Recent developments in statistical prediction of seasonal Atlantic basin tropical cyclone activity

Abstract

Statistical forecasts of Atlantic basin seasonal hurricane activity have been issued since 1984 by the Tropical Meteorology Project at Colorado State University (CSU) headed by William Gray. Since these initial forecasts were developed, considerable improvements in data and statistical techniques have led to improved amounts of skill in both hindcasting and forecasting of seasonal Atlantic basin hurricane activity by CSU as well as other forecast groups including the National Oceanic and Atmospheric Administration and Tropical Storm Risk. Statistical seasonal forecasts derive their skill from atmospheric and oceanic parameters that span the globe. Recent developments in statistical prediction include the development of shorter-period forecasts for the individual months of August, September and October and the issuing of landfall predictions or probabilities. Individual forecast groups generally utilize large-scale atmospheric wind and circulation patterns to issue their landfall forecasts. In addition, landfall probabilities down to the county level for all coastal counties from Brownsville, Texas to Eastport, Maine have recently been made available online. Several forecasting groups are now issuing seasonal forecasts for the Atlantic basin, and it is to be expected that the skill of these forecasts will continue to improve with additional access to improved data and modelling capabilities. © 2007 The Authors Journal compilation © 2007 Blackwell Munksgaard.