1. The possible options for an extended definition (c.f. Newsletter 2)

The extended definition of the CEP must take into account the high frequency components both in polar motion and nutation at the microarcsecond level and must be not dependent on the techniques and strategy of observations.

Definition

The separation between the terrestrial and the celestial parts of the motion as considered for the current definition of the CEP cannot be extended to the high frequency domain, as there is, in such a domain, an overlapping between the motions in the CRS and the TRS. An other kind of separation has therefore to be proposed for extending the definition in such a frequency domain. The definition must clearly specify which part of the motion is considered in the CRS or the TRS on the basis on clear theoretical considerations. Two different approaches have been considered in order to achieve this goal :

A)

a deterministic approach (previously called dynamical), separating the motions according to their physical cause,

B)

a frequency approach separating the motions according to their frequency domain.

Following the deterministic approach, the predictable sub-diurnal terms can be included (options A1, A2) or not (A3) in the nutation model and the predictable high frequency motion in the TRS can be included (A2) or not (A3) in the model for the polar motion

Following the frequency approach, Mathews (1999 a) has proposed a definition (B1) of the CEP which keeps the symmetry in the frequency band between terrestrial motion and celestial motion by extending the definition outside the current frequency interval.

An other option (B2) is to separate the motions according to the known frequencies of the predictable motions in the CRS and TRS. This option corresponds rather to a deterministic approach.

Realization

The procedure (C1) proposed by Mathews for realizing the definition B1 is to estimate simultaneously, in processing the observations, the current celestial pole offsets and current polar motion and the high frequency signal. The estimated high frequency components are diurnal and semi-diurnal retrograde periodic terms in the celestial motion and diurnal and semi-diurnal prograde periodic terms in the terrestrial motion. Such a procedure realizes a definition of the CEP which extends the current one on a symmetric way in the frequency domain (see Brzezinski 1999 for an extensive discussion on the use of these high frequency parameters in polar motion and nutation). However, the estimated high frequency motions do not correspond to the predictable ones, which are prograde within the CRS and both prograde and retrograde in the TRS.

An other proposed procedure (C3) is similar to the previous one but estimating only the current celestial pole offsets in the CRS and the whole predictable high frequency signal (prograde and retrograde diurnal components as well as prograde and retrograde semi-diurnal ones) in the TRS. This corresponds to a deterministic approach .

Other possible realizations are based upon a global analysis (C2) of the current estimates of the EOP (or only the pole coordinates) over a long period in order to extract the high frequency signal.