Hurricane Intensity, Sea Surface Temperature, and Stochastic Variation

Abstract

“Hurricanes” are North Atlantic Ocean (NAO) cyclones that attain a maximum sustained surface-wind speed, vmax, of ≥119 km/h (Bell et al., 2000), which corresponds to SaffirSimpson storm-intensity categories ≥1 (NOAA, 2006). A key climate-change concern involves the question of whether an increasing trend in sea surface temperature (SST)— considered a signal of global warming by being induced or enhanced by increased atmospheric greenhouse-gas concentrations—portends increased tropical cyclone intensities. Historical data, physical theories, and modeling all appear to link increasing SST to increased cyclone activity, intensity, or potential destructiveness in the NAO and other ocean basins (Saunders & Harris, 1997; Emanuel 2000, 2004; Goldenberg et al., 2001; Knutson & Tuleya, 2004; Emanuel, 2005; Trenberth, 2005; Webster et al., 2005; Hoyos et al., 2006; Mann & Emanuel, 2006; Santer et al., 2006; Trenberth & Shea, 2006). Correlations between cyclonic storm intensity, in particular, and increased SST beginning around 1950 to 1970 (Emanuel, 2000, 2004; Goldenberg et al., 2001; Knutson & Tuleya 2004; Emanuel, 2005; Trenberth, 2005; Webster et al., 2005; Hoyos et al., 2006; Mann & Emanuel, 2006; Santer et al., 2006; Trenberth & Shea, 2006) are claimed to support the “SST hypothesis” that local SST directly affects cyclonic intensity, just as predicted by models of storm-related heat-transfer dynamics (Emanuel, 2000, 2004; Knutson & Tuleya, 2004). For example, tropical cyclone frequency, duration, and intensity over the past 35 years showed an increased number and proportion of hurricanes reaching Saffir-Simpson categories 4 or 5 (most notably in the North Pacific, Indian, and Southwest Pacific Oceans; less so in the NAO), but a decreased number of cyclones and cyclone days in all ocean basins except the North Atlantic during the past decade (Webster et al., 2005). This conclusion is consistent with those from studies that focused instead on recent trends toward increased hurricane destructive potential, largely reflecting hurricanes of category 4 or 5 (Emanuel, 2005; Trenberth, 2005). Other studies have questioned a direct, causal link between SST and hurricane intensity, and support an alternative, “extreme-value hypothesis” that recent strong hurricanes like Katrina are simply extreme samples from an essentially stationary intensity distribution that has not increased appreciably over at least the last half century (Landsea et al., 1996; Michaels et al., 2005; Landsea et al., 2006; Bogen et al., 2007). Among factors that are thought to influence hurricane activity—such as multidecadal oscillations in oceanic thermohaline circulation, upper tropospheric high-pressure regions, dips in tropospheric vertical wind