On the use of HF as a reference for the comparison of stratospheric observations and models