Paleomagnetism of the Duke Island, Alaska, ultramafic complex revisited

The Duke Island ultramafic intrusion was emplaced into the Alexander terrane immediately preceding development of a regional mid-Cretaceous thrust belt. Paleomagnetic samples were collected from exposures of ultramafic rock with cumulate layering northwest of Judd Harbor and northwest of Hall Cove. Thermal demagnetization results were analyzed using principal component analysis to isolate the characteristic remanent magnetization. Site-mean characteristic directions determined from 16 sites fail the fold test at 95% confidence, indicating that cumulate layering attitudes were highly contorted at the time of magnetization, at least on a scale of tens of meters. Variations in cumulate layering attitudes probably resulted from the combined effects of thermal convection phenomena during crystallization and deformation following crystallization but prior to magnetization. Analysis of cumulate layering over larger structural domains indicates that kilometer-scale deformation produced southwest plunging folds within the Hall Cove and Judd Harbor bodies. Bogue et al. [1995] proposed that a compound structural correction involving unplunging of fold axes followed by unfolding of average cumulate layering could restore cumulate layering to horizontal. However, using the full set of 21 site-mean paleomagnetic directions from Duke Island (16 from the current study and 5 from Bogue et al. [1995]), the compound structural correction yields mean paleomagnetic directions from the Judd Harbor and Hall Cove areas that are statistically distinguishable at 99% confidence. This result indicates that even on the kilometer-scale, cumulate layering within the Duke Island ultramafic intrusion was neither coplanar nor horizontal at the time of magnetization. Observations of cumulate layering in other ultramafic intrusive rocks indicate that this layering can significantly depart from horizontal by 10 degrees -20 degrees even on the kilometer scale. Therefore use of cumulate layering of ultramafic rocks as a proxy for paleohorizontal is not justified, and paleomagnetic directions from the Duke Island ultramafic intrusion cannot be used to infer the Cretaceous paleolatitude of the Insular superterrane.