This document begins with a complicated but important Figure from Creager (JGR October 10, 1999, page 23129), and continues with a series of Abstracts from the December 1999 AGU Meeting. At the end, are a couple of 1998 AGU Abstracts.

[Map view of ray paths in the inner core, projected to the Earth's surface. Ray paths are shown solid or dashed, if the angle between ray and spin axis (North-South direction) is less than or greater than 35 degrees. Paths are shown with circles or plus symbols, if the BC - DF residual is positive (fast DF) or negative, with symbol size proportional to the residual. The largest circle represents 4.5 s. Note that it is not just paths from South Sandwich Islands to Alaska, which are fast: anisotropy is more pervasive. Note too the occurrence of numerous small circles in the eastern hemisphere (say, more than 30 degrees east of Greenwich), indicating weak anisotropy for much of this region, as first reported by Tanaka and Hamaguchi (JGR 1997) in contrast to the dominance of stronger anisotropy in the western hemisphere. This Figure, from Creager (JGR, October 1999).]

Almost all the following papers were given at the December 1999 meeting of the American Geophysical Union.

[Comments in square brackets following each paper, are by Richards. Papers presenting new evidence that supports the case for inner core rotation are U21B-05 by Song, and U21B-07 by Vidale, Dodge and Earle.]

Elastic Structure of the Inner Core

Paper number: U21B-02
Author: Creager, K C
Author: Ouzounis, A
University of Washington, Geophysics Program, Box 351650, Seattle, WA 98195 United States

Abstract: This paper reviews seismological constraints on the elastic structure of the inner core. Nearly 2000 high-quality differential travel times of compressional waves that pass through the core are collected from eight different studies. The ray sampling of the inner core is inadequate to uniquely constrain the detailed three-dimensional variation of elastic anisotropy. However, a model that is consistent with existing travel-time observations has strong (>3%) anisotropy in the western hemisphere from depths of 100 km to its center. The symmetry direction for hexagonal anisotropy is near the spin axis and represents high wave speeds. The deepest 600 km appears anisotropic at all longitudes, but the upper 400 km of the quasi-eastern hemisphere (40 degrees to 160 degrees E) appears to have far weaker anisotropy (less than or equal to 1%). Anisotropy in the upper 50 km appears to be very weak if it exists at all, but only a handful of paths of differential times are currently available to constrain the uppermost inner core. The isotropic (Voigt) average appears to not vary with longitude, suggesting there is no need for lateral variations in chemistry or temperature which would produce density variations that drive flow. A model with an isotropic upper inner core separated from an anisotropic lower inner core by a discontinuity at depths near 100 km in the western hemisphere and reaching depths of 400 km in parts of the eastern hemisphere can explain many of the features in the differential travel-time observations, as well as the recent suggestion that a discontinuity exists within the inner core. In addition to the large-scale 2% wave-speed variations at 1000 km length scales, differential times observed at stations across the Alaska Seismic Network show the same level of variation (2%) at length scales of a few hundred km laterally as well as radially. Finally, 1-Hz scattered energy and attenuation have been interpreted as being caused by variability in wave speeds with an rms level of a few percent at length scales as small as 3 km.

[Related to a detailed paper published as Creager, K. C., 1999. J. Geophys. Res., 104, 12127--12139. The Figure at the top of this page shows the observations of BC - DF residuals.]

Origin of Anisotropy in the Earth's Inner Core

Paper number: U21B-04
Author: Lee, K
Author: * Karato, S
Author: Lawlis, J D
University of Minnesota, Department of Geology and Geophysics, 108 Pillsbury Hall, Minneapolis, MN 55455 United States
Author: Long, M
Rensselaer Polytechniqnic Institution, 110 8th Street, Troy, NY 12180 United States
Author: Jung, H
University of Minnesota, Department of Geology and Geophysics, 108 Pillsbury Hall, Minneapolis, MN 55455 United States

Abstract: Understanding the origin of inner core anisotropy is critical to extracting useful information from this observation, but difficulties in two areas have hampered in pursuing this exercise. For this purpose, one must first identify the macroscopic process(es) by which the anisotropic structure may have been formed and second investigate the microscopic processes and determine the physical properties that causes seismic anisotropy. After reviewing several macroscopic processes including growth-induced texture formation, texture formation due to the flow caused by growth anisotropy, or by gravitational interaction with mantle, we conclude that the flow caused by the Maxwell stress (stress due to the magnetic field) is the most likely cause for anisotropic structure formation. Because of the finite electrical conductivity, the magnetic field penetrates to a certain depth in the inner core. This field exerts stress (the Maxwell stress) to the inner core materials. The inner core materials therefore flow. This flow will distort the surface and would be terminated when gravitational force balances with the force that causes flow, if there is no reaction at the inner-outer core boundary. However, because the inner-outer core boundary is defined by the melting temperature and the kinetics of melting-solidification are likely to be fast compared to the kinetics of solid state flow in the inner core, this distortion cannot go beyond the position that is determined by the melting point of iron. Therefore, the flow will continue as far as the Maxwell stress is present. We have solved the Stokes equation to determine the flow pattern corresponding to a given Maxwell stress. The relation is general and applies to any geometry of he Maxwell stress, but we calculated the flow field for the two simple cases: symmetric and anti-symmetric toroidal magnetic field. The flow patterns are shown to be sensitive to the geometry of the magnetic field, implying that one can in principle place constraints on the magnetic field from seismic anisotropy in the inner core. The time scale of flow is ~10 eta/A where eta is viscosity and A is the magnitude of the Maxwell stress. For a reasonable estimate of eta and A, this gives time scales of ~1-100 my which are comparable to the time scale of geomagnetic polarity. The actual type of anisotropy corresponding to a given flow geometry depends on a number of physical properties of inner core materials. One of these important mineral physics issues is the nature of LPO in a given flow field. Assuming that the inner core is made of hcp iron, we started an experimental project to investigate the LPO in hcp metals as a function of temperature. The idea behind this is the notion that the nature of LPO in anisotropic crystals often changes with temperature due to the change in microscopic mechanisms of deformation (change in easy slip systems etc.). We use Zn as an analog materials and synthesized and deformed, recrystallized at different temperatures. The results of these experimental studies will be combined with macroscopic modeling of flow to better understand the inner core anisotropy.

[Discusses the likely origin of the Lattice Preferred Orientiation (LPO) which may be the underlying cause of anisotropy in the inner core.]

Inner core super-rotation: facts and artifacts

Paper number: U21B-05
Author: Song, X E-mail:
Univ of Illinois, Department of Geology, Urbana, IL 61801 United States

Abstract: Observational evidence for a differential inner core rotation has recently been reported from time-dependent seismic observations. The determination of the differential rotation, however, is hindered by potential biases from the mantle heterogeneity and possible systematic earthquake location errors. Here we discuss time-dependent observations newly obtained from South Sandwich Islands earthquakes recorded at an array of stations (over 100) in Alaska over the past nearly half a century and Alaskan earthquakes recorded at the South Pole station over 37 years. The earthquakes have been relocated using the Joint Hypercenter Determination technique. The observed differential PKP BC-DF travel times at College, Alaska and three other stations in Alaska and at the South Pole station all show robust changes with time. With the dense samples of different time periods, we apply a joint-inversion technique that allows us to separate the time-dependent inner core structure from the time-independent mantle biases. The results provide strong support for a differential inner core rotation. The rotation rate is constrained to a tight range of 0.3-1.1 degrees per year faster than the mantle even when the uncertainty in the tilt of the anisotropy axis of the patch of the inner core that was sampled is considered. A westward inner core rotation can be ruled out. Surprisingly, the inner core rotation seems resolvable even from 8-years' data observed at Alaskan stations in 1990s (available in digital forms) alone in such a joint inversion, raising the hope that variable inner core rotation, as expected from the geodynamo, may indeed be available. A recent series of papers by Souriau (and co-workers) discussed problems with the evidence of differential inner core rotation, which include biases from event mislocations, biases from mantle heterogeneity, and uncertainty in the tilt of the anisotropy axis. We show that although these potential biases would affect the determination of the rotation rate, they are unlikely to account for the observed time-dependence in the BC-DF times. The time-dependence of BC-DF times on event magnitudes claimed is found to be an artifact.

[Describes three datasets in support of easterly differential rotation of the inner core, that are additional to the original evidence presented by Song and Richards in their 1996 Nature paper (issue of July 18):
  1. Extension of original data set (South Sandwich Islands earthquakes recorded at station COL in Alaska) by two decades;
  2. Addition of Alaskan earthquakes observed at station SPA (South Pole)
  3. The use of more than 100 stations in Alaska, to quantify the effects of near-source structure, near-receiver structure, and mantle heterogeneities.
    Finds evidence for travel-time change in PKIKP, just for data acquired in the decade of the 1990s.]

Rotation of the Inner Core from a new Analysis of Free Oscillations

Paper number: U21B-06
Author: Laske, G
Author: Masters, G
Author: Gilbert, F
IGPP, Scripps Institution of Oceanography, 9500 Gilman Dr., La Jolla, CA 92093-0225 United States

Abstract: Differential rotation of the inner core (IC) has recently been inferred by several body-wave studies with most agreeing that a superrotation may exist with a rate between 0 and 3 degress per year. The wide range of inferred rotation rate is caused by the sensitivity of such studies to local complexities in structure which have now conclusively been demonstrated to exist. Free-oscillation "splitting functions" are insensitive to local structure and are therefore better candidates for estimating differential rotation. We use a new technique to determine splitting functions that allows us to solve for the most general form of the splitting matrix without knowledge of the earthquake sources. This technique is based on the autoregressive property of combinations of seismograms for each event allowing the splitting matrix to be estimated in a non-iterative, one-step process. We apply this technique to recordings of recent "great" earthquakes and get extremely reliable estimates of splitting matrices for modes which are sensitive to the inner core. These matrices can be used to compute theoretical ones for past earthquakes assuming different relative rotation rates for the inner core. In a hypothesis test we investigate which rotation rate best fits the data of past "great" earthquakes. We find that IC differential rotation is essentially zero over the last 20 years implying that the IC is most likely gravitationally locked to the mantle.

[In my opinion, the technique of this paper is not yet sensitive enough to detect the current best estimate of the inner core's rate of differential rotation --- a few tenths of a degree per year, eastward.]

Inner-core scattering: a seismic phase that reveals fine-scale structure and inner core rotation at roughly 0.1 degree per year

Paper number: U21B-07
Author: * Vidale, J E
Dept. of Earth and Space Sciences, UCLA, Box 951567, Los Angeles, CA 90095-1567
Author: Dodge, D A
Lawrence Livermore National Laboratory, L-205 7000 East Avenue, Livermore, CA 94550
Author: Earle, P S
Dept. of Earth and Space Sciences, UCLA, Box 951567, Los Angeles, CA 90095-1567

Abstract: We observe 200 s of seismic waves scattered in the inner core following the expected arrival time of the reflection from the inner-core boundary. Our data are LASA recordings of 12 earthquakes and 4 nuclear tests in the distance range 60 degrees to 70 degrees. The 2% amplitude ratio of these scattered waves with the core-mantle boundary reflection PcP may be explained by 3% stiffness variations with a scale length of 3.5 km across the outer half of the inner core. Modeling of these observations is presented in the poster of Earle et al. (this meeting). These variations are most likely caused by pods of partial melt in a mostly solid matrix, or by variation in orientation or strength of anisotropy. This fine scale structure makes the inner core one of the most heterogeneous places in the planet.

Two of the nuclear explosions we analyze (9/27/71 and 8/29/74) occurred separated by less than 1 km at the southern Novaya Zemlya test site. Waves scattered within the inner core that arrive at LASA with back-azimuths east of the great circle path arrive up to 0.1 s later in 1974 than 1971. Arrivals from the west of the great circle path are up to 0.1 s earlier in 1974 than 1971. In contrast, PKKP and P'P' arrivals, which arrive with nearly as steep angles of incidence and with ample scattered raypaths, but do not encounter the inner core, do not show time shifts that depend on back-azimuth. The simplest interpretation is that the inner core turned 0.3 degree, or at a rate of 0.1 degree per year.

[This paper offers a new and very sensitive method for detecting inner core rotation, and finds a low rate of rotation in the eastward direction. The method, based on the observation of waves scattered from inhomogeneities, is associated with scattering theory described in another AGU abstract (see below, U22A-09). It requires pairs of of events very close together in space, separated in time, and large enough to generate inner-core scattered waves.]

The Role of Gravitational Coupling on Numerical Simulations of the Geodynamo and Inner-Core Rotation

Paper number: U21B-10
Author: Buffett, B A
University of British Columbia, 2219 Main Mall, Vancouver, BC V6T 1Z4 Canada
Author: Glatzmaier, G A
Department of Earth Sciences, UC Santa Cruz, Santa Cruz, CA 95064 United States

Abstract: Mass anomalies in the Earth's mantle distort the equilibrium shape of the inner core, producing topography on the inner core boundary. Gravitational forces on the aspherical inner core may be strong enough to lock the rotation of the inner core to that of the mantle. However, relative rotation is permitted if the boundary topography adjusts during rotation to remain nearly fixed with respect to the mantle. Predictions of inner-core rotation from numerical simulations of the geodynamo have previously assumed that the inner core rotates freely in response to electromagnetic torques. In this study we present high-resolution numerical simulations in which the influences of gravitational forces on the inner core are included. We adopt a representative estimate for the topography on the inner-core boundary (100 m peak-to-peak) and allow for viscous relaxation of the inner-core shape, assuming an effective viscosity of 5 x 1016 Pa s for the inner core. The predicted rotation rate of the inner core relative to the mantle is nominally 0.02 degree/year, which is nearly an order of magnitude slower than the rotation rate predicted in the high-resolution test case when gravitational coupling is turned off. Strong differential velocities develop across the inner-core boundary when free-slip conditions are imposed on the flow. Lines of magnetic force that cross the inner-core boundary are stretched around the inner core, amplifying the azimuthal component of the field. Release of compositional buoyancy from the inner-core boundary can lift and twist the azimuthal field, further enhancing the generation of the magnetic field. We compare the predicted magnetic field with that in the test case (without gravitational coupling) to quantify changes in the vigour of dynamo action near the inner-core boundary. We also present preliminary calculations where no-slip conditions are imposed at the inner-core boundary to explore the combined effects of gravitational, electromagnetic and viscous coupling on inner-core rotation and generation of the magnetic field.

[Extends the work of Buffett, B. A., 1996. Nature, 388, 571--573, which was the first to explore the possibility of an inner core viscosity low enough for the inner core-outer core boundary to have topography that is locked in place with repect to the solid mantle, but still allowing material throughout the volume of the inner core to rotate.]

On Detection of Inner Core Rotation using Doublets

Paper number: U22A-02
Author: Li, A
Lamont-Doherty Earth Observatory, Columbia University, P.O. Box 1000, Palisades, NY 10964 United States
Author: Richards, P G
Lamont-Doherty Earth Observatory, Columbia University, P.O. Box 1000, Palisades, NY 10964 United States

Abstract: Recent studies of the rotation of the Earth's inner core, based on changes of travel-time difference between core phases PKP(BC) and PKP(DF) [PKP(BC)-PKP(DF)], tend to constrain the rotation rate to be small (somewhat less than 1 deg/year eastward). To the extent this result is contested, there is a continuing need to strengthen the evidence. Part of the reason for scatter in previous estimates of travel-time change is the uncertainty of contributions coming from event mislocation and heterogeneity of the Earth's structure, especially at the seismic source region. To strengthen the evidence for PKP(BC)-PKP(DF) travel-time change associated with seismic ray paths through the inner core, we have begun a search for doublets in active seismic regions. Doublets are event pairs, whose seismic waveforms are highly similar, and which occurred at different times but essentially at the same spatial position. Thus the ray paths from doublet events to seismic stations are almost totally overlapping, and heterogeneity of the Earth's structure outside the inner core is canceled out. The high similarity of waveforms enables us to measure the PKP(BC)-PKP(DF) travel-time change with very high precision by using cross-correlation. By this method, rotation of the inner core would be the proposed explanation of detected PKP(BC)-PKP(DF) travel-time change. We have used short period and broad band digital seismograms to search for doublets among seismic events from 1982 to 1998 in the South Sandwich Islands region. We compute the cross-correlation coefficients of P wave records at the same station from different events to check the waveform similarity, and have found events with high cross-correlation coefficients that appear to be doublets. Finally we can compare the PKP(BC)-PKP(DF) times of doublet events to detect any change. We will report our measurements of PKP(BC)-PKP(DF) travel-time change for these events.

[The paper as actually presented found two examples of doublets. However, doublets in the South Sandwich Islands are hard to find, and typically are not large enough to result in amplitudes of PKIKP at station COL that permit the arrival time to be picked with high confidence.]

Inner Core Anisotropy

Paper number: U22A-04
Author: * Dziewonski, A M
Author: Su, W
Harvard University, 20 Oxford street, Cambridge, MA 02138 United States

Abstract: Several complexities in the properties of the inner core and the mantle led to substantial confusion with respect to the level of anisotropy or even its existence. The assumption that the anisotropy is axially symmetric led to estimates of anisotropy as high as 3.5%. This was principally based on observations for the path from South Sandwich Islands to Alaska, which is highly anomalous. This anomaly is localized and it is not even clear whether it is related to anisotropy, since it is sampled only in one direction. Another observation which seemed to support 3% level of anisotropy throughout the inner core were differential travel times PKPAB - PKPDF; it was shown recently that these anomalies are principally caused by the large scale structure near the core--mantle boundary. Subsequently, the need for introduction of anisotropy has been questioned.

We present evidence based on the travel times of PKPBC and PKPDF from ISC Bulletins that there is a strong evidence for anisotropy at a level of 1--1.5%. The Cylindrical Anisotropy Stacks (CAS) for the distance range 150 -- 153 degrees show distinctly fast travel times for paths nearly parallel to the axis of symmetry, while the effect is absent in the corresponding stacks for PKPBC. The anomalies for PKPDF plotted as a function of azimuth and of location of the bottoming points in 10 degree caps are relatively constant for nearly polar azimuths, but show large variation for those close to the equator. However, a similar variation is observed for PKPBC, which means that the strong Y22 signal from the lower mantle might not have been completely removed by correcting for 3-D lower mantle structure derived principally from the P-waves bottoming in the mantle. This could be indicative of a large scale anisotropy in the lower mantle.

[Confirms that global data on PKIKP arrival times support the case for anisotropy in the inner core, though at a somewhat lower level of anisotropy than values originally advocated.]

Constraints on Anisotropy and Discontinuities of the Uppermost Inner Core

Paper number: U22A-06
Author: Ouzounis, A
Author: Creager, K C
University of Washington Geophysics Program, Box 351650, Seattle, WA 98195-1650 United States

Abstract: Thousands of high quality differential travel-time observations suggest that inner core anisotropy is very strong (3-4%) in the western hemisphere from depths of 200 to at least 600 km, but much weaker in the eastern hemisphere ( < 1% ). The strength of anisotropy in the upper 100 km appears to be weak, but is not as well constrained. We estimate the differential times of the interfering PKIKP and PKiKP phases at epicentral distances near 130 degrees using waveform modeling of record sections recorded at arrays. We use independently determined source-time functions from P-waves at different stations. The PKIKP rays at this distance turn about 30 km into the inner core, so this geometry is well suited to imaging the outermost inner core. Rays sampling the inner core at an angle of 30 degrees with respect to Earth's spin axis produce PKiKP - PKIKP differential travel-time residuals of -0.2 and +0.1 s for paths to the Pacific Northwest Seismic Network (western hemisphere) and Kirghiz Network (eastern hemisphere) respectively. These small residuals, along with paths measured by other groups suggest that anisotropy in the outermost inner core is weaker than 1% if it exists at all. Recent studies suggest the existence of a discontinuity in inner core anisotropy. For the two pole-parallel paths we have examined, the coda of the PKIKP and PKiKP phases is low in amplitude and does not contain distinct arrivals that are predicted by models with 3% velocity discontinuity at depths between 200 - 300 km. On the other hand, models with shallower, weaker discontinuities cannot be ruled out as they are expected to produce reflections easily missed in the immediate vicinity of the higher amplitude PKiKP arrival.

[Confirms and builds on the main observation of Tanaka and Hamaguchi (JGR, 1997), that the inner core has a degree-one pattern of inhomogeneity in its anisotropy, with anisotropy being significantly stronger in a pseudo-western hemisphere than in a pseudo-eastern hemisphere. See Figure at the top of this page.]

Fine-scale Inner-core Structure Inferred Through Modeling High-frequency Scattered Energy

Paper number: U22A-09
Author: * Earle, P S
Author: Vidale, J E
Author: Knopoff, L
Dept. of Earth and Space Science, UCLA, Box 951567, Los Angeles, CA 90095-1567

Abstract: We recently detected 1-Hz inner-core scattered seismic waves (Vidale et al. abstract, this meeting) near 70 degrees in stacks of Large Aperture Seismic Array (LASA) data from 12 earthquakes and 4 nuclear tests. The signal initiates at the predicted PKiKP arrival-time and its amplitude increases for 50 seconds before steadily decreasing to the noise level over the next 150 seconds. The scattered wavetrain has a slowness between -0.02 and 0.02 s/km which is consistant with an inner-core origin. At this source-receiver range, PKiKP is predicted to be vanishingly small and we can barely observe it. The emergent onset and long duration of the signal indicates a scattering origin and its high-frequency content implies fine-scale structure. We have found models that explain the inner-core scattering observations. Scattering envelopes are calculated assuming single-scattering, a PREM velocity structure, and an inner-core Q of 240. Good agreement to the data is found for a model having 3% RMS variations in lambda, mu, and density, all with 3.5-km scale-lengths (assuming an exponential distribution and a Poisson solid) spread uniformly throughout the inner core. The steady 50 s increase in amplitude is explained by the contribution from an increasing volume of scatterers to later parts of the wavetrain and the following 150 s decrease depends on the depth distribution of the scatters and inner-core attenuation. We will explore model space to determine the range of elastic-parameters, density, scattering depth-profiles, and attenuation structures, that provide acceptable fits to the data.

[This paper describes the work of modeling the observed inner-core scattered coda to PKiKP. It is such coda waves that underlie the method of paper U21B-07.]

Is the Inner Core of the Earth Partially Molten?

Paper number: U22A-18
Author: Singh, S C
Départment de Seismologie, Institut de Physique du Globe de Paris, 4 Place Jussieu, Paris, F-75252 France
Author: Taylor, M A
Bullard Laboratories, University of Cambridge, Madingley Road, Cambridge, CB3 0EZ United Kingdom
Author: Montagner, J
Départment de Seismologie, Institut de Physique du Globe de Paris, 4 Place Jussieu, Paris, F-75252 France

Abstract: Our information about the inner core comes primarily from seismological observations. We have estimates for the P-wave velocity and it's directional dependence in the inner core, S-wave velocity, and P- and S-wave attenuation. The P-wave velocity appears to vary with depth between 11.04 and 11.26 km/s, while the S-wave velocity is less well constrained, but thought to be less than 3.65 km/s. The P-wave velocity along the pole axis has been observed to be 3 - 4% higher than that in the equatorial plane. Studies of seismic attenuation suggest that at least the upper part of the inner core has high attenuation with quality factors of 200 - 400 for P-waves, and 100 - 200 for S-waves. It is hard to come up with models which simultaneously explain the seismic anisotropy, very low shear wave velocity, and high attenuation. Although the presence of fluid has been appealed to in a qualitative manner to explain individual observations, here, for the first time we make quantitative estimates of the geometry and amount of fluid that would be required to satisfy all of the above observations. We use an effective medium theory to calculate the stiffnesses and resulting seismic velocities and anisotropy of a composite material comprising a solid iron matrix with embedded spheroidal inclusions of melt between the iron crystals. A model of fluid flow between isolated inclusions driven by seismically induced pressure gradients is used to make estimates of the corresponding attenuation in such a medium. We show the effect of changing both the proportion of melt and the aspect ratio of the spheroidal inclusions. We find that a 5 - 10% volume fraction of elongated fluid inclusions, aligned in the equatorial plane, produce behaviour which is consistent with all of the seismic observations. Possible origins for such fluid include dendritic growth of iron, or a mixture of elements which exist in lesser proportion, but are liquid at inner core pressures and temperatures.

[Shows that the theory of fluid-filled cracks can explain observed P-wave anisotropy and low S-wave speed. This paper was later published as Singh, S. C., Taylor, M. A. J. and Montagner, J. P., On the Presence of Liquid in Earth's Inner Core, Science, 287, pp2471-2474, 2000.]

All the above Abstracts are from the December 1999 AGU Meeting. The one below, is from the Spring 1998 Meeting.

A Local Anomaly in the Inner Core

Paper number: S32C-6, from the 1998 Spring meeting (Boston)
Author: Dziewonski, Adam M
Author: Su, Wei-jia
Dept. of Earth and Planetary Sciences, Harvard University, Cambridge, MA 02138

Abstract: 12 years after Morelli et al. (1986) proposed that the PKIKP travel time anomalies can be explained by the cylindrical aniotropy of the inner core with the axis of symmetry aligned with the axis of rotation of the Earth, it is becoming clear that there are significant departures from this simple picture. Su and Dziewonski (1995) described some of them by recognizing a 'statistically significant' tilt of the symmetry axis and detection of substantial longitudinal variation of CAS residuals.
Our reassessment of these results indicates that the perceived tilt of the symmetry axis is an artifact of superposition of a strong local anomaly upon a smooth global pattern. High-pass filtration of the CAS residuals leaves only one region with a substantial amplitude, which has a character of a filtered delta-function. The rays giving rise to this anomaly ccan be traced to the path from the South Sandwich Islands to Alaska. If all PKIKP rays bottoming within a circle of 15 degrees radius from a point 8 degrees N, 75 degrees W are removed from the analysis, there is no indication of a detectable tilt of the symmetry axis away from the rotation axis.

The origin of this regional anomaly is, of course, of great interest, since none of the current scenarios of the origin and evolution of the inner core could begin to explain its existence.

It should be noted also that the principal evidence of Song and Richards (1996) for differential rotation of the inner core comes from rays traversing the local anomaly described here. Creager (1996, 1997) has recognized the significance of the steep lateral gradient of the residuals across Alaska and proposed a differential rotation rate significantly lower from those of Song and Richards (1996) or Su et al. (1996). Our new analysis of the temporal variation of the total pattern of the inner core anomalies implies a zero, or non-detectable, rotation rate.

[This paper concerns the use of ISC data on absolute arrival times of the PKIKP phase. Su Dziewonski and Jeanloz (Science, vol. 274, 1883--1887, 1996) had claimed a fast rate of inner core rotation using such data. Here, their method is acknowledged as implying "a zero, or non-detectable, rotation rate." ]

top of this page
to my main page on inner core rotation