NS43A-01 INVITED
High-Resolution Seismic Reflection to Monitor Change
High-resolution seismic reflection has proven a valuable tool detecting changes in fluid composition, rock petrophysical properties, and structures critical to reservoir production management and groundwater protection in Kansas. Surface seismic reflection is not a method that lends itself to direct detection and delineation of boundaries between different fluid compositions in porous media. However, time-lapse seismic does appear to have been successful identifying areas where calculated changes in seismic characteristics (specifically velocity) are greater than 10% at a miscible CO$_{2}$ flood in Russell County, Kansas. Empirically a 10% change in seismic velocity has proven to be the minimum practical threshold where signal emerging from the noise can be interpreted with any degree of confidence. This change in velocity occurs when the saturation of injection CO$_{2}$ exceeds 30% of the total pore fluid at this site. To evaluate the potential of high-resolution seismic reflection to monitor the injection in a miscible CO$_{2}$ enhanced oil recovery pilot study in a 900 m deep 5 m thick oolitic carbonate petroleum reservoir, a 4-D seismic reflection program was undertaken that includes 12 different 3-D surveys over 6 years. The first 3 years (8 surveys) were designed to specifically address the potential application of this method to enhanced oil recovery. The last 3 years (3 surveys) are intended to evaluate the effective of seismic in providing the assurances necessary for CO$_{2}$ sequestration. Collapse structures related to karst features and anthropogenic leaching resulting from faulty bore fluid containment have posed serious threats to the quality of groundwater above the Hutchinson Salt Member of the Permian Wellington Formation in central Kansas. High-resolution seismic reflection played a key role in characterizing the preferential growth of a sinkhole resulting from the dissolution of the Hutchinson Salt in Pawnee County, Kansas. Salt leaching was instigated by confinement failure of an oil field brine disposal well. In 1998, legacy 2-D seismic data showed the subsurface extent of collapse was approximately an order of magnitude larger than the sinkhole. A consistent pattern of growth, elongated parallel to the anticlinal structure responsible for the oil field, was interpreted on 2004 time-lapse 2-D data. Confinement of several aquifers overlying the salt was compromised when the 300 m of rocks overlying the salt collapsed, forming the sinkhole. This breach in confining layers provided a pathway to the salt for unsaturated brine fluids. Radial growth of the dissolution feature has slowed consistent with volumetric spreading of the dissolution front. The migration of the brine away from the dissolution front and out of the Hutchinson Salt interval has been relatively consistent in spite of changes in source waters. High-resolution seismic monitoring has a great deal of potential to monitor changes in fluid and structures, but requires a high degree of scrutiny and attention to detail for effective application.
NS43A-02 INVITED
Evaluating the Detectability of Streaming Potential Anomalies Caused by Structural Defects in the Core of an Embankment
Internal erosion, defined as the loss of fine-grained material due to seepage forces, can compromise the stability of an earthfill dam and cause its ultimate failure. Monitoring of dam performance has become of critical importance, particularly as the structures age and design methods evolve. Conventional monitoring of the hydraulic regime using piezometers and weirs provides sparse sampling, and these methods may not be sufficient to detect the onset of internal erosion. Consequently, a comprehensive investigation tool is needed to complement these methods. The self-potential (SP) method has been used to delineate anomalous zones that correspond with areas of preferential seepage flow in embankment dams. Observed responses have been attributed to the electrokinetic phenomenon of streaming potential, where electrical current flow is generated by fluid flow through porous media. We have developed a rigorous 3-D finite volume algorithm that calculates the SP field induced by a simulated hydraulic regime, based on the principles of coupled flow. In the current study, we apply the algorithm to investigate the anomalous SP response resulting from seepage through pervious defects in the core of an embankment. While the SP signal due to seepage through the intact embankment is large (in the range of tens of milliVolts), the anomalous responses caused by the presence of pervious defects are close to noise levels for typical measurements made at surface.
NS43A-03 INVITED
4D GPR Experiments--Towards the Virtual Lysimeter
In-situ monitoring of infiltration, water flow and retention in the vadose zone currently rely primarily on invasive methods, which irreversibly disturb original soil structure and alter its hydrologic behavior in the vicinity of the measurement. For example, use of lysimeters requires extraction and repacking of soil samples, and time- domain reflectometry (TDR) requires insertion of probes into the soil profile. This study investigates the use of repeated high-density 3D ground penetrating radar surveys (also known as 4D GPR) as a non-invasive alternative for detailed visualization and quantification of water flow in the vadose zone. Evaluation of the 4D GPR method was based on a series of controlled point-source water injection experiments into undisturbed beach sand deposits at Crandon Park in Miami, Florida. The goal of the GPR surveys was to image the shape and evolution of a wet-bulb as it propagates from the injection points ($\sim$0.5 m) towards the water table at 2.2 m depth. The experimental design was guided by predictive modeling using Hydrus 2D and finite-difference GPR waveform codes. Input parameters for the modeling were derived from hydrologic and electromagnetic characterization of representative sand samples. Guided by modeling results, we injected 30 to 40 liters of tap water through plastic-cased boreholes with slotted bottom sections (0.1 m) located 0.4 to 0.6 m below the surface. During and after injection, an area of 25 m2 was surveyed every 20 minutes using 250 and 500 MHz antennas with a grid spacing of 0.05 x 0.025 m. A total of 20 3D GPR surveys were completed over 3 infiltration sites. To confirm wet-bulb shapes measured by GPR, we injected 2 liters of "brilliant blue" dye ($\sim$100 mg/l) along with a saline water tracer towards the end of one experiment. After completion of GPR scanning, a trench was excavated to examine the distribution of the saltwater and dye using TDR and visual inspection, respectively. Preliminary analysis of the 4D GPR data shows the shape and size of the imaged wet-bulb closely match the modeling and excavation results. Wetting front propagation was on the order of 0.2-0.5 m/hr. GPR detected retained water 17 hr after completion of the infiltration. In addition to pattern detection, attempts for establishing mass balance are underway. In summary, preliminary results from controlled field experiments show the 4D GPR method has the precision, resolution, and repeatability required for non-invasive detection of water-flow patterns and dynamics in the vadose zone, hence could potentially serve as a virtual lysimeter.
<a href='http://www.3dgpr.info'>http://www.3dgpr.info</a>
NS43A-04 INVITED
Designing Resistivity Geometries for Long Term Monitoring of Hydrologic Processes
Electrical resistivity methods have been used for many years to detect the presence and movement of fluids in the subsurface. In recent years, due to advancements in both hardware and software, there have been a number of attempts to use resistivity imaging more quantitatively for site characterization and for monitoring hydrologic processes. There are a number of important issues inherent to the resistivity imaging method that must be carefully considered to use of this method for hydrogeologic applications; a major issue is choosing an electrode geometry prior to working in the field that not only allows for maximum resolution of the target of interest, but is also able to provide the best possible resolution through time. Resolution of electrical resistivity methods is generally highest near the electrode locations and decreases away from the array. However, a problem is that the resolution of electrical resistivity tomography is dependent on the electrical conductivity of the subsurface, and this can change both in time and space when monitoring time-varying hydrologic processes. Development of geometries and that maximize sensitivity through time is of key importance for imaging these processes and making quantitative estimates of hydrologic processes. Additionally, the number of data we can collect in the field is limited due to the temporal scale of the hydrologic process of interest. In this work, we present a new method for designing resistivity geometries for long-term monitoring. To do so, we determine, using all prior information as well as our understanding of the hydrologic process we wish to image, the optimal location of electrodes on the surface and/or in wells, so as to maximize measurement sensitivity in the subsurface region of interest under time constraints on data collection.
NS43A-05
Continuous Monitoring of Vertical Resistivity Probes in the Transition Zone; Field Observations of Resistivity Hysteresis and Other Phenomena
Four vertical geophysical probes containing electrodes for resistivity measurements and thermistors for temperature measurements were installed in a row perpendicular to a pond in Van Buren County, Michigan at distances of 2 m, 3 m, 5 m, and 8 m from the edge of the pond at the time of installation. Probes one and two contain sixteen electrodes and three thermistors. Probes three and four contain thirty-one and thirty-three electrodes respectively and five thermistors each. These probes are designed to monitor primarily the resistivity at the groundwater/vadose zone interface near a surface water body. Additional installations at the site include four monitoring wells and four Time Domain Reflectometry (TDR) access tubes, each paired with the probe installations. The system was installed in the summer of 2005. Data collection has been continuous since December 2005. Data collection at the site consists of Pole-Pole resistivity measurements from each electrode pair with an a-spacing of five centimeters, temperature readings from the thermistors on the probes as well as two surface thermistors, and precipitation measurement from an on-site rain gauge. A CR1000 Campbell Scientific, Inc. datalogger, with four multiplexers, is used to collect and store the data. Data are collected every four hours under normal (non-precipitating) conditions. Each rain event triggers a three-minute sampling regime that lasts four hours after the cessation of rain. Manual water level measurements from each monitoring well and TDR measurements from each TDR access tube are made weekly to complement the automated resistivity measurements. Preliminary analysis shows significant resistivity changes in the vadose zone through precipitation events greater than 10mm total rainfall. Resistivity hysteresis has been observed in the transition zone at two of the four probes during imbibition of the larger precipitation events. There is a smooth, but non-linear increase of resistivity values above the top of the water saturated zone following the minimum resistivity values caused by each precipitation event.
NS43A-06
Electrical Resistivity Monitoring of Voids: Results of Dynamic Modeling Experiments
Remote, non-invasive detection of voids is a challenging problem for environmental and engineering investigations in karst terrain. Many geophysical methods including gravity, electrical, electromagnetic, magnetic, and seismic have potential to detect voids in the subsurface; lithologic heterogeneity and method- specific sources of noise, however, can mask the geophysical signatures of voids. New developments in automated, autonomous geophysical monitoring technology now allow for void detection using differential geophysics. We propose automated collection of electrical resistivity measurements over time. This dynamic approach exploits changes in subsurface electrical properties related to void growth or water-table fluctuation in order to detect voids that would be difficult or impossible to detect using static imaging approaches. We use a series of synthetic modeling experiments to demonstrate the potential of difference electrical resistivity tomography for finding (1) voids that develop vertically upward under a survey line (e.g., an incipient sinkhole); (2) voids that develop horizontally toward a survey line (e.g., a tunnel); and (3) voids that are influenced by changing hydrologic conditions (e.g., void saturation and draining). Synthetic datasets are simulated with a 3D finite-element model, but the inversion assumes a 2D forward model to mimic conventional practice. The results of the synthetic modeling experiments provide insights useful for planning and implementing field-scale monitoring experiments using electrical methods.