Perturbed Orbits and Apsidal Precession

Scrollable formula. Use Left and Right arrow keys to reveal hidden content.

Scrollable table. Use Left and Right arrow keys to reveal additional columns.

Scrollable figure. Use Left and Right arrow keys to inspect hidden content.

A weak perturbation does not usually turn a Kepler ellipse into a single new ellipse. It makes the instantaneous Kepler elements vary. Fast variations can average to zero over an orbit, while a small nonzero average accumulates secularly; apsidal precession is the angular drift of periapsis produced by such a mismatch of radial and azimuthal periods.

Required background. Classical and Variational Mechanics provides action principles, canonical variables, and mechanical conservation laws; Newtonian Gravity provides the gravitational potential, Poisson equation, and inverse-square force law. Helpful background. Calculus and Linear Algebra provides multivariable differentiation, linear maps, inner products, and basis changes; Ordinary Differential Equations provides initial-value problems, linearization, and asymptotic solution methods.

Osculating elements as changing coordinates

Write the relative equation as

\ddot{\mathbf r} =-\frac{\mathcal G}{r^3}\mathbf r +\epsilon\,\mathbf f(\mathbf r,\mathbf v,t), \qquad |\epsilon|\ll1.

The perturbing acceleration $\mathbf f$ may be conservative or dissipative, stationary or time dependent. Those distinctions determine which unperturbed integrals can still be used.

At each time, a state $(\mathbf r,\mathbf v)$ with nonzero angular momentum determines a tangent Kepler conic. Away from the singularities of the chosen element chart, describe it by six elements $C_A(t)$ . To make all six constant in the unperturbed problem, use a phase constant such as the mean anomaly at a fixed reference epoch, rather than the instantaneous mean anomaly that already advances at the Kepler rate. Variation of parameters imposes the tangency condition and trades the second-order vector equation for six first-order equations,

\dot C_A=\epsilon\,F_A(C_B,t).

With the Lagrange tangency condition, this is an exact osculating reparametrization. Other admissible supplementary conditions give exact element descriptions of the same physical trajectory, but those elements are generally nonosculating. The perturbative approximation enters when $F_A$ is evaluated on the unperturbed orbit and truncated in $\epsilon$ . Efroimsky 2006, Abstract and §1.1, pp. 1–3 explains this distinction between the Lagrange constraint and other variation-of-parameters gauges.

Two balance equations reveal the physics without writing every planetary equation. With specific energy $\varepsilon=v^2/2-\mathcal G/r$ and specific angular momentum $\mathbf h=\mathbf r\times\mathbf v$ ,

\dot\varepsilon=\epsilon\,\mathbf v\cdot\mathbf f, \qquad \dot{\mathbf h}=\epsilon\,\mathbf r\times\mathbf f.

A perturbation perpendicular to $\mathbf v$ does no instantaneous work but can torque the plane. A radial perturbation exerts no torque but can change the eccentricity and apsidal phase.

Periodic and secular changes

For an element $C$ changing slowly compared with the Kepler period $T$ , define the orbit average

\langle\dot C\rangle =\frac{1}{T}\int_0^T \dot C\,dt.

If this average vanishes, the first-order change may remain oscillatory. A nonzero average instead produces a secular drift. Over a time short enough that the averaged drift changes little, its local approximation is

C(t)\simeq C(0)+t\,\langle\dot C\rangle_{C(0)}.

The subscript means that the orbit used in the average is held at its initial elements. On the secular timescale, typically $t\sim T/|\epsilon|$ when the perturbation has that relative size, the elements can change by order unity. One must then evolve the averaged equations with their dependence on the changing elements, rather than extrapolate one constant slope. Averaging is not legitimate near every resonance: a phase that is normally fast may become slow when frequencies nearly satisfy an integer relation, and its contribution need not cancel.

JPL Solar System Dynamics, “Orbital Elements” and “Ephemeris Development” (accessed 2026) uses this operational interpretation: osculating elements encode position and velocity at an epoch, while numerical ephemerides propagate the full perturbed equations.

Apsidal precession from a weak inverse-cube term

Consider the specific potential

\Phi(r)=-\frac{\mathcal G}{r}-\frac{\alpha}{r^3}, \qquad \left|\frac{\alpha}{\mathcal G r^2}\right|\ll1.

Binet’s equation is

u''+u=\frac{\mathcal G}{h^2}+\frac{3\alpha}{h^2}u^2, \qquad u=\frac1r,

where primes denote derivatives with respect to $\phi$ . For a nearly circular orbit $u=u_0+\delta u$ , the constant part fixes the corrected circular radius. Linearizing the oscillatory part gives

\delta u''+\kappa^2\delta u=0, \qquad \kappa^2=1-\frac{6\alpha u_0}{h^2}.

One radial oscillation takes azimuth $2\pi/\kappa$ , so the periapsis advance per radial period is

\Delta\varpi =2\pi\left(\frac1\kappa-1\right) \simeq\frac{6\pi\alpha u_0}{h^2}.

Positive $\alpha$ gives prograde precession in this model. The check $\alpha\to0$ restores $\kappa=1$ and a closed Kepler ellipse. This calculation illustrates the frequency mismatch; it is not a derivation of any particular relativistic correction, whose equations and parameter mapping must be obtained from the chosen spacetime model.

What can go wrong

A secular term in an element can be partly a consequence of the element convention or supplementary condition. State which elements and gauge of variation are used before assigning physical meaning.
First-order theory fails after a sufficiently long time if the accumulated correction becomes order unity. Multiple-scale or canonical perturbation methods then reorganize the expansion.
Orbit averaging erases short-period behavior and can fail near resonances, separatrices, close encounters, or impulsive forces.
A precessing projected ellipse does not prove the perturbation is central. Nonspherical, third-body, dissipative, and relativistic effects can all move the apsides.
The sign of $\Delta\varpi$ depends on the definition of advance and on the orientation of $\phi$ ; define both before comparing formulas.

The separation of periodic from secular perturbations and its application to perihelion precession are developed systematically in Stewart 2014, §5.2, pp. 96–98, and §5.3, pp. 98–101, PDF.

Transfer task

Suppose a weak tangential drag is always antiparallel to $\mathbf v$ . Determine the signs of $\dot\varepsilon$ and of the angular-momentum magnitude $\dot h$ .

Solution

Write $\mathbf f=-\gamma\mathbf v$ with $\gamma\gt0$ . Then

\dot\varepsilon =-\epsilon\gamma v^2\lt0, \qquad \dot{\mathbf h} =-\epsilon\gamma\mathbf h.

For $\epsilon\gt0$ , both orbital energy and $h=|\mathbf h|$ decrease. Their relative rates are not generally equal, so the eccentricity need not remain fixed.

Continue to Restricted Few-Body Dynamics, Resonances, and Stability for a setting where rotating-frame equilibria, resonant phases, and nonlinear behavior coexist.

References

Efroimsky, Michael. “Gauge Freedom in Orbital Mechanics.” arXiv:astro-ph/0603092v1, 2006. Abstract and §1.1, pp. 1–3.
Jet Propulsion Laboratory Solar System Dynamics. “Description of Orbits and Ephemerides”, “Orbital Elements” and “Ephemeris Development,” accessed 2026.
Stewart, Iain W. Advanced Classical Mechanics. MIT OpenCourseWare 8.09, Chapter 5 “Perturbation Theory,” Fall 2014. §5.2, pp. 96–98; §5.3, pp. 98–101, PDF.