IRI ENSO Forecast
IRI Technical ENSO Update
Published: February 16, 2017
Note: The SST anomalies cited below refer to the OISSTv2 SST data set, and not ERSSTv4. OISSTv2 is often used for real-time analysis and model initialization, while ERSSTv4 is used for retrospective official ENSO diagnosis because it is more homogeneous over time, allowing for more accurate comparisons among ENSO events that are years apart. During ENSO events, OISSTv2 often shows stronger anomalies than ERSSTv4, and during very strong events the two datasets may differ by as much as 0.5 C. Additionally, the ERSSTv4 may tend to be cooler than OISSTv2, because ERSSTv4 is expressed relative to a base period that is updated every 5 years, while the base period of OISSTv2 is based on a slightly older period and does not account as much for the slow warming trend in the tropical Pacific SST.
Recent and Current Conditions
In January 2017, the NINO3.4 SST anomaly, which had been near or slightly cooler than -0.5 C since the middle of 2016 (making for a borderline or weak La Niña SST condition), warmed back to neutral. For January the SST anomaly was -0.32 C, and for Nov-Jan it was -0.43 C. The IRI’s definition of El Niño, like NOAA/Climate Prediction Center’s, requires that the SST anomaly in the Nino3.4 region (5S-5N; 170W-120W) exceed 0.5 C. Similarly, for La Niña, the anomaly must be -0.5 C or less. The climatological probabilities for La Niña, neutral, and El Niño conditions vary seasonally, and are shown in a table at the bottom of this page for each 3-month season. The most recent weekly anomaly in the Nino3.4 region was 0.1, at an ENSO-neutral level. The SST farther east has increased to above-average levels. Most of the pertinent atmospheric variables also returned to neutral patterns, with the exception of the convection anomalies in the central and western tropical Pacific, which continued to suggest a weak La Niña. The lower-level trade winds and upper level westerly winds have been largely near-average, and the Southern Oscillation Index (SOI) has been near-average during January and early February. Subsurface temperature anomalies across the eastern equatorial Pacific have increased to near-average. Overall, given the SST and the atmospheric conditions, the diagnosis of ENSO-neutral is clearly most appropriate.
Expected Conditions
What is the outlook for the ENSO status going forward? The most recent official diagnosis and outlook was issued one week ago in the NOAA/Climate Prediction Center ENSO Diagnostic Discussion, produced jointly by CPC and IRI; it stated that the ENSO conditions have returned to neutral during January, and that ENSO-neutral is the most likely condition through May 2017. The latest set of model ENSO predictions, from mid-February, now available in the IRI/CPC ENSO prediction plume, is discussed below. Those predictions suggest that the SST is most likely to be in the ENSO-neutral range from February-Apr season forward through most of the first half of 2017, but with increased uncertainty from around May onward, when El Niño development becomes a possibility.
As of mid-February, 96% of the dynamical or statistical models predicts neutral ENSO conditions for the initial Feb-Apr 2017 season, while 4% predicts El Niño conditions. At lead times of 3 or more months into the future, statistical and dynamical models that incorporate information about the ocean’s observed subsurface thermal structure generally exhibit higher predictive skill than those that do not. For the May-Jul 2017 season, among models that do use subsurface temperature information, no model predicts La Niña conditions, 61% predicts El Niño conditions, while 39% predicts neutral ENSO. For all model types, the probabilities for La Niña are 6% or less for for all predicted seasons from Feb-Apr through Oct-Dec 2017. The probability for El Niño conditions is near 5% for Feb-Apr and Mar-May, then rises to near 25% for Apr-Jun, and approximately 50% from May-Jul through the final season of Oct-Dec. Chances for neutral ENSO conditions exceeds 90% for Feb-Apr and Mar-May, is near 75% for Apr-Jun, and between approximately 40 to 55% from May-Jul through Oct-Dec.
Note – Only models that produce a new ENSO prediction every month are included in the above statement.
Caution is advised in interpreting the distribution of model predictions as the actual probabilities. At longer leads, the skill of the models degrades, and skill uncertainty must be convolved with the uncertainties from initial conditions and differing model physics, leading to more climatological probabilities in the long-lead ENSO Outlook than might be suggested by the suite of models. Furthermore, the expected skill of one model versus another has not been established using uniform validation procedures, which may cause a difference in the true probability distribution from that taken verbatim from the raw model predictions.
An alternative way to assess the probabilities of the three possible ENSO conditions is more quantitatively precise and less vulnerable to sampling errors than the categorical tallying method used above. This alternative method uses the mean of the predictions of all models on the plume, equally weighted, and constructs a standard error function centered on that mean. The standard error is Gaussian in shape, and has its width determined by an estimate of overall expected model skill for the season of the year and the lead time. Higher skill results in a relatively narrower error distribution, while low skill results in an error distribution with width approaching that of the historical observed distribution. This method shows probabilities for La Niña at less than 10% from Feb-Apr through Jul-Sep 2017, increasing slightly thereafter, reaching nearly 20% by Oct-Dec. Probabilities for ENSO-neutral are near 95% for Feb-Apr 2017, falling steadily to 55% by May-Jul, and down to near 35% by the final Oct-Dec season. Probabilities for El Niño are less than 5% for Feb-Apr, rise to about 25% by Apr-Jun and to approximately 45-50% for Jun-Aug through the final season of Oct-Dec. A plot of the probabilities generated from this most recent IRI/CPC ENSO prediction plume using the multi-model mean and the Gaussian standard error method summarizes the model consensus out to about 10 months into the future. The same cautions mentioned above for the distributional count of model predictions apply to this Gaussian standard error method of inferring probabilities, due to differing model biases and skills. In particular, this approach considers only the mean of the predictions, and not the total range across the models, nor the ensemble range within individual models.
In summary, the probabilities derived from the models on the IRI/CPC plume describe, on average, a very high likelihood for neutral ENSO conditions for Feb-Apr. ENSO-neutral is predicted to remain the most likely of the three possibilities throughout around Jun-Aug, after which El Niño becomes more likely than neutral through the final season of Oct-Dec. Although most likely, the chances for El Niño only reaches near 50% during Jul-Sep through Oct-Dec. Chances for La Niña are below 10% through the first half of 2017, and only increase slightly later in the year, remaining less than 20% throughout. A caution regarding this latest set of model-based ENSO plume predictions, is that factors such as known specific model biases and recent changes that the models may have missed will be taken into account in the next official outlook to be generated and issued in early March by CPC and IRI, which will include some human judgement in combination with the model guidance.
Climatological Probabilities
| Season |
La Niña |
Neutral |
El Niño |
| DJF |
36% |
30% |
34% |
| JFM |
34% |
38% |
28% |
| FMA |
28% |
49% |
23% |
| MAM |
23% |
56% |
21% |
| AMJ |
21% |
58% |
21% |
| MJJ |
21% |
56% |
23% |
| JJA |
23% |
54% |
23% |
| JAS |
25% |
51% |
24% |
| ASO |
26% |
47% |
27% |
| SON |
29% |
39% |
32% |
| OND |
32% |
33% |
35% |
| NDJ |
35% |
29% |
36% |
IRI/CPC Mid-Month Model-Based ENSO Forecast Probabilities
| Season |
La Niña |
Neutral |
El Niño |
| FMA 2017 |
2% |
94% |
4% |
| MAM 2017 |
1% |
85% |
14% |
| AMJ 2017 |
1% |
72% |
27% |
| MJJ 2017 |
4% |
55% |
41% |
| JJA 2017 |
7% |
47% |
46% |
| JAS 2017 |
9% |
43% |
48% |
| ASO 2017 |
11% |
38% |
51% |
| SON 2017 |
16% |
36% |
48% |
| OND 2017 |
19% |
33% |
48% |
