| dc.description.abstract | This thesis deals with modeling and inversion of electromagnetic data for geophysical applications. The main topic is modeling and inversion of marine Controlled Source Electromagnetic (mCSEM) data, but joint inversion of both mCSEM and MagnetoTelluric (MT) data is also considered. Since electromagnetic offshore hydrocarbon prospecting became feasible about a decade ago, a large number of surveys have been acquired, and thus there has been an increased interest in imaging the subsurface using electromagnetic data.
For imaging the subsurface using inversion algorithms, the forward modeling in a given model is essential. It is the engine of an inversion scheme, and accurate and flexible modeling methods are thus needed. In recent years, the attention to the effect of conductivity anisotropy has increased, and it has been investigated through both modeling and inversion. Anisotropy has a significant effect on data acquired for electromagnetic hydrocarbon prospecting, and studying problems related to this is essential to ensure a better understanding of the acquired data. Furthermore, as hydrocarbon exploration moves into more and more complex areas, several data types are acquired. It is thus interesting to study the properties of simultaneous inversion of several data types, and to see what information can be extracted from integrating several data types. These are topics which are treated in this thesis, and it is into this context this thesis fits.
The thesis consists of a short introduction to the mCSEM andMT methods, and four self-contained papers:
• 2.5D EM modeling in TIV conductivemedia and the effect of anisotropy in normalized amplitude responses demonstrates 2.5D modeling of TIV anisotropic mCSEM data using integral equations. Further, we show how changes in anisotropy can affect normalized amplitude responses. We show that the anisotropy has a large influence on the normalized amplitude plots, and we thus argue that such plots should not be used for data interpretation. When dealing with real data, the model you actually normalize against is not known and a confident interpretation is almost impossible to obtain. It is not a quantitative measure, and it is relative to something uncertain. However, such attributes serve a purpose for synthetic sensitivity studies and data QC, and we demonstrate the first application on a stylistic model from a real survey. The paper is published in the Journal of Geophysics and Engineering, 2013, 10, 015006. A preliminary version of this work was presented at the 2009 annual EAGE meeting in Amsterdam.
• TIV Contrast Source Inversion of mCSEM data treats the inverse conductivity problem of determining the subsurface conductivity structure, using what is known as the Contrast Source Inversion method. We extend previous versions of this method to TIV anisotropic media, and add a model regularization to the scheme. We investigate problems that can occur if TIV anisotropy is not taken into account in inversion, and show how they can lead to wrongful conclusions. This paper is published in Geophysics, 2011, 76, F65-F76.
• On Joint inversion of Marine CSEM and MT data explores the advantages and disadvantages of doing simultaneous inversion using bothmCSEM and MT data, contrary to a sequential approach using the MT result as prior information in mCSEM inversion. We do not find a great advantage of doing joint inversion in our examples, but it is nevertheless an useful tool for checking data consistency between several data types in a given model. Moreover, we suggest doing MT inversion using field components, rather than the conventional impedances. The procedure is demonstrated on a real data example. This paper is submitted to Geophysics.
• AdiscontinuousGalerkin method for modelingmarineControlled Source Electromagnetic data presents a new method for mCSEM modeling. It utilizes the discontinuous Galerkin method, which has several interesting properties for geophysical applications, to model the mCSEM experiment. Its fully discontinuous nature naturally accomodates the discontinuous field components and the point source approximation. It is also very flexible with respect to geometry. We demonstrate the properties and flexibility of this method in the 2.5D setting. This paper is in preparation for Journal of Computational Physics.
Although electromagnetic theory is a mature research topic, the application to hydrocarbon prospecting demands new methods and theoretical insights. Especially the size of the problem puts restrictions on the methods to be used due to computational resources, and both survey planning and reliable interpretation relies on accurate modeling and inversion algorithms. The main objective of this thesis is to contribute to an increased understanding of electromagnetic field propagation in conductive media, and to introduce methods to study the forward and inverse scattering problem for geophysical applications. | nb_NO |
