A METHOD FOR ESTIMATING ERRORS IN CALCULATED STRAINS

Accurate strain estimation requires tests to identify whether natural distributions of strain markers satisfy the assumptions of the method used. To evaluate the capabilities of different strain calculation and statistical methods for this purpose, simulated ellipse distributions of varying axial ratios and orientations are strained and analyzed. If bedding and cleavage cannot be measured, statistical tests on strained ellipse distributions do not produce useful results because of the inseparable effects of strain, initial ellipticity and initial preferred orientations. Tests on data destrained using averaged ellipse matrix values can indicate initial preferred orientations and the nature of the axial ratio distribution. The most useful tests are the chi-square tests of the ellipticity and orientation distributions of the destrained data. The Rayleigh test cannot identify initial preferred orientations in this type of destrained data. © 1990.