H33M-01
Flow versus transport connectivity: What can we infer from pumping and tracer tests
Description of solute transport connectivity between different aquifer locations is one of the major challenges for making reliable transport predictions in risk analysis. For instance, accurate representation of the pathway that connects the contamination plume with a given exposure location will largely dominate model predictions. Unfortunately, most aquifer characterization tools provide information on heterogeneity at different scales but little direct information on how well connected are two different locations of the aquifer with respect to transport. Pumping tests provide direct information on flow connectivity as it largely affects the aquifer response. Nonetheless, this information is seldom used for characterizing transport connectivity partially because of the lack of theoretical basis. In this presentation, we address the problem of the joint characterization of connectivity through pumping and tracer tests at a given location. To achive this, we present a general theoretical framework with companion numerical simulations that explains the relationship between the head response in pumping tests and transport connectivity.
H33M-02
A data domain correlation approach for joint inversion of time-lapse head, concentration, and electrical resistivity data
Timelapse electrical resistivity tomography (ERT) has been used extensively to monitor fluid movement in the subsurface. The underlying basis of this method is that changes in fluid conductivity caused by the movement of conductive (or resistive) fluids result in a change in the subsurface electrical potential distribution measured during the ERT experiment. As fluid transport is controlled in large part by hydrogeologic properties such as porosity and hydraulic conductivity, time-lapse ERT data can (theoretically) be used to constrain inverse estimates of these properties. The primary obstacle inhibiting the use of ERT data directly for inverse estimates of hydrogeologic properties is the lack of a generally applicable transform relating bulk conductivity to fluid conductivity. In general, while such transforms exist, they will be formation dependent. In order to eliminate the necessity of a petrophysical transform in estimating hydrogeologic properties from timelapse ERT data we have developed an approach which utilizes a proxy predicted transient change in electrical conductivity, which is highly correlated to observed transient changes in electrical potential. This correlation is based upon the correlation between changes in fluid conductivity and changes in bulk conductivity. We formulate the inversion to find a hydraulic conductivity distribution that maximizes the correlation between this proxy predicted data and the observed potential data (in addition to hydraulic head and fluid conductivity measurements). Thus, our joint inversion is formulated to estimate hydraulic conductivity conditional to head, fluid concentration, and resistivity measurements, where the resistivity term of the objective function is based on maximizing the correlation between the proxy and observed ERT data. In our paper we will describe the proxy ERT data formulation and demonstrate the correlation between the proxy and observed ERT data. We will also present synthetic examples of our approach.
H33M-03
Characterization of Fractured Reservoirs Using a Combination of Pressure and Self-potential Transient Data
Fractured geothermal reservoirs are often represented computationally as "MINC" double-porosity media. In such reservoir descriptions, global mass exchange between adjacent macroscopic computational grid blocks takes place mainly through the "fracture zone". Inter-block flow through the "matrix region" is relatively unimportant and is usually neglected in most MINC treatments. But this approximation is inappropriate when calculating the global "drag current" caused by electric charges moving with the flowing fluid due to electrokinetic coupling. Since the magnitude of the drag current density is proportional not to the permeability but to the porosity of the medium, the contribution of the drag current through the matrix region to the total global current between adjacent macroscopic grid blocks is not negligible, and in fact usually predominates under steady-state conditions. This property of the drag current brings about much more pronounced differences in the "self-potential transients" between competing "fractured/MINC" and "porous-medium" descriptions of the same reservoir than is the case for pressure transients. Combining continuous pressure and self-potential measurements may therefore provide a means for better characterizing fractured reservoirs. Probable methods to measure the transients of "macroscopic" electric potential, which reflects the contribution of the drag current through the matrix region, will also be discussed.
H33M-04
Combined Hydraulic Tomography - Self Potential - Electrical Resistivity Tomography for Parameter Estimation in the Unconfined Aquifer at the Boise Hydrogeophysical Research Site
A series of combined hydrogeophysical tests were conducted during a two-week period in June 2007 in the unconfined coarse fluvial aquifer at the Boise Hydrogeophysical Research Site (BHRS). The tests included three component experiments -- hydraulic tomography (HT), self potential measurements (SP), and electrical resistivity tomography (ERT) -- which will be used, both individually and jointly, to image subsurface heterogeneity. The major aim of the tests is the estimation of the hydraulic conductivity (K) distribution from hydrologic responses and from SP anomalies under both transient and steady state pumping conditions. The ERT results provide independent information both for the electrical conductivity distribution (which may correlate with subsurface hydrologic parameters) and for the use of this distribution in the SP analysis. For HT, the dipole (pumping and injection, conservative) test configuration was used to achieve steady state as rapidly as possible and also to generate two high-gradient disturbance zones per test, rather than one as in conventional pumping tests. The investigated volume is the central region of the BHRS which is roughly 30m x 30m x 20m and is known to have laterally varying thickness-averaged K variation as wells as layered and patchy heterogeneity. This area was instrumented with offset SP and ERT electrode grids at the surface, transducers in wells, and also SP electrodes in seven wells including the pumping and injection wells; river boundary conditions also were monitored. Preliminary reviews of raw data by respective method teams suggest (pseudo) hydrologic steady state was reached at most observation wells, SP anomalies are clearly observable (but asymmetrical for most tests), and ERT data are repeatable and capture water table drawdown and buildup anomalies. In addition, stage variations in the Boise River adjacent to the BHRS propagated throughout the site and were observed at wells. Although flow rates in the river are controlled, aquifer responses to changes in river stage are analogous to natural stream- aquifer interactions, and the magnitude and timing of these responses at wells can provide additional information for inverse modeling. http://cgiss.boisestate.edu/~billc/EPA3/ht-sp.html
H33M-05
Adaptive Management of Remediation Systems Under Uncertain Hydraulic Conductivity and Plume Distribution
The optimal design and the management of pump-and-treat (PAT) remediation systems is generally tackled with the aid of combined simulation-optimization models to rank alternatives while considering management objectives and constraints. Since this process is typically carried out in an environment of uncertainties, our ability to determine cleanup policies that are cost optimal and reliable at the same time is in fact limited. In this work, we present a stochastic optimal control framework for assisting the management of the PAT cleanup of polluted shallow aquifers. Hydraulic conductivity distribution and dissolved contaminant plume location are considered as the uncertain parameters. The framework considers the subdivision of the cleanup horizon in a sequence of stress periods over which the pumping policy implemented at each stage is dynamically adjusted based on new information that has become available in the previous stages. In particular, we study the idea of monitoring the cumulative contaminant mass extracted from the installed recovery wells, and using these measurements to generate conditional realizations of the hydraulic conductivity field. These realizations are thus used to obtain a more accurate evaluation of the initial plume distribution, and modify accordingly the design of the PAT system for the remainder of the remedial process. The study indicates that measurements of contaminant mass extracted from pumping wells retain valuable information about the plume location and the spatial heterogeneity characterizing the hydraulic conductivity field. However, such an information may prove quite soft, particularly in the instances where recovery wells are installed in regions where contaminant concentration is low or zero. On the other hand, integrated solute mass meausurements may effectively allow for reducing parameter uncertainty and identifying the plume distribution if more recovery wells are available, in particular in the early stages of the cleanup process.
H33M-06
Using Transient Temperature-Depth Data to Constrain Groundwater Flow Velocities in Thermal Aquifers of Long Valley Caldera, California
Temperature-depth profiles are a fundamental tool for studying groundwater flow in porous media and fracture zones. They have been used to measure vertical and horizontal groundwater flow velocities in zones of significant permeability, ascertain changes in mean annual surface temperature, estimate geothermal energy potential, and measure magma emplacement rates and eruption hazard. The vast majority of studies that rely on temperature- depth profiles assume steady-state temperature and groundwater flow conditions. Here, we present an analysis of temperature data from two hot water wells (maximum temperatures 100°C and 129°C) in Long Valley Caldera. Ten resistance temperature devices (RTDs) installed in each well continuously measure temperature at discrete depths, from 24m to 76m in one well and from 67m to 200m in the other. A previously developed model (S. Ge, JVGR 1998) using steady-state temperature-depth profiles in fracture zones to constrain groundwater flow velocity was employed. It assumes a characteristic linear conductive profile outside thermal fluid flow zones and a concave profile within. Here, a modified approach is applied to profiles from the two study wells, which allows for deviation from the characteristic form as well as incorporation of observed transient temperature fluctuations. These variations range from approximately 0.5°C on a daily scale to 1°C on a scale of weeks to months.
H33M-07
Nearest neighbor classification for facies delineation
Geostatistics have become the dominant tool for probabilistic estimation of properties of heterogeneous formations at points where data are not available. Ordinary kriging, the starting point in development of other geostatistical techniques, has a number of serious limitations, chief among which is the intrinsic hypothesis of the (second order) stationarity of the underlying random field. Attempts to overcome this limitation have led to the development of ever more complex flavors of kriging. We pursue an opposite strategy that consists of finding the simplest possible technique that is adequate for the task of facies delineation. Guided by the principle of parsimony, we identify Nearest Neighbor classification (NNC) as a viable alternative to geostatistics among deterministic techniques. We demonstrate that when used for the purpose of facies delineation, the NNC, which has no fitting parameters and operational assumptions, outperforms indicator kriging, which has several parameters.
H33M-08
Hydrostructural Characterization of Fracture Networks
Over the past 30 years, research in underground laboratories for radioactive waste has led to the development of integrated site investigation and modeling methods for fracture networks. These activities began with the Stripa Project in central Sweden from 1977 to 1992 and have continued worldwide. Experiments on the scale of 100- 200 meter blocks have demonstrated the effectiveness of integrating testing during drilling, pressure monitoring, geologic description, flow logging, pressure transient testing, and groundwater chemistry to define fracture network geometries, particularly with respect to the identification of major features, background fractures, and compartmentalization. Major features are those large fractures or fracture zones that control the flow at the scale of interest, and must be simulated as deterministic features. Background fractures are defined stochastically, and provide connectivity between deterministic features. Based on the experience of block-scale investigations, it is possible to develop a clear picture of hydraulic networks using an integrated structural geologic, hydraulic, and hydrochemical approach. Although fracture network characterization requires a good geologic description of fractures and fracture zones from core and image logging, not all geologic features are water-conducting. Identifying water-conducting fractures begins with measurements of flow during drilling and flow logging immediately afterwards to identify significant conducting features. Major flow features must be hydraulically isolated using multiple point piezometer systems, if subsequent investigation methods are to be successful. Once installed, the pressure responses in the piezometers to subsequent drilling provide key information on connectivity and compartmentalization. Generally with three holes are sufficient to develop initial conceptual models of the major, controlling features. Subsequent boreholes test these geometric hypotheses and provide bases for their refinement. Confirmation of the network geometries comes from pressure transient testing. Although flow logging is very useful for locating conductive fractures, their steady-flow analysis methods do not allow for skin effects or accurate determination of hydraulic properties. Transient tests using single hole and interference data provide accurate hydraulic properties of the networks as well as insights to the network geometry and its boundaries. Geochemical data provide further constraints on connectivity and compartmentalization. The hydro-structural characterization of block scale experiments at the Äspö Hard Rock Laboratory provides an example of this integration approach.