore new modeling approaches to explain observations of NAPL behavior in heterogeneous soils. In the modeling component, the ideas involving stochastic differential equations (SDE) used previously to model single-phase flow are extended to two phase flow. This approach results in a nonlinear SDE describing the position of the non-wetting phase fluid particle. The control of non-wetting phase particles across an interface is made using a jump term which derives from the Ito formula for cadlag semimartingales and is based on capillary diffusivity and the pressure-saturation curves of the sands forming the interface. The experiments were conducted in two-dimensional test tanks. The focus was to characterize the test material, develop experimental methods and conduct spill simulations in test tanks. The tanks were packed to represent different configurations of heterogeneity. The packing produced interfaces between different combinations of test sands, which allowed for the investigation of critical flow, fingering and pooling phenomena that occur at the transition zones of the heterogeneity field. The test soils that were used in the packing were accurately characterized using a new method based on Time Domain Reflectometry to obtain the constitutive relations needed in flow modeling. An X-ray photon attenuation method was used to determine porosity and saturations. This automated X-ray system was installed on a movable gantry that allowed for the continuous tracking of the NAPL saturations during migration and after entrapment. Three original ideas in our SDE model of multiphase flow are described which contribute to the relationship between the mathematical theory and the physical theory. The first idea is the inclusion of the jump term to model the capillary end effect. The second is what we call "anisotropic dispersion", where both the mathematical theory and the physical theory contribute to each other. In the case of the physics, the physical experiments clearly demonstrated the anomalous dispersion of the NAPL along the interface between two different sands, which dictated to the mathematical theory that a modification to the traditional diffusion model was needed. The mathematical theory, on the other hand, responded with the Girsanov-Meyer change of measure, which added a scaled dispersion term to the drift component of the SDE, allowing the model to simulate the anomalous dispersion effect. Finally, the third idea that we implemented was based on an approach that takes the flow to be controlled by pore-scale processes. Again, the modification to the macro-scale model was driven by the experimental observations, which many times showed the NAPL plume finding unexpected channels and exhibiting other types of instabilities commonly called "fingering".
H23G-1695
Thermal convection and vertical permeability in karstic networks
Karstic networks constitute highly heterogeneous aquifers. Modelling such aquifers at the 1000 yr time-scale is a fully coupled problem involving flow and reaction of meteoric waters with the carbonated matrix. A better knowledge distribution of groundwater in karstic area is required for water policy in these regions. Yet, the localization of caves in a karstic network cannot be presently predicted. Dissolution by meteoric waters in carbonated formations has been recently explored by accurate models. The drainage networks produced by these models are mainly controlled by the initial heterogeneity distribution (fault or joint, for example), by the regional state of stress and by boundary conditions. The main extension of these networks is mainly along the horizontal, from the top of the formation toward an outlet near a river. However, vertical voids such as shafts or vertically organized series of caves are present in several karstic networks. Shaft are believed to form by roof collapse and to connect superposed networks induced by base level changes, for example during the Messinian crisis in the Mediterranean areas. Roof collapse is however not the only mechanism able to form vertically organized structures and any mechanism involving gravity may be involved. Cooling of upwelling waters has been known for long to form so-called hydrothermal karst, due to the reverse solubility behaviour of carbonated minerals (i.e. their solubility increases with decreasing temperature). Moreover convection inside a fracture zone produced thin vertical upwelling currents which are able to produce vertically organized dissolution zones , at the opposite of round shaped zone expected from convection inside an homogeneous medium. Here we explore how thermal convection inside a fracture can induce chimney-like dissolution zones, by mean of coupled flow and reaction models designed for a series of theoretical cases. Then the conditions required to induce significant dissolution in vertically organized zones inside a limestone formation are discussed.
H23G-1696
Modeling of Thermal Runoff From an Asphalt-Paved Plot in the Framework of the Mass Response Functions
During hot summer months, impervious surfaces within urban areas may store significant amounts of thermal energy, which may be rapidly transferred to stream waters during runoff events. Modeling of heat release from impervious areas to stream waters thus represents a first, necessary step to quantify possible negative impacts of increased stream water temperature on nearby aquatic ecosystems. In this paper, a stochastic Lagrangian approach is developed to simulate heat transfer from an impermeable surface to runoff. The approach is based on the framework of the mass response functions (MRFs), which was originally developed for modeling non-point source pollutant transport in watersheds. The MRF approach has been adapted to describe heat transfer from impervious surfaces to runoff by coupling a heat balance at the asphalt/water interface and a one-dimensional heat diffusion equation within the asphalt. The model incorporates a simplified, physically based description of all the heat fluxes possibly affecting the ensuing thermal response of impervious areas (e.g., solar radiation, evaporation). The model was applied to an artificial asphalt-paved plot of 90 m2 where it was able to accurately reproduce the temperature variation of the asphalt surface and runoff during an artificially produced rainfall event. The effect of the heat diffusivity on the surface temperature response to rainfall input was also examined, showing that the effect could be significant depending on vertical temperature distributions of the plot.
H23G-1697
The Concept of Tensorial Connectivity-Tortuosity for Multiple Fluids to Describe Saturation- Dependent Anisotropy in Hydraulic Conductivity
Natural geological formations are heterogeneous and often anisotropic. Numerous researches have found that the anisotropy in hydraulic conductivity is dependent on fluid saturation. The tensorial connectivity-tortuosity (TCT) concept has been used to describe the hydraulic conductivity tensor of unsaturated soils with saturation- dependent anisotropy. In this investigation, this concept is extended as the concept of tensorial connectivity- tortuosity for multiple fluids (TCT-MF) in unsaturated porous media. The hydraulic conductivity of each fluid in an anisotropic porous medium under unsaturated conditions is described in the form of a symmetric second-order tensor. The TCT-MF theory applies to the generalized hydraulic conductivity model and compatible types of saturation-pressure formulations. The model shows that the anisotropic coefficient of any one of the fluids is dependent only on the saturation of the fluid being considered but independent of the saturation of other fluids. The TCT-MF concept was tested using numerical experiments of infiltration in synthetic Miller-similar soils with four levels of heterogeneity and four levels of anisotropy. The synthetic soils were generated to be anisotropic by allowing the saturated hydraulic conductivity to have different correlation ranges for different directions of flow. The numerical experiments of infiltration of two liquid phases, i.e., water and the nonaqueous phase liquid (NAPL) carbon tetrachloride, were conducted. The results show that, similar to water in a two-fluid (air-water) system, NAPL retention curves in a three-fluid (air-NAPL-water) system were independent of flow direction but dependent on soil heterogeneity; the connectivity-tortuosity coefficients were functions of both soil heterogeneity and anisotropy. The TCT-MF model accurately described the unsaturated hydraulic function of each fluid of the anisotropic soils and can be combined into commonly used relative permeability functions for use in multifluid flow-and-transport numerical simulations.
H23G-1698
Fractal Viscous Fingering in Fracture Networks
We have used two very different physical models and computer codes to study miscible injection of a low- viscosity fluid into a simple fracture network, where it displaces a much-more viscous "defending" fluid through "rock" that is otherwise impermeable. The one code (NETfLow) is a standard pore level model, originally intended to treat laboratory-scale experiments; it assumes negligible mixing of the two fluids. The other code (NFFLOW) was written to treat reservoir-scale engineering problems; It explicitly treats the flow through the fractures and allows for significant mixing of the fluids at the interface. Both codes treat the fractures as parallel plates, of different effective apertures. Results are presented for the composition profiles from both codes. Independent of the degree of fluid-mixing, the profiles from both models have a functional form identical to that for fractal viscous fingering (i.e., diffusion limited aggregation, DLA). The two codes that solve the equations for different models gave similar results; together they suggest that the injection of a low-viscosity fluid into large- scale fracture networks may be much more significantly affected by fractal fingering than previously illustrated.
H23G-1699
General Thermodynamic Parameterization for multi-component multiphase flow
We present a general parameterization of the thermodynamic behavior of multiphase, multi-component systems. The phase behavior in the compositional space is represented using a low dimensional tie-simplex parameterization. For example, these are tie-lines and tie-triangles for two- and three-phase systems, respectively. This parameterization improves the robustness of the phase behavior representation (e.g., phase identification) as well as the efficiency of various types of compositional computations. We demonstrate this Compositional Space Parameterization (CSP) framework for several multiphase multi- component porous media flow problems. Large-scale compositional simulation in highly heterogeneous reservoirs, involving a large number of components, is one type of applications. In the standard compositional simulation approach, an Equation of State (EoS) is used to describe the phase behavior. For each gridblock, given the temperature, pressure and overall compositions, the EoS is used to detect the phase state (e.g., one, two, or more phases), and if multiple phases are present, calculate the phase compositions. These EoS computations can dominate the overall simulation cost. We compare our adaptive CSP approach with standard EoS based simulation for several challenging problems of practical interest. The comparisons indicate quite clearly that the CSP strategy is more robust, and that it leads to an order of magnitude gain in computational efficiency. Another type of applications is an equilibrium flash calculation of systems with a high number of phases (e.g. three or more). The complexity and strong nonlinear behaviors associated with such systems pose serious difficulties for standard techniques. Here, we describe an effective tie-simplex parameterization for such systems at a fixed pressure and temperature. The preprocessed data can be used in conventional EoS based calculations as an initial guess to accelerate convergence.
H23G-1700
Two-Phase Fluid Flow in a Porous Medium Under the Influence of an Oscillatory Body Force
We study the effect of applying an oscillatory body force to a two-phase flow in a porous medium. This is done to get a better basic understanding of how seismic stimulation of an oil reservoir works. We study the behavior at pore scale. The system is both studied in simulations and experiments. The experiments are done in a quasi two-dimensional Hele-Shaw cell containing a porous medium made up of circular cylinders. The simulations are done using a Lattice Boltzmann algorithm which reproduce all the essential properties of a real fluid. Having the exact same porous medium in simulations and experiment makes it possible to do a closer comparison between the dynamics in simulation and experiment. By initially doing a two-dimensional simulation and directly comparing with the experiments allows us to see if this captures the governing behavior of the fluid flow in the system. This is an ongoing study, but preliminary results are in concurrence with our initial assumption that the invading fluid fills a larger fraction of pore space under the influence of an external oscillatory force than in an ordinary invasion process.
H23G-1701
Morphological Changes of the Invading Structure in a 2D Multiphase Flow Driven by Oscillatory Pressure
The interest in investigation of a multiphase flow in porous media has become increasingly popular due to its applicability in oil and environmental engineering. The observed increase of oil recovery in the production wells affected by seismic activity has suggested alternative ways in stimulating removal of residual oil in reservoirs. We experimentally investigate a horizontal flow of two immiscible fluids driven by oscillatory pressure in a small scale quasi two-dimensional porous synthetic medium. The wetting phase (water-glycerol solution) is withdrawn from the porous matrix at a constant speed creating a pressure drop between the phases and initiating the flow. To replicate the seismic activity the pressure in the non-wetting phase (air) is being oscillated. The withdrawal is performed at a slow speed providing that the flow stays within the limits of the capillary regime. The amplitudes and the frequencies of the oscillations are controlled. Systematically changing the amplitude and the frequency different morphologies of the invading cluster are achieved. The change in the shape of the invading structure is greatly due to the elasticity of the system as it responds to the pressure perturbations. By controlling the elasticity we are able to mimic the nature of porous material in oil reservoirs. The difference in the morphological characteristics results in the extraction of various amounts of the non-wetting phase. Ultimately these findings can be used in enhancing oil production or for the soil and ground water remediation routines (e.g. removal of non-aqueous phase liquids).
H23G-1702
Laboratory Experiments for Seawater Intrusion into Freshwater Aquifer with Heterogeneity
It is important for safety assessment of high-level radioactive waste geologic disposal to understand groundwater flow in deep underground accurately. Especially, groundwater flow in the coastal area is considered to be quite complex that involves density and hydraulic gradient driven flow of freshwater and seawater. In order to understand the behavior of seawater intrusion into freshwater in deep underground, we constructed a laboratory equipment, eMini-MACROf (MAss transport Characterization in host ROck). Mini-MACRO consists of three parts: a sandbox (0.5m x 0.25m x 0.1m) and a reservoir tank on each side containing saltwater simulating seawater and freshwater, respectively. Seawater intrusion experiments are conducted using glass beads (sub- millimeter in diameter) and colored saltwater in the sandbox with a transparent face plate to allow visual observation. We created several cases of experimental conditions to observe the seawater intrusion behavior into two-layered stratum against various hydraulic gradients and densities of saltwater resembling the so-called Henry Problem. We confirmed that the results using this equipment match numerical results under simple heterogeneous condition. These results contribute to the better understanding of seawater intrusion behavior and to increasing confidence in modeling methodology of groundwater flow and mass transport in deep underground through comparison with numerical analysis. We believe that it is crucial for the safety assessment of geologic disposal to integrate this knowledge.
H23G-1703
Case studies on the fresh-saline water interactions in volcanic rock with numerical modeling
General Geology in volcanic island, such as Ohau and Jeju Island in South Korea, is highly complicated. And the complexity of geology can provide the difficulties when we try to identify the hydrogeological systems in the volcanic island. For 4 years since January 2004, we have surveyed the hydrogeological features on Jeju Island; the spatial distribution of springs, discharge rates, water budget including direct runoff and fresh-saline groundwater interaction in near coastal area. Among these features, the fresh-saline groundwater interaction in near coastal area is one of the important factors to establish the public water supply management in Jeju Island. Numerical studies were conducted to investigate the fresh-saline groundwater interaction, considering channeling effect in Jeju Island with FEFLOW package. We applied 8 cases of conceptual models and 4 types of inland boundary conditions; constant head/constant flux, time-varying head/constant flux, constant head/time- varying flux and time-varying head/time varying flux. But in this study, we did not consider the oceanic tide. In cases of four conceptual models among 8, we assign the orthogonal channel network and different hydraulic conductivities to channel and matrix block respectively; 10 times higher in channel, 100 times, 1,000 times and impermeable matrix block. And in the other four, it was assumed that the whole media are homogeneous. The cases considering channeling effect with constant head and constant flux boundary condition show some different features in the fresh-saline groundwater interface. In 4 homogeneous cases, the stepwise shape of the interface did not observed while, in 4 channel-considering cases, the sharp and stepwise interface can be observed. When time-varying boundary conditions were applied, homogeneous cases showed the small amplitude of interface variations and slow responses. But 4 channel-considering cases showed totally different features. System responses are relatively fast and the amplitude of interfacial response is higher. These results from channel-considering cases are well matched with the monitoring data in Jeju Island.
H23G-1704
Modeling Variable-density Groundwater Flow in Coastal Aquifer: Submarine Groundwater Discharge (SGD), Tides Influence and Near-shore Circulation
Two-dimensional variable density simulations are performed for a coastal area of the northeastern Gulf of Mexico by applying finite difference model SEAWAT2000. For modeling, piezometers are installed along the coastline perpendicularly and sensors in particular piezometers keep providing hydraulic head and salinity observations by every 10 minutes in company with the changes of sea level. The influence of oscillations of dispersive mixing freshwater-saltwater interface is discussed to the estimation of SGD in the aquifer and through the sea floor. Due to insufficient information of coastal aquifer, several trial models were set first to investigate the influence of heterogeneity. Modeling investigations are also made to research the suggested mechanisms of SGD, such as tidal pumping, near-shore circulation or seasonal exchange and the phenomena related to them. Meanwhile, under three periods: stable sea-level, regular tides and storm, uncertainties of the transient models are analyzed.
H23G-1705
Characterization of fracture aperture for groundwater flow and transport
This paper presents experiments and numerical analyses of flow and transport carried out on natural fractures and transparent replica of fractures. The purpose of this study was to improve the understanding of the role of heterogeneous aperture patterns on channelization of groundwater flow and dispersion in solute transport. The research proceeded as follows: First, a precision plane grinder was applied perpendicular to the fracture plane to characterize the aperture distribution on a natural fracture with 1 mm of increment size. Although both time and labor were intensive, this approach provided a detailed, three dimensional picture of the pattern of fracture aperture. This information was analyzed to provide quantitative measures for the fracture aperture distribution, including JRC (Joint Roughness Coefficient) and fracture contact area ratio. These parameters were used to develop numerical models with corresponding synthetic aperture patterns. The transparent fracture replica and numerical models were then used to study how transport is affected by the aperture spatial pattern. In the transparent replica, transmitted light intensity measured by a CCD camera was used to image channeling and dispersion due to the fracture aperture spatial pattern. The CCD image data was analyzed to obtain the quantitative fracture aperture and tracer concentration data according to Lambert-Beerfs law. The experimental results were analyzed using the numerical models. Comparison of the numerical models to the transparent replica provided information about the nature of channeling and dispersion due to aperture spatial patterns. These results support to develop a methodology for defining representative fracture aperture of a simplified parallel fracture model for flow and transport in heterogeneous fractures for contaminant transport analysis.
H23G-1706
Solute transport in a physically and chemically heterogeneous aquifer model
Solute transport in heterogeneous soil has been intensely studied for the past 30 years. Theories have shown the dependence of effective transport parameters on heterogeneity, and mathematical models have been developed to explain anomalous concentration observations. Yet, little experimental data are available to validate these theories and models under conditions of high heterogeneity and chemical interaction between soil and solute, which are most likely to occur in the field. In an attempt to address this lack of data, we performed a tracer test in a physically and chemically heterogeneous aquifer model set up in an intermediate-scale laboratory test bed. The hydraulic conductivity (K) field follows a three-facies Markov Chain transition probability model. It is highly anisotropic and is characterized by high permeability contrasts between facies (up to 1/7000), resulting in an overall variance of ln(K) close to 16. Steady-state flow is established by imposing a constant flux over the upstream boundary and by fixing the head at the downstream boundary. A numerical flow model developed using MODFLOW 2000 was calibrated based on total discharge and pressure data recorded at 19 locations within the test aquifer. The model allowed the estimation of material permeabilities and highlighted the high-velocity preferential pathways and zones of low velocity. Rhodamine WT (RWT) was injected as a tracer in the upstream inflow reservoir. RWT is known to be composed of two isomers present in equal weight proportion, but having different sorption properties. Concentrations were monitored at 17 locations and at the outlet boundary. Recorded breakthrough curves (BTC"s) are highly skewed, with early arrivals and heavy late-time tails. Numerical transport models are set up using MT3DMS and RT3D. Qualitative comparison of experimental BTC's with the numerical results highlights that the early breakthroughs are caused by (1) rapid advection in preferential pathways; and (2) the presence of an inert isomer. The heavy tails are caused by (1) slow advection and diffusion through low-permeability zones; (2) the presence of a sorbing isomer; and (3) nonlinear sorption. These results and findings are presented and discussed.
H23G-1707
Exploring the Depth and Nature of Flow Systems in Fractured Igneous and Metamorphic Bedrock Aquifers
In rocks possessing highly discontinuous and heterogeneous flow paths, such as fractured igneous and metamorphic rocks, regional scale flow concepts are not as easily applied due to the lack of continuity in flow paths. It is traditionally assumed that flow in these bedrock systems decreases with depth due to a stress- induced decrease in permeability. However, recent hydrogeologic studies have suggested that deeper crustal materials are likely more porous then previously assumed and that critically stressed fractures are the most permeable. In the Northeast U.S. most bedrock consists of these rocks and is more frequently being relied upon as a source of drinking water for both private and public water supplies. It is of essential importance to understand the nature and depth of flow systems in these settings for determining the sustainability and vulnerability of water supplies. Previous workers have completed an extensive fracture mapping study of the Nashoba Terrane, a bedrock terrain located in the densely populated suburbs of Boston, MA, in an attempt to link the surface expression of fractures with the hydrogeologic properties of the subsurface. They suggest a series of hydrostructural domains, based on fracture orientation and density, which govern direction and quantity of flow. We present geophysical and hydrologic data collected in fifteen deep (> 100 meters) boreholes in this domain, in conjunction with a series of discrete fracture network (DFN) models that incorporate the field data, to test the structural domain hypothesis, and to quantify the depth and nature of flow in the terrain. DFN models are used to explore the relationship between flow and fracture orientation, employing a simple relationship between fracture aperture and depth. The field data suggests that most flow occurs in the upper 40 meters of shallow crust, but some flow is found deeper. Field data and DFN models suggest that these fractures are preferably oriented with respect to the assumed local stress field and support the hypothesis that even at shallow depths, permeable fractures are controlled by the local stress state.
H23G-1708
Three-dimensional discrete fracture network simulations of flow and particle transport based on the Laxemar site data (Sweden).
We study particle transport in a 3D DFN scenario based on the Laxemar site characterisation data in Sweden, which is a candidate repository site for high level radioactive waste in the Swedish nuclear waste management program. The site characterisation data has revealed several interesting geometric and hydraulic fracture properties, such as power-law distributed fracture sizes and transmissivities. Our study involves investigating the relationship between the resulting Eulerian flow field at a segment (sub- fracture) scale with Lagrangian trajectories at the characteristic (model domain) transport scale. We present results from a new technique for upscaling particle transitions obtained from Eulerian flow statistics to predictions of tracer discharge at the characteristic transport scale, based on previously developed methods used for 2D DFN's. This includes a mapping algorithm for transforming Eulerian into Lagrangian flow statistics without a priori knowledge of network connectivity, and by retaining the correlation between the water residence time τ and the hydrodynamic control of retention β we present accurate tracer discharge predictions. These results are illustrated using the unlimited diffusion model, and for some hypothetical tracers with properties designed to capture the behaviour of many common radionuclides. Finally we emphasise the importance of capturing the early arrival and peak of tracer breakthrough curves, i.e. to capture the bulk of the tracer mass arrival, in order to make accurate and conservative predictions.
H23G-1709
Diffusion, Dispersion, and Uncertainty in Anisotropic Fractal Porous Media
Motivated by field measurements of aquifer hydraulic conductivity (K), recent techniques were developed to construct anisotropic fractal random fields, in which the scaling, or self-similarity parameter, varies with direction and is defined by a matrix. Ensemble numerical results are analyzed for solute transport through these 2-D "operator-scaling" fractional Brownian motion (fBm) ln(K) fields. Contrary to some analytic stochastic theories for monofractal K fields, the plume growth rates never exceed Mercado's (1967) purely stratified aquifer growth rate of plume apparent dispersivity proportional to mean distance. Apparent super-stratified growth must be the result of other demonstrable factors, such as initial plume size. The addition of large local dispersion and diffusion does not significantly change the effective longitudinal dispersivity of the plumes. In the presence of significant local dispersion or diffusion, the concentration coefficient of variation CV={σc}/{\langle c \rangle} remains large at the leading edge of the plumes. This indicates that even with considerable mixing due to dispersion or diffusion, there is still substantial uncertainty in the leading edge of a plume moving in fractal porous media.
H23G-1710
Global and local probability density function of non-reactive solute concentrations in heterogeneous porous formations
Because of our inability to map the hydraulic properties of heterogeneous formations in detail, most models of solute transport treat solute concentrations as a stochastic variable usually providing only the ensemble mean and in same cases the concentration standard deviation. However, they all miss to give the local concentration probability density function (pdf), which describes the variability of solute concentrations entirely at a given location and is essential in risk analysis where confidence intervals and probability of exceeding threshold values are required. Therefore, we developed a new model for the pdf of the local concentration of conservative tracers migrating in heterogeneous aquifers. Our model accounts for dilution associated with pore scale dispersion, mechanical mixing within the sampling volume, and spreading due to formation heterogeneity. We modeled local concentration dynamics with an Ito Stochastic Differential Equation (SDE) that under the hypothesis of statistical stationarity leads to the Beta pdf for solute concentrations. This model is fully characterized by the solute concentration first two moments, which are the same pieces of information required for standard geostatistical techniques and predicted by stochastic solute transport models. Additionally, we show that in the absence of pore-scale dispersion and for point concentrations the pdf model converges to the binary distribution of Dagan [1982, Water Resour. Res. 18(4), 835-848], while it approaches the Normal distribution for sampling volumes much larger than the characteristic scale of the aquifer heterogeneity. Furthermore, accurate numerical simulations show that our model performs well regardless of aquifer heterogeneity and captures the smoothing effect of the sampling volume and the associated reduction of the probability of exceeding large concentrations. Moreover, we demonstrate that the same model with spatial moments of the solute concentration replacing the statistical moments can be applied to estimate the probability of exceeding a given concentration irrespective of the location within the plume providing a model of global uncertainty. Application of this model to point and vertically averaged bromide concentrations from the first Cape Cod tracer test and to a set of numerical simulations confirms the above findings and for the first time we demonstrate the superiority of the Beta model to both Normal and Log-Normal pdf in interpreting field data. Furthermore, we show that characterization of local concentrations as normally or log-normally distributed may result in severe underestimates of the probability of exceeding large concentrations.
H23G-1711
Dispersion Analysis Using Particle Tracking Simulations Through Heterogeneity Based on Outcrop Lidar Imagery
Solute plumes are believed to disperse in a non-Fickian manner due to small-scale heterogeneity and variable velocities that create preferential pathways. In order to accurately predict dispersion in naturally complex geologic media, the connection between heterogeneity and dispersion must be better understood. Since aquifer properties can not be measured at every location, it is common to simulate small-scale heterogeneity with random field generators based on a two-point covariance (e.g., through use of sequential simulation algorithms). While these random fields can produce preferential flow pathways, it is unknown how well the results simulate solute dispersion through natural heterogeneous media. To evaluate the influence that complex heterogeneity has on dispersion, we utilize high-resolution terrestrial lidar to identify and model lithofacies from outcrop for application in particle tracking solute transport simulations using RWHet. The lidar scan data are used to produce a lab (meter) scale two-dimensional model that captures 2-8 mm scale natural heterogeneity. Numerical simulations utilize various methods to populate the outcrop structure captured by the lidar-based image with reasonable hydraulic conductivity values. The particle tracking simulations result in residence time distributions used to evaluate the nature of dispersion through complex media. Particle tracking simulations through conductivity fields produced from the lidar images are then compared to particle tracking simulations through hydraulic conductivity fields produced from sequential simulation algorithms. Based on this comparison, the study aims to quantify the difference in dispersion when using realistic and simplified representations of aquifer heterogeneity. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy's National Nuclear Security Administration under contract DE-AC04-94AL85000.
H23G-1712
Power Law Tailing in Air-spargingand Invasion-percolation Model for Calculating Diffusion in the Liquid Phase
Power law tailings are often seen in removal of pollutants from soil or ground water. We have found that the power law tailing can also be seen in laboratory scale column air-sparging experiments. For this, we propose the following hypothesis. Introducing gas to the water saturated glass bead column causes some fractal-like gas flow path, which causes the power law distribution of distance that solute has to travel until it reaches the gas- water interface, this causes the power law tailing. In order to prove this, we constructed a model for removal of solute in water by diffusion to gas/water interface based on the invasion percolation. In a simple cubic bond lattice, invasion percolation is performed, assuming that the invading phase is gas and remaining phase is water. Then, set the initial solute concentration in water uniformly unity, set the boundary conditions as the concentration of the solute at the gas/water interfacial nodes is zero (or for random walk, set the interface as absorbing ends). Then diffusion process was calculated with two different methods: finite difference method for solving the partial differential equation based on Fick's law, and the random walk method. The results of two different methods are qualitatively identical. Near the percolation threshold of gaseous phase, the removal rates revealed power law tailing, and the slope in double-log removal late vs. time curves was about -1.0. At higher gaseous content, away from the threshold, this kind of power law behavior was not seen. The removal rate of solute from the lattice filled with water and the top and bottom face are set as zero concentration or absorbing, we got the slope in double-log plot of -0.5, this was in good accordance with the theory. Theoretically, at the threshold of gas phase, water phase in three dimensions is away from threshold, therefore, the water phase cannot be seen as fractal. We consider, in the short range of time, the dangling of water phase contributes to the power-law like distribution of diffusion distance of solute, and thus causes the power law tailing. In the presentation, we discuss the effect of gaseous content, lattice size on the removal rate and the distribution of diffusion distance.
H23G-1713
Simulation of transport under random retardation using CTRW: Impact of small scale spatial heterogeneity on effective transport.
We study the impact of small scale spatial heterogeneity on effective transport. Small scale heterogeneity is represented explicitly by a spatial random retardation factor that is fluctuating on a length scale much smaller than the observation scale. Using continuous time random walks (CTRW) as a coarse graining tool, the heterogeneity distribution can be mapped directly on the distribution of transition times. The latter renders an explicit effective transport model that is completely defined by the small scale heterogeneity distribution. We conduct a series of random walk simulations in order to validate the theoretical findings. Numerical experiments in d=1, 2 and 3 spatial dimensions study spatial and temporal aspects of solute transport in terms of spatial distributions as well as solute breakthrough curves. These findings are used for the interpretation of actual field transport measurements. The experimental findings are in agreement with theoretical predictions.
H23G-1714
Laboratory Assessment of Variable Retardation of a Reactive Solute in Stratified Geomedia
The migration of solutes through heterogeneous geomedia is of interest to a variety of disciplines. The movement of a reactive solute relative to water is characterized by a retardation factor (Rf); the ratio between the travel times of the reactive solute and the bulk solution. Although the retardation factor is often treated as constant, for heterogeneous media it can be a function of the travel distances. We have conducted a series of tracer breakthrough experiments using a small two-dimensional flow cell with idealized layered chemical heterogeneity constructed from naturally metal (hydr)oxide coated sands of different facies from an Atlantic coastal formation near Oyster, VA. Reactive tracer pulses from1 from one pore volume through step breakthrough were applied to simulate different travel distances with flow directions both parallel and perpendicular to layering. Method of moments determination of mean breakthrough time demonstrates that Rf is an anisotropic property and it decreases with travel distance (i.e., smaller applied pulse of tracer) when flow is parallel to layering. These results highlight the need for detailed geologic delineation of the subsurface for accurate prediction of contaminant breakthrough and risk analysis.
H23G-1715
Continuous Time Random Walks: Spatio-Temporal Coupling and Long Jumps
We study continuous time random walks (CTRW) for a fully coupled joint transition length and time distribution ψ(\vs,t) using highly efficient space-time random walk simulations. We study the competition between long jumps and long waiting times and the resulting transport behavior using a model of Levy Walks with a certain velocity v-distribution Φ(ξ), ξ \equiv 1/v. Such models can serve as an effective description of transport in random fracture networks. The spatial and temporal aspects of the transport behavior of these CTRW models is studied in terms of first arrival time distributions and spatial solute distributions and their moments. These results are compared to different forms of the uncoupled case ψ( s,t) = p( s)ψ(t). The advantage of using random walk simulations for coarse-grained heterogeneous media1 will be shown.
H23G-1716
Quantifying the impact of the velocity field organization on the solute transport properties
We propose a framework to quantify the velocity field organization in heterogeneous media for modelling solute transport. The Lagrangian velocity correlation properties are analyzed from numerical simulation of transport in heterogeneous permeability fields. The two-point Lagrangian correlation function is shown to be insufficient in general to quantify the velocity field spatial organization. An alternative method is proposed based on the Lagrangian velocity correlation properties conditioned on the solute particle initial velocity. It allows capturing the effect of the different spatial organization of high and low velocity zones. For the permeability fields studied, we find that low velocity zones are associated with larger Lagrangian correlation times than high velocity zones. We define a random walk model that can simulate solute transport with velocity dependent correlation and discuss its properties and applicability.
H23G-1717
Quantitative relationship between soil structure and preferential flow and transport in contrasting soils and land uses
The elusive relationship between preferential flow dynamics and soil structure is a main factor prohibiting accurate estimation of flow and transport in structured soils. It is essential to develop quantitative relationships between soil structure (especially pore network) and soil hydraulic properties. Recent advance in computed tomography (CT) provides a promising tool to non-invasively observe soil structure. Two soils with contrasting textures and structures (Hagestown and Morrison) and land use (crop and pasture) were investigated in this study. Five soil cores (10 cm in diameter and 30 cm in length) were taken for each land use-soil type combination. The soil columns were scanned using X-ray CT to obtain soil structural parameters (pore size distribution, pore length density, pore hydraulic radius, pore continuity, pore connectivity fractal dimension, and lacunarity). The chemical breakthrough experiments were then conducted on each soil column and the parameters related to preferential flow and transport were reversely estimated (including dispersion coefficient, mass transfer coefficient, and the fraction of macropores). The resultant two sets of the parameters are being analyzed to quantitatively relate soil structure to preferential flow and transport. This study is expected to enhance our understanding of fundamental processes and quantitative relationships for predicting preferential flow and transport at the soil pore to column scales.
H23G-1718
Evolution of Nanometer-scale Pores During Mineralization in Rocks
Nanometer-scale pores in geological media can change the effective solubility of minerals, allowing highly supersaturated fluids to exist within the porous medium; this process, termed pore-size controlled solubility (PCS), stems from the surface tension associated with crystals growing in rigid pores. In many ways analogous to the capillary pressure at a liquid-vapor interface within a pore, this surface tension gives rise to an excess pressure within the crystal, which can be related both to solubility and pore size. Crucially for geological media, theory predicts that minerals can precipitate when a fluid flows from small pores into larger ones, and such a mechanism could account for the preferential mineralization often observed in high-porosity zones and fractures within rocks. Here, we use numerical simulations to demonstrate how PCS could affect mineralization patterns in rocks adjacent to pressure-dissolution interfaces (stylolites). We show that in systems with constant solubility, nanometer-scale pores will close rapidly due to their high specific surface area; by contrast, when the PCS mechanism is included in the model, macro-porosity is filled first, with nano-scale pores remaining open throughout the simulations. In addition, our initial observations in stylolite-bearing rocks suggest that nano-scale pores could control the evolution of porosity in many geological systems.
H23G-1719
Can more data make a model worse? A transport experiment in an exhaustively sampled sandstone slab
An advantage of numerical models over analytic methods is their ability to include field measurements, such as hydraulic conductivity (K) at measured locations. Poor model fit with observed conservative transport is typically attributed to lack of K measurements. Previous research on a 30 x 30 x 2.2 cm slab of Massillon sandstone has shown that despite including the finest K measurements (every 0.33 cm), traditional numerical models fail to capture significant details of observed solute transport. The average K values still cannot adequately represent the sub-measurement scale heterogeneities within the sandstone. The resulting spatially averaged K field violates the assumptions of the local advection-dispersion equation (ADE). We present a new method which uses statistically similar Monte Carlo realizations within typical numerical models that can reproduce of the early breakthrough and late tails seen in natural aquifers. Using MODFLOW and a particle tracking code, ensemble results using a local ADE and no conditioning points capture the observed tracer transport better than models using the 8,649 K measurements directly. The results imply the need for a non-local transport model which can account for the sub-grid heterogeneities, while still honoring the K measurements. Increasing model discretization using an anisotropic fractal scaling model while including all K measurements may also help improve model fit with the experimental results.
H23G-1720
Simulations of non ergodic transport induced by inhomogeneous flow fields in heterogeneous porous formations of hierarchical sedimentary architecture
Dealing with subsurface flow and transport in naturally heterogeneous formations neither statistical homogeneity nor ergodicity are usually met in real world applications. If the stationarity is obeyed by Lagrangian velocity field the ergodicity of transport can be ascribed to the only size of the injection area, but this is not the case in all situations where flow field doesn't meet statistical homogeneity. In this case not only the size of the area where the concentration is injected, but also its location as well as the spatial variability of mean flux play a relevant role in the expected plume evolution. Among other causes, the hierarchical sedimentary architecture of porous formations combined with the finite size of the domain may be considered as one of the main causes of the flow field inhomogeneity and of the subsequent evidences of the non ergodic transport in real cases. To give a better understanding of this phenomenon synthetic 2-D cases of isotropic log conductivity field with integral scale λ1 whose expected value is assumed as a periodic function of the space coordinates of prescribed amplitude are investigated. The wave length of the mean log conductivity fluctuation is λ2, properly chosen to satisfy the relationship λ1 < λ2 < L, being L the characteristic dimension of the finite domain. The time evolution of the spatial moments of the plume driven by a statistically inhomogeneous steady state random velocity field is analyzed by varying the amplitude of the periodic mean flow and its wave length λ2 and by taking into account different sizes of injection area. These moments are achieved by space- time integration of the velocity field covariance structure derived according to the first-order Taylor series expansion by the stochastic finite element method. The discussion of the results leads to a better understanding into the ergodicity lack that affects solute transport in heterogeneous aquifers, by giving a comparative measure of the relevance of two possible causes, that is the flow field inhomogeneity due to the hierarchical sedimentary architecture of porous formations combined with a limited domain and the injection area of finite size.
H23G-1721
Identifying an Optimal ERT Measurement Set to Estimate Contaminant Plume Mass
Heterogeneous hydrologic properties lead to complex pathways for solute transport and spatially nonuniform distributions of solute mass through time. This nonuniformity presents challenges for characterizing solute mass distributions using conventional point sampling. Furthermore, it is unclear how relatively few point measurements can be combined to give reasonable estimates of the total solute mass. These challenges have led to increased interest in the use of geophysical (especially electrical) methods to characterize solute distributions through space and time. Although individual measurements of electrical resistance are rapid, collecting sufficient data to constrain the estimation of the electrical resistivity structure can be expensive. The objective of this study is to determine whether an optimal set of ERT measurements can be identified for the purpose of estimating the total solute mass in a plume through time. An optimal measurement set produces an estimate of the solute mass with the required accuracy and precision using the least number of measurements. Underlying this effort is the hypothesis that the larger support volume and smaller per-measurement cost of ERT measurements can improve estimates of total solute mass. However, these improvements will only be realized if the design of the ERT network explicitly considers the effects of subsurface heterogeneity on the spatial sensitivity of individual ERT measurements and on the entire ERT survey. We compare this approach to ERT survey designs that maximize the sensitivity of the ERT measurements without consideration of state dependent sensitivity. Finally, we show how genetic algorithms can be used to efficiently design optimal ERT surveys and how the optimal surveys used together with coupled flow, transport and ERT modeling improve plume characterization in heterogeneous media.
H23G-1722
Multi-spectral Colloid Transport Imaging at the Meso-scale
Light-transmission is a relatively inexpensive approach to visualization which can be applied at the meso-scale for observation of subsurface flow and transport affected by heterogeneities such as textural layering, textural inclusions, preferential flow paths, and transient infiltration. A light-transmission system was developed to allow for near-real-time collection of multiple signals for the detection of colloid concentration, water content, and water flow paths throughout a 70cm x 50cm x 1cm porous media system. Fluorescent latex colloids, fluorescent tracer dye, various grades of translucent silica sand, and a specific set of LEDs and bandpass filters are employed in this system to allow for the collection of multi-spectral data which may be used to evaluate water flow paths and transport behavior of colloidal particles under spatio-temporally heterogeneous subsurface conditions.
H23G-1723
Simulation of Porous Flow and Transport in Heterogeneous Media Using Fractal Interpolating Functions
Analogous to linear interpolation functions, fractal interpolation functions can be defined for fractally heterogeneous, multi-scale media. A fractal interpolating function (fif) is a continuous function that passes through a set of data points, and has the additional property that it is the attractor of an iterated function set (ifs). Each member of the ifs maps the entire domain into a subdomain, with the union of the subdomains covering the entire domain. Fractal (power law) behavior is a reasonable approximation to permeability scale dependence, at least over a range of scales. For media whose smaller-scale heterogeneity can be approximated by a fractal model, fifs provide a way to capture sub-gridscale properties and fluxes within a coarser finite difference or finite element grid. Advantages of this approach are that it can be incorporated into conventional finite difference or finite element codes, 2-D and 3-D applications are straightforward, and the solution can be recovered to any resolution desired. Several 2-D and 3-D examples illustrate the usefulness of this approach for transport in heterogeneous media.
H23G-1724
Effects of adaptive mesh refinement on estimation of hydraulic properties of aquifer
The hydraulic tomography is emerging technique for subsurface characterization. In order to analyze the field scale subsurface problems, we need to discretize the domain in finite number of elements. For the case of uniform grid pattern over the whole domain the elements under high head gradient are numerically treated same as the elements with smaller head gradient. In this study we analyze the effects of mesh refinement on the estimations of properties. We developed the adaptive mesh refinement algorithm and applied to the part of domain where high gradient is expected for the forward model and also for the sensitivity calculations in inverse algorithm (SSLE). The hydraulic tomography added with adaptive mesh refinement is applied to a synthetic case and to a sequential aquifer test at actual well field and results are discussed.
H23G-1725
Hydraulic Tomography Study Involving the Singular Value Decomposition Method
An oscillatory pressure signal has been generated in the field; and, the amplitude decay and phase shift have been monitored between source and receiver wells in a hydraulic tomography study. Hydraulic tomography studies traditionally require great amounts of time and computing power to process the large amount of data. Modeling of theoretical data indicates that the hydraulic conductivity values can be represented by spatially averaged straight ray paths. This is a good approximation except near the upper and lower boundaries, since the straight ray path method assumes an infinite aquifer. The assumption of a straight ray path greatly reduces the time and computing power needed for data processing, as no nonlinear iterative methods are necessary. The area between wells is divided into rectangular elements. The equations describing the ray paths are solved using Singular Value Decomposition (SVD) for the hydraulic conductivity distribution. The SVD method can resolve differing hydraulic conductivity values for elements of about one meter square or larger. Instability problems arise at the top and bottom of the area because there are fewer ray paths through these regions than through the center. The instability problem is avoided by collapsing the top row of elements into a single element and the bottom row into a single element. After collapsing the rows, the center values are still resolvable into one meter square blocks. Data collected in the field have been processed using the SVD method. The hydraulic conductivity values obtained from the SVD program are within the range of values typically obtained by other methods at the site. The SVD method shows promising results for a relatively simple analysis of oscillatory hydraulic tomography data. This work is funded by the Strategic Environmental Research and Development Program (SERDP) as project ER-1367.
H23G-1726
Three-Dimensional Groundwater Modeling for Tunnel Construction in Faulted and Fractured Bedrock
This paper describes the 3-D groundwater flow model developed for the San Francisco Public Utilities Commission's Irvington Tunnel No. 2 Project. The tunnel will be constructed through the East Bay hills east of Fremont in Alameda County, California. The proposed tunnel is located within about 200 feet and parallel to the existing Irvington Tunnel, which was built in the 1930's. Both tunnels are approximately 3.5 miles long, and have a ground cover ranging from about 100 to 750 feet along the alignment. The geologic formations along the tunnel include cretaceous and tertiary sandstone, shale, and claystone, separated by four inactive faults and a number of minor shear zones. The faults and shear zones contain varying thicknesses of crushed, broken, sheared, and/or soil-like material. Many of these zones provided large groundwater inflows during construction of the first tunnel. The 3-D groundwater flow model was developed using finite-element modeling code FEFLOW, which is capable of simulating discrete features such as fault zones, tunnels, and mine workings. Both, the existing tunnel and the proposed tunnel were incorporated into the model. The modeling objectives were to predict the groundwater inflows to the tunnel during its construction, and to assess the potential hydrological effects of the new tunnel construction on the local groundwater system. Groundwater level variations and flow conditions were modeled during tunnel construction and during the groundwater recovery phase following tunnel lining and completion. The model was first calibrated to simulate the groundwater response reported during construction of the first tunnel in the 1930's. The calibration was carried out using the parameter optimizer PEST. The model calibration was refined using a classical calibration process, focusing on characterizing the hydraulic properties (mainly conductivity and storativity) for the key geologic features. After satisfactory model calibration, the model was used to evaluate the groundwater response during the construction of the new tunnel for several alternative tunnel excavation and lining system scenarios. The predicted groundwater recovery following tunnel completion was also evaluated. The results were used by the design team to establish the tunnel excavation method and lining system for final design. This paper will present details on the methodology of modeling tunnel construction, modeling groundwater flow in fractured zones, and model development. The model setup, calibration, and predictive simulation results are also discussed.
H23G-1727
Log-scale permeability estimation from thin sections and its impact on the effective permeability of the formation of Donghae-1 gas field, Block VI-1, South Korea
Donghae-1 is an offshore gas reservoir located at the Block VI-1, 60km from Ulsan harbor, South Korea. The reservoir sand interval ranges depths from 2,330m to 2,620m, and is divided into five formations, B1 to B5. There are many thin shale layers in the formations. There are a few lab measurements of permeability and they showed a reasonable porosity-permeability relation for the entire reservoir. However, investigating thin sections revealed that formations can have a different lithology each other, and there are lithological differences even within a formation. This requires more careful and detailed investigation. We first obtained blue-epoxy saturated thin sections from available core samples at every 10cm depth, then estimated porosity and permeability from each thin section by the computational rock physics method, which estimates permeability by lattice-Boltzmann flow simulations on stochastically created 3D pore structures from thin section images. The numerically estimated permeabilites were compared to the lab measurements for quality control purposes. This enabled us to have different porosity-permeability relation in each formation. We even try to have multiple relations within a formation for B4 and B5. The multiple porosity-permeability relations were applied to the entire reservoir intervals using the well-log data and we could obtain well-log scale permeability estimation. This new methodology can give a better estimation for effective permeability for the reservoir and can be applied to the reservoir containing log-scale layers with measurable lithological differences.
H23G-1728 [WITHDRAWN]
The Effect of Open-Framework Gravels in Hyporheic Braided River Sediments on Ground Water Flow Velocity
Preserved networks of interconnected, clast-supported, large diameter pore spaces provide conduits for rapid ground water transport. These conduits are found within near-horizontal beds of coarse open-framework sediments naturally deposited and modified by braided rivers. The pore networks provide highly conductive pathways that dominate water flux in fluvial bars and bar complexes and become interconnected at scales of 10's to even 1000's of meters. They are here hypothesized to provide rapid flow through extensive hyporheic exchange networks in the braided river-flood plain setting. At two field sites in the Nyack floodplain in Montana, coarse, open-framework gravel beds have been determined to exist within braided river bar sediments, duplicating what others have found in other braided river settings. A number of small vertical-interval techniques were used to detect ground water velocities and hydraulic conductivities developed in closely-spaced wells along the flow-lines. Geophysical techniques, including borehole GPR and electrical resistivity tomography, are being used to image braided bar architecture for comparison with the hydrologic findings, and to extend the characterization of the sediments beyond the wells. The demonstrated integration of the geologic, geophysical and hydrologic results allows the detection of the link between sedimentological conditions and hyporheic exchange mechanisms in the braided river setting.
H23G-1729
A Stereolithography Pore-Throat Model
A new experimental, heterogeneous pore-throat model has been designed and fabricated using stereolithography (SL). In SL production, a laser cures a thin layer of photo-sensitive resin on the surface of a vat of liquid resin; a moveable platform then submerges the cured layer and a new layer is cured on top of the previous one, creating a physical model from a computer generated model. This layered fabrication of a computer generated model has enabled the production of an experimental porous medium with improved fluid resistance properties, as compared to previously studied, constant-height etched cells. A uniform distribution of throat widths was randomly placed throughout the pore-throat matrix and the throat height of each throat was assigned to increase the range of viscous and capillary resistances within the physical model. This variation in both throat height and width generated a porous medium with fairly low porosity (43%), permeability (~400 D), and wide range of geometric resistance properties. Experimental, two-phase immiscible drainage studies in the porous flowcell were performed. Analysis of the captured images was performed with open-source image processing software. These analysis techniques utilized the capability of both ImageJ and the Gnu Image Manipulation Program to be customized with ancillary codes. This enabled batch procedures to be created that converted the original grey-scale bitmaps to binary data sets, which were then analyzed with in-house codes. The fractal dimension, Df, (measured with box-counting) and percent saturation of these experiments were calculated and shown to compare favorably to fractal predictions and previous flowcell studies. Additionally, using the computer generated pore-throat geometry, a computational fluid dynamics model of two- phase flow through the porous medium was created. This model was created using FLUENT code and the Volume of Fluid method. The percent saturation of the less-viscous invading fluid determined with the computational model was shown to be within 2% of comparable experimental studies and the Df of the invading fluid mass was shown to be within 4% of the experimental studies. This computational model was then expanded to study different fluid and flow properties, highlighting the potential of computational fluid dynamics to study two-phase displacement under conditions similar to geological reservoir conditions.
H23G-1730
A Multi-Point Geostatistical Inverse Model for the Identification of Connected Permeable Subsurface Structures
Depositional features such as fluvial channel bodies often provide structural pathways for fluid and solute transport in the subsurface. In this work, we analyze complex aquifer test response data from an alluvial fan deposit in northern California. The subsurface architecture is characterized by discrete channel deposits embedded within finer-grained and less permeable floodplain deposits. We adopt a discontinuous facies framework to represent the aquifer structure. The distribution of channels is modeled using a sequential stochastic simulation technique based on multiple-point geostatistics. We couple the geostatistical simulation model with a groundwater model and apply an inverse method to find channel locations that are consistent with the hydrologic observations. The inversion routine, which informs simulations of the aquifer structure, relies on direct perturbation of nodal conditional pdfs that describe the probability of each facies. Results indicate that in response to pumping, rapid hydraulic head declines in a distant area of the aquifer system are facilitated by the presence of vertically interconnected channel deposits. The inverse model, guided by the dynamic head- response data, produces a similar channel structure at each successful realization.
H23G-1731
An Enhanced Method Using MODFLOW to simulate Groundwater Extraction/Injection through Wells Penetrating Multiple Aquifers
In a MODFLOW simulation of groundwater extraction and/or injection through a well penetrating multiple aquifers, the conventional method is to represent this type of well with a group of single-layer wells, each open to one of the model layers penetrated by the multi-layer wells Each of the single-layer wells have an individual rate specified for each stress period. Using this method, the total extraction and/or injection rate must be allocated among the individual layers. A common method of doing this is to divide the extraction and/or injection rates in proportion to the layer transmissivities. This partitioning has to be implemented by the user externally to MODFLOW for each multi-layer well and for each stress period in the MODFLOW well package. This approach fails to take into account the interconnection between various layers penetrated by the well, and is thus an incomplete solution to the problem. In both theory and practice, the extraction and/or injection rates through those well-penetrated aquifers also depend on the storativity and hydraulic head. Using transmissivity alone to partition the extraction and/or injection rates is only appropriate if both storativity and hydraulic head are very close for all well-penetrated aquifers. In cases where the top well-penetrated aquifer is unconfined, this method does not work because the storativity in an unconfined aquifer is often much higher than that of a confined aquifer. In addition, for transient flow simulations, the transmissivity of an unconfined aquifer changes with time due to the groundwater table variations resulting from extraction and/or injection (induced by constant or time-variable extraction/injection rates) or other time-variable boundary conditions such as seasonal groundwater recharge. This paper presents an enhanced method to resolve the above problem. With the new method, only the total extraction and/or injection rate of the well needs to be specified, and MODFLOW automatically allocates the extraction and/or injection rates for each well-penetrated layer during the simulation, based on the transimissivity, storativity, and hydraulic heads of each well-penetrated model layer. In addition, the interconnection between all well-penetrated layers is implemented. Test runs of example cases provide the comparison of flow and transport simulation results between using the conventional and the enhanced method, and simulation errors resulted from conventional methods are discussed.
H23G-1732
Analytical Solution for Time-drawdown Response to Constant Pumping from a Homogeneous, Confined Horizontal Aquifer with Unidirectional Flow
An exact analytical solution to the ordinary one-dimensional partial differential equation is derived for transient groundwater flow in a homogeneous, confined, horizontal aquifer using Laplace transformation. The theoretical analysis is based on the assumption that the aquifer is homogeneous and one-dimensional (horizontal); confined between impermeable formations on top and bottom; and of infinite horizontal extent and constant thickness. It is also assumed that there is only a single pumping well penetrating the entire aquifer; flow is everywhere horizontal within the aquifer to the well; the well is pumping with a constant discharge rate; the well diameter is infinitesimally small; and the hydraulic head is uniform throughout the aquifer before pumping. Similar to the Theis solution, this solution is suited to determine transmissivity and storativity for a two- dimensional, vertically confined aquifer, such as a long vertically fractured zone of high permeability within low permeable rocks or a long, high-permeability trench inside a low-permeability porous media. In addition, it can be used to analyze time-drawdown responses to pumping and injection in similar settings. The solution can also be used to approximate the groundwater flow for unconfined conditions if (1) the variation of transmissivity is negligible (groundwater table variation is small in comparison to the saturated thickness); and (2) the unsaturated flow is negligible. The errors associated with the use of the solution to unconfined conditions depend on the accuracies of the above two assumptions. The solution can also be used to assess the impacts of recharge from a seasonal river or irrigation canal on the groundwater system by assuming uniform, time- constant recharge along the river or canal. This paper presents the details for derivation of the analytical solution. The analytical solution is compared to numerical simulation results with example cases. Its accuracy is also assessed and discussed for confined and unconfined conditions.
H23G-1733
Effect of Aperture Field Variability, Flow Rate, and Ionic Strength on Colloid Transport in Single Fractures: Laboratory-Scale Experiments and Numerical Simulation
A good understanding of the physico-chemical processes (i.e., advection, dispersion, attachment/detachment, straining, sedimentation etc.) governing colloid transport in fractured media is imperative in order to develop appropriate bioremediation and/or bioaugmentation strategies for contaminated fractured aquifers, form management plans for groundwater resources to prevent pathogen contamination, and identify suitable radioactive waste disposal sites. However, research in this field is still in its infancy due to the complex heterogeneous nature of fractured media and the resulting difficulty in characterizing this media. The goal of this research is to investigate the effects of aperture field variability, flow rate and ionic strength on colloid transport processes in well characterized single fractures. A combination of laboratory-scale experiments, numerical simulations, and imaging techniques were employed to achieve this goal. Transparent replicas were cast from natural rock fractures, and a light transmission technique was employed to measure their aperture fields directly. The surface properties of the synthetic fractures were characterized by measuring the zeta-potential under different ionic strengths. A 33 (3 increased to the power of 3) factorial experiment was implemented to investigate the influence of aperture field variability, flow rate, and ionic strength on different colloid transport processes in the laboratory-scale fractures, specifically dispersion and attachment/detachment. A fluorescent stain technique was employed to photograph the colloid transport processes, and an analytical solution to the one-dimensional transport equation was fit to the colloid breakthrough curves to calculate the average transport velocity, dispersion coefficient, and attachment/detachment coefficient. The Reynolds equation was solved to obtain the flow field in the measured aperture fields, and the random walk particle tracking technique was employed to model the colloid transport experiments. The images clearly show the development of preferential pathways for colloid transport in the different aperture fields and under different flow conditions. Additionally, a correlation between colloid deposition and fracture wall topography was identified. This presentation will demonstrate (1) differential transport between colloid and solute in single fractures, and the relationship between differential transport and aperture field statistics; (2) the relationship between the colloid dispersion coefficient and aperture field statistics; and (3) the relationship between attachment/detachment, aperture field statistics, fracture wall topography, flow rate, and ionic strength. In addition, this presentation will provide insight into the application of the random walk particle tracking technique for modeling colloid transport in variable-aperture fractures.