DATA WEIGHTING SCHEMES FOR LARGE SCALE AEM INVERSION

Abstract

PreviousNext You have accessSymposium on the Application of Geophysics to Engineering and Environmental Problems 2013DATA WEIGHTING SCHEMES FOR LARGE SCALE AEM INVERSIONAuthors: Leif CoxBurke MinsleyMicheal ZhdanovDavid SunwallLeif Cox USGS; TechnoImagingSearch for more papers by this author, Burke Minsley USGS;Search for more papers by this author, Micheal Zhdanov USGS; TechnoImagingSearch for more papers by this author, and David Sunwall TechnoImagingSearch for more papers by this authorhttps://doi.org/10.4133/sageep2013-108.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail Abstract Inverse methods attempt to reconstruct a geophysical model which satisfies a set of geophysical data while conforming to some geologic prejudice. Since measured data always contain some level of noise, one goal of any inversion algorithm is to fit the measured data to a level such that the geologic signal is fit, but the noise is not. There several different methods for obtaining the optimal fit such as generalized cross validation, L-curve analysis, and chi squared cutoff. All of these methods assume that one global measure not only applies to the entire survey domain, but generally to every part of the survey domain. However, this is often not the case with large scale surveys. The rate of convergence and number of iterations required to adequately fit observed data is largely dependent on the geologic complexity and how close the initial model is to the final model. In large scale surveys, there may be areas where the geology is well approximated by a homogenous half-space, others areas with variable overburden, and still other areas with strong contrasts and great geologic complexity. If one uses a simple global cutoff without regard for the varying geologic regimes, it will likely lead to regions of over-fit data and regions of under-fit data. Since this issue is largely one of data fit, we suggest an adaptive data weighting scheme which takes into account the data fit over regions at the end of each iteration, and then adjusting the data weights to ensure that each geologic regime or area is fit to approximately the same level. We demonstrate the technique with a RESOLVE data set acquired near Ft. Yukon, Alaska. We compare several different methods of adaptive data weighting with the conventional chi squared cutoff approach. Permalink: https://doi.org/10.4133/sageep2013-108.1FiguresReferencesRelatedDetails Symposium on the Application of Geophysics to Engineering and Environmental Problems 2013ISSN (online):1554-8015Copyright: 2013 Pages: 821 publication data© 2013 Published in electronic format with permission by the Environment and Engineering Geophysical SocietyPublisher:Environmental & Engineering Geophysical Society HistoryPublished Online: 28 May 2013 CITATION INFORMATION Leif Cox, Burke Minsley, Micheal Zhdanov, and David Sunwall, (2013), "DATA WEIGHTING SCHEMES FOR LARGE SCALE AEM INVERSION," Symposium on the Application of Geophysics to Engineering and Environmental Problems Proceedings : 649-649. https://doi.org/10.4133/sageep2013-108.1 Plain-Language Summary PDF DownloadLoading ...