HR: 1340h
AN: H13C-1347    [Abstracts]
TI: Source inversion of self-potential data with compactness constraints
AU: * Minsley, B
EM: minsley@mit.edu
AF: Earth Resources Laboratory, Dept. of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, E34-356, 42 Carleton St., Cambridge, MA 02139 United States
AU: Sogade, J
EM: sogade@erl.mit.edu
AF: Earth Resources Laboratory, Dept. of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, E34-356, 42 Carleton St., Cambridge, MA 02139 United States
AU: Morgan, F D
EM: morgan@erl.mit.edu
AF: Earth Resources Laboratory, Dept. of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, E34-356, 42 Carleton St., Cambridge, MA 02139 United States
AB: Self-potential (SP) measurements sample the electric field that results from a source term in Poisson's equation together with the earth resistivity structure and appropriate boundary conditions. These sources can be generated by various subsurface flows, typically due to hydraulic, thermal, or chemical gradients. We utilize forward and inverse modeling using the transmission network analogy to determine the 3D self-potential source distribution from measured SP and resistivity data. Self-potential source inversion is a linear problem, though it is complicated by ill-conditioning and non-uniqueness common to potential field problems. The linear operator is composed of the Green's functions defined by the survey geometry and estimated resistivity structure. Inversion stability often relies on regularization that imposes a flatness or smoothness constraint on the model that is sometimes physically inappropriate. Our approach utilizes a minimum support regularization operator that favors a class of solutions that fit the SP data with sources that are spatially compact. This iterative method produces multiple source models with increasing compactness but similar data misfit, and is an effective way to examine the non-uniqueness in this type of problem where there is an inherent tradeoff between the source shape, magnitude, and location. Prior knowledge of the expected compactness of the source can then be used to select a physically appropriate model. Similar forms of regularization have proven successful in other potential field problems such as gravity and magnetics, as well as electroencephalographic (EEG) imaging in the medical community. A significant benefit of these methods is the resolution of targets at depth from surface measurements alone. We also discuss the incorporation of sensitivity information in the inversion, imperfect knowledge of the resistivity structure, and the effect of noisy data using synthetic and field examples.
DE: 0520 Data analysis: algorithms and implementation
DE: 0903 Computational methods: potential fields (1214)
DE: 0925 Magnetic and electrical methods (5109)
DE: 1835 Hydrogeophysics
DE: 3260 Inverse theory
SC: Hydrology [H]
MN: Fall Meeting 2005