PREDICTIVE SKILLS OF SEASONAL TO ANNUAL RAINFALL VARIATIONS IN THE USAFFILIATED PACIFIC ISLANDS - CANONICAL CORRELATION-ANALYSIS AND MULTIVARIATE PRINCIPAL COMPONENT REGRESSION APPROACHES

Drought and flooding are recurrent and serious problems in the U.S. Af filiated Pacific Islands (USAPI). Given the agricultural and water-dep endent characteristics of the USAPI economies, accurate forecasts of s easonal to interseasonal rainfall variations have the potential to pro vide important information for decision makers involved in resource ma nagement issues and response strategies related to drought and flood e vents. Climatology of rainfall and outgoing longwave radiation (OLR) c ycle in the USAPI and the response of OLR to the El Nino-Southern Osci llation (ENSO) are addressed. Boxplot and harmonic analyses indicate t hat the annual cycles in rainfall and OLR are generally strong in USAP I except those stations close to the equator Northern USAPI have posit ive (negative) OLR anomalies during El Nino (La Nina) winters. Two sta tistical models, canonical correlation analysis (CCA) and a relatively new method called multivariate Principal Component Regression (PCR), are employed to forecast rainfall variations in 10 USAPI stations. Sea surface temperatures (SSTs) in the Pacific Ocean are used as predicto rs for both models. The results of this study indicate that both model s are potentially useful in predicting seasonal rainfall variations in the USAPI region, especially in winter (DJF) and spring (MAM). CCA cr oss validation shows that at one and two seasons lead JFM is the most accurately forecast period in the northern USAPI stations, with averag e skills of 0.53 and 0.41, respectively. However, the authors' analysi s indicates a problem of lower predictive skill in summer (JJA) and fa ll (SON). One reason might be associated with the so-called spring bar rier in predictive skill in the tropical ocean-atmosphere system. Anot her reason might be associated with the tropical cyclone activity duri ng these seasons. Predictions using the PCR model yield similar predic tive skill. Though simpler than He and Barnston's model in term of the number of predictor variables used, the authors' CCA and PCR provide comparable skills.