Abstract
A method for introducing bogus observations in the analysis of tropical cyclones has been tested for a period in 1991. The impacts have been evaluated for analyses, first-guess forecasts during data assimilation, and medium-range forecasts. In almost all cases, the inclusion of bogus data leads to a clear improvement in both the position of the cyclone center and the intensity and structure of the associated vortex. The intensity is generally improved in the forecasts and the track errors are reduced in the short-range forecasts. However, medium-range forecasts show a mainly unfavorable impact of this bogusing method in terms of predicted cyclone tracks for the cases investigated.
