HR: 14:25h
AN: G42B-04    [PDF]
TI: Statistical distributions for water vapor parameters
AU: * Elosegui, P
EM: pelosegui@cfa.harvard.edu
AF: Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, MS-42, Cambridge, MA 02138 United States
AU: * Elosegui, P
EM: pelosegui@cfa.harvard.edu
AF: Institut Estudis Espacials Catalonia, Gran Capita 2-4, Barcelona, 08034 Spain
AU: Davis, J L
EM: jdavis@cfa.harvard.edu
AF: Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, MS-42, Cambridge, MA 02138 United States
AB: In assessing the accuracy of atmospheric information obtained from GPS observations, it is customarily (and conveniently) assumed that water vapor parameter estimates follow a Gaussian distribution. This assumption is clearly inappropriate for atmospheric information of interest such as wet zenith delays and wet delay gradients. Wet zenith delay estimates, for example, can present very asymmetric (skewed) distributions, and wet delay gradients are known to be nearly zero for most of the time accompanied by instances of relatively large values. A Gaussian distribution is a poor descriptor of such parameters. It is thus critical to determine the statistical basis of these atmospheric parameters for assessing the accuracy of GPS-based meteorological and climatological inferences and predictions. Comparisons of root-mean-square and other statistics that are based on a Gaussian assumption will lead to erroneous conclusions regarding the significance of these statistics. In this presentation, we will describe our theoretical approach and experimental studies to determine the statistics of water vapor parameters.
DE: 1200 GEODESY AND GRAVITY
DE: 3300 METEOROLOGY AND ATMOSPHERIC DYNAMICS
DE: 3360 Remote sensing
SC: Geodesy [G]
MN: 2003 Fall Meeting