Constructed Analogue Prediction of the East Central Tropical Pacific SST for the remainder of 1999 and 2000

contributed by Huug van den Dool

Climate Prediction Center, NOAA, Camp Springs, Maryland

Because natural analogues are highly unlikely to occur in high degree-of-freedom processes, we may benefit from constructing an analogue having greater similarity than the best natural analogue. As described in van den Dool (1994), the construction is a linear combination of past observed anomaly patterns in the predictor fields such that the combination is as close as desired to the initial state (or 'base'). Here, we forecast the future SST anomaly in the Niño 3.4 region (5N-5S, 120-170W) of the tropical Pacific. We use as our predictor (the analogue selection criterion) the first 5 EOFs of the global SST field at four consecutive 3-month periods prior to forecast time. Data extending from 1955 to the present are used for a priori skill evaluation.

For a given base time (previous ones extending back to 1956, or the current real time forecast ending with JJA99), a linear combination is made of the first 5 EOFs of global SST from all years (1956-1997) excluding the base year), so as to match the SST pattern of the base time. This is done using multiple regression, with each year's SST state as a predictor to which a weight is assigned, determined by inverting the 42 X 42 (available years) covariance matrix. These weights are then applied to the subsequently occurring Niño 3.4 SST in the predictand period for these years past, forming the forecast for the base year's predictand period. Note that the predictand is not involved in the construction process. The constructed analogue is the same linear combination for all leads, i.e the weights are persisted, and can be applied to predictands other than Niño 3.4.

Additional detail about the constructed analogue method (van den Dool 1994) shows that constructed analogues usually outperform natural analogues (such as they are) in specification mode (i.e. "forecasting" one meteorological variable from another, contemporaneously). This advantage may also be expected to occur in real forecasting, as long as the (linear) construction does not compromise the physics of the system too much. A constructed analogue yields a single linear operator derived from data by which the system can be propagated forward in time. This is methodologically related to POP and linear inverse modeling. The skill of the constructed analogue method in forecasting SST is discussed in van den Dool and Barnston (1995).

The current constructed analogue forecasts for Niño 3.4 out to 1.5 years lead are shown in Fig. 1, using data through Aug 1999. The expected cross-validated skill is also shown (dashed;right-hand scale). The SST anomaly observed during JJA99 is plotted as the earliest "forecast" value. For the early leads the observed SST for JJA enters into the plotted forecast with a 2/3 and 1/3 weight, respectively, providing continuity with the known initial condition (JJA) thru ASO.

A closer look at the skill of the constructed analogue method was provided by Fig. 2 in the June 1996 issue of this Bulletin (p. 73). Skill averaged over 1956-present has gone up considerably with the addition of the 97/98/99 events added to the verification sample. The skill is competitive with those of other empirical as well as dynamical methods (Barnston et al. 1994). An evaluation over 1996-98 (Barnston et al 1999) shows CA, CCA and CLIPER to be the clear frontrunners among the empirical methods and continuing to be competitive with dynamical methods, such as the NCEP and COLA models. Forecasts for late fall through winter tend to be most skillful, while summer forecasts have lower skill. While skill (dashed line in Fig. 1) generally decreases with lead time, the dependence on the target season is sometimes a stronger factor, causing what seems a 'return of skill' with increasing lead.

Last winter's strong cold La Niña had decreased in magnitude in terms of degree C until about July but has increased since then, in agreement with the seasonal cycle in standard deviation. From last winter onward the SST has been at about -2 standard deviations all the time. CA forecasts Niño 3.4 anomalies to increase again to -1.7 C by late 1999 while maintaining a roughly -2 standard deviations anomaly. Clearly, next winter is forecast to be another cold event, only a tad weaker than last year's cold event. At this point of the annual cycle skill for forecasts for next winter is >0.8 correlation. Beyond JFM CA calls for a rapid return to near zero anomalies which would then persist all the way out to winter 2000/2001. However, skill is down to .25 correlation by summer.

Although the forecast is for negative Niño3.4 it is too simplistic to say that the forecast is for the opposite of a warm event. Inspection of the climate state in terms of SST-EOFs shows that winter 97/98 was extreme in EOF#1, while winter 98/99 appeared to have peaked in EOF#2, 3 and 4. Put another way: Niño3.4 is a compromise index to describe warm events with SST anomalies from the dateline to the S. American coast and cold events that have SST anomalies both east and west of the dateline. Perhaps next winter, a very likely cold event, will be different still, and using a single index is not doing the job.

Table 1 provides information about the role of each of the past years in the construction process for the current forecasts. The inner product (IP) shows the degree of similarity (or, if negative, similarity to opposite) of this year's predictor periods to those of the other years on the global domain. On the other hand, the weights (Wt) shows the contribution of each year's pattern to the constructed analogue. The inner products and the weights, while similar, are not proportional, because co-linearity among years is accounted for. This is because, for example, two past years having the same kind of similarity are unnecessary; only one of them may have been assigned the appropriately high weight, leaving the other with little to contribute. The weights have changed very little from 3 months ago, see June issue, as they should if CA is to be skillful in making forecasts. Indeed, CA has been above average in accuracy on both the onset, maturing and continuation of the current cold event.

The most important positive (+) and negative (-) contributors to the description of the global SST over the last 4 seasons (SON98 to JJA99; denoted as 1999) are, in chronological order, 1956(+), 1957(-), 1961(+),1966(--), 1968(--), 1971(+), 1978(-), 1980(-), 1986(--), 1988(++), 1989(++), 1993(-),1994(-) and 1996(++). An interdecadal variability in the analogue weights time series (e.g. negatives before 1980, positives in 1980s and 1990s) is suggested in the weights and more clearly in the inner products. The current global SST has a positive correlation with the global SST fields in all years since 1980 (except three). Although ENSO clearly dominates the interdecadal variability at this time, the trends are still potent. The years 1966, 1968 (denoting the SON67-JJA68 period) are very heavily negatively weighted. 1988 carries the highest positive weight. The year now labeled 1984 has had high positive weight for almost two years, but is waning. This means that the 1982/83 warm event plus the 1.5 years following have had similarities with the 97/98/99 cycle, but are now diverging. While the ENSO situation definitely enters into the analogue selection (more strongly so at the moment than generally), non-ENSO (remember, global SST EOFs are used - the CA knows nothing specifically about NINO3.4) processes other than ENSO also determine the weights and the resulting forecast.

All anomalies refer to 1961-90 base period.

References

Barnston, A.G., H.M. van den Dool, S.E. Zebiak, T.P. Barnett, M. Ji, D.R. Rodenhuis, M.A. Cane, A. Leetmaa, N.E. Graham, C.F. Ropelewski, V.E. Kousky, E.A. O'Lenic and R.E. Livezey, 1994: Long-lead seasonal forecasts-Where do we stand? Bull. Amer. Meteor. Soc., 75, 2097-2114.

Barnston, A. G., M. H. Glantz and Yuxiang He, 1999: Predictive skill of statistical and dynamical climate models in SST forecasts during the 1997/98 El Niño episode and the 1998 La Niña onset. Bull. Amer. Meteor. Soc., 80, 217-243.

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. Proceed-ings of the 19th Annual Climate Diagnostics Workshop, Nov. 14-18, 1994, College Park, Maryland, 416-419.

Table 1. Inner products (IP; scaled such that sum of absolute values is 100) and weights (Wt; from multiple regression) of each of the years to construct an analogue to the sequence of 4 consecutive 3-month periods defined as the base (currently the string SON98, DJF98/99, MAM99 and JJA99). Years are labeled by the middle month of the last of the four consecutive predictor seasons. 1998 is not yet used as a candidate analogue because long lead forecasts are not possible beyond the latest observations.

Year	IP	Wt	Year	IP	Wt	Year	IP	Wt	Year	IP	Wt
56	0	15	67	-5	-7	78	-6	-18	89	5	29
57	-4	-19	68	-5	-35	79	-7	4	90	4	7
58	-3	2	69	-1	5	80	-3	-16	91	1	7
59	-4	-4	70	1	2	81	0	-6	92	-1	-4
60	-3	10	71	2	16	82	0	2	93	-2	-21
61	0	15	72	-2	2	83	0	13	94	-1	-23
62	1	-4	73	2	3	84	5	6	95	1	2
63	1	1	74	1	6	85	3	2	96	4	24
64	-1	-3	75	0	-2	86	0	-21	97	1	12
65	-4	-7	76	-1	6	87	0	0
66	-7	-24	77	-6	-8	88	3	30

Fig. 1. Time series of constructed analogue forecasts (solid line) for Niño 3.4 SST based on the sequence of four consecutive 3-month periods ending in Aug 1999. The dashed line indicates the expected skill (correlation) based on historical performance for 1956-mid99. The x-axis represents the target period. The left y-axis (solid line) shows the SST forecast; the right y-axis (thin dashed line) shows the skill. The observation is shown instead of the constructed analogue specification for the initial state JJA 1999, and this observation also contributes by decreasing amounts to the JAS and ASO99 plotted values (see text).