Atmospheric Forecasting

Abstract

Numerical experiments were conducted to assess the impact of Omegadropwindsonde (ODW) data and Special Sensor Microwave/Imager (SSM/I) rain rates in the analysis and prediction of Hurricane Florence (1988). The ODW data were used to enhance the initial analysis that was based on the National Meteorological Center/Regional Analysis and Forecast System (NCEP/RAFS) 2.5° analysis at 0000 UTC 9 September 1988. The SSM/I rain rates at 0000 and 1200 UTC 9 September 1988 were assimilated into the Naval Research Laboratory's limited-area model during model integration.Four numerical simulations were performed to assess the impact of ODW data and SSM/I rain rates on numerical forecasts of Hurricane Florence (1988). A nested three-pass Barnes scheme objective analysis was used to enhance the NCEP/RAFS 2.5° analysis because of the relative high resolution of the ODW data near the center of Florence. The NCEP/RAFS 2.5° analysis and the ODW enhanced analysis were then used as the initial conditions for the numerical prediction experiments. We found that the forecast error of the intensity was reduced from 14 to 4 mb with the assimilation of the ODW data. It is also clear that the ODW data reduced the forecast error of the landfall time in the control experiment by 5 hours (down from 9 to 4 hours), and the landfall location error by 95% (down from 294 to 15 km). The impact of the ODW data was more pronounced in the first 24 hours of the integration and gradually decreased beyond 24 hours. This result is in agreement with Franklin and DeMaria (1992), in which a barotropic, nested, spectral hurricane track forecasting model was used to determine the impact of the ODW data. The advantage of using a three-dimensional primitive equation model with the inclusion of baroclinic and physical processes, as in this study, is that the model can simulate the hurricane structure in more detail (DeMaria et al., 1992).However, a major limitation of the ODW observations is that they do not provide information above 400 mb because of the altitude limitations of the NOAA WP-3D, while satellite-derived winds can provide some upper-level information . Detailed observations of the three-dimensional structure of the upper levels of tropical cyclones are extremely scarce. Dropsonde data collected from 200 mb to the surface by the recent TCM-90 experiment (Elsberry and Abbey, 1991) may be proven to be very useful in depicting the upper-level structure of tropical cyclones (Merrill and Velden, 1995).The rainfall rates retrieved from the SSM/I data at 0000 UTC and 1200 UTC 9September 1988 were assimilated into the NRL limited-area model initialized with the NCEP/RAFS 2.5° analysis (Exp. Control+SSMI) and the ODW enhanced analysis (Exp. ODW+SSMI). Results indicate that the assimilation of the SSM/I rain rates (Exp.Control+SSMI) reduced the 24-hour forecast errors in the landfall location in the control experiment by about 43% (down from 294 to 169 km), and the landfall time by 5 hours (down from 9 to 4 hours). At mean time, the 24-hour minimum SLP and ...