Inversion and Resolution Analysis of Electromagnetic Data: Examples from Geothermal and Watershed Characterization

Publication Date


Type of Culminating Activity


Degree Title

Doctor of Philosophy in Geophysics



Major Advisor

Partha S. Routh, Ph.D.


Raw data from geophysical experiments have limited use for quantitative interpretation of the subsurface. To make quantitative inferences we must apply an imaging operation to our data. Knowledge of the governing physics for a particular experiment allows us to construct subsurface images via an inverse modeling operation. Because time and resource constraints render it impossible to collect perfect geophysical data, our geophysical images are imperfect representations of the true subsurface.

Image resolution analysis is a process that aids in determining how well we are able to resolve the subsurface with a given set of geophysical measurements. Resolution analysis can be formulated so that image interpretation is more intuitive and more accurate compared with interpretation strategies that do not specifically account for model uncertainty. We propose and demonstrate several approaches for determining the resolving capability of geophysical inversion. First, a simple lD linear inverse problem is used to develop the methodology and demonstrate the utility of resolution analysis. The resolution analysis procedure is then applied progressively to more complex inverse problems, including field data examples. The generic approaches to resolution analysis described herein can be applied to a wide variety of practical problems.

Using case study examples, we demonstrate that image resolution analysis enhances our capability to interpret field data inversion results. For example, in a geothermal reservoir imaging application, we show that resolution analysis is useful in constraining reservoir depth and thickness and for determining preferential fluid flow paths within the reservoir. In an aquifer storage and recovery application, we use resolution analysis to demonstrate the feasibility of using CSEM techniques to monitor long-term changes in water saturation.

We present another case study example to demonstrate how time-lapse geophysical data provide a means to monitor changing water saturation in the subsurface. Repeatability of the geophysical measurements becomes a major difficulty in analyzing data from time-lapse surveys because varying noise levels in the data complicate direct comparison of static inversion results. Using synthetic data examples, we demonstrate generic approaches to time-lapse data analysis that do not strictly require repeatability of the measurements. We then apply these time-lapse inversion strategies to field data from a small high-elevation watershed. We successfully delineated the regions of changing water saturation and imaged the fracture orientations and fluid flow paths within the watershed.

Files over 30MB may be slow to open. For best results, right-click and select "save as..."