From my point of view, the velocities for a set of points computed on ITRF2000 will be very similar to those calculated in ITRF2008. This can give rise to try and use a euler pole estimated using ITRF2000 to transform another set of velocities defined in ITRF2008.
However, the euler poles computed for a set of points will be strictly different for both geodetic frames. If an euler pole calculated for a geodetic reference frame is used to transform a set of points obtained in another reference frame, a component error for this fact will be entered. This component will be larger as more local is the set of points.
If finally a euler pole estimated using ITRF2000 to transform another set of velocities defined in ITRF2008 is used, the magnitude of this component error must be evaluated in order to avoid its propagation to the new values.
Thank you Juan for the answer. May be the transformation parameter used to convert observations from one ITRF to another may help. If we convert the velocities used to define the pole to the new ITRF and then re-estimate the pole, we can use it for the transformation. Its just an idea which need verification.
Different set of geodetic point on a same frame can provide different euler pole, also additional measurements on the same geodetic points may change the velocity of a point therefore the Euler pole. Your question is subjective and lot of precautions need to be taken before concluding.
thank you all for your answers.... one thing is clear that, since GPS network solution strongly depend on how we are constraining it (i.e. the ITRF), the pole determined using ITRF2000 cannot be simply used to transform velocities estimated in ITRF2008. ITRF2008 is much more improved than ITRF2000. So better we make the frame and the pole definition uniform and then transform.
The Euler Pole estimation correspond to the Helmert-type velocity transformation model using only rotational rates. Euler Pole Parameters have one-by-one dependency to rotational rates about the three axes.
In my opinion, use the Helmert type transformation model using shift and rotational rates (not scale).