Hurricane Tracking in the Atlantic and Associated Insurance Losses

The characteristics of tropical cyclones (TCs) like intensity, formation and track vary on many time scales. The distributions of TCs are hypothesized to be modulated by natural phenomena, e.g., the Atlantic Multidecadal Oscillation (AMO), the meridional overturning circulation (MOC), El Niño Southern Oscillation (ENSO) and by global warming. The IPCC 2007 suggested that a reduction is likely due to global warming. Thus a possible impact of such a reduction on SSTs and therefore on hurricanes is of interest. Both wind speed and pressure drop would increase when the sea surface warms, visible in the extraordinary hurricane season 2005. Although controversial, trends in the intensity have been noted potentially linked to changes in the tropical SSTs associated with anthropogenic global warming. However, a basic problem is to separate global warming changes and changes caused by natural climate variability. For analysing hurricane characteristics related to climate variability a normal tracking tool for midlatitudes has been adapted. First, the 6-hourly ERA-40 reanalyses data, i.e., sea level pressure, wind speed and relative vorticity from 1957-2002 with a horizontal resolution of 1.125o x 1.125o are used as an input for the adapted storm tracking tool. Based on this method we identify hurricanes and study Atlantic hurricane characteristics i.e. track and intensity. To evaluate the ERA-40 hurricane tracks the best track data from the National Hurricane Center are used as reference. Then, the identified hurricanes in ERA-40 are grouped with respect to climate modes of natural variability, e.g., the MOC, ENSO and AMO. A similar investigation strategy is applied to time slice experiments of the ECHAM 5 atmospheric model. Two simulations are available: a control run for the period 1960-90 and an A2 scenario for the period 2070-2100. The obtained tracks and hurricane characteristics are then used to create probabilistic hurricane event sets which will enter in the insurance loss model of the Swiss Re to estimate the associated losses and changes in those losses on the US coast, caused by hurricanes.