“Optimal” Conditions for Rapid Intensifications of Tropical Cyclones with Limited Factors
Rapid Intensifications (RI) of tropical cyclones (TC) provide major error sources in the challenging task of TC intensity forecasting. There are many factors affecting the RI processes of TCs, and identifying the combination of conditions most favorable to RI development is very time consuming when using traditional statistical data analysis methods. Data mining techniques have been implemented in the analysis of RI of TCs, and a simpler combined condition is identified, which gives a higher RI probability than a more complex condition from ordinary statistical analysis. Moreover, the data mining technique can be used to identify the “optimal” RI conditions when the number of affecting factors is given. The variation of RI probabilities with the factor numbers leads to a saturation stage, and individual cases are traced back for the cases with the globally most favorable RI conditions. In this paper, we will report the most recent findings through the data mining technique based on the data for SHIPS (Statistical Hurricane Intensity Prediction Scheme), an operational statistical-dynamical hurricane intensity forecasting model.
Yang, R., Tang, J., Kafatos, M., (2008) “Optimal” Conditions for Rapid Intensifications of Tropical Cyclones with Limited Factors. 28th Conference on Hurricanes and Tropical Meteorology, American Meteorological Society, Orlando, FL, 28 April – 2 May 2008.