On the use of Synoptic Composites for Distinguishing Rapid and non-Rapid Intensification in Tropical Cyclones

Forecasting rapid intensification (RI) of tropical cyclones is still a challenge due to the limited understanding of the meteorological variables that are necessary for predicting RI. This research aims to identify large-scale synoptic controls that are relevant for RI in the North Atlantic basin. To identify these controls, all North Atlantic basin tropical cyclones from 1979-2009 were split into two groups; those that underwent RI (24 mb deepening in 24 hours) and those that did not. Base-state meteorological variables from the global NASA MERRA dataset were obtained and centered on the National Hurricane Center track point, 24-hours prior to the lowest central pressure in each cyclone. Each group was summarized using a T-mode rotated principal component analysis (RPCA) that revealed the most dominant synoptic patterns within each group. This allowed for finding events similar in synoptic structure, and allowed for a better representation of the atmospheric profile. Events with similar synoptic structure were revealed by the RPC loadings, so a k-means cluster analysis was used to group RPC loadings into clusters of types of RI and non-RI cyclones. The original MERRA fields for events in these clusters are averaged, revealing synoptic composites of RI and non-RI storms. To assess the distinctions portrayed by the composites of each group, permutation tests on the original MERRA fields will be completed; revealing which variables are most distinct between the two types of intensification. It is expected that potential temperature, divergence, relative humidity, and vorticity will be statistically significant variables associated with a RI tropical cyclone.