14A.4 Estimating The Probability Of Rapid Intensification of Tropical Cyclones in the Atlantic and Eastern Pacific Basins John Kaplan* Hurricane Research Division, NOAA/AOML Miami, Florida Mark DeMaria RAMMT/ NOAA/NESDIS Fort Collins, CO 1. Introduction While there has been some improvement in operational tropical cyclone (TC) intensity forecasting skill in recent years, TC intensity forecasts remain considerably less skillful than that TC track forecasts (Simpson et al. 2003). More importantly, the capability of predicting rapid intensification (RI) remains inadequate as evidenced by the unexpected RI of several recent Atlantic (e.g., Keith (2002), and Lilly (2002)) and East Pacific (Kenna (2002)) hurricanes. This lack of skill has prompted the National Hurricane Center (NHC) to rank improving the capability to forecast episodes of rapid intensification (RI) as one of their highest forecast priorities (Rappaport, personal communication). Since current operational intensity prediction models have not yet demonstrated the ability to reliably predict RI events, Kaplan and DeMaria (2003) have developed a simple RI index that can be used to estimate the probability of RI for a 24-h period using output from the SHIPS model. The RI index has been provided in real-time to forecasters at the NHC for the past three Atlantic hurricane seasons as part of the Joint Hurricane Testbed (JHT). Although, the RI index has been designated for operational implementation by the NHC commencing with the 2004 hurricane season, it was originally developed for the Atlantic basin. Thus, an analogous index is being developed for the East Pacific basin. This paper will discuss the derivation of the East Pacific RI index and will compare its performance to the Atlantic version. 2. Data The data employed in this study were obtained from the 1989-2002 SHIPS and NHC HURDAT databases. The former contains synoptic and storm-scale predictors evaluated every 12h for all tropical and subtropical cyclones in the Atlantic and Eastern Pacific basins. The SHIPS synoptic atmospheric predictors were evaluated using the t=0 h NCEP Aviation analysis fields while the synoptic oceanic predictors were evaluated using the most recent weekly Reynolds sea-surface temperature analysis. The storm-scale inner-core predictors were * Corresponding author address: John Kaplan, NOAA/AOML/HRD, 4301 Rickenbacker Causeway, Miami, FL 33149. E-mail: John.Kaplan@noaa.gov evaluated using GOES infrared imagery (Zehr 2000). The NHC HURDAT file, which contains TC positions and intensities every 6 h from 1886 to the present, was employed to evaluate the climatological and persistence predictors. Fig. 1. The cumulative frequency distributions for the Atlantic and Eastern Pacific basins. 3. Results Figure 1 shows the 24-h over-water intensity change (∆V24) distributions for tropical and subtropical cyclones that formed in the Atlantic and Eastern Pacific basins from 1989-2002. The distributions indicate Eastern Pacific cyclones exhibit a larger range of ∆V24 than do Atlantic cyclones. Specifically, the maximum (minimum) ∆V24 observed in the Eastern Pacific basin is 46 ms -1 (-44 ms -1 ) while the maximum (minimum) intensity change observed in the Atlantic basin is only 33 ms -1 (-26 ms -1 ). Following the methodology of Kaplan and DeMaria (2003), RI was defined as the 95 th percentile of all ∆V24 values. This corresponds to a ∆V24 of ~16 ms -1 and ~19 ms - 1 in the Atlantic and Eastern Pacific basins. However, for the purpose of this study we have defined RI as a ∆V24 of >15.4 ms-1 (30 kt) and >18 ms -1 (35 kt) for the Atlantic and Eastern Pacific basins, respectively. Thus, the actual percentage of systems undergoing RI in the Atlantic and Eastern Pacific basins is about 6%. Figure 2 shows the locations of the RI cases in both basins. Interestingly, the RI cases in the Atlantic basin occur over a much broader area than those in the Eastern Pacific basin. Specifically, the Atlantic RI cases are most prevalent in