Regional Increases in Landfall Frequency and Intensity of Atlantic Hurricanes in a Stochastic Model Forecast

Abstract

Good assessments of hurricane landfall risk that are highly localized to particular geographic locations are needed for a range of applications, such as for insurance pricing, disaster planning and mitigation. Risk assessments are traditionally performed using the full historical record, even though hurricane activity is well-known to exhibit nonstationarity. The scientific literature addressing periodicity and secular changes in hurricane activity is usually focused on broad metrics, such as basin-wide frequency, which have only limited utility to local risk assessment. Stochastic hurricane models are often used for this purpose but are usually created to mimic the long-term average of historical data. This study will present results from a novel stochastic model capable of making predictions of near future shifts in hurricane activity, with a focus on geographical shifts in landfall frequency and intensity. The existence of multidecadal variability in hurricane activity in the Atlantic, alternating between phases of high and low frequency, is well-known (Delworth and Mann 2000; Goldenberg et al. 2001; Elsner et al. 2004; Jewson and Penzer 2006; Holland and Webster 2007; Klotzbach and Gray 2008). The current high-activity phase began in 1995. Multidecadal phases of hurricane frequency are almost certainly linked to multidecadal phases of increased or decreased sea surface temperature (SST) anomalies in the main development region (MDR). Fluctuations in MDR SST have alternatively been classified as the Atlantic Multidecadal Oscillation (AMO) (Delworth and Mann 2000; Goldenberg et al. 2001; Gray et al. 2004; Knight et al. 2006; Enfield and Cid-Serrano 2009) or the Atlantic Meridional Mode (Kossin and Vimont 2007; Vimont and Kossin 2007). In addition to the multidecadal variability, some researchers have found an increasing trend in frequency (Emanuel 2005; Mann and Emanuel 2006; Emanuel 2008) and intensity (Webster et al. 2005; Elsner 2006; Elsner et al. 2008) linked with anthropogenic climate change. Recent papers have also emphasized the importance of the tropical mean SST, which is dominated by the tropical Indo-Pacific (IP), rather than SST strictly local to the developing storm (Vecchi and Soden 2007; Swanson 2008).