Cardiff University | Prifysgol Caerdydd ORCA
Online Research @ Cardiff 
WelshClear Cookie - decide language by browser settings

Paleostresses deduced from striated faults: Are they up to scratch?

Orife, Tobore and Lisle, Richard John 2005. Paleostresses deduced from striated faults: Are they up to scratch? Geophysical Research Abstracts 7 , 04058.

Full text not available from this repository.

Abstract

A major restriction on the application of fault-slip analysis methods is the availability of a collection of faults whose displacements where induced by the resolved shear stresses associated with a single stress tensor, and where the slip on individual faults are not influenced by the slip on other nearly faults. How does the prospective user of these methods know that the available data meet these requirements? To help answer this, the performance of fault-slip methods were examined when used to analyze unsuitable data; namely, faults and slip lineations with randomly chosen orientations. A favourite criterion for judging data quality is the average discrepancy between the orientation of actual lineation on each fault and the lineation theoretically predicted from the best-fit tensor. However in this work it is found that random faults also yield small angular discrepancies in conditions where 8 or less faults are used. This criterion is only useful for large samples of faults. Another test is to use the existence of tensors that are compatible with a given data set. For random data, there is possibility that tensors can be found that are capable of explaining the lineation orientation. For example, the existence of compatible stress orientations deduced from the right dihedron method is no proof of data meeting the assumptions of the method. The probability of finding such tensors depends on the tolerance used to assess fit, and the total number of trial tensors used. However a more useful check is to use the proportion of trial tensors that are found to fit random datasets. This proportion decreases rapidly with sample size. In sample sizes greater than five faults the expected proportion of tensors fitting is very small (<1%). Statistical tests are proposed. A powerful test on data quality is found to be one based on the consistency of the best-fit tensors obtained from sub-samples of the data set. A bootstrap method is able to detect poor quality data from a lack of consistency in the resulting tensors. This study emphasizes the dangers of palaeostress determinations from small numbers of faults. All of the tests of quality increase in power as the number of faults in the sample increases.

Item Type: Article
Date Type: Publication
Status: Published
Schools: Earth and Ocean Sciences
Subjects: Q Science > QE Geology
Additional Information: Poster
Publisher: European Geosciences Union
ISSN: 1029-7006
Related URLs:
Last Modified: 04 Jun 2017 02:10
URI: http://orca-mwe.cf.ac.uk/id/eprint/10404

Actions (repository staff only)

Edit Item Edit Item