The study also shows that, in a few stations, the mean differences of in situ surface pressure derived ZHD with respect to both ECMWF-025 and ERA Interim estimates reveal non negligible values of several millimetres (four stations have a bias with an absolute value larger than 5 mm). For VMF1, 17 stations have biases with an absolute value larger than 5 mm and 6 with values larger than 1 cm. Only with these data, without additional information, it is not possible to ascertain these biases to the in situ measurements or to local small scale systematic errors in the ECMWF and VMF1 models, and these deserve a future analysis. A global comparison of ZWD values, estimated from GNSS-derived ZTD using the ZHD computed as described in this paper, with independent values of ZWD calculated from measurements of microwave radiometers on board altimeter missions such as Jason-1 and Envisat, is under analysis and will allow to draw conclusions about the origin of these biases. Using eleven IGS sites and a period of 1.5 years, Kouba (2009b) reported that the ZHD of the gridded VMF1 compare with the site-dependent VMF1 ZHD with RMS of 3 mm, subject to some biases and discontinuities of up to 4 cm, which are likely due to different strategies used in the generation of the site-dependent VMF1 data. The results by Kouba (2009b) cannot be directly compared with the results presented in this study, since they refer to comparisons between ZHD derived from two ECMWF products, with different spatial resolutions. These results further corroborated with the analysis presented in this study show that the precision of gridded VMF1 zenith hydrostatic delays should be sufficient for accurate a priori ZHD mapping in most of precise GNSS and very long baseline interferometry. However, for the purpose of separating the dry and wet components of the total tropospheric delay, a more precise source of ZHD such as ECMWF-025 or ERA Interim is advisable.