Relation between the Celestial Reference System and the Terrestrial Reference System of a rigid Earth

A relation between the Celestial Reference System (CRS) and the Terrestrial Reference System is established theoretically by solving the equations of motion of a rigid Earth under the influence of the Sun and the Moon up to the second order perturbation. The solutions include not only nutation including Oppolzer terms but also the right ascension of the dynamical departure point (DP), as well as the wobble matrix.We have found that the kinematical definition of the Non-Rotating Origin NRO (for which our term is DP) given by Capitaine, Guinot and Souchay (1987) is not entirely equivalent to that included in the solutions of the equations of motion but shows perturbation, in particular when this is taken on the instantaneous equator. Besides this serious fault, we feel little merit in taking the DP as reference: (1) Unnecessary spurious mixed secular terms appear which come from the geometrical configuration that the DP leaves far and far from the ecliptic. (2) the DP moves secularly as well as oscillating with respect to space; this literally contradicts the term NRO, or is at least misleading. (3) It does not free us from the precession uncertainty to adopt DP as reference, since we cannot avoid virtual proper motions in terms of the current CRS. (4) No terms ignored hitherto are introduced, even if we take the DP properly chosen, i.e., on the equator of the celestial ephemeris pole. The transformation is only mathematical. There is no sufficient reason to take it instead of the equinox, which is observable in principle, as reference at the cost of the labor of changing all the textbooks, ephemerides, data and computer software now existing.