H23A-1101 1340h
Estimating field-scale soil hydraulic properties and petrophysical models through joint GPR/hydrological measurement inversion
As ground-penetrating radar (GPR) travel times are highly sensitive to transient and non-uniform water distributions, they are potentially quite useful for inferring soil hydraulic parameters. In this research, multiple-offset cross-borehole GPR travel times are used jointly with additional hydrological measurements to estimate field-scale soil hydraulic parameters through inversion. Our approach allows for estimation of 1) the soil hydraulic parameters, 2) the parameters describing the petrophysical model (the constitutive model relating the dielectric constant to the porosity and water saturation), and 3) spatial correlation model parameters of the permeability field. A synthetic example involving the point injection of water and the simultaneous collection of nearby borehole neutron probe and GPR measurements is first considered to examine the impact of inaccurate petrophysical models on the estimation of soil hydraulic parameters. Errors can be introduced when applying a petrophysical model to a situation with conditions different from those for which the model was derived (e.g., when non-site-specific or laboratory-scale petrophysical models are applied to field-scale measurements). Our synthetic study suggests that small errors in the petrophysical model cause substantial errors in the soil hydraulic parameter estimates. However, we show that these errors may be overcome through joint estimation of the petrophysical model itself and the soil hydraulic parameters. Finally, the approach is applied to a GPR-neutron probe data set collected at the Hanford DOE site in Washington, allowing us to draw conclusions regarding the strengths and weaknesses of the approach in a real-world, 3-D setting. This work was supported in part by the U.S. Dept. of Energy under Contract No. DE-AC03-76SF00098.
H23A-1102 1340h
Effects of Finite Antenna Length on Crosshole GPR Tomography
Over the past decade, crosshole ground-penetrating radar (GPR) tomography has become an important tool for the estimation of subsurface moisture content. In theory, in order to produce the highest resolution images possible with this technique, rays covering a wide range of angles between the boreholes are necessary. In practice, however, including high-angle ray data in crosshole GPR inversions often results in poor data fitting and tomographic images with obvious errors. The reasons for this discrepancy between theory and practice have not been adequately addressed in the literature, and usually high-angle rays are discarded in crosshole GPR inversions to avoid problems. Unfortunately, this leads to tomograms with poor horizontal resolution. Here, we investigate whether the failure to account for the finite length of borehole radar antennas plays some role in the difficulties encountered with high-angle ray data. Specifically, we look at whether, for large vertical offsets between the radar antennas, energy traveling between the antenna tips sometimes arrives before that traveling between their centers. This will result in geometrical inversion artifacts because standard inversion algorithms assume that all first arrival energy travels between the antenna centers. Through numerical modeling, we have found that, in both vadose and saturated zone environments, the velocity of a current pulse along a borehole GPR antenna can be significantly faster than the velocity through the earth between the boreholes. This results because the borehole antenna wire is not embedded directly in the earth, but rather surrounded immediately by the antenna insulation and borehole filling material. Consequently, for high source-receiver angles, energy propagating up the transmitter antenna, across to the receiver antenna tip, and up the receiver antenna, can sometimes arrive before that propagating directly between the antenna centers. Using our modeling results, we have simulated crosshole GPR with realistic, finite length antennas, and subsequently inverted the synthetic data with the standard assumption that rays join the centers of the antennas. We have found that not accounting for coupling between the tips of our borehole antennas can lead to significant errors in the resulting tomographic images of moisture content. We are currently working on an inversion algorithm that allows for the successful incorporation of high-angle rays into crosshole GPR inversions, and thus higher resolution moisture content estimates.
H23A-1103 1340h
Ground Penetrating Radar Response to thin layers: Example from Waites Island, South Carolina
Ground penetrating radar is commonly utilized in stratigraphic interpretation of depositional environments. In many settings, the thickness of layers (mm scale) is much smaller than radar wavelengths (tens of cms). The presence of thin layers in the shallow subsurface can pose difficulty for identification of reflecting surfaces in the radar profiles. A sequence of such layers will result in reflection patterns that are the result of complex interference of the reflecting radar energy. No single reflection event can be assigned to each reflecting contact. In order to better understand radar wave behavior in response to thin layer patterns characteristic of sedimentary environments, we have developed numerical models to simulate GPR reflections at frequencies between 50MHz and 500MHz. Several suites of FDTD (finite-difference-time-domain) models have been run with different possible layer patterns, simulating thin layers ranging in thickness from 1 mm to ~Íš 1/4 or 1/3 of the radar wavelength. Models are also compared against GPR surveys carried out at Waites Island, South Carolina. Sediment cores from this barrier island contain some good examples of thin layers in the form of beach laminations. On Waites Island, the presence of magnetite as a constituent heavy mineral in the mm-thick laminations clearly plays a significant role in causing radar reflections.
H23A-1104 1340h
What are the Physical Causes of GPR Reflections in an Oolitic Carbonate Environment? Excavation, TDR, and Modeling Results
Ground Penetrating Radar (GPR) can produce sub-meter resolution 3D images of internal anatomy and moving water in the Miami Oolitic Limestones. However, the exact physical causes of GPR responses are still unclear. In order to reach our long-term goal of extracting hydrologically relevant and quantitative parameter volumes from 3D and 4D GPR data we have initiated an integrated study involving full-resolution 3D GPR imaging, excavation, fresh outcrop inspection, sample analysis, TDR measurements and synthetic GPR modeling. Just before the excavation for a multistory building began, we acquired a full-resolution 3D GPR survey covering an area of 16 m x 14 m with a 0.05 m x 0.10 m grid using 250 MHz antennae, reaching the watertable at a depth of 4.5 m. The entire rock volume of 1000 cubic meters was subsequently excavated, allowing the exceptional opportunity to directly compare 3D radar data with a 3D rock cube. Qualitative comparisons between the 3D GPR results and the excavated rock faces show that the technique is valid; the GPR shows the main stratigraphic boundaries seen in the real rock. However, some reflections and attenuated zones in the radar data do not fit with any visible structure. To help explain these discrepancies we took large (0.05 cubic meter) samples from within the rock volume that corresponded with either an anomaly location or a known stratigraphic reflector, and analyze their petrophysical characteristics under laboratory conditions. The samples were cut into cubes and tested at a range of saturations from dry (0%) to fully saturated (100%) using 10cm long Time Domain Reflectometry (TDR) probes, which provide accurate values of dielectric constant and allow us to investigate its relationship with water saturation. We find that the relationship between dielectric constant and water saturation depends on sample stratigraphy, porosity, mineralogy, and permeability. This result suggests that each different oolite sample produces a characteristic GPR response. To upscale these results from the TDR scale (0.002 cubic meter) to 3D GPR volume (1000 cubic meter), we use two synthetic models that are compared with field data. One of the models shows how GPR can detect and quantify water saturation and fluid flow. This finding indicates that we can estimate water saturation and fluid flow from 4D GPR data. The second model indicates that sink holes, a common feature in this geological environment, can be detected in GPR surveys as attenuated data surrounded with time-shift anomalies. We compare these models with the original 3D GPR results to help us to better interpret these data. The results from this study will be used to design future 4D GPR experiments, which will enable us to more fully understand the detailed structure and quantify the hydogeological behavior of oolitic carbonates.
H23A-1105 1340h
Using Borehole Ground Penetrating Radar to Monitor Transient Flow during Pumping of an Unconfined Aquifer
The contribution of the unsaturated zone to the response of an unconfined aquifer to pumping remains an area of active debate and research in hydrology. It is generally accepted that water is released from storage above the water table relatively slowly and that this delayed drainage contributes to the unique shape of pumping test responses. However, the magnitude of this contribution is in question. Furthermore, it is not clear whether moisture release curves determined under static conditions can be used to characterize the vadose zone response. In this investigation, we use borehole ground penetrating radar (BGPR) to monitor the volumetric water content throughout the vadose zone and into the shallow saturated zone during pumping in an unconfined aquifer. We show that BGPR is uniquely capable of providing water content profiles with sufficient depth and temporal resolution to characterize the vadose zone response under these transient conditions. The measurements quantify both the rate and timing of delayed drainage. In addition, the results indicate that the moisture retention characteristics, measured in the subsurface before pumping, do not adequately describe the transient response of the vadose zone. The results of this study are a promising demonstration of the current and future utility of geophysical methods to help address long standing questions in the field of hydrology.
H23A-1106 1340h
Using Relative Gravity Measurements to Monitor Transient Infiltration
Ground based microgravimetry has shown great promise for monitoring changes in water storage in deep vadose zones. The ability of the method to make noninvasive, nondestructive measurements of subsurface water storage makes it ideal for watershed-scale hydrologic monitoring. However, in many environments with deep vadose zones, areas associated with large changes in water storage are also associated with large changes in water content throughout the vadose zone. This presents a potential limitation to gravity methods due to the spatial sensitivity of the methods, which decreases as the inverse depth of the location of water content change squared. In this study, we examine the effects of changes in water storage in the vadose zone on the response of surface based gravimeters. Specifically, we model the response of gravimeters to infiltration and drainage beneath an ephemeral stream. We show that without proper analysis, changes in water storage in the vadose zone can cause very large errors in the interpretation of gravity measurements. However, with more complete analysis and higher temporal resolution of the gravity signal than is commonly acquired, gravity measurements can be corrected and can even be used to characterize effective unsaturated flow parameters. The results of this study point to the importance of combining geophysical and hydrologic understanding in the use and interpretation of hydrogeophysical methods.
H23A-1107 1340h
Gravity Monitoring of the Weber River Aquifer Storage Project
Repeated precision gravity measurements provide an economical way to track aquifer storage changes through time. In early 2004, the Weber River Water Conservancy District in northern Utah began an aquifer storage and recovery pilot project by infiltrating river water into a depleted aquifer. We are tracking the infiltrated water by measuring gravity changes over the aquifer through time. A network of 28 stations around the infiltration location was established, with an additional station in the nearby mountains for a stable reference. Gravity surveys are conducted at approximately two week intervals; monthly rapid-static GPS campaigns monitor ground deformation across the network. Gravity monitoring commenced in Feburary 2004, to establish a baseline before infiltration and investigate the magnitudes of natural signals and measurment noise. Infiltration commenced six weeks after the start of monitoring and by early July 2004, nearly 750 000 m$^3$ of water were infiltrated; gravity changes at the infiltration site reached a peak of $\sim$100 $\mu$Gal. Gaussian integration of the peak gravity signal is consistent with the total volume of infiltrated water. Continued monitoring during infiltration tracked the horizontal migration of infiltration water south and west of the site, consistent with known hydraulic gradients. Infiltration ended in July 2004 and gravity measurements show a declining recharge mound, with the peak decreasing to $\sim$60 $\mu$Gal one month later. The spatial and temporal changes in gravity will be used to refine and enhance reservoir modeling around the infiltration site.
http://thermal.gg.utah.edu/~gettings
H23A-1108 1340h
Gravity Monitoring of Ground-Water Storage Change in the Southwestern United States
Repeat measurements of absolute gravity have been made since 1998 to estimate changes in ground-water mass as part of ground-water budget estimates in arid and semiarid regions of the Southwestern United States. The absolute acceleration of gravity is measured twice each year at 16 stations to an accuracy of about plus or minus 2 microGal, or about 5 cm of water. Observations are normally done for the purpose of providing gravity control for relative gravity surveys of networks of stations across wider areas. Other data incorporated into the ground-water budget estimates include precipitation, water levels, moisture content in the unsaturated zone, surface water runoff, and ellipsoid heights using the Global Positioning System (GPS). Gravity and water-level changes are correlated for stations measured in the Basin and Range Physiographic Province near Tucson, Phoenix, Casa Grande, and Sierra Vista, Arizona. Decreasing gravity and water levels in the Tucson area since the summer of 1998 are likely related to predominant drought conditions and decreases in ground-water storage following above average winter precipitation and recharge during the El Nino of 1998. Increases in gravity at stations in the upper and middle Verde Valley Watershed in central Arizona since the fall of 2000 do not correlate well with declining streamflows and water levels and may be caused by temporary increases in soil moisture following wet winters. There have been no significant observed gravity changes at two stations in the El Paso, Texas, area since the initial observations during the summer of 2003, even though ground-water pumping in the area has been heavy.
H23A-1109 1340h
Quantifying Saturation-Dependent Anisotropy in Soil Hydraulic Conductivity
The anisotropy in unsaturated hydraulic conductivity is saturation-dependent. Accurate characterization of soil anisotropy is very important in simulating flow and contaminants (e.g., radioactive nuclides in Hanford) transport. A recently proposed tensorial connectivity-tortuosity (TCT) concept describes the hydraulic conductivity tensor of the unsaturated anisotropic soils as the product of a scalar variable, the symmetric connectivity tortuosity tensor, and the hydraulic conductivity tensor at saturation. In this study, the TCT model is used to quantify soil anisotropy in unsaturated hydraulic conductivity. The results show that the anisotropy coefficient, A, is independent of soil water retention properties. At a certain saturation, A can be characterized by the ratio of the saturated hydraulic conductivities and the difference in the tortuosity-connectivity coefficients in orthogonal directions. The model was tested using directional measurements of unsaturated hydraulic conductivity of undisturbed soil cores. Results show that the TCT model can describe different types of soil anisotropy, previously ignored in other models. The TCT model can be used to describe either monotonic increases or decreases in A with saturation and allows the principal direction of hydraulic conductivity to rotate when saturation varies. The Pacific Northwest National Laboratory is operated for the U.S. Department of Energy by Battelle under Contract DE-AC06-76RL01830.
H23A-1110 1340h
Hydrostratigraphic Characterization Using High-resolution Borehole Moisture Logs and Grain-size Distribution Statistics
Small-scale heterogeneities typical of natural soils are known to impact field scale flow. Model predictions of flow and transport under such extreme variations in hydraulic and geochemical properties often show large discrepancies from field observations, perhaps because small-scale variations are often ignored. In this study, we develop pedotransfer functions, based entirely on the statistics of the particle-size distribution, to predict hydraulic and geochemical properties at multiple scales. Grain-size statistics, namely mean diameter, d$_{g}$, sorting index, S$_{o}$, and the Fredle index, F$_{i}$, are derived from samples taken on a coarse sampling interval. A relationship between the size-statistics and specific retention is coupled with high-resolution neutron moisture logs to predict size statistics on intervals as small as 3 inches. Hydraulic and geochemical properties, reflecting the small-scale heterogeneity, are then generated using the pedotransfer functions. Flow and transport of reactive contaminants using the STOMP simulator show very good agreement with observations from a waste site at Hanford. Hydrostratigraphic models ignoring the small scale heterogeneity were unable to match field observations. The Pacific Northwest National Laboratory is operated for the U.S. Department of Energy by Battelle under Contract DE-AC06-76RL01830.
H23A-1111 1340h
Characterization of Physical and Hydro-Geological Properties of Kanamaru Research Site in Japan
Establishing the comprehensive knowledge of applicability of the methods for investigating hydraulic properties of low permeability geologic strata is an urgent issue for supporting regulation of geological disposal of nuclear waste in the near future. As a beginning of this work, a systematic examination of various kinds of techniques for hydro-geological surveys has been started in Kanamaru Research Site in Japan. This paper briefly introduces the research plan and preliminary results obtained from the first year of investigation. The survey techniques include borehole excavation, borehole imaging, gamma-ray, caliper, acoustic, electrical resistivity and density loggings, permeability tests and flow direction measurement using a single borehole, permeability tests and flow direction measurement using multi boreholes, etc. Preliminary findings can be summarized as follows: (1) The stratigraphy at the survey area consists of topsoil, debris sediments, sandstone, mudstone, conglomeratic sandstone, mudstone, arkose sandstone, and granite. High uranium concentrations are detected at lower portion of the conglomeratic sandstone by gamma-ray logging. (2) The survey area is located at a slope inclined from the north to the south, and the dominant groundwater flow is considered to be in the direction form the north to the south. And the downward flow was also recognized by the flow direction measurements and water quality logging. (3) Hydraulic conductivities derived from permeability tests using a single borehole were in the range of 5E-10 to 1E-7 m/s. The hydraulic conductivities of the same lithology derived from different boreholes varied, and the discrepancies were up to an order. This result indicates that the formations in the survey area have hydraulic heterogeneity in both the vertical and horizontal directions. (4) On the whole, stratum with fast velocity of elastic wave showed large resistivity and low permeability. The degree of correlation between the hydraulic and physical properties was different for each stratum.
H23A-1112 1340h
Quantification of Meso-Scale Porosity From Borehole Wall Images. Example From the ALIANCE Campos Experimental Site, Mallorca, Spain.
The porosity distribution in carbonates from reefal origin is controlled both by primary and diagenetic processes (fracturing, dolomitisation and dissolution). The study and characterisation of porosity in reservoirs therefore requires a multi-scale and multidisciplinary approach. The analysis of mm-scale digital borehole wall images provides a reliable and practical method to identify mesoscale (vuggy to karstic) porosity and/or fracture porosity, and a new method providing a quantitative analysis of such images is presented here. This new method is based on the analysis of borehole acoustic and optical images. These images have been collected from the MC2 corehole that penetrates the Miocene reefal carbonates of the Campos Coastal site (Mallorca, Baleares, Spain). These sites were drilled in the framework of the Advance Logging Investigations of Aquifers iN Coastal Environment (ALIANCE) European Community Program aimed at developing new techniques to investigate salt-water intrusion in coastal regions. This new method (i) is insensitive to potential eccentering bias of the borehole image logging tools, (ii) analyses vugs and background based on the greyscale colour of each digital images and (iii) calculates the percentage of vuggy/fracture porosity and its attributes in terms of pores size, shape and orientation distributions. Experimentation and comparison with classical porosity determinations (triple weight and Mercury (Hg) porosity) on core, image analysis of slabbed-core samples and thin sections showed that a simple greyscale threshold would not realistically differentiate between greyscale contrast of vugs and background. A moving average calculated around the circumference of the borehole and expressed as the background greyscale intensity is identified as a baseline from which to identify a local greyscale threshold for vugs. Once identified, the characterization of the pores intersected by the borehole as a function of depth and azimuth allow (i) to correlate porosity with lithology (natural gamma ray, velocity and electrical conductivity) as well as hydrological profiles (fluid temperature, pH, dissolved 02), (ii) to characterize lithological units, (iii) to identify the diagenetic processes leading to the present transport properties of the structure and (iv) to evaluate the impact of present brine intrusion several km into the island.
H23A-1113 1340h
Detecting Seepage Through a Natural Moraine Dam Using the Self-Potential Method
We conducted a self-potential (SP) survey of the Dana Lake moraine dam to investigate seepage flow through a natural moraine. Seepage erosion, along with overtopping, are the two primary mechanisms that cause moraine dams to fail. Specifically, localized seepage paths can entrain dam matrix material and may lead to undermining and partial collapse. Our survey included 200+ self-potential measurements spaced between 2 to 10 meters apart. The Dana Lake moraine dam is located in the Sierra Nevada, California. The moraine is likely between 4000 and 11,000 years old, and impounds approximately 3 x 10$^{5}$ m$^{3}$ of water. The moraine also has two small lakes (unofficially East and West lakes) at its downstream toe, which are fed in part by seepage from Dana Lake. Intriguingly, West Lake is primarily sand and silt bedded while gravel and boulders dominate the bed of East Lake, despite them being ~40 meters apart. We concentrated our study on the west side of the moraine dam because 1) we observed active seepage flow on that side and 2) the lake bed material difference between East and West lakes suggests seepage may be removing fines from the western side of the moraine. Preliminary analysis of our self-potential data indicates nonuniform seepage flow through the moraine. On the western side, a broad positive SP anomaly is spatially coincident with observed seepage outlets and a sub-moraine bedrock knob visible at the surface. Additionally, a large negative anomaly coincides with the edge of the bedrock knob and may represent a vertical seepage path. On the eastern side of the moraine, groundwater flow appears to originate at the lake margin and follows an arcuate path towards West Lake. Other localized positive SP anomalies indicate possible seepage outlets. Future work includes forward modeling to resolve seepage flow rates and resulting stresses for use in a full seepage erosion analysis of the moraine dam.
H23A-1114 1340h
Relationship between spatial distribution of the streamwater chemistry and catchment-scale geomorphic characteristics: Conceptual model approach focusing on subsurface and groundwater interaction
In order to clarify the mechanisms determining the spatial distributions of streamwater chemistry, a conceptual model explaining the relationship between spatial variability of geochemical solute concentration among the streams and subsurface-groundwater mixing in each subcatchment was proposed. The model assumes that the stream solute concentration is controlled by mixing ratio of subsurface and groundwaters. The variability of solute concentration among subcatchments is potentially generated by difference in the ratio of catchment size of subsurface and groundwaters. The model was applied to the field data taken from two watershed having different geologic settings (granitic and sedimentary bedrock) in central Japan. Variability in geochemical solute concentrations was largest among smallest subcatchments in the most headwaters, and decreased with increasing of subcatchment size. Subsequently the solute concentrations converged into a constant value at the specific catchment scale. Two watersheds have a different converging subcatchment scale. Simulations based on these two field cases suggest that the size of geomorphologically minimum subcatchments (zero-order catchments) at the most headwaters and the variability of solute concentrations among these subcatchments are key parameters determining the size of the converging catchments. Both tow key parameters were larger in the sedimentary rock watershed than the granitic rock watershed, and the size of converging subcatchment was also larger in the sedimentary watershed. It is considered that the difference in geological setting affects the difference originally in the geomorphological characteristics such as valley shape and size of minimum subcatchments at the most headwaters. Consequently, this difference causes the variation in the size of converging subcatchments. The size of converging subcatchment can be treated as an expression of the representative elementary area (REA). The field data suggest the possibility for identifying REA by measuring the geochemical solute concentrations along the stream orders. Moreover, the model approach in this study imply that the determining mechanisms of REA can be generalized by 1) the size of minimum subcatchments at the most headwaters, 2) the variability of geochemical solute concentration among minimum subcatchments, and 3) size of converging subcatchments.
http://www.bluemoon.kais.kyoto-u.ac.jp/site-eg.html
H23A-1115 1340h
Comprehensive Geophysical Investigation over a Former Radioactive Waste Site, Hanford, Washington
Several geophysical methods were combined to characterize a technetium-99 (Tc-99) plume beneath a former radioactive waste site, including a magnetic gradiometry survey, a broadband multi-frequency electromagnetic survey and a high resolution resistivity (HRR) survey. The 50 acre site was previously used for the disposal of fission products formed during uranium processing, where high volumes of liquid mixed waste were discharged to unlined trenches over a two year period from 1956 to 1958. The magnetic and electromagnetic surveys identified buried infrastructure used during the construction of the disposal site. The HRR survey, which consisted of electrical resistivity measurements along transects parallel and perpendicular to a trench in the southern portion of the site, identified an electrically conductive plume coincident with Tc-99 data obtained from a nearby borehole. The data were used to confirm assumptions used in an unsaturated flow model, including recharge and hydraulic properties.
H23A-1116 1340h
Combining Core and Geophysical Data Using Spatial Statistics to Infer the Permeability Structure of Glacial Sediments
Geophysical surveys, such as resistivity and electromagnetic surveys, are a valuable tool in characterization of the permeability structure of aquifer confining layers for the purpose of pollution vulnerability assessment. However, these geophysical techniques suffer from a range of limitations associated with sampling volume or data inversion process. For example, when characterizing glacial sequences comprising sands and clays, resistivity inversions often mis-place the depths of layer boundaries, and fail to relatively detect thin layers at depth, or those below conductive overburden. These problems seriously compromise our ability to generate accurate permeability fields directly from geophysical data. We present an approach aimed at overcoming these limitations by combining geophysical data with core data using co-kriging. Normally, geophysical datasets are accompanied by relatively sparse direct geological data (e.g. borehole logs). These data can be used to identify where geophysical datasets fail to adequately reflect geological structure; they can also identify areas where the structure is more adequately reflected in the geophysical data. Here, we present an attempt to use the spatial structure of areas of a glacial cover sequence where good geophysical data are available to interpolate geological structure between sparse boreholes. The spatial structure of the geophysical data is first characterized using semi-variograms; co-kriging of the spatially dense geophysical and sparse borehole data enables the interpolation of a more complete geological cross section. The permeability fields produced by this approach are compared with those generated by direct conversion of resistivity profiling data, in order to ascertain the significance of the proposed approach for groundwater vulnerability assessment.
H23A-1117 1340h
Using Tracer Test Data To Estimate Field Scale Correlation Length of Permeability
The spatial permeability distributions of natural media are strongly heterogeneous. In a geostatistic model, the heterogeneity is often characterized by its stochastic properties: mean, variance and correlation length. The correlation length can only be obtained in field by large amount of measurements. In many field conditions, these measurements are not available, so the correlation lengths are assumed based on limited data. In this study, a novel approach is developed to estimate the correlation length of a medium permeability from field tracer test data and limited permeability measurements. The underlying idea of this approach is that the more accurate estimation of geostatistic properties will lead to a more fast and accurate convergence of measurement and simulation results. Our result shows this approach can well capture the medium permeability's correlation length. The influence of the amount of permeability measurements on the accuracy of estimation is also discussed.
H23A-1118 1340h
Delineating Hydraulic Conductivity with Direct Push Electrical Conductivity and High-Resolution Slug Testing
Direct push technology continues to make advances in efficiently measuring water related parameters in unconsolidated sediments. A direct push subsurface profiling technique used during field investigations measures the electrical conductivity (EC) of sediments and fluid surrounding the EC probe. The EC geophysical method is typically used for gross lithologic definition. When numerous direct push EC profiles are completed, a general impression of vertical and lateral variation of subsurface lithology can be inferred. Unfortunately, these EC profiles do not directly measure the hydraulic conductivity, even though the profiles may indicate the presence of fine-grained material such as silt and clay, which are known to affect the hydraulic conductivity. The direct push EC vertical profiles can be obtained quickly and efficiently over an extended area. These EC profiles can be examined for regions that display a subsurface EC response with a potentially interesting behavior of the hydraulic conductivity. At selected locations, 5cm (2 inch) PVC monitoring wells with appropriate screen lengths can be installed by direct push techniques. We have developed equipment and techniques for performing high-resolution slug tests efficiently in 5cm (2 inch) wells. Correlation of the EC response and the high-resolution slug test results can aid in developing a 3-D picture of the hydraulic conductivity distribution at a given site. In this paper, we present the results of such a correlation for a well located near the Geohydrologic Experiment and Monitoring Site (GEMS) at the University of Kansas in the Kansas River valley. During the installation of this well with direct push equipment, it was discovered that the EC log indicated a prominent but relatively thin silt-clay layer at depth, which is somewhat unusual for this area. We routinely perform high-resolution slug testing efficiently over intervals as small as 7.5cm (3 inches) to 15cm (6 inches); therefore, we decided to see if high-resolution slug testing could accurately delineate this layer of apparent reduced hydraulic conductivity. Results show good correlation between the EC profile and the high-resolution slug test profile. Coupling direct push EC profiling with high-resolution slug testing may provide an efficient way to establish a detailed representation of the hydraulic conductivity distribution in a given area.
H23A-1119 1340h
Investigation of Cross-Correlations between Fracture Frequencies and Results of the Vertical Flowmeter Test
In, general, it is one of the well-known phenomena in fractured media that there is no explicit correlation between fracture densities and bulk hydraulic conductivity. However, in most cases, the meaning of fracture does not have a unique sense in hydrogeology. Fracture simultaneously means as bedding plane, tectonic joint, cooling joint, weathering joint and so on. All types of these fractures have different origins and different properties. In addition, there are a plenty of the conceptual models for fractured media with the hydrogeological sense; equivalent porous medium model, discrete model, dual porosity model, discrete percolation model, continuum percolation model, fractal model, linear programming methods and so forth. Which factors should be characterized is totally dependent on which conceptual model would be applied. The applied conceptual model is continuum percolation model in this study, and the purpose of this study is the investigation of cross-correlations between fracture frequency and the variation of vertical flow rate in borehole to check up how the REV of fractured medium can be determined in the field scale. We measured the relative vertical flow rates in two testing boreholes under the natural condition and the artificially stressed conditions, and collected the information of detectible fractures in the boreholes with a acoustic televiewer logging. And then we evaluated the change of vertical flow rates in surveyed boreholes and categorized the surveyed fractures based on their depths and orientations, and counted the frequency of fractures along with the borehole depth (1m interval) with various widths of surveyed windows, 1$\sim$10m. Using the results of frequency survey, cross-correlations between the changes of flow rate and the fracture frequency with various widths of windows were calculated. Results of this analysis show that 7m width of surveyed window has the highest cross-correlation. Even if more studies and investigations should be needed, 7m width can provide the important information to determine the REV of fractured medium in the tested site.
H23A-1120 1340h
Sled-Mounted Geophone Arrays for Near-Surface (0-4m) Seismic Profiling in Highly-attenuating Sedimentary Facies: Atchafalaya Basin Indian Bayou, Louisiana
Towed land-geophone seismic arrays have the potential to increase markedly the efficiency for collecting near-surface (0-100m) high-resolution seismic data, but viable cases are few and have been limited to a narrow range of near-surface sedimentary facies. During November 2003 through June 2004 we conducted extensive seismic tests with traditional geophones mounted on low-cost $\Pi$-shaped sleds. We targeted human habitation surfaces within the upper few meters of a crevasse splay complex in the Atchafalaya Basin study area, Indian Bayou Wildlife Management Area, Louisiana, U.S. For seismic-to-core correlation, sealed, continuous test cores were run through a multi-sensor to test for magnetic susceptibility, bulk sediment density and electrical resistivity. We compared 24-channel seismic data using a variety of seismic source-receiver combinations. Sources comprised a 12-gauge pipe-gun, a 0.22 caliber-powered piston gun, an accelerated weight drop, and a small claw hammer. Commercial blanks, 2g-black-powder, and primer-only shells were fired by the pipe gun. Receivers included 100-Hz vertical-, and 14-Hz-horizontal-component geophones. For comparison, both ground-planted and geophones mounted on wooden and iron sleds 0.3 and 1.2m long respectively. Geophones mounted on steel sleds produced data of adequate quality. Whereas traditional ground-planted geophones showed better data quality, time and cost efficiency make mounted phones more feasible for regional studies as traditional arrays are prohibitively expensive. Because of the high seismic attenuation, only horizontal-component geophones mounted on heavy (9-kg) steel sleds provided useful data, although the shallowest reflection observed in the shear wave data came from a boundary at ~ 19m depth, too far below the target depth of 4-5 m. Instead, we forward-modeled refraction traveltime data to derive the acoustic and SH velocity structure.
H23A-1121 1340h
Using High-Resolution Seismic Reflection Techniques to Image Sand Channels Within the Potomac Aquifer System in New Castle County, Delaware
Preliminary results are presented from an ongoing project designed to use high-resolution seismic reflection imaging methods to examine the stratigraphy of the Potomac aquifer system in Delaware. The Potomac Formation is a heterogeneous, non-marine, Cretaceous unit that includes important fresh water aquifers. This study aims to use seismic reflection data to characterize the sand bodies that are woven though a section predominantly composed of silts and clays. The sands are believed to be paleo-river channels, whereas the silts and clays are overbank deposits. The seismic reflection data will provide the geometrical characteristics of the sand channels, which can provide important constraints for groundwater modeling or contaminant transportation studies in the Potomac aquifer system. We are currently conducting the seismic reflection survey using a 24-channel Geometrics Strataview seismograph and a sledgehammer source. Three profiles covering a total of 1.6 kilometers have been collected to date. These seismic lines were collected using a geophone and shot spacing of 5 meters and a source offset of 30 meters. The data were collected on farms located immediately south of the Chesapeake and Delaware Canal in a region of open land area where Potomac sand channels are likely. After preliminary processing, reflections in the data can be seen as deep as 350 ms, which is possibly basement. Most coherent reflections are above 300 ms. Strong reflections correlate between profiles at both 185-200 ms and 240-260 ms, which are possible sand bodies within the Potomac aquifer system. These reflections are generally continuous, with some lateral variation. Frequency content is reasonably high, up to about 200 Hz. Continuing work will include additional lines within the study area allowing for improved delineation of the lateral extent and stratigraphic relationships of the channel intervals in three dimensions.
H23A-1122 1340h
Hydrogeophysical characterization of bedrock fracture orientations using azimuthal seismic refraction tomography
Seismic anisotropy is a commonly analyzed attribute, mainly in oil exploration and crustal studies. Anisotropy interpretations can also be useful, however, in near-surface applications (upper 200 m) to aid in defining field-scale hydrologic parameters such as bedrock fracture orientation due to the link between seismic anisotropy and fracture anisotropy. A series of seismic refraction tomography profiles were collected in July of 2002 and August of 2003 at Birch Hill on Fort Wainwright in Fairbanks, Alaska, as part of a larger environmental site characterization study to better define the geologic setting and identify structures that may influence groundwater flow and contaminant migration. These data are also used in an effort to explore the usefulness of seismic anisotropy analysis as a tool in hydrogeophysical applications. The intersections of these profiles were analyzed using two different plots (polar and Cartesian) of velocity versus the azimuth of profile orientation. When identifying seismic anisotropy (if it exists) at a particular intersection, the velocities at multiple depths in the subsurface are compared among the intersecting profiles to determine if a particular azimuth indicates higher velocities. Site-wide seismic anisotropy trends were determined by comparing all the results from multiple intersections of over 30 profiles. Interpretable results were obtained from this study. Through the analysis of the intersections, tentative fracture anisotropy trends at the Birch Hill site were obtained. Correlations exist between the known fracture orientations and the anisotropy data.
http://www.geophysics.buffalo.edu
H23A-1123 1340h
Evaluating Electric Resistance Tomographs of a Controlled Infiltration Event in a Complex Vadose Regime Using Stochastic Models
Linking field observations to field-scale models of subsurface flow and transport processes remains a challenge, particularly in the vadose zone. The LLNL Vadose Zone Observatory (VZO) experiments gathered Electrical Resistance Tomography (ERT) data to remotely monitor the infiltration process of water and brine from near-surface to an 18-m deep water table over several days. ERT observations suggested, surprisingly, that infiltration might rapidly reach the water table within only hours, which is a concern where the vadose zone has been assumed to act as a barrier to contaminants. To better understand this behavior, we are integrating several LLNL 3-D computer models (listed in parenthesis below) together to investigate the combined complexity of geologic heterogeneity (TSIM), variably saturated flow and transport (NUFT), flow of electrical current due to ERT (ParFlow), and ERT inversion (MultiBH). This integrated simulation approach, involving stochastic models, permits linkage between the ERT observations and 3-D flow and transport process models. While our simulations compare reasonably to some types of monitoring data (e.g. gypsum blocks), they do not exhibit rapid infiltration to the water table as indicated by the ERT inversions based on field data. The models indicate that strong preferential pathways, such as fracturing or vertically oriented sand lenses, may be required to explain the rapid communication between the infiltration point and the water table. Alternatively, recent bench-scale lab experiments involving resistivity measurements of plumes infiltrating a partially saturated sand-filled test section performed as part of this project (Parekh, et al., 2004, Rensselaer Polytechnic Institute) suggest the existence of air-water interfacial conduction processes strongly decreasing the formation resistivity in the unsaturated zone (see also R. Knight, Geophys., 56, p. 2139-2147, 1991) that are not simply associated with a resistivity decrease resulting from the downward migration of a brine tracer. This work was performed under the auspices of the U.S. Department of Energy by University of California, Lawrence Livermore National Laboratory under contract No. W-7405-Eng-48.
H23A-1124 1340h
Time-lapse Electrical Resistivity Anomalies due to Contaminant Transport Around Landfills: Forward Modeling Results
This paper presents theoretical geo-electrical responses associated with polluted groundwater flow in porous media around landfills. Forward computation has proven an important role in all stages of geo-electrical surveys. Previous forward studies have emphasized the space-dependence of electrical conductivity, but largely, they ignored its transient feature resulting from the spreading-out of contaminant plumes in directions transverse to the flow path as well as in the longitudinal flow direction. In this research, we adopt a novel `combined' hydrogeological and geo-electrical modeling approach as follows: first using the finite element computer package FEFLOW to simulate contaminant transport and groundwater flow; then establishing the geo-electrical structure based on an empirical relationships between contaminant concentration and water conductivity as well as Archie's law; and finally undertaking the forward computation of geo-electrical anomalies using the finite difference computer packages DCIP2D/DCIP3D. Numerical case studies are conducted for a variety of landfills with different hydrogeological and geological conditions, from which optimal mapping and monitoring configurations are determined.
H23A-1125 1340h
Improved Site Characterization Using Microbiological Community Profiles from Landfill-leachate Contaminated Groundwater and Artificial Neural Networks
Microbiological profiles and general water quality were sampled from groundwater monitoring wells surrounding a leaky municipal landfill in northeastern New York. Microbial samples were analyzed using polymerase chain reaction and gel electrophoresis, and water quality was tested for pH, temperature, redox, turbidity, and specific conductance. The bacterial profiles for each sample location were incorporated into an artificial neural network for the purpose of improving site characterization using multiple types of data. A classification artificial neural network was trained to predict the water quality based on the microbial profile by detecting relationships and structure between samples. Additional site data was used for testing and validation of the network
H23A-1126 1340h
Estimating Vertical Surface Movement using Artificial Neural Networks
Artificial Neural Networks (ANNs) are used to predict vertical surface movement when soils expand and contract with changes in soil moisture caused by climatic conditions. Temperature and rainfall data, soil property data, and soil moisture measurements are used for training ANNs to simulate the movement of spread footings at a field site in Arlington, Texas. A research team from Texas A&M University surveyed the footing movement monthly over a two-year period. The performance of the ANNs is evaluated by comparing the predictions to the observed movements. Data for temperature and rainfall are available for each month of the two-year field study, but soil moisture data (on which soil shrinking/swelling is predominantly dependent) is only available for seven of those months. Therefore, two ANNs were used in series. The first ANN estimates soil moisture for the months when data are not available, while the second ANN uses both measured and estimated soil moisture as training input to estimate vertical movement.
H23A-1127 1340h
Full waveform elastic inversion in a space frequency domain formulation: a powerful geotechnical tool for superficial reconstruction
The superficial weathered zone, few hundreds meters thick, presents high variable and complex near-surface structures. This leads to an energetic seismic ground roll and therefore hide information coming from deeper areas. Moreover near-surface anomalies may characterize potentially dangerous structures as cavities or their surrounding altered media. Therefore detecting heterogeneities in near-surface areas and quantifying their physical properties will be of great help for seismic imaging and for natural hazard assessment. Since heterogeneities are located in near-surface areas, both surface and body waves induce complex footprints in seismic data. The propagation of 2D P-SV is performed thanks to a frequency domain modeling. This frequency formulation takes into account attenuating behavior and efficiently takes benefit of multisource and multireceiver configurations. A new finite-difference stencil of second order using rotated derivatives axes (Saenger et al., 2000) simulates surface waves very precisely and remains stable nearby the free surface and/or rapidly-varying zones. It will be the forward problem kernel of our approach. We follow the matrix formalism of Pratt et al. (1998) and perform a linearized inversion in the least-square sense, since heterogeneities of reasonable amplitudes towards the surrounding medium are considered, leading us to resort to the Born approximation. We use the gradient method to perform the full waveform inversion for elastic waves. In this formulation we take only the Hessian diagonal part and use a parabolic approximation to find the stepping in the gradient direction. Our selected inversion takes into account kinematic and amplitude information for waves coming from various reflection angles at different offsets. This allows to recover local parameters as P wave and S wave velocities from dense seismic experiments. Applications to realistic synthetic configurations illustrate the potentiality of the method when both backward and forward scatterings are encountered. Influences of data sampling, data geometry and data redundancy are, of course, critical but the initial model is a very sensitive key input for successful convergence to the minimum of our misfit function, taking into account the complexity of waves interaction and propagation. From these illustrations, we highlight the importance of data introduction into the inversion tool in order to avoid non-global minimum. Saenger E. H., Gold N. & Shapiro S. A., 2000. Modeling the propagation of elastic waves using a modified finite-difference grid. Wave Motion, 31, 77-92. Pratt G., Shin C. & Hicks G.J., 1998. Gauss-Newton and full Newton methods in frequence space seismic waveform inversion. Geophys. J. Int., 133, 341-362.
H23A-1128 1340h
A Bayesian Approach to Estimate Aquifer Heterogeneity with Generalized Parameterization
The research proposes a generalized parameterization (GP) method and Bayesian estimation for parameter heterogeneity characterization and identification in groundwater modeling. GP unifies zonation and interpolation through a set of weighting coefficients, and is capable of creating a zonation structure, a continuous distribution, or a mixed structure. GP shows greater flexibility not only in manipulating the highly complex spatial distribution but also in identifying the parameter structure. With GP, parameter structure identification seeks to identify the parameter dimension, parameter pattern, parameter values as well as the values of the weighting coefficients simultaneously through a set of basis points. Additionally, this study develops an embedded genetic algorithm (GA) for solving the structure identification problem. A Bayesian estimator that estimates the basis point values as well as the values of the weighing coefficients is embedded in the GA, which searches for the best basis point locations. We demonstrate the inverse methodology by a numerical example in which the distributed transmissivity in a two-dimensional confined aquifer is identified. We calculate the Jacobian matrix by the adjoint state method. With GP, we have successfully identified the transmissivity structure with four basis points that results in a good fitting in groundwater heads and captures the non-smooth characteristic as well as the trend of the true transmissivity field. We compare GP with Voronoi tessellation (zonation) and natural neighbor interpolation. Results show that GP outperforms the other two parameterization methods in that GP identified the transmissivity field with a smaller parameter uncertainty along with a sufficiently small fitting residual and without over parameterization.
H23A-1129 1340h
Multiconstituent Reaction Identification in Groundwater Modeling
The selection of appropriate reaction terms for coupled advection-dispersion-reaction (ADR) groundwater systems is a difficult enterprise. Choosing a proper analytical form is crucial to the accurate prediction of system behavior. The decision can be daunting in the absence of accurate a priori biological, chemical, and geophysical properties of the site. Moreover, once the functional form is chosen, the analyst must then calibrate that reaction function parameters. The calibration process itself can be challenging, especially in the case of complex empirical reaction models. Our research investigates the utility of avoiding the direct choice of a reaction function. Instead of picking a particular analytic reaction term, we construct a reaction function as best as possible given available data. The reaction term is developed by conjoining an optimal set of hyperplanes which approximate the functional geometry of the complex, multiconstituent reactions that the existing measurements suggest. The methodology employs a genetic algorithm (GA) to identify increasingly complex sets of junction nodes in the reaction space, sequential quadratic programming (SQP) to compute an ordinate at each node, and Delaunay triangulation to combine this information into a reaction surface of intersecting hyperplanes. The GA and the SQP ally with the ADR simulation itself to provide the most elementary functional approximation which satisfies the prediction and reliability requirements of the analyst. Results will be shown which demonstrate the utility of this multidimensional inverse modeling approach. Analyses will also be presented which compare an integral goodness of fit metric that indicates which member of a family of analytic functions the reaction surface approximation most closely resembles. In so doing, this methodology serves to provide both the most reliable reaction function which the data allow and to indicate to the researcher which analytic form is most likely to be present in the system studied.
H23A-1130 1340h
Application of Genetic Algorithms To Identify Optimal Groundwater Monitoring Well Locations in 3D
Monitoring groundwater aquifers for possible sources of contamination is an important aspect of water resources management. The design of monitoring networks has been one of the key concerns of researchers who deal with the management of groundwater quality. Optimal monitoring network design can be beneficial to both groundwater simulation as well as optimization modeling. This paper discusses the applications of various optimization techniques from traditional to global methods for the solution of groundwater monitoring network and groundwater quality management problems. In order to solve optimization-based groundwater management models, various mathematical programming techniques such as linear/nonlinear programming, mixed-integer programming, differential dynamic programming, stochastic programming, as well as global optimization methods such as Genetic Algorithms are used by researchers to obtain optimal solutions for groundwater management. The resent study will also discuss a state of the art method, which combines simulation of groundwater flow and transport with genetic algorithm optimization. In order to ensure that the optimal management strategy is physically acceptable, a simulation model is necessary to simulate the system behavior. The simulation model basically provides solutions that satisfy the equations governing the relevant processes in the system. Thus the simulation models can be used for checking the feasibility of a management strategy. Once the optimization model is formulated, a suitable mathematical programming technique such as genetic algorithm is applied to obtain the optimal solution. This approach not only accounts for the complex and non-linear behavior of the groundwater system, but also identifies the best monitoring strategy under a specific objective function with several constraints. The solution identifies the best location of monitoring wells.
H23A-1131 1340h
Electrometrical Methods Application for Detection of Heating System Pipeline Corrosion
Coated steel underground pipelines are widely used for the petroleum and gaze transportation, for the water and heat supply. The soils, where the pipelines are placed, are usually highly corrosive for pipe's metal. In the places of crippling of external coating the corrosion processes begin, and this can provoke a pipe breakage. To ensure the pipeline survivability it is necessary to carry out the control of pipeline conditions. The geophysical methods are used to provide such diagnostic. Authors have studied the corrosion processes of the municipal heating system pipelines in Saint-Petersburg (Russia) using the air thermal imaging method, the investigation of electromagnetic fields and spontaneous polarization, measurements of electrode potentials of metal tubes. The pipeline reparation works, which have been provided this year, allowed us to make the visual observation of pipes. The investigation object comprises a pipeline composed of two parallel tubes, which are placed 1-2 meters deep. The fact that the Russian Federation and CIS countries still use the direct heat supply system makes impossible any addition of anticorrosion components to circulating water. Pipelines operate under high pressure (up to 5 atm) and high temperature (designed temperature is 150°C). Tube's isolation is meant for heat loss minimization, and ordinary has poor hydro-isolation. Some pipeline construction elements (sliding and fixed bearings, pressure compensators, heat enclosures) are often non-isolated, and tube's metal contacts with soil. Hard usage condition, ingress of technical contamination cause, stray currents etc. cause high accidental rate. Realization of geophysical diagnostics, including electrometry, is hampered in a city by underground communication systems, power lines, isolating ground cover (asphalt), limitation of the working area with buildings. These restrictions form the investigation conditions. In order to detect and localize isolation (coat) defects authors successfully use the excitation-at-the-mass method measurement together with the measurements of magnetic and electrical components of electromagnetic field. However, the electrical contact between a tube and the soil, as well as the presence of zones of isolation defects is not the direct indicators of corrosion focus places. Authors use the spontaneous polarization method to investigate electrical fields, caused by natural electromotive forces of electrochemical origin. Different types of EM and SP anomalies have been detected. After statistical study and visual observations of extracted pipes, the relations between such anomalies and pipeline condition have been obtained. The places of underground pipeline coat destruction can be specified by complex of geophysical investigations. Also, it is possible to detect the intensity of destruction and corrosion processes in real time.