| Summary: | Three climate state variables the El Niño/Southern Oscillation (ENSO), peak (August-October) sea surface temperatures (SST) over the main development region (MDR) of the tropical North Atlantic, and the North Atlantic Oscillation (NAO) are hypothesized to influence variations in Atlantic annual tropical cyclone (TC) counts on interannual and longer timescales. An unconditional distribution of TC counts is inconsistent with the distribution of a fixed-rate random Poisson process. A Poisson regression model conditioning TC counts on the three state variables accounts for much of the non-random variations in observed TC counts. Upon validation over the historical record, the model is employed to project the impact of future climate changes on TC counts. In contrast to some recent studies, the model suggests that anthropogenic climate change may lead to a substantial increase in annual TC counts. Spatial and frequency patterns of Florida TC landfall and vulnerable populations also are examined over 1988-2007, particularly in the context of different ENSO phases. A maximum in calculated TC wind damage exists in east-central Florida during El Niño-like TC seasons and maxima in damage exist in south Florida and on the western edge of the Florida panhandle during La Niña-like seasons. Overall, risk of higher wind damage and counts of vulnerable population increase across the entire state during La Niña-like (non-El Niño) seasons.
|
|---|