Two satellites, two predictions, checked against orbit

A prediction that is never checked is an opinion. These are two we can check, one against a completed re-entry and one against a satellite still in orbit.

VERONIKA. We first ran the analysis in November 2024. It put re-entry at 4 June 2025 under strong space weather, 8 October 2025 under average conditions, and 6 April 2026 under weak. The satellite came down on 15 August 2025.

The average case was 54 days from the actual date, about 20 percent of the lifetime the satellite had left when we ran it. With solar activity running high, we recalculated in April 2025, four months before the end. That run landed three days from the actual date, about 2.5 percent of the remaining lifetime.

That is the shape of the result we care about. The first run was already inside the range published work treats as good, and the rerun, with the space weather that had actually happened, was closer again.

LASARSAT. We predicted its decay in mid-April 2025, from a semi-major axis of 495 km. On 18 September 2026 the measured value was about 410 km, against about 406 km predicted for that date.

At its current decay of about 300 m a day, that 4 km gap is about 13 days behind the prediction after seventeen months, which is about 2.5 percent of the time elapsed. Read the other way, the satellite lost 85 km over those seventeen months and the prediction missed by 4 km, which is 4.7 percent of the decay. Same result, two readings, and the second is why the first is worth stating.

We state that against the decay rather than against altitude on purpose. Measured against the whole orbit the same 4 km would look smaller, and it would look smaller however badly the prediction had done, because most of the altitude is still there. What a decay prediction gets right or wrong is the part that decays.

Both numbers land in the same place. Veronika’s April 2025 rerun came in at about 2.5 percent of the lifetime then remaining, and Lasarsat is at about 2.5 percent of the time elapsed since its prediction, against the 10 to 20 percent the published benchmarks treat as normal for long-horizon work.

FOR CONTEXT. Long-horizon lifetime prediction is generally accurate to roughly 10 to 20 percent of remaining lifetime. King-Hele put the practical limit at about 10 percent; the IADC re-entry campaign analyses by Pardini and Anselmo report errors of that order; and the 20 percent window is the one ESA and Aerospace Corporation work to. In every one of them the largest single source of error is the solar activity forecast, which is also what moved Veronika between our two runs.

THE LIMITS. Lasarsat has not re-entered. It is a checkpoint and not a completed prediction, and the only thing that will settle it is whether the satellite stays inside the band we stated. Veronika is settled, and it is one object. Two cases are two cases; they are not a validation campaign, and we are not presenting them as one.

Both runs were made on our own mission dynamics engine.

What the analysis is, and what arrives in it, is on the Orbit Decay Report.

Have an object with a compliance date coming? BOOK A CALL (opens in a new tab)→