Abstract
Assimilation of SSM/I-retrieved rainrate has been shown to be beneficial in the numerical prediction of tropical cyclones with a limited-area model (Peng and Chang, 1996). However, due to the near sun-synchronous orbits and the narrow swath, SSM/I coverage of the tropics is not complete. There are many instances that one or more tropical cyclones are not covered by SSM/I swaths; therefore, rainrates cannot be retrieved for assimilation into the prediction models. In this study, the effect of the assimilation of rainrate inferred from cloud top brightness temperature (BT) is assessed. The algorithm inferring rainrates from infrared (IR) cloud top brightness temperature based on concurrent retrieval of the Special Sensor Microwave/Imager (SSM/I) rainfall of Manobianco et al. (1994) was applied in a numerical prediction experiment for Typhoon Flo of 1990. It was found that assimilating IR rainfall rates produced a positive impact on both track and intensity forecasts of Typhoon Flo, qualitatively similar to those in Peng and Chang using the SSM/I-retrieved rainrates. The analytic algorithm relating the IR rainrates to the SSM/I rainrates became less accurate when concurrent IR and SSM/I brightness temperatures were not available. The accuracy of the IR rainrates and improvements in forecast skill both decreased in a sensitivity experiment where IR rainrates were obtained by an obsolete relationship of IR brightness temperature and rainfall rates.
