Astronomy 6570

Physics of the Planets

Planetary Rotation, Figures, and Gravity Fields

Topics to be covered:

1. Rotational distortion & oblateness

2. Gravity field of an oblate planet

3. Free & forced planetary precession

Rotational distortion

Consider a spherical, rigid planet rotating at an angular rate .

Point P on the surface experiences a centrifugal acceleration:

ac=

2r sin x

= 2x x

By symmetry, a similar point located on the surface in the yz plane feels an acceleration

ac=

2y y,

so that an arbitrary point on the surface at position (x, y, z) feels an acceleration

ac=

2x x + y y( )

It is convenient to think of this centrifugal acceleration in terms of a centrifugal potential, Vc, such that

ac= V

c.

Clearly we must have

Vc=

1

2

2x

2+ y

2( )

=1

2

2r

2 sin2

Now consider an ocean on the surface of the rotating planet. the fluid experiences a total potential

VT

r,( ) =GM

r+V

cr,( ).

In equilibrium, the surface of a fluid must lie on an equi-potential surface (i.e., the surface is

locally perpendicular to the net gravitational and centrifugal acceleration). Assuming the distortion

of the ocean surface from a sphere to be small, we write

rocean

= a + r ( ), a = requator( )

so VT

surface( )GM

a+

GM

a2r

1

2

2a2 sin2 2a sin2 r = constant

the last term is << the second (see below), so

r const. +2a4

2GMsin2

As expected, the ocean surface is flattened at the poles. The oblateness of the ocean surface is defined as

=r

eqr

pole

req

i.e., 2a3

2GM

1

2q

Note that the dimensionless ratio q is just the ratio of centrifugal acceleration at the equator to the

gravitational acceleration. Our analysis will only be reasonally correct so long as q << 1.

Some observed* values of q and for planets:

We see that q << 1 as we assumed above, and that and q

2 are comparable in size, but that in general is 1.5 - 2 times greater

than predicted. Why?

The answer is that we have neglected the feedback effect that the rotational distortion has on the planet's gravity field, which is no longer

that of a sphere.

[Aside: Rotational disruption.

Clearly a planet becomes severly distorted if q approaches unity. This sets an upper limit on the rotational velocity of a planet. Roughly,

max

GM

a3

1

2

2 G( )1

2

where is the planet's average density. For the Earth, = 5.52 g cm3, so

max

1.2 10 3 rad sec 1

or Pmin

1.4 hrs.

For Jupiter,

Pmin

2.9 hrs.]

*values in ( ) are uncertainties in the last figure.

Planetary Gravity Fields

Because of rotational flattening, most planets (but not most satellites) may be treated to a good approximation as

oblate spheroids; i.e., ellipsoids with 2 equal long axes and 1 short axis. A basic result of potential theory is that the

external potential of any body with an axis of symmetry can be written in the form

VG

r,( ) =GM

r1 J

n

R

r( )

n

Pnμ( )

n=2

where M = total mass

R = equatorial radius

Jn= dimensionless constants

Pn(μ) = Legendre polynomial of degree n

μ = cos

(Note that there is no n = 1 term, as long as the origin of co-ordinates is chosen as the body's center of mass.)

The first few Pn's are:

P0μ( ) = 1

P1μ( ) = μ

P2μ( ) =

1

23μ2 1( )

P3μ( ) =

1

25μ3 3μ( )

P4μ( ) =

1

835μ4 30μ2 + 3( ).

The Jn's reflect the distribution of mass within the body, and must be determined empirically for a planet. J

2 is by

far the most important, and has a simple physical interpretation in terms of the polar (C) and equatorial (A) moments of inertia:

J2 =

C A

MR2

[In general, Jn is given by the integral

Jn=

1

MRn

rnP

nμ( )

1

1

0

R

r,μ( )2 r2dμ dr

where r,μ( ) is the internal density distribution. Since Pnμ( ) is an odd function for odd n, J

3= J

5= J

7= = 0 for a planet

whose northern and southern hemispheres are symmetric. In fact, only the Earth has a measured non-zero value of J3.]

Relation between rotation, J2, and oblateness

Let us now return to the question of the oblateness of a rotating planet, including the effect of the non-spherical gravity field.

Recall that the centrifugal potential is given by

Vc=

1

2

2r

2 sin2

=1

3

2r

2P

2μ( ) 1( )

We can ignore the term which is independent of μ, so Vc has the same angular dependence as the J

2 term in the planetary gravity field.

The total potential at the surface of the planet is

VT

r,( ) =VG

r,( ) +Vc

r,( )

=GM

r+

GMR2

r3

J2+

1

3

2r

2P

2μ *

where we neglect J4, J

6 etc.

As before, we require the surface to be an equipotential (this is true even for solid planets) and write r = R + r ( ) :

r = const. J2+

1

3q R P

2μ( )

Relation of Jn to rotation rate

The connection between J2, J4, J6, … and the rotation rate is not straightforward to derive, involving a self-

consistent solution along the lines:

So long as the rotation parameter, q, is small, it can be shown that

Where the proportionality constants are usually of order unity. Since q << 1 generally, the higher-order Jn’s

rapidly become small. For J2, the above relation is usually written

With k2 defined this way, we have

For a uniform density planet it can be shown that

Examples of Planetary Jn’s and k2’s:

J

nq

n 2

J2=

1

3k

2q (*Note that

VJ

2

Vc

= 3J

2

q= k

2, which is the real definition of k

2.)

k2= 3 2

so J2=

q2

.

J2

J3

J4

k2 = 3J2/q

J4/q2

Note that the expression for Jn on the previous slide

implies that Jn becomes smaller as the mass of the planet

is more concentrated toward the center. Based on the

values for k2 in the Table, we conclude that Jupiter is the

most centrally-concentrated, and Mars the least. Even

Mars, however, has a k2 appreciably less than that of a

uniform-density planet, i.e., 1.50.

q = 3J

2, i.e., V

J2

=Vc, when k

2= 1

and k2 is known as a 2nd order Love number.

Weighting functions for

Jn: rn+2 (r) (Uranus)

Since P2μ( ) varies from 1.0 at the poles μ = ±1( ) to 0.5 at the equator μ = 0( ) the planet's oblateness is

=r

eqr

pole

req

i.e., =3

2J

2+

1

2q

(This replaces our earlier result, =1

2q.)

For our hypothetical uniform-density planet,

J2=

1

2q, so

= 5

4q.)

In general, in terms of the Love number k2= 3

J2

q, we have

= 1+ k2( )

q

2

h2

q

2, k

2 is the "response coefficient" of the planet to V

c

where we introduce another Love number, h2. As we have shown, for a planet in hydrostatic equilibrium,

h2= 1+ k

2

Let's compare the observed oblateness with this result (note that if both J2and q are known for a planet, no further

knowledge of the planet's density distribution is needed to predict ):

Let’s return to the question of the planet’s equilibrium shape:

Voila! – almost perfect agreement with our predicted values, confirming our assumption of hydrostatic equilibrium.

In the last line we give the polar moment of inertia factor, C

MR2

, derived from these observational values. This is an

important constraint on planetary interior models, and is obtained from the approximate formula, derived by Clairaut

h2

5

1+ 52

154

i C

MR2( )

.

(Note that this yields the correct value of h2= 5 2 for a uniform density planet with c = 2 5 MR

2; but gives h2= 20 / 29

for C = 0, rather than the correct value of 1.)

Review of Love numbers:

centrifugal potential, Vc=

1

2

2r 2 sin2

=1

3

2r 2 P2

cos( ) 1

gravitational potential (quadrupole term), V2=

GMR2

r3J

2P

2cos( )

potential Love number, k2

V2

V3

r = R( )

= 3GM

2 R3J

2

i.e., J2=

1

3k

2q

hyrdrostatic figure: =1

2q 1+

V2

Vc

=1

2q 1+ k

2( )

1

2q h

2

where the shape Love number h2= 1+ k

2

=1

2q +

1

2q k

2q =

1

2q +

3

2J

2

In general, k2 and h

2 depend on the internal density profile r( ). For a reasonably uniform planet, Clairaut &Radau showed that:

h2= 1+ k

2

5

1+ 52

154

CMR2( )

2

if = constant, C

MR2=

2

5 and we have:

h2=

5

2, k

2=

3

2

R2

J2

1

3k

2q

f1

2h

2q

J2

f

2

3

k2

h2

=2

3

k2

1+ k2

=2

5 for uniform planet

Figure after Dermott (1984)

D-R relation: J

2

f=

8

15

5

61

3

2

C

MR2

2

Experimental Determination of , J2, and Rotation Rate,

Interior rotation rate:

Magnetic field rotation via radio emission

Surface/atmospheric rotation rates:

Feature tracking (surfaces & clouds)

Doppler Shifts (radar)

Gravity Field: GM, J2, J4, etc.

Natural satellites:

Periods & semi-major axes GM

Rings

resonance locations GM, J2

Spacecraft:

Doppler tracking GM, J2, J4

Oblateness, :

Solid surface/cloud tops:

Imaging (Earth-based or s/c)

Surface pressure variations

Radio occultations

Upper atmosphere

Radio occultations (p ~ 100 mb)

Stellar occulations (1 μb < p < 1 mb)

& vs. a J

2, J

4

& vs. a J

2, J

4, J

6

(1) Mercury’s spin is

in a 3:2 resonance

with its orbit.

(2) Venus’ spin (solar

day) is near a 5:1

resonance with

Earth conjunction,

though this may

be coincidence.

*Cassini results suggest a

longer, and variable, period

-Voyager measurements (1979 – 1981)

Note that “0” wind speed is measured wrt the radio (System III) rotation period, which is

assumed to match the magnetic field and thus the interior rotation period.

Lewis 1995

Sanchez-Lavega (2005) Science 307, 1223.

Hammel et al., (2001) Icarus 153, 229

Hammel et al., (2000) Icarus 175, 534

F. Stacey: “Physics of the Earth”

to 0 J4( )

=GM

R3

1

2 R

a

7

2 3

2J

2

15

4J

4

R

a

2

=GM

R3

1

2 R

a

7

2 3

2J

2

9

4J

2

2+

15

4J

4

R

a

2

where a = geometric mean radius

R = planet's equatorial radius

M = planet's mass

Lindal et al. (1985), A.J. 90, 1136.