Geophys. J . Int. (1991) 106, 699-702 RESEARCH NOTE The Earth's equatorial principal axes and moments of inertia H. S. Liu and B. F. Chao Geodynamics Branch, Goddard Space Flight Center, Greenbelt, M D 20771, USA Accepted 1991 February 26. Received 1990 October 22 s2, Key words: Earth's moment of inertia, Earth's rotation. INTRODUCTION Let the three principal moments of inertia of the Earth be A, B and C , where A < B < C . Let the corresponding principal axes be called the a, b and c axes, and the geographical coordinate axes be x, y and z. Disregarding internal relative motions, the principal axes are fixed in the geographical coordinates. Apart from a slight polar motion in the Earth's rotation, the z axis coincides with the c axis (Fig. l ) , whose orientation and the corresponding value of C are well determined. The present paper focuses on the other two principal axes, which lie on the xy, or equatorial plane, and their corresponding 'equatorial' principal moments of inertia. Specifically, where are the a and b axes in the equatorial plane? And what can one say about A and B? The study of the Earth as a triaxial body is a classical problem; the pre-space era history of investigations using astrogeodetic and gravimetric methods were reviewed by Izsak (1961) and Heiskanen (1962); the results were crude and in discord. Progressively better results were obtained when it was realized that the orientation of the a and b axes and the difference B - A can be determined from the Earth's external gravitational field coefficients (the Stokes coefficients) of degree 2 and order 2, which were inferred from artificial satellite orbit observations (e.g., Izsak 1961; Blitzer ef al. 1962). These and subsequent studies that did discuss, or include (usually in passing) a discussion of the matter fell short of providing explicit derivation and expressions (e.g., Goldreich & Toomre 1969; Lorell 1972; Wagner 1973; Kinoshita 1977; Bills & Ferrari 1978; Kaula 1979; Bills, Kieffer & Jones 1987; Chao & Gross 1987; Chao & Rubincam 1989). Moreover, as far as we are aware, textbook coverage of the subject has been skimpy (e.g., Heiskanen & Moritz 1967; Baker & Makemson 1967; Jeffreys 1976; Torge 1980; Hubbard 1984), if treated at all. Although far less significant than the orientation of the c axis and the difference C - A , the quantities under present consideration are not without dynamical implications. Chao (1989) has pointed out that the orientation of the a and b axes serves as an important constraint on laterally heterogeneous density models for the Earth, bearing on tectonic movements, mantle convections, and perhaps core configuration and activities. Indeed the difference B - A (and other Stokes coefficients for that matter) can serve as such constraints as well. Moreover, Liu & Chao (1990) argue that the luni-solar torque exerted on the Earth's rotation owing to the non-zero B - A (e.g., Woolard 1953) approaches the threshold of detection envisioned for the near future (cf. Dickey & Schutz 1989). On other (terrestrial) planets for which data are available, these effects can be much larger in magnitude although any geophysical discussion becomes more speculative. At any rate we believe it is worthwhile to devote a short paper to the subject. FORMULA The external gravitational field of a planetary body, satisfying the Laplace's equation, is customarily given by x ( C , cos mA + S, sin mA), (1) 699 Downloaded from http://gji.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016 SUMMARY Let the Earth's equatorial principal moments of inertia be A and B, where A < B, and the corresponding principal axes be a and b. Explicit formulae are here derived for determining the orientation of a and b axes and the difference B - A using CZ2 and the two gravitational harmonic coefficients of degree 2 and order 2. For the Earth, the a axis lies along the (14.93"W' 165.07"E) diameter, and the b axis lies perpendicular to it along the (75.07"E' 104.93"W) diameter. The difference B - A is MR2. These quantities for other planets are contrasted, and geophysi7.260 x cal implications are discussed. 700 H . S. Liu and B. F. Chao z=c of Z, on gets C,, + is,, = J(x2 - y2 + 2ixy)p d V / K =l ( x a (3) + = ( B - A)ei2"/K . (4) Equating the real and imaginary parts in equation (4) one gets an expression for the (positive) difference in the two equatorial moments of inertia: B - A = KY-, (54 and the east longitude of the principal axis of the least moment of inertia: A = f arg(C,, + is.,). (5b) Note that, because of the factor 1/2 in (Sb), it is arg(C,,+iS,,) (between -n and n), rather than tan-'(S,,/C,,) (between - n / 2 and n/2), that should be used in equation (5b). For arg(C,,+iS,,) between - n / 2 and n/2, the two are the same and either gives the correct A. Otherwise, tan-'(&2/C22) would be off by n/2, hence switching the a and b axes. RESULTS A N D DISCUSSION A complete detailed derivation of the above with explicit normalization factors can be found in, e.g., Chao & Gross (1987). We note in passing that there are five degree-2 Stokes coefficients but six independent components in Z. Consequently, in order to determine Z uniquely one needs an additional piece of information; for instance the astronomically measured precession constant H = (C A ) / C in the case of the Earth, or the physical longitudinal libration parameter y = ( B - A ) / C in the case of the Moon. This, of course, is a manifestation of the classical non-uniqueness in the gravitational inverse problem (e.g., Morse & Feshbach 1953, pp. 1276-1283). In particular, given C,, and S, one can only obtain the orientation of a and b in the equatorial plane and the difference B - A , as follows. Here the complex notation is found to be extremely convenient. We will take the xy plane to be the complex plane, the x and y axes being the real and imaginary axes, respectively. Then from equation (2d,e) and the definition The observed Earth values for degree-2 Stokes coefficicnts of the GEM-TZ model are: C,, = -484.1653(3) X CZ2= 2.4390(2) X and &, = -1.4001(2) X lop6 (Marsh et al. 1990). With the known values of M R 2 and C, equation (5a) gives B-A -- - 7.260 X MR, and the longitudinal libration parameter y = ( B - A ) / C = 2.1946 x (7) With known H (see above) one gets ( B - A ) / ( C - A ) = 1/150; that is, the Earth's equatorial dynamical ellipticity is 150 times smaller than the polar dynamical ellipticity. Equation (7) is about twice as large as the value quoted by Scheidegger [1982, p. 197, which, incidentally, erroneously takes (7) to be the cause of the Earth's free wobble; Scheidegger 1983, personal communication]. Similarly, from equation (5b) and to the accuracy of the C,, and S, coefficients, A = -14.93". (8) That is, the a axis of the Earth lies along the (14.93"W, 165.07"E) diameter, and the b axis lies perpendicular to it along the (75.07"E, 104.93"W) diameter. The latter points Downloaded from http://gji.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016 C,, + i&, where G is the gravitational constant; M and R are the Earth's mass and mean radius, respectively; r = (r, 0, A) gives the radial distance, the colatitude and the east longitude of the field point; and 6, is the 4n-normalized associated Legendre function of degree I and order m. The are the (normalized) Stokes parameters Cf, and S, coefficients that are now routinely determined for the Earth from observations of satellite orbits. Comparing equation (1) with the multiple expansion of U (e.g., Jackson 1975), one can show that the Stokes coefficients are simply normalized multipoies of the Earth's mass density distribution p(r). In particular, the inertia tensor Z of the Earth is related to the degree-2 Stokes coefficients in the following manner: dV/K where K = 2(5/3)'"MR2 for brevity. Invoking the coordinate rotation in the xy plane through an angle A (Fig. l), the integrand in equation (3) becomes, in the new coordinates a and b, ( x iy)' = (a ib)2ei2". Now if the new axes happen to be the principal axes (hence the use of symbols a and b here), then by definition the product of inertia $abp dV vanishes, and the integral in equation (3) becomes ei2" $(a + ib),p dV = ei2"$(a2 - b2)p dV. Then, by the definition of A and B, + Figure 1. The relation between the principal axes (u, b, c) and the geographical axes (x, y, 2). Angle A is the east longitude of a, the principal axis of the least moment of inertia. + iy)'p Earth’s principal axes and moments of inertia + accumulate (e.g., Dickey & Schutz 1989). It should be noted that these amplitudes are not much smaller than the semi-diurnal ocean tide contributions in the rotational speed variation (Yoder, Williams & Parke 1981; but see Baader, Brosche & Hovel 1983). The separation of the two effects in the observation is therefore desirable in studying the behaviour of the ocean tides. ACKNOWLEDGMENTS We would like to thank B. G. Bills for discussions and comments. REFERENCES Baader, H. R., Brosche, P. & Hovel, 1983. Ocean tides and periodic variations of the Earths rotation, J . Geophys., 52, 140-142. Baker, R. M. L. & Makemson, M. W., 1967. Asrrodynamics, 2nd edn, Academic Press, New York. Balmino, G., Moynot, B. & Vales, N., 1982. Gravity field model of Mars in spherical harmonics up to degree and order eighteen, J. geophys. Res., 87, 9735-9746. Bills, B. G. & Ferrari, A. J., 1978. Mars topography harmonics and geophysical implications, J. geophys. Res., 83, 3497-3508. Bills, B. G., Kieffer, W. S. & Jones, R. L., 1987. Venus gravity: A harmonic analysis, J. geophys. Res., 92, 10 335-10351. Blitzer, L., Boughton, E. M., Kang, G. & Page, R. M., 1962. Effect of ellipticity of the equator on 24-hour nearly circular satellite orbits, J. geophys. Res., 67, 329-335. Chao, B. F., 1989. Comment on ‘Moment of inertia of three-dimensional models of the Earth’ by T. Tanimoto, Geophys. Res. Lett., 16, 1075. Chao, B. F. & Gross, R. S., 1987. Changes in the Earth’s rotation and low-degree gravitational field induced by earthquakes, Geophys. J.R. astr. SOC., 91, 569-596. Chao, B. F. & Rubincam, D . P., 1989. The gravitational field of Phobos, Geophys. Res. Left., 16, 859-862. Chao, B. F. & Au, A. Y., 1991. Temporal variation of the Earth’s low-degree gravitational field caused by atmospheric mass redistribution: 1980-1988, J. geophys. Res., 96, 6569-6575. Dickey, J. 0. & Schutz, R. E., 1989. Earth rotation and references studies, EOS, Trans. A m . geophys. Un., 43, 1062. Goldreich, P. & Toomre, A , , 1969. Some remarks on polar wandering, J. geophys. Res., 74, 2555-2567. Heiskanen, W. A,, 1962. Is the Earth a triaxial ellipsoid?, J. geophys. Res., 67, 321-327. Heiskanen, W. A. & Moritz, H., 1967. Physical Geodesy, Freeman, San Francisco. Hubbard, W. B., 1984. Planetary Inferiors, Van Nostrand, Reinhold, New York. Izsak, I. G., 1961. A determination of the ellipticity of the Earth’s equator from the motion of two satellites, Asfr. J., 66, 226-229. Jackson, J. D., 1975. Classical Electrodynamics, 2nd edn, Wiley, New York. Jeffreys, H., 1976. The Earth, 6th edn, Cambridge University Press, Cambridge. Kaula, W. M., 1979. The moment of inertia of Mars, Geophys. Res. Lett., 6, 194-196. Kinoshita, H., 1977. Theory of the rotation of the rigid Earth, Celest. Mech., 15, 277-326. Kohnlein, W., 1966. The geometric structure of the Earth’s gravitational field, in Geodetic Parameters for the 1966 Smifhsonian Institution Standard Earth, vol. 111, eds Lundquist , C. A. & Veis, G., SAO Spec. Rep. 200, Cambridge, MA. Downloaded from http://gji.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016 precisely to the great geoid low in the Indian Ocean, as might be expected. Note that, for the real, non-rigid Earth, any change in p(r) due to mass redistribution will cause (6) and (8) to change. Chao & Gross (1987) have studied the effect of earthquakes, while Chao & Aug (1990) studied for the atmospheric fluctuations; the changes are found to be too small to detect. Chao & Gross (1987, p. 585) have presented the correct relations (5a,b) but inadvertently left out the factor 1/2 in (5b) in figuring the a axis orientation as quoted there. Chao (1989) made the same error. This bears on Tanimoto’s (1989) model predictions based on heterogeneous Earth models obtained from seismic tomography: Tanimoto’s predictions, as a result, agree somewhat less satisfactorily with the observations than claimed. One can express the equator in terms of an equivalent geoid ‘ellipse’-equivalent in the sense of its rotational dynamics (see below). Equation (8) gives the orientation of the semi-major axis of this ellipse, while the difference in the semi-major axis and the semi-minor axis is given by fl(Cq, S:2)1’2R= 69.4 m. The latter, obtained by using Bruns formula on equation (l), agrees with that found by Kohnlein (1966; see also Baker & Makemson 1967; Torge 1980). This ellipse, however, does not correspond to a ‘best-fitting’ geometrical figure for the equator, simply because there are many other Stokes coefficients that also contribute significantly to the equatorial geoid figure. In fact, it is known that the (sea-level) difference between the western Pacific Ocean geoid high (near the a axis) and the Indian Ocean geoid low (along the b axis) reaches some 180 m. A similar phenomenon occurs on Mars but on a much grander scale, dominated by the great Tharsis rise: C,, = -84.79 x lop6, S,, = 48.64 x (Balmino, Moynot & Vales 1982). Thus, Martian a axis lies along the (75.loE, 104.9”W) diameter [this corrects an error in Bills & Ferrari (1978) which mislabelled the a and b axes due to the n / 2 ambiguity mentioned above; Bills 1990, personal communication]. It points to the centre of the Tharsis topographic high. Martin (B - A)/MR2 is about 2.52 X corresponding to a difference in the geoid semi-major and semi-minor axes of almost 1.3 km. Its (B - A ) / ( C - A ) ratio is as large as about 118. The Moon, Venus and Phobos are the only other planetary bodies for which the said parameters have been determined or inferred (see Kopal 1966; Bills et al. 1987; and Chao & Rubincam 1989, respectively). Unlike the Earth (but Goldreich & Toomre 1969) and Mars, these slowly rotating bodies are truly triaxial bodies in the sense that the quantities C - A , C - B, and B - A are of the same order of magnitude. The Earth’s y (equation 7), although small, is not without dynamical consequences. The axial torque exerted by the Sun and the Moon due to the non-zero y generates variations in the Earth’s rotational speed at semi-diurnal tidal periods. The maximum peak-to-peak amplitude amounts to some 0.075 milliarcsecond in (longitudinal) orientation, or 0.005ms in the Universal Time (UT1) over every 12hr (Liu & Chao 1990). Since the signal-to-noise ratio (in the frequency domain) of coherent harmonic signals increases as the square root of the number of observations, these semi-diurnal signals may become detectable in the near future as more accurate observations 701 702 H . S. Liu and B. F. Chao Kopal, Z., 1966. An Introduction to the Study of the Moon, Reidel, New York. Liu, H. S. & Chao, B. F., 1990. Semidiurnal longitudinal libration in the Earth’s rotation, EOS, Trans. Am. Geophys. Un., 71, 1271. Lorell, J., 1972. Estimation of gravity field harmonics in the presence of spin-axis direction error using radio tracking data, J . Astronaut. Sci., XX, 44-54. Marsh, J. G. et al., 1990. The GEM-T2 gravitational model, J. geophys. Res., 95, 22 043-22 071. Morse, P. M. & Feshbach, H., 1953. MefhodF of Theoretical Physics, McGraw-Hill, New York. Scheidegger, A. E., 1982. Principles of Geodynamics, SpringerVerlag, New York. Tanimoto, T., 1989. Reply to Chao, Geophys. Res. Letf., 16, 1076-1077. Torge, W., 1980. Geodesy, Walter de Gruyter, New York. Wagner, C. A., 1973. Does AZ2 vary?, J. geophys. Rex, 78, 470-475. Woolard, A. E., 1953. Theory of the rotation of the Earth around its center of mass, Astr. Pap. Am. Ephemeris XV, 1. Yoder, C. F., Williams, J. G. & Parke, M. E., 1981. Tidal variations of Earth rotation, J. geophys. Res., 86, 881-891. Downloaded from http://gji.oxfordjournals.org/ at Pennsylvania State University on May 17, 2016
© Copyright 2026 Paperzz