Following are excerpts from an article by Anthony Barnston, who is Chief Forecaster for Climate and ENSO Forecasting at the International Research Institute for Climate and Society (IRI). Prior to joining IRI, Mr. Barnston was a NOAA (National Oceanic and Atmospheric Administration) forecaster for 17 years. He has authored atlases, reports and journal papers on weather and climate, many of which were about statistical diagnosis of large-scale circulation patterns and on empirical climate prediction. This item contains his views of the current El Nino situation and is not an official NOAA commentary.--Bryce
As the strong El Nino begins weakening later this winter and spring 2016, some clever folks may wonder whether La Nina conditions might develop in the second half of the year.
...sometimes La Nina does occur the year after a significant El Nino, like after the El Nino events of 1997-98, 1972-73 and 2009-10. But it doesn't always happen, such as after the events of 1991-92 and 2002-03.
Can we estimate how likely a switchover is from an El Nino to a La Nina for the following year? And does it depend on how strong the El Nino is? I'll try to answer this question using relationships based on past data from 1950 to present.
I start by rating each year's ENSO state--that is, if it's El Nino, La Nina, or neither, and how strong. Then I classify each year's average anomalies of at least 0.5 degree but less than 1.0 degree Celsius as being a weak El Nino, at least 1.0 deg C but less than 1.5 deg C a moderate El Nino, and 1.5 deg C or greater a strong El Nino.
Given that El Nino only occurs every few years, we have a fairly short historical record from which to draw inferences, which is why the analysis...needs to be put in context alongside other forecast tools, including computer models and the official ENSO (El Nino Southern Oscillation) forecasts.
(BA--following is a discussion on the outcome of seven moderate and three strong El Nino events since 1950.)
P[L1] D[0x0] M[300x250] OOP[F] ADUNIT T
In six out of the ten cases, the sea surface temperature was -0.5 deg C or cooler than average, satisfying a definition for La Nina. For the three strongest events, all resulted in La Nina based on this method of classification. But how convincing is this? Would we want to put money on a La Nina prediction?
When looking at the results for weak, moderate, and strong El Nino separately, I find an average sea surface temperature anomaly of -0.15 deg C the year after the 11 weak El Ninos, -0.40 deg C after the 7 moderate El Ninos, and -1.17 deg C for the 3 strong El Ninos.
These averages do suggest that stronger El Nino events have a higher likelihood for a La Nina the following year. But not so fast! It's always possible that the pattern I found is simply a coincidence. To test the likelihood that these mean differences reflect true underlying differences, I threw all 21 cases together (11, 7, and 3 cases for weak, moderate and strong El Nino, respectively), and computed the correlation between the strength of El Nino during the first year and the sea surface temperature state the following year.
The resulting correlation is -0.31, which translates to a very weak tendency for the sea surface temperature the following year to be colder when the El Nino the first year is stronger. We can use this relationship to make a rough statistical prediction of the ENSO state next year.
It's a safe bet that the sea surface temperature departure this year will be greater than 1.5 deg C even though we don't yet have 3-month data through January-March 2016. According to this analysis, the best guess for the 2016-17 ENSO state would be within the weak La Nina category.
But the fly in the ointment in this "forecast" is...large uncertainty... the probability of getting La Nina for 2016-17 is 66 percent, leaving a 34 percent chance for falling short of the La Nina threshold.
The large uncertainty of this method is why forecasters don't just look at the past to predict the future, but also take into account other prediction tools, including state-of-the-art computer models that consider a more comprehensive set of features relevant to ENSO prediction.
Aside from doing this number crunching on the historical observations, there are accepted physical reasons for expecting a tendency toward La Nina the year after a significant El Nino. One of these is the delayed oscillator theory, introduced in 1988 by Suarez and Schopf.
The theory says that the low-level westerly wind anomalies, a hallmark of El Nino, not only trigger eastward-moving oceanic Kelvin waves at the equator, but also westward-moving waves just north and south of the equator (called Rossby waves). While Kelvin waves are pushing warm water east, these Rossby waves move cooler subsurface water toward the west. They then bounce off the western side of the tropical Pacific (around Indonesia) and have a return trip, traveling eastward near the equator.
On their eastward trip, these waves also promote cooler water, and can neutralize or reverse El Nino around 6 months after the westerly wind bursts. This cool pulse interrupts the positive feedback mechanism responsible for the growth of an El Nino, ending El Nino and promoting La Nina development.
Since stronger El Nino events often involve stronger westerly wind anomalies, these events tend to trigger stronger Rossby waves and stronger tendencies for El Nino to decay and possibly reverse after peaking at the end of a calendar year.
Based on the statistics derived from the historical data and on the more physical basis as described by delayed oscillator theory, the CPC/IRI team is expecting some cooling coming up in 2016-17.
The NOAA forecast does hint toward La Nina, but we will need to wait another couple of months to see what it says about seasons closer to the usual peak in late autumn 2016.
The full article with graphics and footnotes is here: https://goo.gl/…
© Copyright 2016 DTN/The Progressive Farmer. All rights reserved.