Michael D. Knoll, Ph.D.

The electrical resistivity tomography (ERT) experiment is just one of numerous geophysical imaging techniques which has great possibility of aiding within the non-invasive, nearly continuous estimation of fabric characteristics in near-surface ecological and engineering applications. To fully realize this potential will need a increased knowledge of numerous regions of the ERT experiment. This dissertation presents some advancements for that ERT experiment in the theoretical and modeling context, and poor time-lapse imaging in the injection/withdrawal experiment created for aquifer portrayal then when a proxy for pump-and-treat removal effort in a unconfined alluvial aquifer.

The theoretical and modeling efforts show: 1) Source-specific boundary conditions for subsurface electrodes are affordable and may provide significant precision gains when dealing with computational limitations of multi-scale record grids. 2) The main-secondary separation of potential allows you to derive a scattering series within the sparse differential domain obtaining a convergence qualifying qualifying qualifying criterion that creates about. both magnitude and distribution of heterogeneity of electrical conductivity. For linearization connected while using the Frechet derivative to obtain valid, the perturbation in electrical conductivity needs to be tiny as based on convergence within the scattered series. A derivation is presented that enables efficient calculation of charge accumulation across contrasts in electrical conductivity by equating accrued surface charge density for a way to acquire scattered potential within the differential domain. 3) Data error because of electrode mislocations can considerably contaminate ERT data along with the reconstructed electrical conductivity.

A technique is provided to calculate systematic data error connected with electrode mislocations and to estimate the resulting artifacts within the reconstructed electrical conductivity images. Both data error and model artifacts are experiment- and model-dependent.

The experimental factor relating to this dissertation concentrates on the conclusion results of numerous procedures for incorporating time-lapse information towards the inversion of ERT data collected with the injection and consecutive withdrawal in the saline proxy contaminant in a unconfined aquifer. Solute mass is actually under-predicted. However, time-lapse regularization creates the largest adjustments to model size to lead to estimates of solute mass which are nearest to actual solute mass. No matter regularization plan, estimates within the difference in mass between experimental stages are usually accurate than estimates of total solute mass at any particular experimental stage. The quantity of analysis index (VOl) is brought to define regions of reliability for the ERT models. Resolution within the ERT models is quantified when using the point-spread function (PSF) which requires advancement of a dependable computational method of the big three-dimensional problem. Both measures of solution evaluation, the VOl and PSF, are model- or time-dependent. The PSF exhibit reasonable resolution, but in addition significant localization errors which are, by yet, not well understood.

Oldenborger, Greg Arthur, “Advances in Electrical Resistivity Tomography: Modeling, Electrode Position Errors, Time-Lapse Monitoring in the Injection/Withdrawal Experiment and Solution Evaluation” (2006). Boise Condition College Theses and Dissertations. Paper 387.
scholarworks.boisestate.edu/td/387