T54A-01 INVITED 16:00h
The Anisotropy of the Inner Core: A Review
The anisotropy of the inner core has been identified about 20 years ago from both body wave propagation and eigenmode splitting. The P-wave velocity is about 3% higher for waves travelling parallel to Earth rotation axis than for waves travelling parallel to equatorial plane. Since this discovery, many controversial results about the inner core anisotropy properties have been published, and even the existence of anisotropy itself has been questioned. In this review, I will present and discuss the main results concerning the inner core anisotropy properties, in particular: 1) The radial distribution of the anisotropy, with the absence of anisotropy at the top and the probable existence of a deep inner-core with a different anisotropy orientation; 2) the problematic longitudinal pattern of the inner core anisotropy; 3) the possibility of an anisotropy in attenuation. The possible artefacts due to mantle heterogeneities and to liquid core will be discussed, as well as the compatibility between eigenmodes and body waves. Alternative explanations with inner core heterogeneities will also be considered. Finally, the resolution of the models will be discussed, as well as the modification of the radial and anisotropic mean models when P-wave sensitivity kernels are taken into account.
T54A-02 INVITED 16:15h
What do Seismic Observations tell us About the Origin of Elastic Anisotropy in the Inner Core?
Three observations have guided the search for the origin of elastic anisotropy in the inner core. The most robust observation is that the orientation of anisotropy is aligned with the rotation axis of the Earth. This result suggests that the anisotropy is controlled by processes in the fluid outer core because rotation is expected to have a strong influence on fluid flow, heat flow and the structure of the magnetic field. A variety of mechanisms that depend on conditions in the outer core have been proposed to explain the elastic anisotropy. However, it is not presently clear whether any of these mechanisms are sufficient to produce the required alignment of crystals. Alternative explanations rely on the rotation of the inner core relative to the forcing that causes crystals to align. For example, periodic gravitational forcing from either the mantle or tides could conceivable produce a cylindrical fabric when the deformation is averaged over many revolutions. External control over the orientation of anisotropy raises serious problems when attempting to explain the hemispherical variations in travel-time anomalies. While it possible to freeze-in local variations in the degree of crystal alignment during solidification of the inner core, it is unlikely that these variations would persist in the same geographic locations over the age of the inner core. Fluctuations in the outer core and gradual rotation to the inner core should eliminate longitudinal variations in the source of anisotropy. The most plausible explanation for hemispherical variations involves processes internal to the inner core that provide a positive feedback on the development of anisotropy. The effects of crystal alignment on rheology or thermal conductivity in the inner core can alter the deformation or thermal structure, producing a feedback on the development of anisotropy. The final observation concerns the depth dependence of anisotropy in the inner core. A gradual increase in anisotropy with depth is expected because strain must accumulate before crystal alignment develops. The apparent lack of anisotropy from the top of the inner core suggests that the solidification texture is weak, and that crystals are buried by inner-core growth before substantial strain accumulates. Abrupt changes in the strength of anisotropy with depth are more difficult to explain, but provide important constraints if confirmed by observations. Recent suggestions of changes in the orientation of anisotropy with depth may indicate that the inner core is capable of true polar wander.
T54A-03 INVITED 16:30h
Elastic anisotropy and texture of iron at high pressure
Understanding the elastic anisotropy and texture of relevant core materials at high pressure is crucial to interpreting seismic observations of inner core anisotropy. Conventional measurements of elastic moduli require large, single crystals which have not yet been synthesized for hcp iron. However, with the development of new synchrotron techniques, we are able to get equivalent information via the combination of texture studies and directional phonon measurements. Radial diffraction conducted at Sector 16 of the Advanced Photon Source, Argonne National Laboratory, was used to determine the preferred orientation in polycrystalline iron samples under the presence of differential stress. Here a diamond anvil cell with a large ($150\deg$) side opening and x-ray transparent gasket (e.g. beryllium or amorphous boron) allowed access to the complete strain field. The intensity variations around the resulting Debye-Scherrer rings for different reflections were used to determine pole figures. We then conducted inelastic x-ray studies at BL35XU of SPring-8 on the same, textured iron samples at high pressure and measured phonon energy as a function of momentum transfer. The initial slope of that curve was used to calculate the acoustic velocities, allowing determination of elastic anisotropy in a well-characterized textured hcp iron sample.
T54A-04 16:45h
Transition From Isotropy to Anisotropy in the Upper Inner Core: the Role of Anisotropic Thermal Conductivity and Horizontal Convection
We investigate models of the radial structure of the western hemisphere of Earth\'s inner core. Differential travel times of PKiKP $-$ PKIKP indicate that the outermost inner core is isotropic. On the other hand, observed travel-time residuals of PKP$_{BC}$ $-$ PKIKP and PKP$_{AB}$ $-$ PKIKP increase systematically from 1 to 6 s as a function of increasing ray turning depths for ray paths that are parallel to Earth\'s spin axis. Rays perpendicular to the spin axis typically have slightly negative residuals. These observations suggest the outermost inner core is nearly isotropic and that strong anisotropy exists deeper in the inner core. We invert these times for models characterized by an outer isotropic layer and a deeper anisotropic layer separated by a transition zone with thickness varying from 0 to 150 km. Because of the strong velocity gradients, anisotropic ray tracing through the inner core is required to infer accurate models. We find that models with an isotropic layer ranging from 150 to 300 km thick all adequately fit the travel-time data. Models with discontinuities and linear gradients up to 150 km thick cannot be distinguished by travel times alone. On the other hand, synthetic seismograms demonstrate that amplitudes of direct and reflected arrivals are very sensitive to the width of the transition zone. Robust stacks of waveforms across regional arrays do not exhibit coherent arrivals, suggesting either that the transition is broad, or that sharp transitions are not coherent over large spatial scales. The elastic anisotropy is most likely caused by the systematic alignment of crystals. If the thermal conductivity of iron is anisotropic then within regions of well-aligned crystals, conductive heat flux will be direction dependent. Simple models show that 10 percent anisotropy in thermal conductivity in the lower inner core, below a 200-km thick isotropic upper inner core, would produce latitudinal temperature differences at the base of the isotropic layer of about 2 degrees K. Scaling inner core parameters to the analogous problem of horizontal convection(flow driven by variations in heat flux along a horizontal boundary), which has been studied in the laboratory and numerically, suggest that this mechanism may play an important role in aligning iron crystals. For example, scaling laws suggest the overturn time is of the order 500 Ma.
T54A-05 16:57h
Is There an Innermost Inner Core?
Ishii and Dziewonski (2002) suggests that the inner core has a simple constant anisotropy except the innermost inner core (IMIC, about 300 km or so in radius), which possesses a different form of anisotropy. The primary data constraining the innermost inner core are PKP(DF) arrival times from the ISC. Although abundant, absolute arrival times have large picking errors and are subject to contamination of heterogeneous upper mantle. There are also strong evidence for significant lateral variations in anisotropy in at least top 500 km of the inner core. Because of the extremely small volume of the innermost inner core, a relatively small uncorrected structure of mantle and the rest of the inner core can be falsely mapped into this region. Here we examine this issue using a large collection of high-quality differential PKP AB-DF and BC-DF travel times. We examine our data in three different ways and our preliminary results are generally consistent: the form of anisotropy in the IMIC indeed seems different. However, our biggest difference between the IMIC and the shallower depth is at the EW direction where the IMIC is faster by about 3% in the innermost 400 km or so; near $45\deg$ from the EW direction, where Ishii and Dziewonski (2002) shows the biggest difference, our velocity is comparable. The three different data processing methods used are as follows. (1) We invert for 3D anisotropic structure by dividing the inner core into layers and sectors. (2) We correct the AB-DF data at distances 173 to $180\deg$ for the average anisotropic models inferred from the data at slightly smaller distances. (3) We construct the ''travel-time curves'' (differential PKP residual vs. distance) of equatorial paths with similar ray directions (within 10 or $15\deg$). Because the lateral variation of the inner core velocity at equatorial directions is small (except perhaps the top 100 km of the inner core), the velocity change with depth would be indicated directly by the change of the curvature of the travel-time curve. We observe an change in the slope of the travel-time curve at distance greater than about $160\deg$ for rays with less than $10\deg$ from the EW direction, a smaller change in the slope for rays around $20\deg$ from the EW direction, and little change in the slope for rays around $40\deg$ from the EW direction, suggesting a change in anisotropy in the IMIC, rather than the inaccuracy of the 1D radial reference model. So far we have not found evidence of waveform triplication for EW paths at near antipodal distances.
T54A-06 17:09h
Structure of the Deep Inner Core From Antipodal PKPPKP Waves
For about two decades, seismic studies have demonstrated that the inner core is acoustically anisotropic. However, until the present day the amplitude and radial dependence of inner core anisotropy remain unknown and somewhat controversial. Apart from previously observed complexity in geometry, some recent results suggest changes in the anisotropic properties in the deepest part of the inner core, close to the planetary center. While there is a large number of seismological studies focusing on the uppermost part of the inner core, there is a disproportionate number of studies concerning the bulk of the inner core, especially its center. One reason for this is highly the attenuative nature of the inner core for compressional waves. Another reason lies in the inadequate sampling of the inner core by PKP waves, whose travel times are traditionally used to study inner core properties. In order to sample the central regions of the inner core, PKP waves must be nearly antipodal, and with the spatial distribution of large earthquakes and current configuration of seismographic stations worldwide, this is difficult to achieve, except for paths nearly parallel to the equatorial plane. Thus, the center of the inner core remains unsampled by near-polar PKP paths, which makes interpretation about anisotropic properties near the planet's center, at minimum, very challenging. Here we present our efforts to study the center of the inner core using near-antipodal PKPPKP waves. These phases are generated by large earthquakes or explosions and travel through the inner core, reflect from the free surface of the Earth and travel back through the inner core to a recording station near the source. Our preliminary results show that these waves are extremely difficult to observe at very short epicentral distances (less than 10 degrees). However every new observation is precious, as it represents unique spatial sampling of the inner core. We will present findings from our search on a global scale, utilizing both earthquakes and explosions and analyses in time and frequency domains, from both individual and array records. We will discuss our findings in light of current perception of inner core anisotropy and the physical state of the inner core in general.
T54A-07 17:24h
Inner Core Anisotropy as Seen With Normal Mode Data and the Neighbourhood Algorithm
A model space search technique applied to core-sensitive normal mode splitting measurements reveals a robust pattern of P-wave and S-wave anisotropy in the inner core. In the classical inverse approach, regularization and normal mode data quality have a significant influence on the final model, whereas the chosen mantle model was less crucial. This is the most likely explanation for discrepancies among existing seismological models of inner core anisotropy. We employed a forward modelling approach, the neighbourhood algorithm, to obtain all possible models of inner core anisotropy compatible with carefully chosen free oscillation data. Our models reveal a relatively complex radial dependence of the anisotropy, with a likely change of sign in the parameters in the deepest part of the inner core. The parameter describing P-wave anisotropy changes sign around 800~km depth, while S-wave anisotropy is small in the upper two-thirds of the inner core and becomes negative at greater depths. Our results are in agreement with all observed travel-time anomalies of rays travelling through the inner core and are the first models derived from normal mode data alone that have significant anisotropy in the deepest part of the inner core. They predict between four and six seconds of travel-time anomalies for rays travelling at epicentral distances of $170^\circ$. Possible interpretations of these models include a phase change of iron in the centre of the core.
T54A-08 17:36h
Constraints on inner core structure from normal mode coupling
Since early last century it has been known that the Earth has an inner core and normal mode observations provided first seismological evidence for its solidity in the early 1970's. During the last two decades, observations of anomalous splitting of core sensitive modes has been interpreted in terms of a significant amount of anisotropy in the inner core. Thus, normal mode data, in addition to body wave data, place important comstraints on inner core structure and have been used to make fundamental discoveries of different types of inner core structure. We study the anisotropic, attenuation and shear wave structure of the inner core using normal mode data. Inner core structure has traditionally been studied using splitting of isolated modes. Here we extend this method to include cross-coupling between core sensitive modes that are closely spaced in frequency. We find that several inner core modes are strongly coupled due to degree two structure, in particular mantle shear wave velocity and inner core anisotropy. This provides new constraints on models of inner core anisotropy and implies that we cannot ignore cross-coupling for core sensitive modes. Certain closely spaced pairs of core modes are also very sentive to small changes in inner core shear wave velocity. A small change in inner core shear wave velocity, within the permissable bounds of normal mode data and the observation of PKJKP, makes certain modes visible that were previously identified as `high Q modes', for example $_{10}S_2$ and $_{11}S_2$. Thus, there is no need for a very high Q in the inner core as has been suggested in several previous studies and review papers.
T54A-09 17:48h
Identification of PKJKP using data from a Broad-band Seismic Array
The solidity of the inner core, first proposed by Inge Lehmann in 1936, was confirmed using seismic normal mode observations more than three decades ago. Since then, significant progress has been made in documenting the sharpness of the Inner Core Boundary (ICB), and in the inner core, the existence of P-wave anisotropy, as well as various features of the P-wave velocity and attenuation and some constraints on the solid-state mineralogical texture. However, direct evidence for the solid inner core, PKJKP (traversing the inner core as a shear wave), is still a topic of debate. Very few studies have addressed this issue, and each of them has generated some level of skepticism about the reliability of the observations. Using stacked broadband records from the high quality Graeffenberg network in Germany, we here document the observation of a high signal to noise phase on the vertical component, whose arrival time and slowness are in agreement with predictions for PKJKP from the seismic reference model PREM. The high quality of the observed waveform provides us a unique opportunity to verify whether what we observed is a real PKJKP phase. For this we use two distinct approaches: (1) comparison with unambiguous synthetic PKJKP phases without any interference from mantle phases. The resulting vespagram is very compatible with the observed one. (2) waveform modeling. Taking both solid and liquid inner core into account, we discuss the possible interfering phases (including unidentifiable mantle multiples) in detail. Our results show that the most likely influence on our observed PKJKP is from a mantle multiple. However, its slowness is positive, whereas the slowness of our observed PKJKP is negative, and so it can clearly be distinguihsed from PKJKP. This observed PKJKP waveform also allows us to estimate $Q_\beta$ in the central part of the inner core. We find a somewhat higher value than that of PREM, indicating an increase of $Q_\beta$ with depth in the inner-core, in agreement with what is generally observed for $Q_\alpha$.