A statistical classification of cyclone tracks in high latitudes of the Southern Hemisphere: Preliminary results

LUCINI, MARÍA MAGDALENA; PULIDO, MANUEL ARTURO; PELOZO, SILVIA

Montreal, Canadá

Congreso; MOCA 09- IAMAS/IACSO/IACS Joint Assembly; 2009

We apply an automatic feature tracking technique to detect and follow cyclones in the high latitudes of the Southern Hemisphere using ERA 40 data (ECMWF). Cyclone tracking provides an alternative approach to the usual Eulerian studies of the storm-tracks. In order to characterize the storm track variability, cyclone trajectories are classified statistically using two clustering techniques: Mixed-Gaussians and K-means. The variabity of the cyclone track clusters for different years is discussed. We also classify the genesis of cyclones by means of a clustering technique. In this preliminary study we focus mainly on the reliability of the statitistical techniques: a careful sensitivity study to the different parameters of the techniques is presented. For the optimal set of parameters we show that both clustering techniques give in general similar results. We discuss differences in the cyclone clusters between the pre and after satellite era.