It is important to understand the nature of an occultation observation. It is a recording of the alignment of the star, the limb of the moon, and the observer - at a particular time. Any one or more of those elements can be investigated and improved from occultation observations - with the focus usually being on the component that is worst-known. Thus during the 18th century, time was a huge unknown. During the 19th century accurate time became readily available, and the lunar ephemeris was the largest unknown. During the 20th century:

• the issue of the Earth's irregular rotation was of great interest. RGO became the central collection point for lunar occultation observations specifically to determine the existence of deltaT
• with the irregular rotation of the earth being established, improving the lunar ephemeris became a priority
• the Watts charts were created so that lunar occultations could be used to better establish observer longitudes, particularly across oceans
• baily beads at solar eclipses were observed and analysed to determine long-term variations of the solar diameter, based on an assumption that the edge of the moon was accurately known

In all these investigations, the data that can be obtained from a lunar occultation usually related to the element that was worst known. For example, 20 years ago the biggest uncertainty for a lunar graze was the catalogue position of the star, and star catalogue positions were an important outcome. With Hipparcos and Tycho2, the uncertainty in the star's position became much less than the uncertainties in the limb of the moon, and the position of the moon.

In recent years there have been a number of improvements that change the focus on the data to be extracted from a lunar occultation.

Firstly, the lunar ephemeris has been considerably improved as a result of lunar laser ranging. Secondly, the Hipparcos, Tycho2 and UCAC2 catalogues have reduced the uncertainty in the position of stars. This has meant that for the last 10 years or so the largest uncertainty has been in the lunar limb.

The preliminary Kaguya results that Mitsuru has demonstrated indicate that Kaguya is capable of producing a much more accurate lunar profile than the Watts charts - although the angular resolution is quite coarse. So where does this leave lunar occultation observations?

Assuming the Kaguya data is capable of generating reliable profile corrections around the limb of the moon, the component that has the greatest uncertainties changes. Most notably, the stellar reference frame established by Hipparcos is open to some fundamental questions. As Mitsuru Soma has written:

It is said that the positions and proper motions in the Hipparcos
Catalogue refer to the International Celestial Reference System (ICRS),
which is defined by the coordinates of extragalactic radio sources.
The claimed accuracy is 0.25 mas/year.  However, since Hipparcos was not
able to observe extragalactic radio sources (the Hipparcos program did
include the brightest quasar, 3C273, but it was too faint to contribute
to the link), the link had to be made indirectly, and therefore it is
possible that the Hipparcos reference frame has some rotation with
respect to ICRS.  In fact, the direct comparison of the proper motions
between Hipparcos and FK5 (Feissel and Mignard, 1998, A&A 331, L33)
gave inconsistent results with the precession error of -3 mas/year of
the FK5, which had been independently obtained from VLBI, lunar laser
ranging, and proper motion analyses.  Now that the lunar positions in
the latest JPL planetary and lunar ephemerides, such as DE405/LE405,
have the mas level accuracy with respect to the ICRS, the problem can
be resolved by analyzing lunar occultations using such ephemerides and
the Hipparcos Catalogue.  In 2000 Soma presented a preliminary result
of such an analysis at the IAU Colloquium 180 held at the U.S. Naval
Observatory, which suggested much larger errors in the Hipparcos
reference frame than its claimed accuracy.  If errors of lunar limb
corrections are really improved by the observations of the Japanese
lunar explorer Kaguya, it is expected that the results of such analyses
are also much improved.

To put this into context. With a PAL 25fps video camera running at a 1-frame resolution, the event time is established to 0.04secs. With the relative motion of the moon to the star being typically about 0.3"/sec, video easily provides a relative positional accuracy of 0.01". This is smaller than the catalogue uncertainties of UCAC2 (typically around 0.03"), and is at a level of precision where systematic errors in the Hipparcos reference frame can be investigated.

In addition, the Kaguya data does not have the same spatial resolution as lunar graze observations - such that Kaguya is unlikely to replace grazes in Baily Bead analysis.

Additionally, there are groups of observers in Australasia and North America who are starting to use lunar occultations to measure double star details - separation, position angle, and the relative magnitude (which is poorly known for most close double stars.)

So - are lunar occultation observations now obsolete? A very definite NO. All that is changing is the nature of the investigations that can be made from the observations. And this is just a continuation of the changes in focus that have occurred over the last 380 years of observations.

David Dunham
Dave Herald
Mitsuru Soma
***************************************

Dave Herald
Canberra, Australia