Earth Planets Space, Vol. 52 (No. 11), pp. 885-892, 2000
E. Robert Kursinski1, Sean B. Healy2, and Larry J. Romans1
1MS 238-600 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109-8099, U.S.A.
2Numerical Weather Prediction Division (Room 338), The Met Office, London Road, Bracknell RG12 2SZ, U.K.
(Received January 20, 2000; Revised August 28, 2000; Accepted August 28, 2000)
Abstract: We present results of combining occultation refractivity profiles from GPS/MET with ECMWF global analyses in a 1DVar framework in order to separate the wet and dry contributions to refractivity and assess their impact on the analyzed temperature, surface pressure and specific humidity fields. We find significant zonal mean temperature, surface pressure and humidity differences between the 1DVar solutions and the ECMWF analyses reflecting biases between the GPS refractivities and ECMWF analyses. Large profile-to-profile temperature discrepancies in the tropical lower stratosphere are due to waves not represented in the analyses. The 1DVar solution is generally drier than ECMWF particularly in the southern subtropics. Lack of moisture above 300 hPa in the present model caused the solution to make large adjustments in low latitude surface pressure and tropospheric temperatures to increase upper troposphere densities and compensate for the missing upper level moisture. The discrepancies between the solution and the background and observational data sets represent roughly a 2-sigma level of agreement rather than the 1-sigma level desired in a 1DVar solution. Given the simplicity of our error covariances, our results are promising as a first step. In the future, the error covariances need to be refined and, in particular, to vary with location.