Three-dimensional variations in inner core anisotropy

Geophysics Program, Box 351650,
University of Washington,
Seattle, WA, 98195-1650, (206) 685-2803
kcc@geophys.washington.edu

poster/oral:

We analyze 1775 high-quality observations of differential PKP travel times: PKPBC-PKPDF and PKPAB-PKPDF. These include only the highest quality picks chosen from about 10,000 seismograms for data of the IRIS FARM (Global Seismic Network) and for four events to the Alaska Seismic Network (ASN) and one to the Northern California Seismic Network. Times are estimated using cross- correlation of digital seismograms filtered to short period (1 Hz). We compare the : PKPBC-PKPDF observations with predictions from radial models PREM, PREM2, IASP91, and AK135 and find that AK135 fits the data the best, though PREM2 also does an adequate job, while trends with distance for IASP91 and PREM significantly misfit the globally distributed observations. Among PKPBC-PKPDF rays whose directions through the inner core are more than 35° from the spin axis, those that turn in the western hemisphere are 0.7 s less, on average, than those that turn in the eastern hemisphere. Among rays less than 35° from the spin axis those that turn in the western hemisphere average more than 3 s, while the those in the eastern hemisphere show much more scatter and average 1.5 s. PKPAB-PKPDF exhibit similar trends suggesting that these large-scale variations continue to depths as great as 600 km into the inner core. PKPBC-PKPDF time anomalies from the 1000-km aperture ASN vary systematically from 2 to 4 s providing compelling evidence that the anisotropy varies laterally by 30% and in depth by 40% over length scales of 300 km. Changes in ray path length account for the remaining variation in travel times. Anisotropy reaches values of at least 4% averaged along paths exceeding 1000 km in length. Taken as a whole these observations argue strongly for significant 3-D variation in anisotropy in the inner core. They can be explained by a model in which the intrinsic properties of the inner core do not vary laterally, but the degree of alignment does. Apparently most of the western hemisphere is well aligned (fast velocity parallel to the spin axis), while crystal alignment in the eastern hemisphere is sporadic. We suggest that whatever mechanism is invoked to explain the alignment, it must include a feedback mechanism such that once a region becomes well aligned it helps organize the alignment of crystals around it. The strong alignment of the western hemisphere to depths of 600 km argues that if alignment occurs during growth of the inner core, the strong feedback mechanism has been working since the inner core was half its current radius (one eighth its current volume).