Validity of ephemeris time

Question

In the wikipedia article for ephemeris time as first adopted in 1952, it does not explain clearly how or why it was superseded. For example, it reads:

In 1976 the IAU resolved that the theoretical basis for its current (1952) standard of Ephemeris Time was non-relativistic, and that therefore, beginning in 1984, Ephemeris Time would be replaced by two relativistic timescales intended to constitute dynamical timescales: Terrestrial Dynamical Time (TDT) and Barycentric Dynamical Time (TDB).[28] Difficulties were recognized, which led to these being in turn superseded in the 1990s by time scales Terrestrial Time (TT), Geocentric Coordinate Time GCT(TCG) and Barycentric Coordinate Time BCT(TCB). [16]

So, from this it appears that ephemeris time was never superseded and it continues to be evolved with a different theoretical basis?

From reading Robert Newton's books such as "Ancient planetary observations and the validity of ephemeris time", it would seem that ephemeris time is invalid because of non-secular accelerations in the motion of Mercury and Venus.

So, I guess my question is: does current astronomical theory consider ephemeris time to be invalid or valid, and if it considers it to be valid, then what is current evaluation of Newton's calculations and their implications for the validity of ephemeris time?

Answer 1

It depends on what one means by "Ephemeris Time." Officially, the Newtonian-based Ephemeris Time has been superseded by the relativistically-correct Barycentric Dynamic Time (TDB) and Barycentric Coordinate Time (TCB). However, the concept of what was meant by Ephemeris Time carries on to this day. The intent of Ephemeris Time was to use the solar system as a timekeeping device. This is what TDB and TCB do, but in a relativistic rather than Newtonian setting.

TCB is a theoretical time scale whose clocks are infinitely removed from the solar system, are co-moving with the solar system barycenter, and that tick at the rate of one tick per 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the caesium-133 atom (at a temperature of 0 K). (Note well: This is the same definition currently used for an SI second.) That TCB clocks are well removed from the solar system means that a second measured by an atomic clock on the surface of the Earth ticks a tiny bit slower than does a TCB clock.

These slightly different tick rates was perceived as a problem by some. The solution was eventually to redefine TDB as a scaled version of TCB such that, on average, TCB ticks at the same rate as atomic clocks on the surface of the Earth. Atomic clocks on the surface of the Earth tick faster / slower than would TCB clocks based on whether the Earth is further from / closer to the Sun than average. But averaged over the course of years, TCB and TAI remain more or less in sync.

Regarding the more or less: TDB is updated every year. As technology improves, these subtle updates change the timing of past events. TAI does not do this. As technology improves, older atomic clocks are replaced with newer ones. Atomic clocks have a shelf life of a decade or so. This can result in discontinuities in atomic time. This is not a problem with TDB as it is re-issued annually. This also means that TDB is independent of atomic time.