Salstein et al. (2008) compared the surface pressure fields from the ECMWF and NCEP atmospheric operational analyses to each other and to a set of land and ocean observations for a 5-year period (2001–2005). The two analyses were found to fit the observations equally well, a result that indicates similar uncertainty levels for both ECMWF and NCEP. Other authors, Ponte and Dorandeu (2003) using a smaller number of observations for an earlier period (1993–1995) attribute less accuracy to the NCEP analyses when compared with ECMWF. The results by Salstein et al. (2008) suggest that changes in the data assimilation systems over the last decade have considerablyreduced any major differences between ECMWF and NCEP pressure analyses, particularly in places like the Southern Ocean. Being recognised that the air tides cannot be properly modelled when performing time interpolation from NWM 6h grids such as those from ECMWF, the impact of the air tides on ZHD computation was assessed. It was found that the effect is latitude dependent, being maximum at the equator with values ranging between ±3 mm. The same effect affects the VMF1 grids which also use 6h sampling. Therefore, the air tide correction should also be applied to VMF1. However, the application of the air tide correction to VMF1 grids would hardly have any impact, since these grids possess errors much larger than the magnitude of this correction.