Physically–based hurricane risk analysis

Tropical cyclones (TCs) cause tremendous damage worldwide due to the associated strong winds, heavy rainfall, and storm surge. As the climate warms, these TC hazards may intensify (e.g., Emanuel et al., 2008; Knutson et al., 2010; Lin et al., 2012), making their societal impacts an increasing concern (Mendelsohn et al., 2012). Events have revealed the vulnerability of the USA to severe TCs. In 2011, Hurricane Irene produced more than US$10 billion in damage in the northeastern USA. In 2012, Hurricane Sandy struck the Northeastern Seaboard, killing over 200 people, causing more than US$65 billion in damage, and leaving millions without electric service. Hurricane Katrina of 2005 resulted in more than 1800 fatalities and caused more than US$80 billion in losses (the costliest natural disaster in the US history). To prevent such TC disasters in the future, major advances in TC risk management are urgently needed. Due to uncertainties in future climate and demographic conditions, including TC activity, sea-level rise (SLR), and exposure and vulnerability, effective TC risk management should be based on probabilistic risk assessment. Currently, increases in exposure appear to dominate over the effects of anthropogenic climate change as the leading cause of change in TC damage (Pielke, 2007, Handmer et al., 2012). However, potential changes in storm characteristics and hazards due to climate forcing and related SLR also affect risk (Mendelsohn et al., 2012). Accordingly, changes in the physical state of the atmosphere and ocean also must be taken into account. Thus reliable TC risk assessment, unlike traditional assessments, cannot rely solely on direct statistical analyses of the (quite limited) historical TC records and damage data. Rather, it requires a new physically-based approach that incorporates information on current and projected future climates as well as exposure and vulnerability.