In canonical correlation analysis (CCA), relation-ships between multicomponent predictors and multi-component predictands are linearly modeled. These typically take the form of pattern­to­pattern relation-ships in space and/or time. CCA is designed to minimize squared error in hindcasting linear combinations of predictand elements from linear combinations of the predictor elements.

CCA has been used in the social sciences for many decades, but only in the last 10 years has it begun being used in the atmospheric sciences. For example, Barnett and Preisendorfer (1987) applied CCA to monthly and seasonal prediction of U.S. temperature. Graham et al. (1987a,b) and Barnston and Ropelewski (1992) applied it to predicting aspects of the ENSO phenomenon, and Barnston (1994) forecasted short-term climate anomalies in the Northern Hemisphere. Recently, Barnston and He (1996) explored CCA as a tool for seasonal temperature and precipitation forecasts for Hawaii and Alaska. The skills resulting from the latter two studies, while generally modest, were high enough for the U.S. National Weather Service to use the forecasts on a real-time, operational basis.

Here, CCA is used to predict 3-month total precipitation anomalies in the Pacific Islands out to a year in advance, as described in He and Barnston (1996). It is well known that rainfall in the tropical and subtropical Pacific is strongly related to ENSO (Ropelewski and Halpert 1987, 1996). Therefore it is worthwhile to set up a seasonal prediction system that produces real-time forecasts on a monthly basis for the benefit of agricultural and commercial interests in the Pacific Islands.

The predictor fields used for the forecasts include quasi­global sea surface temperature (SST), Northern Hemisphere 700 mb geopotential height, and the predictand precipitation itself (33 island stations) at an earlier time. Experiments with different subsets of predictors and predictor field weights showed that the most valuable predictor field is SST, with 700 mb heights and prior precipitation somewhat helpful. The SST predictors are therefore given double their natural weight. This helps prevent overfitting to Alucky@ relationships with other predictors over the relatively short (1955-present) period of record. Further details about the skills, the underlying relationships, and the need to weight the SST double are provided in He and Barnston (1996). The set of predictors is configured as four consecutive 3­month periods prior to the time of the forecast, followed by a variable lead time, and then a single 3-month predictand period. The predictand includes 3-month total rainfall at 33 Pacific Island stations within 25^oN-30^oS, including 4 Hawaiian stations. The lead time is defined as the time between the end of the final (fourth) predictor period (i.e., the time of the forecast) and the beginning of the 3-month predictand period. This strict definition contrasts with that in which the shortest lead forecast would be called 3-month lead instead of zero lead.

The expected skill of the forecasts was estimated using cross-validation, in which each year in turn was held out of the model development sample and used as the forecast target. These skill estimates indicated that at 1 month lead time the highest correlation skill across the Pacific Islands occurs in Jan-Feb-Mar at 0.44 (0.29) averaged over all stations north (south) of the equator, and the lowest occurs from September through December at about 0.15 (0.30) for stations north (south) of the equator. At four months lead skills are only slightly lower except for the Jan-Feb-Mar average skill north of the equator which drops significantly to 0.26.

Figure 1a shows standardized precipitation anomaly forecasts for 33 Pacific Island stations for Feb-Mar-Apr 1997 made using data through May 1996 (8 months lead). The geographical distribution of expected skill for this forecast, based on cross-validation, is shown in part (b) in terms of a temporal correlation between forecasts and observations. These precipitation anomaly forecasts for late winter are weak in amplitude, even at stations where skill is moderately high (e.g. Johnston [r=0.47] or Yap [r=0.51]), where the forecasts could have high amplitude if the predictor values were indicative. This makes sense in view of the current ENSO situation. We have had a mild to medium cold episode over the last year, but are now in a period of weak SST anomalies at the time of year when the likely phase of ENSO for winter 1996-97 is already somewhat determined by what we have now. If a strong reversal to warm conditions were in the offing, the transition to the warm side of the mean should have taken place by now. We could still, however, be in somewhat of a transition period and headed for mild warm conditions by November 1996. The implication of the forecast is that winter 1996-97 will be approximately a neutral year, and thus ENSO impacts on rainfall will be weak.

More detailed forecasts for 9 U.S.-affiliated Pacific Island stations, located as shown in Fig. 2, are provided in Fig. 3. In the latter figure, long-lead rainfall forecasts from 1 to 13 seasons lead are shown (solid bars), along with their expected skills (lines). The horizontal axis reflects the lead time, whose corresponding actual target period for this forecast is indicated in the legend along the top of the figure (e.g. 1=Jul-Aug-Sep 1996). The same ordinate scale is used for both forecasts and skills (standardized anomaly and correlation, respectively). The skill curve applies to the target season for the associated lead time of the present forecast. Sometimes a Areturn of skill@ occurs as the lead is increased because a more forecastable target season has been reached. The forecasts and their skills differ as a result of both location differences within the Pacific basin and differences in orientation with respect to the local orography (if any) and subsequent exposure to the prevailing low-level wind flow.

We note that at most of the U.S. affiliated stations no substantial anomalies in either direction are predicted in the next 6 months, but that the forecast for spring 1997 shows a tendency for subnormal precipitation. The expected skill for the boreal winter and spring tends to be higher than that for summer and fall despite the longer lead time. The CCA modes (not shown; He and Barnston 1996) emphasize ENSO as the leading influence on tropical Pacific climate most strongly during the months of Nov-Dec-Jan-Feb-Mar-Apr-May (and even earlier than Nov along the immediate equator near and somewhat east of the dateline). The current forecast indicates that the predicted rainfall deficits for spring 1997 come as a result of an expectation of warm tropical Pacific ENSO conditions arising in late winter 1996-97 and increasing in spring. This scenario warrants close attention over the next few months. If the predictor SST in the tropical Pacific warms appreciably over this boreal summer, the CCA could begin predicting warm ENSO conditions setting in earlier, such as by winter 1996-97. As discussed above, however, given that the Apersistence period@ is beginning and that warm conditions are not already emerging, it appears unlikely that a strong warm event will occur as early as winter 1996-97.

Barnett, T.P. and R. Preisendorfer, 1987: Origins and levels of monthly and seasonal forecast skill for United States surface air temperatures determined by canonical correlation analysis. Mon. Wea. Rev., 115, 1825­1850.

Barnston, A.G., 1994: Linear statistical short-term climate predictive skill in the Northern Hemisphere. J. Climate, 5, 1514-1564.

Barnston, A.G. and C.F. Ropelewski, 1992: Prediction of ENSO episodes using canonical correlation analysis. J. Climate, 7, 1316­1345.

Barnston, A.G. and Y. He, 1996: Skill of CCA forecasts of 3-month mean surface climate in Hawaii and Alaska. J. Climate, 9, accepted.

Graham, N.E., J. Michaelsen and T. Barnett, 1987a: An investigation of the El Nino­Southern Oscillation cycle with statistical models. 1. Predictor field characteristics. J. Geophys. Res., 92, 14251­14270.

Graham, N.E., J. Machaelsen and T. Barnett, 1987b: An investigation of the El Nino­Southern Oscillation cycle with statistical models. 2. Model results. J. Geophys. Res., 92, 14271­14289.

He, Y. and A.G. Barnston, 1996: Long-lead forecasts of seasonal precipitation in the tropical Pacific islands Using CCA. J. Climate, 9, in press.

Ropelewski, C.F. and M.S. Halpert, 1987: Global and regional scale precipitation patterns associated with the El Nino/Southern Oscillation. Mon. Wea. Rev., 115, 1606-1626.

Ropelewski, C.F., and M.S. Halpert, 1996: Quantifying Southern Oscillation-precipitation relationships. J. Climate, 9, 1043-1059.

Figures