HR: 08:00h
AN: H51J-01    [Abstracts]
TI: Fusion of Multiple Levels of Subsurface Information for Imaging and Hydrologic Analysis
AU: * Daniels, J J
EM: jeff@geology.ohio-state.edu
AF: Department of Geological Sciences The Ohio State University, College of Math and Physical Sciences 425 Stillman Hall, Columbus, OH 43210 United States
AU: Illman, W A
EM: walter-illman@uiowa.edu
AF: Department of GeoScience The University of Iowa, 121 Trowbridge Hall, Iowa City, IA 52242 United States
AU: Yeh, T
EM: yeh@hwr.arizona.edu
AF: Department of Hydrology and Water Resources The University of Arizona, John Harshbarger Building 1133 E. North Campus Drive, Tucson, AZ 85721 United States
AU: Parashar, M
EM: parashar@caip.rutgers.edu
AF: Department of Electrical and Computer Engineering Rutgers University, 94 Brett Road, Piscataway, NJ 08854 United States
AU: Hariri, S A
EM: hariri@ece.arizona.edu
AF: Department of Electrical and Computer Engineering The University of Arizona, 1230 E Speedway Blvd., Tucson, AZ 85721 United States
AU: Kruger, A
EM: anton-kruger@uiowa.edu
AF: IIHR-Hydroscience and Engineering The University of Iowa, C. Maxwell Stanley Hydraulics Laboratory, Iowa City, IA 52242 United States
AB: Information on detailed spatial distributions of hydraulic properties, water, contaminants and other geofluids in the fractured vadose zone is important to our understanding and prediction of many natural processes and management of our precious natural resources. Many innovative hydraulic, tracer, and geophysical tomography surveys have been recently developed for this purpose. While these techniques are capable of collecting a vast amount of information of different types, they often lead to information overload, uncertainties, and difficulties in interpretation. A newly developed stochastic information fusion approach combined with dynamic, adaptive, parallel computing technologies is developed to meet these needs. The approach to solving the complex problem of data integration involves the following: 1) development of new generation of forward and inverse methodology, based on local adaptive refinement methods and the stochastic successive linear estimator (SLE), for geophysical surveys, including electrical resistivity tomography (ERT), high-frequency ground penetrating radar (GPR), and mid-range-frequency radar. 2) development of a stochastic fusion technology to assimilate the strengths of the three geophysical technologies to enhance the images of resistivity and dielectric constant variations over large volumes of the subsurface and to quantify their uncertainties; 3) development of a stochastic fusion technology to congregate the ability of geophysical surveys to image fracture patterns from resistivity and dielectric constant anomalies and the ability of pneumatic/gas tracer tomography to estimate air permeability, porosity, and potential connectivity of fractures. The resulting integration of these data provides a detailed view of the 3-D distribution of the pneumatic properties. A similar fusion approach is also applied for hydraulic/tracer tomography and geophysical surveys. The three different fusion processes take an iterative approach to maximize the usefulness of information from the different sensors. Success of the research advances not only estimation theory in general, but also our technologies for characterizing and monitoring the fractured vadose zone. A dynamic, adaptive, autonomic, parallel computing environment that uses local clusters is also developed to facilitate and expedite fusion of the stochastic information collected from geophysical, hydraulic, and pneumatic sensors.
DE: 1800 HYDROLOGY
SC: Hydrology [H]
MN: Fall Meeting 2005