Aberration corrections

When you ask SPICE for “where is body A relative to body B?”, there are (at least) two answers:

• the geometric position (instantaneous, ignoring the speed of light)
• the apparent position (accounting for light travel time and optionally stellar aberration)

SPICE controls this via the “aberration correction” flag (abcorr).

What it is

Two related effects:

• Light time: light takes time to travel between observer and target, so the target’s apparent position corresponds to an earlier (or later) target epoch.
• Stellar aberration: the observer’s velocity changes the apparent direction to the target.

In CSPICE, you select these with strings like:

• "NONE": geometric
• "LT", "LT+S": one-way light time, with optional stellar aberration
• "CN", "CN+S": converged Newtonian light time (iterative), with optional stellar aberration
• "XLT", "XLT+S", "XCN", "XCN+S": transmission-case variants

Why it matters

At solar-system distances, light time is not a rounding error — it can be minutes.

If you’re doing:

• observer-centric visualization (“what does the observer see?”)
• instrument pointing / attitude computations
• high-precision geometry

…you usually want consistent aberration correction choices.

Kernels required

• SPK: required (you need ephemerides for the target/observer)
• LSK: required if you’re converting from UTC strings (e.g. via kit.utcToEt / raw.str2et); not required if you already have ET
• FK/PCK/CK/SCLK: required if you request states in non-inertial or kernel-defined frames

How it maps to tspice

kit.getState({ aberration })

tspice exposes aberration correction as the aberration field on kit.getState(...).

• Default is "NONE".
• The returned StateVector includes lightTime (one-way seconds).

// Requires:
// - LSK (leapseconds) if converting UTC -> ET (e.g. naif0012.tls)
// - SPK(s) that cover the target/observer ephemerides (e.g. de440.bsp)
const et = await spice.kit.utcToEt("2000 JAN 01 12:00:00");

// Geometric state (default aberration: "NONE")
const geom = await spice.kit.getState({ target: "MARS", observer: "EARTH", at: et });

// Apparent state with one-way light time + stellar aberration
const app = await spice.kit.getState({
  target: "MARS",
  observer: "EARTH",
  at: et,
  aberration: "LT+S",
});

console.log(app.lightTime);

raw.spkezr(...) / raw.spkpos(...) (parity)

kit.getState(...) is a structured wrapper around raw.spkezr(...).

const { state, lt } = await spice.raw.spkezr("MARS", et, "J2000", "LT+S", "EARTH");
const { pos, lt: lt2 } = await spice.raw.spkpos("MARS", et, "J2000", "LT+S", "EARTH");

Gotchas

• Be consistent: comparing vectors computed with different abcorr choices is a common source of confusion.
• Transmission (X...) vs reception: SPICE distinguishes “where was the target when the observer received the photons?” (no X) from “where is the target when the observer transmits the photons?” (X...). Pick the convention that matches your model.
• Iterative corrections cost more: CN/XCN are typically slower than LT/XLT.
• lightTime is one-way seconds: it’s not automatically “round trip.”