You are on page 1of 14

The Frequency-dependence of Q and its Measurement at Different Periods

R. Brian White
Department of Earth and Planetary Sciences; Washington University; St. Louis, MO, USA

Abstract. A good measure of the Earth’s anelasticity is the quality factor Q, which can be determined through a number
of methods including seismology, tidal observations, and irregularities in the Earth’s rotation. These types of
deformation of the Earth span a wide timescale from a fraction of a second to many years. Since the forcing
mechanisms for most of these excitations of the Earth sample the planet in a similar way, we expect them to be
sensing the same bulk properties, including Q. Thus, by finding Q with different methods and at different time
scales, we can constrain the frequency-dependence of Q and the Earth’s bulk properties over many decades of
frequency variation. It is assumed that this frequency-dependence is linear and follows a relation of the form Q -1
α
= ω . Most values of α quoted in the literature range from 0.1-0.4. In the past, atmospheric and oceanic
effects could not be separated from tidal observations, skewing the Q values found with that method, but recent
advancements in satellite laser ranging give reasonable Q values more consistent with seismic and wobble
observations. This also offers a promising way to measure Q for other planets when our probes visit them.

1. Introduction
Q is a dimensionless quantity related to the anelasticity of a material. It goes by many names such as

“attenuation” in seismology, “dissipation” in geodesy/astronomy, and “quality factor” in electronics, which is where the

symbol Q is derived. Whatever it is called, Q is a measure of the energy lost per cycle of deformation of a material. It is

inversely proportional to energy, meaning that lower Q values describe higher energy dissipation. A fundamental

definition of Q is given by [e.g. Goldreich and Soter, 1966]

1 dE 1 DE
Q- 1 = -
2pE 0
ò dt dt = - 2p E0
(1)

where -∆ E is the energy lost per cycle due to the anelasticity of the material, and E0 is the peak strain energy stored in the

volume. This assumes a linear process; meaning Q is independent of the amplitude of the excitation source. Q can be

measured at many different periods to probe the composition, state, and structure of the interior of the Earth and other

planets. Understanding its frequency dependence is a much sought-after goal because it will allow for extrapolation of

measured Q values at one period to others.


2

We can understand how Q relates to the volumetric rate of energy dissipation in a material by considering a

stress and strain during a cycle of forced oscillation at a given frequency. The work down against internal friction

within a unit volume of the solid during a full cycle of oscillation is the area of the hysteresis loop (Figure 1). From

consideration of such a stress-strain relationship, the following is true [Dahlen and Tromp, 1998]

Q - 1(w ) = tan f (w ) (2)

Here, φ (ω ) is the frequency-dependent phase lag of the stress behind the strain in the loop. Thus, the more anelastic a

material is, the more open the loop becomes.

Figure 1. Trajectory of stress versus strain during a harmonic cycle of forced oscillation. Dashed lines
represent relaxed and unrelaxed perfectly elastic behavior [after Dahlen and Tromp, 1998].

Since Q is related to the attenuation of energy, it can be written in many forms, depending on the relaxation

mechanism. Several constitutive relations have been suggested to describe the recovery from deformation associated with

such dissipative processes, including the modified Lommitz law [Jeffreys, 1972] and absorption band models [Liu et al.,

1976]. One general expression that is valid for thermoelastic, diffusion, dislocation, and viscous grain boundary relaxation

mechanisms is [Jackson, 1969]

é wt ù
Q - 1(w ) = Aê 2 2ú (3)
ë1 + w t û

Here, A is amplitude,  is relaxation time, and  is frequency. Note that, here, Q is frequency-dependent. For a thermally-

activated relaxation process, the characteristic relaxation time varies with temperature according to [Jackson and

Anderson, 1970]

é E* ù
τ = τ 0 exp ê ú (4)
ë RT û

where 0 is a constant related to the lattice vibration frequency, E* is the activation energy, P is pressure, T is

temperature, and R is the universal gas constant. The dominant mechanism of attenuation in solids at temperatures greater
3

than about half the melting point increases exponentially with temperature and has a form given by [Anderson and

Minster, 1979]

-a -a
é a E* ù
Q - 1(w ) = A (wt ) = A (wt 0 ) exp ê- ú (5)
ë RT û

The constant is generally a value between 1/4 and 1/2, and the quantity E* can be thought of as an apparent activation

energy. If E* is known from creep or diffusion experiments,  can be determined if one knows the temperature and

frequency dependence of Q. Anderson and Minster [1979] showed that this expression is equivalent in form

to the equation of transient creep.

-a -a
é a E* ù
ε( t ) = B( t t ) = B(t t 0 ) exp ê- ú (6)
ë RT û

where B is some amplitude. This formulation is valid for short time scales or small strain

situations, such as seismic wave attenuation, tidal friction, damping of the Chandler wobble, and

post glacial rebound. For most applications, the frequency dependence of Q is simplified to an

expression of the form.

Q - 1(w ) = w a (7)

Thus, the constant α is the key factor needed to understand the frequency dependence of Q. Figure 2 illustrates the wide

variety of time scales associated with deformation of the Earth.

Figure 2. Relation of the characteristic time scale for several geodynamic phenomena to the relaxation
time of the upper mantle. Long time scale phenomena are governed by a higher viscosity ν 1 than short
4

time scale events which are governed by a viscosity ν 2 [after Yuen and Peltier, 1982].

2. Motivation

Lagus and Anderson [1968] showed that the elastic potential energy stored with depth is very similar for both

excitation from the 0S2 mode (period 54 minutes) and the semi-diurnal tide (period 12.42 hours), meaning these two

perturbations sample the Earth in much the same way (Figure 2). Other modes such as 1S11 also sample the Earth in a

similar fashion [Anderson and Minster, 1979].

Figure 3. Elastic potential energy Fréchet kernel vs. depth in the Earth’s mantle for the 0S2 mode (A)
and the semi-diurnal tide (B) [after Lagus and Anderson, 1968].

Work by Jeffreys [1972] and Smith and Dahlen [1981] further showed that the Chandler wobble (period 435 days) also

samples the Earth the same way (Figure 4).


5

Figure 4. Comparison of the Chandler wobble’s shear modulus Fréchet kernel in the mantle (solid)
with the corresponding kernel for the 0S2 mode [after Smith and Dahlen, 1981].

The similarity of the Fréchet kernels for the fundamental spheroidal modes, solid body tides, and the Chandler wobble

tells us that Q measurements derived from each should be measuring the same thing and that any differences among them

can be ascribed to either the intrinsic frequency dependence of Q or other dissipative factors such as atmospheric or

oceanic effects, which are not excited by seismic sources. We will now consider some of the different methods and

frequencies at which Q can be measured for the Earth or other planets.

3. Q-SAMPLING OF THE EARTH BY DIFFERENT MECHANISMS

3.1 Seismic Q

The attenuation and dispersion of seismic energy in the Earth is a direct measure of anelasticity. With the

widespread availability of seismic data, the most extensive body of research on the Earth’s Q has been in this field.

Traditional seismology, including long-period gravimetry, can probe periods of fractions of a second for body waves up to

54 minutes for the Earth’s gravest mode of free oscillation 0S2 mode. Thus, the seismic band tells us about the relatively

short-term response of the Earth to perturbations from seismic waves.

Historically, it has been assumed the Earth is elastic at seismic timescales, meaning the elastic moduli and

seismic velocities are independent of frequency. However, since waves do attenuate and free oscillations decay with time,

we know the Earth is not a perfectly elastic body, and these moduli should vary with frequency. Having frequency-

dependent velocities and moduli mean these quantities are complex. One way to define Q for seismic waves is [Stein and

Wysession, 2002]
6

2w * 2c*
Q - 1(w ) = = (8)
w c

Here, we can see that Q-1 is simply a ratio of the complex to real parts of frequency ω or velocity c. We can also write Q-1

in terms of complex shear and bulk moduli µ and Κ ,

m*
Q -m1 = = Q -s 1 (9)
m

-1 2 2 -1
K * Q p - ( 4b ) ( 3a ) Q m
Q -K1 = = (10)
K 1 - ( 4b 2 ) ( 3a 2 )

Where α and β are the P- and S-wave velocities with corresponding Q p for P waves and Qs for S waves. Anderson et al.

[1965] showed that losses in compression are negligible compared to losses in shear, meaning QK is very small. This

explains why Qp is generally much higher than Qs. Thus, the Q often reported in seismology is Qs.

Since body waves have short periods, observations of their travel times can give us a higher resolution view into

the Earth than other methods. Measuring Q with body waves can be problematic because they are prone to scattering,

reflection, refraction, and source-time errors. Lateral heterogeneities in attenuation, including passage through the low

velocity zone of the upper mantle, further complicate the analysis. However, some body wave phases are amenable to the

study of Q, most notably reflections from the core-mantle boundary. The observation of near-vertical multiple ScS phases

provides a simple and robust measure of the average shear wave attenuation in the mantle. Since the rays travel along

similar paths, the relative uncertainty due to the source, instrument, scattering, or spreading is diminished. The average Q s

≈ 285 found this way is valid from about 10-50 s [Anderson and Hart, 1978; Sipkin and Jordan, 1979]. Similar work

with PcP and PcS phases has helped constrain the velocity and density contrast at the core-mantle boundary and gives Q p

≈ 650 for the mantle [Bolt and Canas, 1985].

Surface waves (T ≈ 10-200 s) and normal modes (T ≈ 50-54 min) probe much longer periods than body waves

and offer more targeted studies of Q since the modes sample a variety of depths within the Earth. Analyses of the decay of

the Earth’s free oscillations have supplied numerous estimates of Q at many periods. This is usually done by spectral

width δ ω o (at half power) of the wobble’s frequency ω o.

dw o
Q- 1 = (12)
wo

where δ ω o is the spectral line width (at half power) of the wobble’s frequency ω o.
7

Often gravimeters and strainmeters are used rather than seismometers for the longest period modes [Agnew et al.,

1976; Agnew, 1981; Hansen, 1982]. Assuming Q is independent of frequency within the relatively narrow seismic band,

Q data can be interpreted in terms of Q which increases with depth within the Earth’s mantle. The following is a recent

compilation of normal mode and surface wave data.

Figure 5. How mode and surface wave observations tell us about the depth-dependence of Q within
the Earth’s mantle. Periods less than 100 s use surface waves, and greater than 100 s uses normal
modes. Different symbols correspond to contributions from different authors [new Reference Earth
Model website, http://mahi.ucsd.edu/Gabi/rem.html].

We can see how modes that sample the upper mantle have a much lower Q than those in the lower mantle, which

appears to have a more constant Q. Since different wavelengths sample the Earth at different depths and since Q is so

sensitive to temperature, pressure, and phase changes which may mask any intrinsic frequency-dependence, these

variations of Q with period here can be attributed to Q’s depth-dependence, assuming the Earth is intrinsically frequency

independent. Data like these have been combined with body wave observations to determine Earth Q models, such as that

used in PREM [Dziewonski and Anderson, 1981] or QL6 [Durek and Ekström, 1996]. A common to feature in all such

models is a low Q asthenosphere and relatively constant or slightly increasing Q within the lower mantle. Q is also low

again at the core-mantle boundary, suggesting it, like the asthenosphere, is close to the solidus.
8

Figure 6. Shear wave attenuation model QL6, which is found using long-period (150 < T < 300 sec)
Rayleigh- and Love-wave attenuation and an extensive set of mode decay observations [Durek and
Ekström, 1996].

3.2 The Chandler Wobble

Unlike the other free oscillations of the Earth, the Chandler wobble (CW) cannot be observed using seismology.

Instead, it is measured using geodetic techniques that include very long baseline interferometry (VLBI), satellite and lunar

laser ranging, and remote sensing satellites such as TOPEX/POSEIDON. If the Earth were perfectly rigid, the CW would

be its only free oscillation [Stacey, 1992]. It is the periodic rotation of the Earth’s rotation axis in inertial space about the

axis of greatest principal moment of inertia. It is equivalent to the free nutation of a spinning top and manifests itself as a

variation in the latitude of astronomical observations and length of day. The CW has an observed period of about 435

sidereal days [Smith and Dahlen, 1981], which is due to a combination of effects (see Figure 7). The period and Q of the

CW, like those of any free oscillation, depend upon the Earth’s physical properties and can tell us about the response of

the Earth at this 14 month period.

Figure 7. To explain the discrepancy of the CW’s observed period 435.2 days with that for a rigid
9

Earth 304.4 days, one must take into account the effects of elasticity, the fluid core, the oceans, and
mantle anelasticity [Smith and Dahlen, 1981].

Figure 8A illustrates the Chandler wobble. First, let’s consider a rigid Earth rotating on an axis ω with angular

momentum vector H. The CW is a counter-clockwise precession of the principal moment of inertial axis C about H,

which is fixed in inertial space. The rotation axis ω will also precess about H in such a way so the three axes C, H, ω

are coplanar, but this effect is so small that it is usually ignored [Smith and Dahlen, 1981]. If the Earth were perfectly

rigid and axisymmetric, the intersection of H with the Earth’s surface would be a circle. However, attenuation curves this

path into an inward spiral, and asymmetries cause the path to be elliptical (Figure 8B) [Smith and Dahlen, 1981].

In the real Earth, there are separate H’s and ’s for the core, mantle, crust, oceans, and atmospheres. The

excitation mechanism of the Chandler wobble remains a mystery, but it is probably due to mass and angular momentum

redistributions in the atmosphere and oceans. As can be seen in Figure 8B, the wobble is not perfectly regular, so the

irregularities must be due to the excitation mechanism [Stacey, 1992]. A measure of the CW’s irregularity is Q, which can

be defined the same way as Equation 12 for the other normal modes [Stacey, 1992]. Estimates of Q from the Chandler

wobble fall in the range of 50-300 [Anderson and Minster, 1979; Smith and Dahlen, 1981; Dickman and Nam, 1998].

Polar Motion: Jan 1995 - Jan 2001

0.7

0.6

0.5
Jan 01
0.4 Jan 95

y 0.3
(arcsec)

0.2

0.1

0
-0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4
x (arcsec)

A B

Figure 8. (A) The CW appears as a counter-clockwise precession of the C axis about H. (B) Path of
the pole from 1995 to 2001, as seen with GPS and VLBI [IERS EOP website,
http://hpiers.obspm.fr/eop-pc].
10

3.3 Tidal Friction

Determination of Q at diurnal (24 hr) and semi-diurnal (12 hr) periods has been notoriously difficult to measure

[see discussion in Wahr and Bergen, 1986]. Gravimetric or geodetic observations of the solid Earth tides can be used to

constrain the behavior of Q within the 4 decade gap between the seismic band and the Chandler wobble. In addition to

mean density and moment of inertia, tidal friction is the only other physical property which can be estimated for planets

and their satellites by the study of their orbits. In the absence of oceans or atmospheres, the tidal energy loss can be

attributed to physical dissipation of energy within the planet and is a measure of the viscosity and temperature of the

interior.

Tidal friction results in an imperfect response of the solid Earth to tidal forces, and hence yields a non-

equilibrium tidal bulge. The tidal bulge exerts a secular torque causing a transfer of angular momentum and energy

between the satellite and planet. The excess spin energy generated this way is dissipated as heat within the body’s interior.

When Ω > n (satellite is above the synchronous height), the tidal bulge is carried ahead of the satellite by an

angle ε (Figure 9A). The semimajor axis a of the satellite will increase while the rotation rate of the planet will decrease.

This is the case for the Moon and Earth, with a lunar retreat of 3.8 cm/yr [Dickey et al., 1994]. When Ω < n (satellite is

below the synchronous height), the tidal bulge lags behind of the satellite by an angle ε (Figure 9B). The semimajor axis

a of the satellite will decrease while the rotation rate of the planet will increase. This is the case for Phobos, which will

someday fall into Mars [Smith and Born, 1976].

Figure 9. A satellite with mass m in circular orbit with mean motion n about a planet of mass M and
rotation rate Ω raises a diurnal tide with frequency 2(Ω -n) on the planet [after Murray and Dermott,
2001].

The tide raised on a perfectly elastic planet would follow the line of centers, meaning there would be no lag in
11

the tidal bulge (  = 0). Thus, knowledge of  can provide a direct measure of the planet’s anelasticity. The relation of

Q to this lag angle is [MacDonald, 1964]

Q - 1 = tan ( 2e ) (13)

Notice the similarity of this expression to that for the stress-strain hysteresis loop in Equation 2.

In the past, terrestrial tidal Q investigations were hampered by the inability to account for atmospheric or oceanic

effects [Lagus and Anderson, 1968]. Goldreich and Soter [1966] made use of secular changes in the orbits of satellites

and planets to estimate Q for bodies in the solar system. They found Q values of 10-500 for the inner planets and the

major satellites of the outer planets and Q’s on the order of 104 for the gas giants, which is expected because fluids have

very little dissipation of energy. However, the small Q values for the some of the terrestrial planets (especially Earth) are

dubious due to atmospheric and oceanic effects. Previous astronomical observations gave a tidal bulge phase lag of ε =

2.16°, which tells us Q for the Earth is about 13 from Equation 13. Smith and Jungels [1970] used strainmeters to

determine ε = 2.9°and Q = 10 and concluded that their results were entirely due to the ocean and atmosphere and has no

bearing on the anelasticity of the solid Earth. For many years, knowledge of the Earth’s solid body Q at tidal timescales

was unknown.

Advances in space geodetic technology such as satellite and lunar laser ranging (SLR and LLR), along with

direct measurement of ocean tides using satellite radar altimetry, have permitted direct detection of the effects of the tidal

friction phenomena within the Earth-Moon system. This can be done because orbit perturbations are sensitive to the

gravitational effect of the ocean and Earth tides while altimetry can measure the ocean tides independently [Ray et al.,

2001]. Ray et al. [1996; 2001] used satellite tracking and altimetry data to obtain ε = 0.20° and Q = 280 for the lunar

semi-diurnal tide. This is in general agreement with Wahr and Bergen’s [1986] theoretical estimate of ε = 0.24° and Q =

235, VLBI observations of ε = 0.29° and Q = 200 [Herring and Dong, 1994; Schuh and Haas, 1998], ε = 0.38° and Q =

150 from gravimeter data [Melchior, 1989], and values of ε = 0.21° and Q = 270 deduced from the Chandler wobble

[Zschau, 1986]. Thus tidal observations can also help constrain Q in the frequencies between seismic and the Chandler

wobble.

3.5 Free Core Nutation

Because of the fluidity of the core, the Earth has another rotational mode in addition to the CW. Free core

nutation (FCN) results when the rotation axes of the core and mantle do not coincide due to flattening at the core-mantle
12

boundary (CMB). This misalignment causes a restoring force at the elliptical CMB that tries to realign them. Because the

Earth is spinning, this force produces a damped wobble in the instantaneous rotation axis about the principal moment of

inertia axis. In the terrestrial reference frame (e.g. with SLR, LLR, and gravimetry), this motion has a period of about one

day and is called the nearly diurnal free wobble (NDFW). Viewed from the celestial frame (e.g. with VLBI), it is called

free core nutation (FCN) and has a period of about 432 days. Both the CW and FCN exist because of the flattening of the

planet, but FCN is a forced motion due to lunisolar tides. The NDFW is particularly important because its period lies

within the diurnal band, thereby creating a resonance between them and enhancing the response of the Earth to the Moon’s

tidal force [Neuberg et al., 1987; VanHoolst et al., 2000]. Thus, observing the NDFW/FCN is a useful way to measure the

CMB flattening and learn about dissipation at this interface [Defraigne et al., 1994].

The NDFW is due to the pressure coupling between the liquid core and solid mantle. Its eigenperiod depends on

the CMB ellipticity and the Earth’s elasticity [Sasao and Wahr, 1981]. This can be measured via very long baseline

interferometery (VLBI) and precise tidal gravimetry. The attenuation of this mode is a direct consequence of damping

mechanisms at the CMB. The periods for it found with gravimetry and VLBI agree well, but the Q factors found with

each methods can very different. In order to get Q from the observed period, one simply needs to take the ratio of the real

to imaginary components of the nutation period, as expressed in Equation 8. Q values found from gravimetry are about

104-105 and from VLBI 105-106 [e.g. Neuberg et al., 1987; Cummins and Wahr, 1993; Florsch and Hinderer, 2000].

These large Q values tell us that FCN is measuring Q of the fluid outer core rather than the solid mantle [Lefftz and

Legros, 1992] and should not be included with analyses of Q from tidal, CW, or seismic observations.

4. Discussion

Some of the most compelling evidence that the dissipation of elastic energy within the Earth is linear (hence,

Equation 7 applies) is a study of the Earth strain tides by Agnew [1981]. He found that for peak strains of 10-8, the Earth’s

response is linear to 1 part in 1000. Strains associated with the tides or the largest earthquakes are at least an order of

magnitude less than this. Thus, one absorption band should apply from fractions of a second to months or years. Bolt and

Canas [1985] estimated 0.4 ≤ α ≤ 1 from PcP and PcS seismic body waves (1-5 s). Jeffreys [1972] used seismic and

nutation data to find α ≈ 0.2. Anderson and Minster [1979] used Chandler wobble, tidal, and modes, to find 0.2 ≤ α ≤

0.33. Smith and Dahlen [Smith and Dahlen, 1981] found 0.04 ≤ α ≤ 0.19 for the Chandler wobble. Using gravimeters,

Baker et al. [Baker et al., 1996] found very little frequency-dependence of Q with α ≤ 0.09. Flanagan and Wiens [1998]

found strong frequency-dependence of Q in upper mantle of the Lau backarc basin. They found 0.1 ≤ α ≤ 0.3, which is
13

also consistent with experimental results [Berckhemer, 1982; Gueguen, 1989; Sato et al, 1989].

Figure 10 summarizes some of our current understanding of the variation of Q with period for the methods

discussed in this paper. The prospect of having one robust linear relation for all of these timescales is promising.

10000
Body Modes Tides, CW
Waves NDFW

1000
α ≈ 0.2

Q 100

α ≈ 0.4

10

1
1.E+00 10.E+00 1.E+02 10.E+02 1.E+04 10.E+04 1.E+06 10.E+06 1.E+08

Period (s)

Figure 10. Compilation of some seismic, tidal, NDFW, and CW Q values found in the literature.

5. References

Agnew, D., J. Berger, R. Buland, W. Farrell, and F. Gilbert, International Deployment of Accelerometers: A Network for Very
Long Period Seismology, Eos Trans. AGU, 57, 180-188, 1976.
Agnew, D.C., Nonlinearity in Rock: Evidence from Earth Tides, J. Geophys. Res., 86 (B5), 3969-3978, 1981.
Anderson, D.L., A. Ben-Menahem, and C.B. Archambeau, Attenuation of Seismic Energy in the Upper Mantle, J. Geophys. Res.,
70, 505, 1965.
Anderson, D.L., and R.S. Hart, Attenuation Models of the Earth, Phys. Earth Planet. Int., 16, 289-306, 1978.
Anderson, D.L., and J.B. Minster, The Frequency Dependence of Q in the Earth and Implications for Mantle Rheolgoy and
Chandler Wobble, Geophys. J. R. astr. Soc., 58, 431-440, 1979.
Baker, T.F., D.J. Curtis, and A.H. Dodson, A New Test of Earth Tide Models in Central Europe, Geophys. Res. Lett., 23 (24),
3559-3562, 1996.
Bolt, B.A., and J.A. Canas, Constraints from Core Reflections on Mantle Q and Density at the Core Boundary, Phys. Earth
Planet. Int., 38, 1-8, 1985.
Cummins, P.R., and J.M. Wahr, A Study of the Earth's Free Core Nutation Using International Deployment of Accelerometers
Gravity Data, J. Geophys. Res., 98 (B2), 2091-2103, 1993.
Dahlen, F.A., and J. Tromp, Theoretical Global Seismology, Princeton University Press, Princeton, NJ, 1998.
Defraigne, P., V. Dehant, and J. Hinderer, Stacking Gravity Tide Measurements and Nutation Observations in Ordre to
Determine the Complex Eigenfrequency of the Nearly Diurnal Free Wobble, J. Geophys. Res., 99 (B5), 9203-9214,
1994.
Dickey, J.O., P.L. Bender, J.E. Faller, X.X. Newhall, R.L. Ricklefs, J.G. Ries, P.J. Shelus, C. Veillet, A.L. Whipple, J.R. Wiant,
J.G. Williams, and C.F. Yoder, Lunar Laser Ranging - A Continuing Legacy of the Apollo Program, Science, 265, 482-
490, 1994.
Dickman, S.R., and Y.S. Nam, Constraints on Q at Long Periods from Earth's Rotation, Geophys. Res. Lett., 25, 211-214, 1998.
14

Durek, J.J., and G. Ekström, A Radial Model of Anelasticity Consistent with Long Period Surface Wave Attenuation, Bull.
Seism. Soc. Am., 86, 144-158, 1996.
Dziewonski, A.M., and D.L. Anderson, Preliminary Reference Earth Model, Phys. Earth Planet. Int., 25, 297-356, 1981.
Flanagan, M.P., and D.A. Wiens, Attenuation of Broadband P and S Waves in Tonga: Observations of Frequency Dependent Q,
Pure appl. geophys., 153, 345-375, 1998.
Florsch, N., and J. Hinderer, Bayesian Estimation of the Free Core Nutation Parameters from the Analysis of Precise Tidal
Gravity Data, Phys. Earth Planet. Int., 117, 21-35, 2000.
Goldreich, P., and S. Soter, Q in the Solar System, Icarus, 5, 375-389, 1966.
Hansen, R.A., Observational Study of Terrestrial Eigenvibrations, Phys. Earth Planet. Int., 28, 27-69, 1982.
Herring, T.A., and D. Dong, Measurement of the Diurnal and Semidurnal Rotational Variations and Tidal Parameters of the
Earth, J. Geophys. Res., 99 (18), 18051-18071, 1994.
Jackson, D.D., Elastic Relaxation Model for Seismic Wave Attenuation in the Earth, Phys. Earth Planet. Int., 2, 30-34, 1969.
Jackson, D.D., and D.L. Anderson, Physical Mechanisms of Seismic-wave Attenuation, Rev. Geohpys. Space Phys., 81 (1), 1-63,
1970.
Jeffreys, H., Creep in the Earth and Planets, Tectonophysics, 13, 569-581, 1972.
Lagus, P.L., and D.L. Anderson, Tidal Dissipation in the Earth and Planets, Phys. Earth Planet. Int., 1, 505-510, 1968.
Lefftz, M., and H. Legros, Some Remarks About the Rotatiosn of a Viscous Planet and its Homogeneous Liquid Core: Linear
Theory, Geophys. J. Int., 108, 705-724, 1992.
Liu, H.-P., D.L. Anderson, and H. Kanamori, Velocity Dispersion Due to Anelasticity: Implications for Seismology and Mantle
Composition, Geophys. J. R. astr. Soc., 47, 41-58, 1976.
MacDonald, G.J.F., Tidal Friction, Rev. Geophys., 2, 467-541, 1964.
Melchior, P., The Phase Lag of Earth Tides and the Breaking of the Earth's Rotation, Phys. Earth Planet. Int., 56, 186-188, 1989.
Murray, C.D., and S.F. Dermott, Solar System Dynamics, Cambridge University Press, Cambridge, United Kingdom, 2001.
Neuberg, J., J. Hinderer, and W. Zürn, Stacking Gravity Tide Observations in Central Europe for the Retrival of the Complex
Eigenfrequency of the Nearly Diurnal Free-wobble, Geophys. J. R. astr. Soc., 91, 853-868, 1987.
Ray, R.D., R.J. Eanes, and B.F. Chao, Detection of Tidal Dissipation in the Solid Earth by Satellite Tracking and Altimetry,
Nature, 381, 595-597, 1996.
Ray, R.D., R.J. Eanes, and F.G. Lemoine, Constraints on Energy Dissipation in the Earth's Body Tide from Satellite Tracking and
Altimetry, Geophys. J. Int., 144, 471-480, 2001.
Sasao, T., and J.M. Wahr, An Excitation Mechanism for the Free 'Core Nutation', Geophys. J. R. astr. Soc., 64, 729-746, 1981.
Schuh, H., and R. Haas, Earth Tides in VLBI Observations, in Proc. 13th Int. Symp. on Earth Tides, edited by B. Ducarme, and
Pâquet, Observatoire Royal Belgique, Brussels, 1998.
Sipkin, S.A., and T.H. Jordan, Frequency Dependence of QScS, Bull. Seism. Soc. Am., 69, 1055-1079, 1979.
Smith, J.C., and G.H. Born, Secular Acceleration of Phobos and Q of Mars, Icarus, 27 (1), 51-53., 1976.
Smith, M.L., and F.A. Dahlen, The Period and Q of the Chandler Wobble, Geophys. J. R. astr. Soc., 64, 223-281, 1981.
Smith, S.W., and P. Jungels, Phase Delay of the Solid Earth Tide, Phys. Earth Planet. Int., 2, 233-238, 1970.
Stacey, F.D., Physics of the Earth, Brookfield Press, Kenmore, Australia, 1992.
Stein, S., and M. Wysession, An Introduction to Seismology, Earthquakes, and Earth Structure, Blackwell Publishing, Oxford,
UK, 2002.
VanHoolst, T., V. Dehant, and P. Defrigne, Sensitivity of the Free Core Nutation and the Chandler Wobble to Changes in the
Interior Structure of Mars, Phys. Earth Planet. Int., 117, 397-405, 2000.
Wahr, J., and Z. Bergen, The Effects of Mantle Anelasticity on Nutations, Earth Tides, and Tidal Variations in Rotation Rate,
Geophys. J. R. astr. Soc., 87, 633-668, 1986.
Yuen, D.A., and W.R. Peltier, Normal Modes of the Viscoelastic Earth, Geophys. J. R. astr. Soc., 69, 495-526, 1982.
Zschau, J., Tidal Friction in the Solid Earth: Constraints from the Chandler Wobble Period, in Space Geodesy and Geodynamics,
edited by A. Anderson, and A. Cazenave, Academic Press, London, 1986.

You might also like