The Impact of SST Bias Correction on North Atlantic Hurricane Retrospective Forecasts

Abstract

Abstract The impact of sea surface temperature (SST) bias correction on seasonal North Atlantic hurricane retrospective forecasts is examined with the Florida State University/Center for Ocean–Atmospheric Prediction Studies (FSU/COAPS) atmospheric global spectral model. The retrospective forecasts cover a 28-yr period (1982–2009). For each year of the retrospective forecast, two, four-member ensembles are developed using predicted SSTs from the National Oceanic and Atmospheric Administration’s Climate Forecast System, version 1 (CFSv1), model. The first ensemble uses the SSTs forecasted from the CFSv1 model and the second ensemble uses the same SSTs with a simple bias correction applied. Seasonal hurricane counts determined using both SSTs are shown to have skill, although the skill is much greater using the bias-corrected SSTs. In addition, a positive significant trend in the hurricane counts is noted using the bias-corrected SSTs, in good agreement with observations while no significant trend is noted using the non-bias-corrected SST. The reason for the enhanced skill is related in part to the magnitude of the mean SST allowing stronger convective activity in the equatorial Pacific to more effectively alter the vertical wind shear in the tropical North Atlantic through changes in the Walker cell.