Abstract
The impact of satellite radiometric soundings on numerical forecasts with the Israel Meteorological Service's three-parameter, haroclinic model is evaluated. Two sets of experiments are performed. One set (26 forecasts) uses experimental VTPR (Vertical Temperature Profile Radiometer) temperature profiles retrieved by the Goddard Institute for Space Studies during January 1974; the second set (13 forecasts) uses operational VTPR temperature profiles transmitted over the global meteorological telecommunications system in January 1976. In both sets of experiments the VTPR soundings are used to enhance the initial analyses over the Atlantic Ocean. Parallel forecasts with and without the VTPR data are verified against operational analyses over a region that includes most of Europe, the Middle East and North Africa. The results indicate that, on the average, there is a small improvement in the forecasts that utilize the satellite temperature soundings. For example, the root-mean-square height errors of the 48 h 500 mb forecast are reduced by ∼4% in the cases utilizing experimental VTPR data and by ∼3% in the cases using operational VTPR data. In both sets of experiments, this average effect is composed of a combination of marginally influenced, positively influenced and negatively influenced forecasts.
