Hello all,

I cross-post because I've seen suggestions on both lists suggesting
that, if you use heliocentric elements close to the time of the flyby,
you can get a decent ephemeris. In particular, since the perigee is at
15.8 February, elements for 16 February "ought" to work Just Fine. Surely,
the thinking goes, the perturbations can't be _that_ much over a few hours.

I just ran a comparison of unperturbed ephemerides using heliocentric
elements for epoch 16 February, versus an integrated ephemeris (good to
within a couple of arcminutes at present). Of course, they agree exactly
at 16 February 00:00 UTC, 4.5 hours after perigee.

Two hours earlier, at 22:00 UTC on 15 February, the error is 8 arcminutes.
An hour before that (and 1.5 hours after perigee), the error is over a degree.
At 20:00 UTC 15 February, the error is almost six degrees.

Geocentric elements fit the "real" ephemeris to within an arcminute or two
at all points during the flyby, even without ephemerides. Either use them,
or use integrated ephemerides.

-- Bill