Experimental designs for determination of the anisotropy of remanence-test of the efficiency of least-square and bootstrap methods applied to metamorohic rocks from southern Poland

Abstract

The collection of samples from two drill cores (serpentinite/ peridotite and gabbro/ lamprophyre) from the Fore-Sudetic Block (SW Poland) has been used to test several experimental designs for determination of anisotropy of anhysteretic remanence. Ellipsoids of anisotropy for ARM were obtained for batches of samples from four levels of drill cores. Standard least - square methods for 3, 9, 12 and 15 independent measurements of ARM were used to determine the anisotropy of remanence ellipsoid for each single sample and to evaluate the measurement errors of parameters of the ellipsoid. The distribution of anisotropy parameters (axes orientation and anisotropy ratios) for batches of samples from four core levels was also evaluated for each of the ARM determination methods. Bootstrap, Jelinek (1978) and Hext (1963) methods were applied to calculate confidence cones for mean anisotropy axes (Constable and Tauxe, 1990) for four tested batches of samples.For highly-anisotropic, low-coercive serpentinite/peridotite rocks time-efficient 3-orientations procedures (Stephenson et al. 1986) may lead to accurate determination of ARM ellipsoid for the tested sample and a batch of samples. The differences of anisotropy axes directions determined by different methods are here of order 3–5°. The anisotropy parameters Pj-T of anisotropy ellipsoid are practically the same for 6- and 15-positions methods and slightly different for 3- positions method. For pyrrhotite bearing gabbros 3-positions method produces the AARM ellipsoids (for single samples and the collection) that is 5–10° apart from results of the other methods and overestimates the anisotropy ratio Pj. That effect is enhanced for lamprophyres. On the contrary to the 6- and 15-positions method (Pj<1.5, weakly grouped anisotropy axes), the 3-positions method gives here the well-grouped, differently oriented axes with Pj of 2–3.5! The correction for suspected GRM effect (Stephenson, 1993) shifts the AARM axes to be in agreement with results of the remaining methods. However, the high anisotropy still persists after GRM correction.