Earth's Normal Modes Research Articles

New determinations of Q quality factors and eigenfrequencies for the whole set of singlets of the Earth's normal modes 0S0, 0S2, 0S3 and 2S1 using superconducting gravimeter data from the GGP network

Abstract We present new modal Q measurements of the 0 S 0 and 0 S 2 modes, and first modal frequencies and Q measurements of 2 S 1 and 0 S 3 modes. The high quality of the GGP (Global Geodynamics Project) superconducting gravimeters (SGs) contributes to the clear observation of seismic normal modes at frequencies lower than 1 mHz and offers a good opportunity for studying the behaviour of these modes. The interest of scientists in the gravest normal modes is due to the fact that they do contribute to a better knowledge of the density profile in the Earth, helping to constrain Earth's models. These modes have been clearly identified after some large recent events recorded with superconducting gravimeters, particularly well-suited for low-frequency investigations. The Ms = 8.1 (Mw = 8.4) Peruvian earthquake of June 2001 and the Ms = 9.0 (Mw = 9.3) Sumatra-Andaman earthquake of December 2004 provide us with individual spectra which exhibit a clear splitting of the spheroidal modes 0 S 2 , 0 S 3 and 2 S 1 , and a clear identification of each of the individual singlets, with a resolution never obtained from broad-band seismometers records. The Q quality factors have been determined from the apparent decrease of the amplitude of each singlet with time, according to a well-suited technique [Roult, G., Clevede, E., 2000. New refinements in attenuation measurements from free-oscillation and surface-wave observations. Phys. Earth Planet. Inter. 121, 1–37]. The results are compared to the theoretical frequencies and Q quality factors computed for the PREM and 1066A models, taking into account both rotation and ellipticity effects of the Earth. The two observed datasets (frequencies and Q quality factors) exhibit a splitting on the observed values different from the predicted one. That seems to point out that some parameters as density or attenuation values used in the theoretical models do not explain the observations. A new dataset of frequencies and Q quality factors of the whole set of singlets of the gravest spheroidal modes is thus under construction. That dataset includes the five individual singlets of the 0 S 2 mode clearly identified on the SG records, the three singlets of the 2 S 1 mode recently observed for the first time by [Rosat, S., Hinderer, J., Rivera, L., 2003b. First observation of 2 S 1 and study of the splitting of the football mode 0 S 2 after the June 2001 Peru earthquake of magnitude 8.4. Geophys. Res. Lett. 30, 21211, doi:10-1029/2003L018304], the 0 S 0 radial mode, and the seven individual singlets of the 0 S 3 mode.

On the resolution of density anomalies in the Earth's mantle using spectral fitting of normal-mode data

The resolution of the 3-D density structure in the Earth's deep interior has long eluded geoscientists. High-quality data from digital seismic instruments emplaced this decade have renewed interest in the measurement of low-frequency Earth normal modes with the goal of extracting heterogeneous density structure. Here we perform a series of synthetic experiments aimed at investigating the resolution of lateral variations in the mantle from normal-mode spectral data. Contamination effects between seismic velocities and density are examined in two ways: (1) by using resolution matrices computed from data kernels and (2) by inverting synthetic spectra computed from realistic input Earth models. The first type of experiment assumes that the non-linearity of the inverse problem is weak. No such assumptions are necessary in the second type of tests which, nevertheless concur in their results with the former. These synthetic tests indicate that density structure retrieved from presently available normal-mode data is not reliable. Contamination of seismic velocity structure into the density model space produces patterns that resemble those obtained from inversions of real data. The resulting density models appear to be dependent on a combination of the starting velocity models and the model parametrization.

Time domain analysis of Earth's long‐period background seismic radiation

A time domain detector of long‐period (200–400 s) surface‐wave energy is developed and used to investigate Earth's background seismic radiation. Several recent studies have presented convincing evidence for an incessant low‐level excitation of Earth's normal modes in the frequency band 2–7 mHz. The detection of these oscillations has been accomplished through analysis of low‐frequency amplitude spectra, which provides good resolution of the excited frequencies but poor resolution of temporal variations and signal amplitude. The new time‐domain method is based on the autocorrelation function. A seismogram is correlated with a copy of itself to which a reverse great‐circle dispersion operator has been applied. The resulting correlogram is labeled the “autogram”. If multiple‐orbit surface waves are present in the seismogram, their energy will be phase equalized and concentrated at zero lag time in the autogram, allowing for identification of coherent radiation. Analysis of vertical‐component seismograms recorded at quiet stations leads to nearly continuous detections of coherent seismic energy using correlation time windows of only a few hours duration. Stacking of autograms from several stations enhances the detections. Five years of vertical long‐period data from the Global Seismographic Network are analyzed. All detections with an amplitude equivalent to that of an earthquake of Mw≥6.0 can be correlated with known earthquakes; no evidence is found for the existence of anomalously slow or “silent” earthquakes in the period band 200–400 s. The background excitation level is remarkably constant, corresponding approximately to the maximum amplitude detection generated by a Mw = 5.75 earthquake. Coherent Rayleigh wave energy is detected 94% of the time analyzed. A seasonal component with a period of 6 months is well resolved, with maxima in mid‐January and mid‐July, consistent with the hypothesis that atmospheric processes are the main cause of the phenomenon. Variations in the background signal level on timescales of several days to weeks are also documented.

Theoretical free-oscillation spectra: the importance of wide band coupling

SUMMARY Illustrative calculations are presented of the effect of the coupling of large groups of Earth's normal modes in a realistic 3-D earth model. In previous work the effect of modal splitting and coupling has been treated either in the self-coupling approximation (SC) or in the group-coupling approximation (GC), also known as quasi-degenerate perturbation theory. In SC a mode is treated as isolated in the spectrum and the theory of degenerate splitting is used to calculate the effect of lateral heterogeneity, rotation and ellipticity on each individual modal multiplet. In GC a small number of modes close together in frequency are treated as a group and the effect of coupling within the group is included. Of course, SC can be considered as a special case of GC in which there is only a single modal multiplet in the group. In principle, modal spectra are affected by coupling among all modes. Given that explicit calculations can be performed only for a finite collection of modes, we are led to consider the question of how large the group of modes considered must be in order to obtain theoretical spectra of sufficient accuracy for a given purpose. The particular purpose that we have in mind is the use of modal spectra to refine 3-D models of earth structure. To do this we carry out calculations in which the results of SC and GC are compared with those of full coupling (FC), by which we mean coupling calculations including large collections of multiplets, for example all modes having frequencies less than a certain upper bound or cut-off frequency. The results indicate that SC and GC often represent a poor approximation to FC. The differences in modal spectra calculated using SC/GC and those obtained using FC are, in many cases, comparable to the differences between observed and theoretical spectra. We compare FC for all 140 spheroidal and toroidal modes up to 3 mHz with GC using the 25 pairs of modes defined by Resovsky & Ritzwoller (1998) and with GC in 33 subgroups of spheroidal and toroidal modes. We find that full coupling is needed to obtain the best synthetic spectra. The matrices for full coupling of 140 modes become large (approximately 2500×2500), but are still tractable.

Cause of continuous oscillations of the Earth

Spheroidal fundamental mode oscillations of the Earth for frequencies between 2 and 7 mHz (millihertz) are observed even on seismically quiet days. Two hypotheses of the cause of these oscillations are investigated: the cumulative effect of small earthquakes and atmospheric pressure variations. The cumulative effect of earthquakes, assuming that earthquakes follow the Gutenberg‐Richter law, is shown to be 1–2 orders of magnitude too small. The observed amplitudes of modes require an equivalent earthquake of magnitude 6.0 everyday, which cannot be achieved by summing up contributions from small earthquakes. The hypothesis of atmospheric excitation is favored because of the discovery of seasonal variations in stacked modal amplitudes for spheroidal modes between 0S20 and 0S40. It is also evaluated by comparing observed modal amplitudes with theoretical amplitudes, derived from a stochastic normal mode theory. The source of excitation is atmospheric pressure variations, which indicate turbulent motion of the atmosphere for the frequency range of interest and are estimated by barometer data. The observed modal amplitudes can be matched by the stochastic normal mode theory, indicating that atmospheric pressure variation is large enough to excite solid Earth normal modes up to the observed amplitudes. Therefore two lines of evidence, detection of seasonal variations and approximate match of overall modal amplitudes, support the hypothesis that the continuous background oscillations are excited by atmospheric pressure variations.

Measuring seismic normal modes with the GWR C021 superconducting gravimeter

In the framework of the global geodynamic project (CGP) that plans to study Earth's normal modes with superconducting gravimeters (SG), we present a study of the C021 SG located in Membach, Belgium, in the seismic normal modes band (periods shorter than 1 h). Our results are based on the spectra associated with the Irian Jaya (February 17, 1996, M W=7.9) and the Baleny Island (March 25, 1998, M W=8.1) earthquakes. We demonstrate that the signal to noise ratio (SNR) of the SG can be improved by a factor of 4 below 1.2 mHz using local barometric records [Zürn, W., Widmer, R., 1995. On noise reduction in vertical seismic records below 2 mHz using local barometric pressure. Geophys. Res. Lett. 22, 3537–3540]. A side-by-side comparison of a long period STS-1 V seismometer with the SG C021 confirms the low noise level of this latter instrument and demonstrates its ability to measure seismic normal modes. Several suggestions are given about the noise sources that could affect SGs in the seismic frequency band.

Energy Budget of Deep‐Focus Earthquakes Suggests They May Be Slip‐Sliding Away

To understand why the Bolivian earthquake of 9 June 1994 has so rapidly become one of the most thoroughly studied seismic events in history, one must know a little about both the event and the significance of when and where it occurred. (See PHYSICS TODAY, October 1994, page 17.) To begin with, the Bolivian quake occurred in the mantle, nearly 640 km beneath the Amazon basin east of the Andes. Such a large “deep‐focus” earthquake—and this one, with magnitude 8.3, was the largest such event ever recorded—can make the planet ring like a bell, exciting overtones of Earth's normal modes of oscillation that are sensitive to inhomogeneities in the core and mantle. Add in the fact that the Bolivian quake occurred after the deployment of several global and local networks of high‐quality digital seismological stations, and it becomes clear that the quake offered researchers an unprecedented opportunity to illuminate the three‐dimensional structure of Earth's mantle and core. (See the box at right.)

Terrestrial Tomography

To researchers who use Earth's normal modes of oscillation to map the threedimensional structure of the planet's interior, the good fortune of having the largest‐ever deep‐focus earthquake occur just after the deployment of global broadband seismographic networks seemed like a dream come true. It was certainly no dream, though, and computers have been crunching data ever since. Barbara Romanowicz, Xiang‐Dong Li and Joseph Durek (University of California, Berkeley) used normal modes excited by recent large deep‐focus quakes in Bolivia and elsewhere to characterize the anisotropy of Earth's inner core. Michael Ritzwoller and Joseph Resovsky (University of Colorado at Boulder) and Jeroen Tromp and Xiong He (Harvard University) used data from several recent deep‐focus earthquakes to show that normal‐mode studies can yield improvements in three‐dimensional models of Earth, especially for the planet's mantle

Could the energy near the FCN and the FICN be explained by luni-solar or atmospheric forcing?

SUMMARY The observed time-series of precession/nutation show residuals with respect to an empirical model based on the rigid Earth theoretical nutations and a frequency dependent transfer function with resonances to the Earth's normal modes. These residuals display energy mainly in the frequency domain around 430 and 500 days in the inertial frame. In this frequency band, the energy is possibly related to two normalmode frequencies: the free core nutation (FCN) and the free inner core nutation (FICN). In this paper, we examine the possibility of obtaining this energy from the resonance effect induced by a luni-solar (or planetary) forcing, or by an atmospheric forcing at a frequency very close to these Earth free nutations. The amplification factor due to the resonance is computed from an analytical formula expressed in the case of a simplified three-layer ellipsoidal rotating earth (with an elastic inner core, a liquid outer core and an elastic mantle), as well as the empirical formula based on the analysis of VLBI observations. For the tidal forcing, the theoretical results do not show any resonance at the level of precision we have examined but it is still possible to find one frequency near the FCN or FICN frequencies which could be excited. In contrast, for the atmospheric pressure the level of energy needed could be obtained from the diurnal pressure, depending on the noise level of the Earth's global pressure. We also show that the combination of three waves can explain the observed decrease of energy with time. While the tidal potential amplitudes are too small, a pressure noise level of 0.5 Pa would be sufficient to excite these waves.

Calculation of static deformation following the Bolivia Earthquake by summation of Earth's normal modes

The vertical ground displacement associated with the 9 June 1994 great Bolivia earthquake is calculated by summation of Earth's normal modes. Because of the great depth of the earthquake, the contributions to the surface displacement field from modes of low radial and angular orders predominate. Convergence is fast, and the truncated mode sum provides a good estimate of the total global displacement field. The vertical displacement is calculated to be on the order of 1 cm within a few hundred km of the epicenter, and 1 mm at distances as great as 1500 km. Normal mode seismograms are calculated for a location near the epicenter, and the static offset is seen to occur primarily between the P and S phases, consistent with the effect of a near‐field term. The predicted seismograms are in qualitative agreement with the recently reported observations of seismically recorded static displacements on the BANJO broadband array [Jiao et al., this issue]. The static offset calculated in this study is smaller than the observations and predictions reported by Jiao et al., probably owing to uncertainties in the observations and in their approximate calculations, which are based on dislocation theory in an elastic halfspace.

Asymptotic normal modes of the Earth-III. Fréchet kernel and group velocity

SUMMARY A comprehensive study of the Frechet kernels and group velocities of the Earth's normal modes is conducted based upon the asymptotic eigenfrequency and eigenfunction analyses that we developed previously. Two different approaches, which employ the eigenfunctions or eigenfrequency equations, yield asymptotically equivalent results for the Frechet kernels and group velocities. The latter approach is considerably simpler, since the need to specify the normal-mode eigenfunctions as well as their radial derivatives is removed, so that the Frechet kernels and group velocities can be obtained from knowledge of the asymptotic eigenfrequencies only. The asymptotic analyses are discussed for all possible ray parameter regimes and ray-path combinations within a crustless version of the earth model 1066A with two discontinuities: a core-mantle boundary and an inner core boundary. The exact and asymptotic numerical results for the Frechet kernels and group velocities are compared for such a model. The comparison shows that the asymptotic Frechet kernels and group velocities are very accurate. The accuracy is better for toroidal modes and relatively high-frequency spheroidal modes.

Surface-wave mode coupling for efficient forward modelling and inversion of body-wave phases

SUMMARY The importance of surface-wave mode coupling in the modelling of body-wave phases by surface-wave mode summation is studied by means of sensitivity kernels obtained with the Born approximation and exact solutions of the Invariant Imbedding Technique. It is shown that, independent of the character of the lateral heterogeneity, surface-wave mode coupling is required to model body-wave phase perturbations and that neglecting intermode coupling, as in the WKBJ method for surface waves, can lead to large biases. Because methods which describe surfacewave mode coupling in an exact fashion are computationally too expensive to use in inversion schemes, the Scalar Exponent Approximation (SEA) is presented, which is a computationally efficient method and takes mode coupling into account. Since, instead of Earth normal modes, surface-wave modes are used, the summation over the angular order l is carried out analytically. This means that the number of modes and mode interactions needed is significantly reduced which assures an efficient manner of modelling. It is shown that the SEA is accurate in modelling body-wave phase perturbations for geophysically realistic configurations. Because, in contrast to the WKBJ sensitivity kernels, mode coupling introduces sensitivity kernels which also depend on the position along the source-receiver path, the SEA requires a larger model parameter set in inversions. A procedure is presented which reorganizes the model parameter set and leads to a reduced set of physically relevant model parameters. The combination of the SEA and the reorganization of the model parameters can be used efficiently in large-scale 3-D inversions which incorporate the important effects of surface-wave mode coupling.

Support for anisotropy of the Earth's inner core from free oscillations

IN 1983, Poupinet et al.1 observed that compressional seismic waves traversing the inner core along a trajectory parallel to the Earth's rotation axis arrive faster than the same (PKIKP) waves travelling in the equatorial plane. They interpreted this observation as revealing prolate topography of the inner-core boundary. In 1986, Morelli et al.2 and Woodhouse et al.3 suggested that inner-core anisotropy could explain both the travel-time observations and the anomalous splitting of some of the Earth's normal modes. Inner-core anisotropy continues to be the preferred explanation for the travel-time anomalies, although there is disagreement about the magnitude of anisotropy4–6. More recent explanations for the anomalous splitting involve topography of the inner-core and core–mantle boundaries as well as lateral heterogeneity of the core7–11. In particular, Widmer et al.11 dismissed a rather complex recent model of inner-core anisotropy12 because it could not explain the splitting of several previously unidentified modes. Here I show that the anomalous splitting of all currently identified modes can in fact be explained by cylindrical anisotropy of the Earth's inner core that is also compatible with the observed PKIKP travel-time anomalies. The resulting model should be regarded as an upper limit to the amount of anisotropy, as lateral heterogeneity also undoubtedly contributes to the splitting.

Anomalous amplification of the Earth's normal modes near the epicenter due to lateral heterogeneity

The broadband seismograms recorded at stations near the epicenter of the Landers, California earthquake of 28 June 1992 exhibit observed amplitudes of the spheroidal modes of the Earth that are nearly one order of magnitude larger than those calculated for a spherically symmetric Earth model. We show that these anomalous observations at stations close to the epicenter can be explained by coupling between the normal modes due to the earth's asphericity. We calculate and sum fully coupled normal modes for three‐dimensional shear‐wave velocity models of the Earth's mantle to calculate synthetic seismograms. Our coupled‐mode synthetic seismograms explain the observed anomalously large amplitude of spheroidal modes and indicate that the coupling between the normal modes due to laterally heterogeneous elastic structure of the Earth causes this anomalous observation. Our results also show that the present three‐dimensional Earth models give qualitative fits to this observation.

Teleseismic Delay Times In A 3-D Earth and A New Look At theSDiscrepancy

SUMMARY When the earliest measurements of the Earth's normal modes became available, it was soon noticed that V, models satisfying this new data set were generally slower than models based on S traveltimes. This S discrepancy has since then been attributed to the influence of dispersion that accompanies attenuation, but it ignores a competing process that may have the same effect: in a heterogeneous earth, seismic rays will bend to take advantage of local high-velocity anomalies. This effect is difficult to quantify because of the computational problems associated with the determination of minimum traveltime paths in heterogeneous earth models in three dimensions. We have developed a new method to find absolute minimum traveltimes even in very complicated earth models and applied this to a large number of rays generated in different random earth models. The S discrepancy is a function of correlation length and amplitude of V, heterogeneity. A standard deviation u of 0.15 km sC', with a correlation length of 200 km in the whole mantle, would give a discrepancy even larger than observed at A > 60 and must be discarded. A u of 0.15 km s-' in the upper mantle and 0.05 km s-' elsewhere explains a discrepancy of 1-2 s, depending on the correlation length. Since the total discrepancy is estimated to be about 4 s, this still allows for effects of attenuation. We conclude that the S-wave discrepancy can be used to set limits to the heterogeneity in the lower mantle at short correlation lengths. With our present-day knowledge of the magnitude of the discrepancy a heterogeneity larger than 1 per cent with correlation length 200 km is unlikely.

Review of the Earth tidal models and contribution of Earth tides in geodynamics

The attraction of the Moon and the Sun on the Earth induces deformations of the Earth body. At the surface of the Earth, these effects are described using tidal parameters which are observed by tidal instruments and estimated using an Earth model. These tidal parameters are indeed computed from a linear combination of the displacement field, the stress field, and the potential due to the mass readjustment associated with the tides. The theoretical tidal parameter values depend crucially on the model used for the interior of the Earth. The system of equations used in the computation contains a stress‐strain relationship. This equation is the only one which changes when considering mantle inelasticity instead of elasticity. For an inelastic mantle, the stress‐strain relationship is a Hooke law with complex and frequency dependent rheological parameters. The numerical values of these rheological parameters depend on the hypotheses formulated on the Earth body. Even after the introduction of mantle inelasticity in the equations, there are still some differences between the theory and the observations. By increasing the complexity of the model, scientists want to decrease the difference between the theory and the observations. This comparison brings answers to very important points in geophysics. In particular, the resonances at the free core nutation frequency and at the free inner core nutation frequency in the tidal response give the values of these Earth normal mode periods and the comparison between the observed and the computed period constrains the core flattening and the jump in the density at the inner core‐outer core boundary, respectively.

Effect of a global plume distribution on Earth normal modes

We show that a global distribution of small scale heterogeneities in the lower mantle induces both frequency splitting and amplitude scatter of low angular order modes. These effects can be used in order to map the global distribution of small scale structures in the mantle, such as plumes. Two independent datasets, consisting respectively of lower mantle modes interaction coefficients and of observed amplitudes of core modes, can then be used to constrain the low angular degrees of the spatial distribution and the mean size of heterogeneities, allowing us to constrain the mantle structure at different scales. If these heterogeneities are assumed to be plumes associated with hotspot activity, the mean size of plumes obtained using the two datasets is then in good agreement with that obtained from geodetical or numerical convection studies.

Searching for slow and silent earthquakes using free oscillations

Our knowledge of Earth structure can be used to distinguish far‐field seismic signals from near‐field noise. In particular, earthquakes excite ground motion at the eigenfrequencies of the Earth's normal modes. By observing the spectral levels in small bands centered on the known frequencies of the normal mode resonance peaks and comparing them to the spectral levels of ambient noise, it is possible to detect earthquakes at low frequencies without knowing anything about them at high frequencies. We have used this fact to develop an earthquake detection algorithm that compares the signal level in the mode bands, as measured by the modulus of zeroth‐order moment of the Fourier spectrum, with that in the noise bands using an F test. The excitation of each mode over a global network of stations and of all the modes over the network is evaluated from a summed‐score statistic using a binomial test and a Markov test, respectively, which are robust with respect to deviations from the Gaussian noise model. We apply the technique to the Earth's fundamental spheroidal modes 0S8‐0S43 as recorded by the IDA network for the 2‐year interval 1978–1979. We calculate the level of excitation at 3‐hour intervals using a time domain integration technique designed to optimize the signal‐to‐noise ratio for each mode. The algorithm detects 50% of all teleseismic earthquakes having seismic moments M0 ≥ 1×1018 N m (Mw ≥ 6.0), and it detects all events having M0 ≥ 3×1018 N m (Mw ≥ 6.3). Slow earthquakes, i.e., those having anomalously large characteristic durations, are observed to be enriched in low‐frequency mode excitation relative to other events of comparable magnitude. Using this criterion, we have found a number of slow earthquakes that had not been previously identified as anomalous. Most of these newly identified slow earthquakes occurred on oceanic transform faults. In addition to the cataloged normal and slow earthquakes, we observe 27 episodes of substantial mode excitation that are reasonably well isolated from significant earthquakes. Such anomalies may represent “silent earthquakes,” events with propagation velocities sufficiently low that they do not generate globally detectable wave trains on high‐frequency seismometers.

Rupture characteristics of the 1982 Tonga and 1986 Kermadec earthquakes

Source time functions are obtained for the December 19, 1982, Tonga and October 20, 1986, Kermadec earthquakes from deconvolution of P waves, using a model including a significant dip of the ocean bottom. In such a geometry, the amplitudes of subsequent water multiples are large and irregular. As a result, we can model these two events as consisting of a single pulse of short duration (23 s in the case of the 1982 event, and 19 s for the 1986 one), despite complex teleseismic waveshapes. This is in contrast to inversions using flat‐layered models, which required very complex source time functions, which were mutually inconsistent at different stations. The anomalously large tsunami generated by the 1982 event is modeled as the result of propagation of the rupture into the sedimentary wedge at the trench, rather than due to a long‐lived complex source time function, the latter being also incompatible with the total lack of excitation of the Earth's normal modes in the low‐frequency limit. For the 1986 event, we derive a source duration (19 s) very short in relation to its moment (8×1027 dyn‐cm). This earthquake, a high stress drop rupture featuring a unique combination of strike‐slip and thrust motion on a plane dipping oceanwards of the trench, is interpreted as expressing internal deformation of the subducting Pacific slab at a depth of 40 km, under the compressional stresses resulting from the collision.

Effect of sharp lateral heterogeneity on the Earth's normal modes

When inverting normal mode data for global large-scale lateral heterogeneity, possible biases due to small-scale structure are commonly ignored. We conducted two experiments. In the first one, we calculated, with a recently developed first order scattering formalism, the effect of a simplified subduction zone model on the eigenfrequencies of the fundamental spheroidal mode branch. The frequency shifts induced by this single subduction zone appear to be smaller by only a factor of 3 to 4 than shifts observed on long-period seismograms, or than shifts induced by existing models of global upper mantle inhomogeneity. The second experiment involved a model of a hot spot plume on the core mantle boundary. Solving the variational problem with many coupling terms included, we calculated the effect of such a plume on the amplitude of some low angular order modes. The results suggest an effect at least as large as that due to available large scale mantle models. It thus appears, that relatively small-scale, sharp lateral structure in different depth ranges of the Earth may have an important effect on normal mode observations and that these effects may not be ignored in inversions of normal mode data.