An improved mean-gravity model for GPS hydrostatic delay calibration

The determination of global positioning system (GPS) heights with submillimeter accuracy needs proper correction of tropospheric delay. In this letter, the modeling of zenith hydrostatic delay (ZHD) was addressed, considering that wet delay must be treated separately. ZHD is traditionally estimated from Saastamoinen's formula using a mean-gravity model and surface pressure observations. The uncertainty in ZHD associated with the mean-gravity model is about 0.3 mm. An improved parametric model is derived here, which yields an uncertainty in the ZHD less than 0.1 mm when the surface altitude lies in the range of 0-9 km. A second parametric model is derived for higher altitudes (such as above radiosonde observations or atmospheric models). Both parametric models depend on latitude, height, and time variables. This dependence is due to the link between the mean gravity and temperature profiles between the surface and -80-km altitude. The uncertainty in the parametric models due to short-term temporal variability of the temperature profiles is shown to produce an uncertainty in ZHD smaller than 0.1 mm.