Premium
Variational retrieval of humidity profile, wind speed and cloud liquid‐water path with the SSM/I: Potential for numerical weather prediction
Author(s) -
Phalippou L.
Publication year - 1996
Publication title -
quarterly journal of the royal meteorological society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.744
H-Index - 143
eISSN - 1477-870X
pISSN - 0035-9009
DOI - 10.1002/qj.49712253002
Subject(s) - liquid water path , numerical weather prediction , precipitable water , wind speed , radiance , meteorology , a priori and a posteriori , radiative transfer , remote sensing , special sensor microwave/imager , environmental science , weather forecasting , computer science , water vapor , microwave , precipitation , geography , physics , brightness temperature , telecommunications , philosophy , epistemology , quantum mechanics
Several regression algorithms have been proposed to retrieve geophysical parameters from the Special Sensor Microwave/Imager (SSM/I) radiances. Their performances are generally limited by a simplified handling of nonlinearities and/or by the poor quality of the a priori information. In this paper, a variational method is proposed for retrieving the atmospheric humidity profile, the wind speed and the cloud liquid‐water path from SSM/I observations over ocean. This method is based on nonlinear optimal estimation theory. The first guess is derived from a European Centre for Medium‐Range Weather Forecasts forecast, and the forecast‐error covariance is used as a constraint. The geophysical variable space is mapped into the radiance space through a radiative‐transfer model which permits an accurate representation of nonlinearities. This method has been applied to several orbits and the results for one of them are presented and discussed. It is argued that the variational approach is a simple optimal way of extracting information from SSM/I radiances, exploiting the high quality a priori information available from a numerical weather‐prediction model. The retrieved humidity profiles are found to compare well with the total precipitable water estimated from a regression algorithm, while avoiding local bias in very dry and very wet conditions. It is also shown that the potential wind speed‐cloud ambiguity is removed through the use of a high‐quality wind speed first guess.