Geodetic Precession and Gravity Probe B: Measuring Curved Spacetime with Gyroscopes

What Geodetic Precession Is

Geodetic precession is a purely relativistic effect: a spinning gyroscope carried in orbit around a massive body slowly changes the direction its spin axis points, even though no torque acts on it in the Newtonian sense. In flat spacetime, an ideal free gyroscope keeps its axis fixed relative to the distant stars forever. Around a mass like Earth, spacetime is curved, and the rule for carrying a vector "as straight as possible" through curved space — parallel transport — no longer returns the axis to its starting orientation after one loop.

The effect was predicted by Willem de Sitter in 1916, only a year after Einstein completed general relativity, and refined by others including Fokker. It is sometimes called the de Sitter effect or the geodetic effect. Crucially, it arises from the static curvature produced by the presence of mass — you do not need the central body to be spinning. That distinguishes it sharply from frame-dragging, which requires rotation.

• No Newtonian torque is involved — the axis "drifts" because geometry itself is curved.
• The tilt accumulates in the plane of the orbit.
• It exists even for a non-rotating central mass.

The Mechanism and Governing Relation

Geodetic precession is a form of holonomy: transport a vector around a closed loop in curved space and it comes back rotated by an angle set by the curvature enclosed. For a gyroscope in a circular orbit, the accumulated rotation per orbit scales with the spatial curvature times the area swept out.

The precession rate is given by

• Omega_geo = (3/2) (GM / c^2 r^3) (x cross v)

where G is Newton's constant, M the central mass, c the speed of light, r the orbital radius, and x and v the position and velocity vectors. The magnitude scales as (3/2)(GM/c^2 r) x (v/r), i.e. the gravitational potential depth GM/(c^2 r) multiplied by the orbital angular frequency. Two-thirds of the effect comes from space curvature and one-third from the relativistic ('Thomas-like') time-dilation term, which is why the coefficient is 3/2 rather than 1. The tighter the orbit and the faster the gyroscope moves, the larger the annual drift.

Characteristic Numbers and a Worked Example

Plug in GP-B's orbit to see the size of the effect. GP-B flew a near-circular polar orbit at 642 km altitude (r about 7,020 km), with an orbital speed of roughly 7.6 km/s and a period near 97.5 minutes. The dimensionless potential is GM/(c^2 r) about 6.3 x 10^-10, and the orbital angular frequency is about 1.07 x 10^-3 rad/s.

• Geodetic rate: (3/2) x (6.3e-10) x (1.07e-3 rad/s) works out to about 6.6 arcseconds per year — precisely the 6606.1 mas/yr predicted.
• Over one year the spin axis tilts by 6.6 arcseconds, about 0.0018 degrees.
• For comparison, the far weaker frame-dragging term is only 39.2 mas/yr, some 168 times smaller.

The scaling law makes the sensitivity clear: fly lower or around a more massive body and the effect grows. Near a neutron star or black hole, where GM/(c^2 r) approaches unity, gyroscope precession becomes enormous rather than a fraction of an arcsecond per year.

How Gravity Probe B Detected It

Measuring an angle of 6.6 arcseconds per year demanded almost absurd precision. GP-B, launched on 20 April 2004, carried four gyroscopes: fused-quartz rotors 38 mm across, coated in superconducting niobium, homogeneous to a few parts per million and round to within about 25 nanometers — the roundest objects ever manufactured, bettered only by neutron stars.

• The rotors spun at up to 4,000 rpm in a vacuum, chilled to about 2.5 K by superfluid helium.
• A spinning superconductor develops a London magnetic moment aligned with its spin axis; ultra-sensitive SQUID magnetometers read the axis direction to milliarcsecond precision.
• The spacecraft was drag-free, floating around a proof mass so the gyros followed a pure gravitational orbit, and a telescope locked onto the guide star IM Pegasi as a fixed reference.

Data were collected from 28 August 2004 to 14 August 2005. Unexpected electrostatic patch potentials on the rotor surfaces produced nuisance torques that had to be modeled out, delaying analysis. NASA announced the final results on 4 May 2011, published in Physical Review Letters: geodetic drift -6601.8 +/- 18.3 mas/yr and frame-dragging -37.2 +/- 7.2 mas/yr, both consistent with Einstein.

Geodetic Precession Versus Its Cousins

Geodetic precession is easy to confuse with related effects, so the distinctions matter:

• Frame-dragging (Lense-Thirring): caused by the rotation of the central body, which drags spacetime around with it (gravitomagnetism). It depends on the body's angular momentum J, not just its mass, and at GP-B was ~168 times smaller than the geodetic term. Geodetic precession would exist even if Earth were not spinning; frame-dragging would not.
• Thomas precession: a special-relativistic kinematic rotation of an accelerated spinning frame. It is actually embedded within the geodetic effect — the '1/2' part of the 3/2 coefficient is Thomas-like.
• Newtonian / classical precession: such as the wobble of a spinning top or Earth's 26,000-year axial precession, driven by real torques from external masses. Geodetic precession needs no torque at all.

The same de Sitter physics also acts on the Earth-Moon system as it orbits the Sun, twisting the lunar orbit by about 19.2 mas/yr, a value confirmed by lunar laser ranging to better than 1%.

Significance, Confirmation, and Open Questions

Gravity Probe B stands as one of the cleanest direct tests of general relativity's prediction that mass and rotation reshape the geometry of spacetime itself. Conceived by Stanford's Leonard Schiff and William Fairbank around 1960 and championed by Francis Everitt, the mission took roughly five decades and about 750 million dollars to fly — one of the longest-running projects in NASA history.

• The geodetic effect was confirmed to about 0.28% precision, a resounding agreement with Einstein.
• The frame-dragging result, at ~19% precision, was noisier because the patch-effect torques were comparable to the tiny signal; the LARES and LAGEOS laser-ranging satellites have since measured frame-dragging to a few percent, complementing GP-B.

Open questions center less on whether GR is right in this weak-field regime — it clearly is — and more on pushing precision to constrain alternative gravity theories (scalar-tensor, PPN parameter gamma) and on measuring analogous precessions in strong-field settings, such as pulsars whose spin axes precess in binary orbits, offering a laboratory where GM/(c^2 r) is far from negligible.

Geodetic (de Sitter) precession versus frame-dragging (Lense-Thirring) precession as measured by Gravity Probe B in Earth orbit.

Property	Geodetic precession	Frame-dragging (Lense-Thirring)
Physical cause	Motion through space curved by Earth's mass	Rotation of Earth twisting spacetime (gravitomagnetism)
Depends on	Central mass M and orbital velocity v	Central angular momentum J (Earth's spin)
GR prediction (GP-B)	-6606.1 mas/yr	-39.2 mas/yr
GP-B measurement	-6601.8 +/- 18.3 mas/yr	-37.2 +/- 7.2 mas/yr
Direction of tilt	In the orbital plane	Perpendicular, along Earth's rotation axis
Relative size	~168x larger	The smaller, harder-to-detect effect