We propose observing GNSS satellites with VLBI operationally in the celestial frame. Basically, the terrestrial frame is fully determined only by the GNSS satellites and the celestial frame is based on the VLBI with AGNs. To combine these two fully independent frames, the only direct solution is to observe AGNs with VLBI in the L-band (L1 and L2 GPS frequencies). All other approaches would introduce local ties and additional effects. This GNSS frequency band is internationally a metrology frequency band outside the interferences with huge LEO constellations of CubeSats operating in the S-band and in the X-band. With the STE-QUEST mission proposal (Svehla et al. 2014), pre-selected (2009-2014) in the ESA Cosmic Vision Programme as well as in GRASP, eGRASP and GENESIS proposals and activities, there is only an indirect link via LEO or a similar orbit between GNSS satellites and AGNs in the celestial frame. This issue is not going to be solved in order to secure orbits of GNSS satellites directly connected with AGNs in the celestial frame. The only way to improve the accuracy of GNSS satellite orbits, and in this way the terrestrial and celestial frame, is to observe GNSS satellites with VLBI relative to the AGNs. LEO orbit is not sufficiently well suited to significantly improving the orbit accuracy of GNSS satellites (gravity, albedo, SRP and thermal re-radiation effects for a very short LEO orbit revolution) nor to be easily observed by all VLBI antennae on the ground. At higher orbit altitudes there are no active satellites placed in the inner Van Allen radiation belts <12000 km that would enable operating such a satellite actively over a longer period of time (e.g. passive LAGEOS satellites or active GPS satellites placed at much higher orbit altitudes). In terms of ICRF frame realization, positions or definitions of AGNs are ”bigger” in the L-band than the X-band, but there are lock-in amplifiers or the existing broadband VLBI receivers (VGOS) that can sample the VLBI signal up to the X-band and that could also cover the L-band simultaneously with the S-band and the X-band. Therefore, the overall cost of such a modification at a VLBI station has been significantly reduced over the years. It is expected that there will be an extension of the celestial frame to higher frequencies such as Q-band and V-band, mainly driven by NASA and ESA for the Delta-DOR technique, but the majority of VLBI stations could only focus on the lower frequency band that could be observed simultaneously in the L-band, S-band and X-band or even up to the Ka-band and used for the realization of reference frames. All GNSS frequencies in the L-band will remain internationally as a metrology frequency band in the future. References: Svehla D, Rothacher M, Hugentobler U, Nothnagel A, Willis P, Biancale R, Ziebart M, Appleby G, Schuh H, Adam J, Iess L and Cacciapuoti L (2014) Terrestrial and Celestial Reference Frame Realization with Highly Elliptical Orbit - The ESA STE-QUEST Mission. Geophysical Research Abstracts, Vol. 16, EGU2014-7934-2.

{"references": ["Svehla D, Rothacher M, Hugentobler U, Nothnagel A, Willis P, Biancale R, Ziebart M, Appleby G, Schuh H, Adam J, Iess L and Cacciapuoti L (2014) Terrestrial and Celestial Reference Frame Realization with Highly Elliptical Orbit - The ESA STE-QUEST Mission. Geophysical Research Abstracts, Vol. 16, EGU2014-7934-2."]}