In this Bulletin we find a fairly large number of forecasts for the east-central tropical Pacific SST for the coming year. Our objective here is to synthesize information from some of the predictive sources into a single objective estimate of the likely evolution of the SST's in the tropical Pacific.

One approach to the problem is to combine, or consolidate, the forecasts of several models into a single forecast. This could be done on the basis of the past behavior of each contributing model, as well as the overlap of information among the models. There are several methods by which this can be done. A common method, and the one used here, is linear multiple regression. In effect, a statistical scheme is used to combine outputs of entire models whose natures themselves may be statistical, dynamical, or a mixture of the two. In this case we use four input models. Two are dynamical: the Lamont-Doherty Earth Observa-tory's simple coupled model (the improved LDEO2; Chen et al. 1995; Cane and Zebiak 1986), and the recently implemented CMP12 NCEP coupled model (Ji et al. 1994). The other two models are statistical: the NCEP constructed analogue (CA) model (Van den Dool 1994; Van den Dool and Barnston 1995), and the NCEP canonical correlation analysis (CCA) model (Barnston 1994). The individual forecasts of each model are shown elsewhere in this Bulletin issue.

To derive the multiple regression equations for each target season for each lead time, histories of the forecasts of each model were obtained. The CCA and CA models have histories covering 1956-1995. The Lamont coupled model has a 1972-95 history, and the NCEP coupled model 1981-95. To circumvent the problem of the differing units and climatologies used, all forecasts were converted to actual C forecasts. The observations were expressed likewise. The regressions are based on forecasts for the Niño 3.4 region (5N-5S, 120-170W), except for the Lamont model, from which we receive forecasts for the Niño 3 region. The Niño 3 forecast histories from the Lamont model were used as a predictor for Niño 3.4 in the equation development. The regression coefficients compensate for the slight differences between Niño 3 and Niño 3.4 to obtain the least squares fit for Niño 3.4. We expect to begin receiving gridded forecast fields from Lamont shortly, and will then be able to use Lamont's Niño 3.4 forecasts directly.

The desired lead times of the consolidated forecasts range from 0.5 months to 12.5 months by 1 month increments, where lead time is defined as the time skipped between the time of the forecast and the beginning of the forecasted (target) period. For example, the forecasts shown here, which are issued in the middle of December 1996, have target periods including Jan-Feb-Mar 1997, Feb-Mar-Apr 1997,..., Jan-Feb-Mar 1998. Three of the four individual models have forecast histories whose leads range to 12.5 months or greater, while one (the NCEP coupled model) has a maximum lead of only 8.5 months. Consolidated forecasts for lead times higher than 8.5 months, therefore, are based only on the other three models. A temporal smoother has been applied to the forecasts to avoid any discontinuity in the forecast time series, however, a noticeable increase in the error estimates occurs when the NCEP model drops from the consolidation.

Because the NCEP coupled model forecast only has a 1981-95 history, the training period for the regression is limited to that period and thus results in greater uncertainty in the coefficients than would be the case if a longer history could be used. When that model is not included in the consolidation process for the longer lead times, the 1972-95 period is used to derive the regression equations, making for a more favorable training sample. Data from three lead times were pooled together to help equation stability and help smooth forecasts from projection to projection. Predictor and predictand data from the season preceding and following the target season were combined to form the regression equation. The first (last) target season shares the equation with the adjacent season.

The consolidated forecast for Niño 3.4 made in mid-December 1996 is shown in Fig. 1. Forecasts are expressed as standardized anomalies relative to the 1961-1990 climatology. The box and whisker intervals for the forecasts indicate the one and two error standard deviations, based on estimated skill following shrinkage of the dependent sample skill results in accordance with the sample size and number of predictors. The compo-nent forecasts are displayed on the same chart for comparison. Note that the Lamont model (CZ) anom-alies are for Niño 3. The observed SST anomaly for the most recent 3-mo period is also shown. The forecast has been adjusted on the last three projections to prevent statistical extrapolation outside the bounds of either the highest or lowest component model forecast at any particular time. In this case, the consolidation forecasts were adjusted upwards slightly to be equal to the CZ model; which had the lowest SSTs of the three models at those projections.

The consolidation forecast indicates that the SST's will remain cool through February, 1997, followed by warming to near normal conditions by about April. Temperatures will remain about normal until late 1997 when a trend toward warmer than normal temperatures is expected.

The error bars abruptly increase in width in the spring of 1997, reflecting the so-called spring barrier noted in SST predictions in the ENSO region. It is worth noting that these forecasts have not been inflated. It is a customary practice in many prediction methods to adjust the forecast variance to be about equal to that of the observations. Inflation would push these marginally skillful forecasts away from the mean, however, here, with the anomalies so close to zero, it is unlikely that this would alter the basic prediction of near normal conditions throughout most of 1997.

Acknowledgments: We are grateful to Stephen Zebiak and Mark Cane of Lamont Doherty Earth Observatory, and Ming Ji and Ants Leetmaa from the National Centers for Environmental Prediction, for providing the forecast histories from their respective dynamical models, as well as their current real-time forecasts.

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

Cane, M., S.E. Zebiak and S.C. Dolan, 1986: Experimental forecasts of El Niño. Nature, 321, 827-832.

Chen, D., S.E. Zebiak, A.J. Busalacchi and M.A. Cane, 1995: An improved procedure for El Niño forecasting: Implications for predictability. Science, 269, 1699-1702.

Ji, M., A. Kumar and A. Leetmaa, 1994: An experimental coupled forecast system at the National Meteorological Center: Some early results. Tellus, 46A, 398-418.

van den Dool, H.M., 1994: Searching for analogues, how long must we wait? Tellus, 46A, 314-324.

van den Dool, H.M. and A.G. Barnston, 1995: Forecasts of global sea surface temperature out to a year using the constructed analogue method. Proceedings of the 19th Annual Climate Diagnostics Workshop, November 14-18, 1994, College Park, Maryland, 416-419.

[Previous Article] [Next Article]