On the consistency of the current conventional EOP series and the celestial and terrestrial reference frames

Precise transformation between the celestial reference frames (crf) and terrestrial reference frames (trf) is needed for many purposes in earth and space sciences. according to the global geodetic observing system (ggos) recommendations, the accuracy of positions and stability of reference frames should reach 1 mm and 0.1 mm year $$^{-1}$$ , and thus, the earth orientation parameters (eop) should be estimated with similar accuracy. different realizations of trfs, based on the combination of solutions from four different space geodetic techniques, and crfs, based on a single technique only (vlbi, very long baseline interferometry), might cause a slow degradation of the consistency among eop, crfs, and trfs (e.g., because of differences in geometry, orientation and scale) and a misalignment of the current conventional eop series, iers 08 c04. we empirically assess the consistency among the conventional reference frames and eop by analyzing the record of vlbi sessions since 1990 with varied settings to reflect the impact of changing frames or other processing strategies on the eop estimates. our tests show that the eop estimates are insensitive to crf changes, but sensitive to trf variations and unmodeled geophysical signals at the ggos level. the differences between the conventional iers 08 c04 and other eop series computed with distinct trf settings exhibit biases and even non-negligible trends in the cases where no differential rotations should appear, e.g., a drift of about 20 $$upmu $$ as year $$^{-1 }$$ in $$y_{mathrm{pol }}$$  when the vlbi-only frame vtrf2008 is used. likewise, different strategies on station position modeling originate scatters larger than 150 $$upmu $$ as in the terrestrial pole coordinates.