Deconvolution of natural internal variability and trends in the SST signal for the analysis of the increase of hurricane intensity

In the light of the evidence presented in recent publications linking the increase in hurricane intensity to increasing sea surface temperatures, it is vital to design methodologies to evaluate whether the change in intensity is explained by the observed positive trend in sea surface temperature (SST) in the last few decades or by internal natural variability of the climate system. In particular since 1970, all basins, where hurricanes form, show a clear increasing trend in SST. Recent modeling studies provide evidence that this global SST trend cannot be reproduced in climate models without the inclusion of anthropogenic greenhouse gases. In contrast, decadal-scale oscillations tend to be specific to each ocean basin and tend to be anti-correlated from one basin to another.

In this work we present a methodology based on Bayesian statistics and information theory in order to provide clues on the relative role of SST trend and natural variability in the increase of hurricane intensity. In detail, we examine the joint statistical distribution of SST and hurricane data sets in the principal ocean basins in which hurricanes occur, considering the raw SST data as well as isolating the SST trend and the observed natural variability. Information theory estimates, such as joint entropy and mutual information, are also considered in the study. Our analysis focuses on the period of 1970 to present. In addition to the analysis of the relationship between SST and hurricane intensity, other variables that have been identified in the literature as key in the formation and intensification of hurricanes are included in the study. Among these variables are vertical shear, stretching deformation, atmospheric moisture, ocean heat content, and easterly wave activity.