Hi Andrew,
> For anyone interested I have just emailed Bill, the copy of the
> rwo file I downloaded a few days back was slightly bigger, it had
> at least one extra column "Ast Cat" before the observatory code
> which is now absent.
Yup, that's the culprit.
The "new" format is marked as version 2 in the header (used to
be always version 1) and the "errmod" is marked as "cbm10" (used
to be blank). The biases, previously always zero, now contain
some non-zero values. And the star catalog code, provided in
column 71 of MPC files, has been added just before the observatory
code in the .rwo files. I should be able to provide an update that
will load in these "version 2" .rwo files. (Though I must admit
that thus far, I've been unable to access such files myself. They
are actually to be preferred to the version 1 format, since they
provide that extra bit of data regarding which star catalog was
used to compute the position.)
Apparently, the new .rwo files incorporate the de-biasing
strategy described in Icarus 210 (2010) 158-181, "Treatment of
star catalog biases in asteroid astrometric observations", by
Steve Chesley, Jim Baer, and Dave Monet.
-- Bill