Reply to comment by Joseph J. Barsugli on “Global warming and United States landfalling hurricanes”

Abstract

[1] In our recent paper [Wang and Lee, 2008] (henceforth WL08), we perform an empirical orthogonal function (EOF) analysis on the global annual mean SST over the past 153 years (from 1854 to 2006). The first three EOF modes, which account for 28.3%, 15.3%, and 5.3% of the total variance in the SST data, may represent global warming, ENSO-like (including the Pacific decadal oscillation), and the Atlantic multidecadal oscillation (AMO), respectively. We then use the first EOF mode to study the relationship among global warming, vertical wind shear in the Atlantic hurricane main development region (MDR) and U.S. landfalling hurricanes. The second EOF mode (ENSO-like) and the third EOF mode (the AMO) are presented in another paper (see Figure 1 of Wang et al. [2008a] for these two modes). [2] In his comment, Barsugli [2009] (henceforth B09) points out that the SST spatial pattern of the first EOF mode is similar to that of an El Nino event and that its time series contains variations on interannual timescale. He thus concerns that the first EOF mode of global warming may include the interannual ENSO signal and therefore the observed relationship between global warming and U.S. landfalling hurricanes (also vertical wind shear in the hurricane MDR) of WL08 may be partly attributed to interannual ENSO’s effect. We appreciate and understand his concern since there is no ‘‘perfect’’ method for separating climate modes on various timescales from observational data such as global warming and ENSO modes. We agree that the first EOF mode contains some interannual variations possibly linked to ENSO and that it is difficult to distinguish anthropogenic climate change from natural lowfrequency variability in observational data. However, here we show that the conclusion of WL08 is still true even after removing interannual signals from the first EOF mode. We also present some additional evidences that support the conclusion of WL08. [3] B09 suggests that interannual variations are first removed by applying an 11-year running mean to the SST data prior to performing the EOF analysis. We agree that this approach may be a reasonable way to remove the interannual ENSO signal although it still keeps lowerfrequency variations of ENSO. We re-perform our EOF analysis following B09’s suggestion. The resulting spatial pattern and time series of the first EOF mode are shown in Figures 1a and 1b (Figure 1a is almost identical to Figure 1b of B09), which accounts for 57.4% of the total variance. In comparison with Figure 1 of WL08, the amplitude of SST in