More Technical ENSO Comments

Current Conditions

The basic ENSO indices were somewhat warmer than neutral in April 2002, but not to the level associated with a minimal El Niño. Anomalously warm SST conditions that had emerged in the equatorial Pacific during the second week of April, with a band of SSTs of 0.5 degrees C or more above average stretching from the coast of South America to west of the dateline, declined somewhat during later April and weakened further during the first half of May. For purposes of this discussion, neutral SST conditions are defined as the sea surface temperatures (SSTs) in the Niño 3.4 region averaged over the Jul-Aug-Sep period (for which model forecasts are highlighted here) being between -0.6 and 0.6 degrees C, while less than -0.6 would be indicative of La Niña and more than 0.6 indicative of El Niño. This definition allows about half of all years to be classified as neutral.

April's overall near-neutral (somewhat warmer than average) condition is reflected not only in the SSTs in the central and eastern tropical Pacific Ocean (with Niño 3.4 SST anomaly of 0.3 degree C, and Niño 3 anomaly of 0.2 degree C), but also in the low level winds across the equatorial Pacific and the Southern Oscillation Index (SOI), which did not show behaviors characteristic of El Niño. A more detailed look at the SSTs shows that in the vicinity of the date line in the central tropical Pacific they have been above average by approximately 1 degree C during April, having continued from earlier months. The waters of the far eastern equatorial Pacific had warmed to more than 1.5 degrees C above average in February (with locally stronger anomalies immediately off the coasts of Ecuador and northern Peru), and during late April and early May have weakened.

The substantially warmer than average sub-surface sea temperature across the tropical Pacific that has been seen earlier this calendar year has weakened during March, April, and early May as some of it came to the surface. This warm subsurface water had been enhanced by strong westerly winds in the far western Pacific in December, resulting in substantially increased SST in early February in the far eastern tropical Pacific. Significant westerly wind bursts in the western and central equatorial Pacific have not been observed since the above-mentioned major wind event. The lack of these wind triggers over the last several months, combined with the failure of the above normal SST in March and April to initiate larger-scale tropical Pacific atmospheric anomalies, have allowed the SST anomalies in the eastern Pacific to lose some of their strength and areal coverage. With no westerly wind bursts detected by mid-May, movement toward El Niño conditions is expected to continue to stall in the short-term. Conditions could still change over the course of the next few months to enhance the likelihood of an El Niño in 2002, as for example with the occurrence of new westerly wind bursts or a slow movement of the atmospheric state towards El Niño conditions without such wind events. The subsurface sea temperature is still weakly anomalously warm. On the other hand, the seasonal window of opportunity is beginning to close, as the cold tongue in the eastern Pacific is re-establishing itself and the ITCZ is beginning its northward migration with the approach of the northern warm season. Both of these factors make El Niño development more difficult after the middle of the calendar year. Thus, with every week between now and the end of June that El Niño conditions are not approached, the probability of one developing decreases.

Expected Conditions - See summary of model predictions

There is some variation among ENSO model forecasts for the coming 9 months. Compared to one month ago, a slightly larger proportion of the models has begun forecasting at least weak El Niño conditions. At long lead times of 6 to 9 months into the future, statistical and dynamical models that incorporate information about the ocean's observed sub-surface thermal structure generally exhibit higher predictive skill than those that do not. Among models that do use sub-surface temperature information, 60% to 70% of them are calling for warming to at least the minimum El Niño level to take place between July and September of 2002. Of the models not indicating appreciable warming, all indicate a continuation of near-neutral conditions; none indicate cold conditions. (Only models that produce a new ENSO forecast every month are included in the above statement.) Overall, the ensemble of the available model forecasts implies a moderately high likelihood for El Niño conditions of weak to moderate strength by the end of northern summer of 2002. If El Niño conditions do develop by that time, a continuation of such conditions would be expected through at least March of 2003, as El Niño events tend to reach their peak magnitude near the end of the year. Out of the 14 ENSO forecast models examined, 9 of them forecast at least a 0.7 degree C positive anomaly in the Niño 3.4 regions for the second half of 2002 (e.g. Sep-Oct-Nov), while the remaining 5 predict neutral SST levels. If we omit the models that do not explicitly use subsurface sea temperature data, the counts become 7 out of 10 versus 3 out of 10.

ENSO forecast models have a relatively modest level of skill for forecasts made during the February to April period. Now that it is May, skills are somewhat higher, although caution is still advised in interpreting the distribution of model forecasts as the actual probabilities for the ENSO conditions expected for the later parts of this year. In addition to the overall skill, the expected skill of one model versus another has not been established using uniform validation procedures; this may also cause a difference of the above probability estimate from the true probability. Compared to one month ago, there has been a slight decrease in the number of models that are forecasting at least weak El Niño conditions developing by the second half of 2002. This is consistent with the observation that during the mid-April to mid-May period, besides the continued lack of participation of the atmosphere, the SST itself has retreated somewhat from its approach to a near-El Niño state that it had attained briefly in the second week of April. Compared to one month ago, the range of likely scenarios has narrowed somewhat, even though the chances of having at least a minimal El Niño versus not having one are roughly equal to one another. In contrast to last month, when there was a non-trivial chance of having a moderate (or greater) strength El Niño, it now appears that the most likely outcome will be either a weak El Niño or a near-neutral condition during the remainder of 2002. Confidence in the forecasts has increased somewhat compared with what they were one month ago, and will increase further in the June forecasts.

Requirements for El Niño Formation

A transition toward warm ENSO conditions is possible, as the equatorial SSTs between the dateline and the South American coast have moved toward, but not reached, levels that rank in the top 25% of the historical climatological distribution for this time of the year. When top-quartile SST levels are reached during northern spring or early summer, El Niño conditions become attainable, as the Pacific ocean-atmosphere system is most conducive to ENSO phase changes at this time of year. Two major atmospheric behaviors are required in parallel with El Nino-associated SST behavior: the easterly trade winds in the central and eastern tropical Pacific must weaken or even reverse direction, and deep convection that is normally limited to the western and central tropical Pacific must migrate eastward over the warmed SST in the east-central Pacific. Such growth in area and intensity of the warm water, and the associated atmospheric changes, occur if the ocean/atmosphere dynamical feedbacks are mutually enhancing, as follows: The convection associated with regions of positive SST anomalies east of the dateline leads to a pervasive weakening of the easterly trade winds. The convection adds heat to the atmosphere, resulting in a weaker east-west equatorial temperature gradient. The weaker trade winds then allow additional warm water from the western Pacific to move eastward, further weakening the temperature gradient, leading to further reduction in the trade winds, etc. Such feedbacks have not yet occurred this year, and with the SST anomalies weakening relative to the stronger conditions seen in the second week of April, the chances of atmospheric participation have decreased as well. Uncertainty lies in whether the above-described sequence of coupling events will take place between the tropical Pacific ocean and atmosphere to a sufficient extent to allow the SST to re-strengthen, rather than return further towards more climatological values. The variation among the forecasts of the different models is a reflection of this uncertainty. The next one to two months represents the preferred time of year in which the climate system is historically most amenable to the growth of El Niño events through the above-described positive dynamical feedback of ocean and atmosphere. While El Niño episodes can develop after June and even occasionally after July, their probability of development drops sharply if they have not shown strong signs of development by the end of June. Thus, within the next month, the certainty will increase as to whether the events leading to a positive ocean-atmosphere feedback are taking place. While a majority of numerical ENSO models point to El Niño development for this year, they may not yet have had the chance to take into account the recent, shorter-term re-strengthening of the easterly trades and the weakening of the tropical Pacific SST state, particularly in the eastern third of the basin. If this is the case, their June forecasts should weaken accordingly.

The Probability of an El Niño in 2002

IRI's assessment is that there is a 55% probability of an El Niño by mid-2002, lasting into early 2003. This assessment is based on the collective forecasts of many computer models of various types, as well as on the experience of the several oceanographers and atmospheric scientists familiar with the ENSO phenomenon. This is well above the climatological probability of about 20-25% (i.e. one El Niño per 4 to 5 years), but is a marked decrease relative to the IRI's April statement in which the probability was estimated at 70%. It clearly still leaves uncertainty, although the range of possibilities has been narrowed compared to that of one month ago, in that the probability of a moderate (or greater) strength El Niño is now reduced. If an El Niño does occur, its likely strength will be weak or possibly the lower portion of the moderate strength category, with anomalies values of greater than 1.3 C in the Niño 3.4 region unlikely. It would likely form during the late May through June period, and last through March 2003.