HR: 14:40h
AN: G42B-05    [PDF]
TI: An error analysis of ground-based GPS slant-path residuals in a simulation study
AU: * Ha, S
EM: syha@ucar.edu
AF: University Corporation for Atmospheric Research, COSMIC Program, P.O.Box 3000, Boulder, CO 80307 United States
AU: Braun, J
EM: braunj@ucar.edu
AF: University Corporation for Atmospheric Research, COSMIC Program, P.O.Box 3000, Boulder, CO 80307 United States
AU: Kuo, Y
EM: kuo@ucar.edu
AF: University Corporation for Atmospheric Research, COSMIC Program, P.O.Box 3000, Boulder, CO 80307 United States
AU: Rocken, C
EM: rocken@ucar.edu
AF: University Corporation for Atmospheric Research, COSMIC Program, P.O.Box 3000, Boulder, CO 80307 United States
AB: The accuracy of a method for obtaining slant-path delays from GPS double difference residuals is assessed in a severe squall line case. To retrieve the absolute value of phase delay along each ray path from GPS double differences, Alber et al. (2000) introduced "zero mean difference assumption" as an additional independent constraint. In this study, we perform an error analysis of the zero mean difference assumption by simulating "true" slant-paths defined by GPS receiver sites and satellite orbits in a high-resolution numerical weather prediction model. Using these slant delays we examine the validity of the so-called "zero mean assumptions". Despite the existence of a strong systematic bias associated with the surface cold front and severe convection, deviation from the zero mean difference assumption is on the order of a few mm in the unit of phase delay. This implies that slant-path residuals contain useful information on the spatial variability even when the zero-mean assumption is violated due to atmospheric conditions. The representative error is also investigated by comparing the vertically integrated zenith hydrostatic delay with the one averaged over all available zenith-mapped slant hydrostatic delays for a certain period of time. In spite of strong horizontal pressure gradient over the entire observation network, the representative error of zenith hydrostatic delay is about 0.1%.
DE: 1294 Instruments and techniques
DE: 3337 Numerical modeling and data assimilation
SC: Geodesy [G]
MN: 2003 Fall Meeting