Seasonal Influences Upon and Long-Term Trends in the Length of the Atlantic Hurricane Season

Utilizing quantile regression applied to the National Hurricane Center's “best track” database for the years 1980 to 2007, Kossin (2008, Geophys. Res. Lett.) identified an approximate two day per year increase in the length of the Atlantic hurricane season to approximately 80% confidence. However, Kossin (2008) considered only a subset of the entire Atlantic basin, encompassing an extended main development region south of 30ºN and east of 75ºW. It is uncertain whether the trend in season length identified by Kossin (2008) holds over the entire Atlantic basin; furthermore, it is uncertain whether this trend, independent of the portion of the basin considered, holds to the present.

Motivated by these uncertainties, we seek to answer two questions:

1) Is the Atlantic hurricane season really growing longer?

2) What are the synoptic- and planetary-scale conditions associated with early-starting and late-ending seasons?

Utilizing quantile regression, we find that there is no trend in Atlantic hurricane season length for the years 1979 to 2013, whether the entire Atlantic basin or the extended main development region subset thereof is considered. Furthermore, we find that there is no trend in Atlantic hurricane season length for the years 1980 to 2007 when the entire Atlantic basin is considered. The season length trend identified by Kossin (2008) is found to result primarily from three abnormally-long seasons – 2003, 2005, and 2007 – at the end of their period of record.

Linear regression of June and November monthly-mean fields of sea surface temperature, 850 hPa relative vorticity, 600 hPa relative humidity, 850-300 hPa vertical wind shear, and 500 hPa geopotential height against the 10th and 90th percentile formation dates is utilized to identify the synoptic- and planetary-scale conditions associated with early-starting and late-ending seasons. To first order, large-scale conditions that favor tropical cyclogenesis in the regions in and along the genesis pathways (e.g., McTaggart-Cowan et al. 2013, Mon. Wea. Rev.) by which cyclogenesis preferentially occurs early and late in the Atlantic hurricane season promote early-starting and late-ending seasons. La Nina conditions generally favor late-ending seasons, but no statistically-significant relationship between El Nino Southern Oscillation phase and the start of the Atlantic hurricane season can be identified.