Atmospheric Sensing Using GNSS RO - GNSS Remote Sensing: Theory, Methods and Applications - page 137

Fig. 6.1 Refractivity difference between RO occultation soundings (COSMIC-LEO#01, in red
and TerraSAR-X, TSX, in blue ) and the near-coincident NCEP analyses (interpolated into the time
and place of the RO soundings) from December 26, 2010 to January 5, 2011. The mean differences
and the standard deviation ( left ) as well as the number of RO soundings reaching various levels in
the atmosphere are shown (Courtesy of C. O. Ao at NASA Jet Propulsion Laboratory)

(or temperature) cannot be independently derived and require the a-priori informa-
tion of temperature (or water vapor) as well as hydrostatic balance assumption. With
the variational methods, e.g., the one dimensional variational (1DVar) to 4DVar
analysis method, temperature, and water vapor profiles can be derived simultaneous
(e.g., Kuo et al. 2000 ; Kursinski et al. 2000b ;Palmeretal. 2000 ; Poli et al. 2002 ).
The errors in the a-priori information will propagate into the derived temperature or
water vapor retrieval.

6.2.6 Experimental Validation of RO Accuracy and Precision

The accuracy of the RO measurements can be determined through comparisons
with independent data (e.g., high-quality radiosondes and model analyses). For
example, Fig. 6.1 shows the mean and standard deviation of the 1 month of COSMC
(only LEO#01 satellite) and TerraSAR-X RO refractivity as compared with the