On Tue 2003-07-01T21:51:09 +0100, Markus Kuhn hath writ: > http://adsabs.harvard.edu/cgi-bin/nph-bib_query?bibcode=1994A%26A...282..663S&db_key=AST
Caution. It is extremely important to note which reference frame is being used. > Thanks. If I understood it correctly, section 5.8.3 gives me with > > L = 100°.46645683 + 1295977422".83429*t - 2".04411*t^2 + 0".00523*t^3 All of the expressions 5.8 are with respect to the mean dynamical ecliptic and equinox J2000. That is a reference frame which is very nearly, but not quite, inertial. As such it does not give the longitude of the sun with respect to the seasons that the Gregorian calendar attempts to mark. The expressions 5.9 are referred to the mean dynamical ecliptic and equinox of date. Those relate more closely to the calendar, and as such they are a much better match to Newcomb's values. > The large number of decimal digits provided for the Newcomb formula in > the Metrologia leapsecond paper had suggested to me that all this is > rather exact science and that these digit are actually significant ... :-( precision vs. accuracy. The precision is necessary in the sense that UT1 is still determined in a fashion which refers to the fictitious mean sun of Newcomb's tables, such as they were. Whenever a new and better model is applied, the boundary conditions are set such that there is no measurable discontinuity in UT1 during the changeover. This practice has been followed for the past century, and was an issue of concern just this last year when the new IAU system (which abolished the ecliptic and equinox) was instituted on Jan 1. This means that whatever errors are present in the previous set of expressions are carried into the initial conditions of the new expressions. As such, the longitude the fictitious mean sun accumulates small errors based on the current amount of imprecision, and the mean time at which the sun actually crosses the Greenwich meridian does slightly deviate from 12:00 UT. This small drift is part of the motivation being used by the leap second haters to justify completely disconnecting civil time from the sun. But it should be pointed out that the calculation of TAI suffers from an analogous drift. The rules for calculating TAI have changed several times, which means that a best extrapolation of a completely uniform timescale back through the history of atomic time would deviate from the published values of TAI. > Is the Simon et al. formula considered good enough to predict the mean > longitude of the sun within half a day for the next few tenthousand > years? The authors make only a statement on the quality of ephimerides > derived for the years 1000-3000. The accuracies are given in table 7 as 6 arcsec for the earth/moon. For ranges outside the years 1000-3000 it would be better to see the referenced works of Bretagnon and Laskar. -- Steve Allen UCO/Lick Observatory Santa Cruz, CA 95064 [EMAIL PROTECTED] Voice: +1 831 459 3046 http://www.ucolick.org/~sla PGP: 1024/E46978C5 F6 78 D1 10 62 94 8F 2E 49 89 0E FE 26 B4 14 93
