Predicting Summertime Caribbean Pressure in Early April

Abstract

A method to predict the June?September (JJAS) Caribbean sea level pressure anomalies (SLPAs) using data available the previous April is described. The method involves the creation of a multiple linear regression equation that uses three predictors. These predictors are the January?March (JFM) North Atlantic (50°?60°N, 10°?50°W) sea surface temperature anomalies (SSTAs), JFM Niño 3.4 (5°N?5°S, 120°?170°W) SSTAs, and the strength of the eastern Atlantic subtropical pressure ridge measured in March between 20° and 30°W. The physical role of each of the predictors in determining the variations of Caribbean SLPAs is discussed. The forecast equation is developed using a training dataset covering the period 1950?95 (46 yr) and is tested upon independent data covering the period 1903?49 where the data availability permits (42 yr). Results suggest that skillful forecasts are possible. The method reduced the interannual variance of the JJAS Caribbean SLPAs by 50% in the developmental dataset and by 40% in the independent dataset. Separate forecasts for the June?July and August?September SLPAs are also developed, tested, and discussed.