Discovering Predictors for Rapid Intensification of Cyclones in Bay of Bengal

Abstract

Accurate forecast of cyclone intensities is important for disaster preparedness in the coastal areas. Intensity prediction becomes especially difficult when they undergo rapid intensification (RI). The inability to predict RI comes from the fact that the physical processes that contribute to the cyclonic intensification are not very well understood. Data mining techniques are being used to find the relationship between change in the environmental variable values, and intensification. Though good at identifying precursors to rapid intensification, data mining techniques are computationally expensive. Moreover, very few studies are conducted for the Bay of Bengal region, despite it being a hotbed of cyclonic systems. In our work, we used a computationally cheaper LSH-SNN algorithm to identify precursors to rapid intensification in Bay of Bengal, and showed its effectiveness in identifying precursors.