El Nino: The Models, the Forecasting, and The Impacts

El Nino has been in the news a lot lately, because most models predict that we’ll have a huge, “El Nino of the Century” this year. Not only is this super cool because it affects a lot of the world, is about climate forecasting, and has made major news, but I have been spending the summer researching the quality of El Nino models.   Therefore, it seemed fitting to write about ENSO, the mathematics behind the forecasts, and the importance of these particular forecasts.

What is ENSO?

ENSO stands for El Nino Southern Oscillation, which refers to a warming or a cooling of sea surface temperatures in the Pacific Ocean. Warmer temperatures cause an El Nino, cooler temperatures cause a La Nina. The exact causes of ENSO are still being researched, although we do know that easterly trade winds weaken, which reduces a current on the coasts of South America, allowing warmer waters to build. These warmer waters are accompanied by a decrease in atmospheric pressure. El Nino events tend to reach their full strength around December.  You can see in this picture where the warmer waters of an El Nino are:

File:El-nino.png

Image from: https://commons.wikimedia.org/wiki/File:El-nino.png

These do not happen every year. El Nino or La Nina events happen around every three to five years, and strength varies depending on many variables, including changes in sea surface temperatures and pressure.

File:Enso jma.png

Image from: https://commons.wikimedia.org/wiki/File:Enso_jma.png

In this graph, red indicates El Nino, blue indicates La Nina – this graph shows how often these vents occur and it shows La Nina occasionally following El Nino.

Why is it important?

ENSO influences precipitation across the world, causing increased rainfall in some areas (yes, including California) and drier areas in others (such as Southern Africa and parts of South Asia). It influences monsoons in South Asia and droughts in Australia. Therefore, forecasting ENSO events is incredibly important.  In the following picture, you can see the effects that El Nino has on the world:

File:El Nino regional impacts.png

Image from: https://commons.wikimedia.org/wiki/File:El_Nino_regional_impacts.png

Some basics about the models.

Before we start talking generally – most of this will only be about the section of ENSO forecasting that I am most familiar with.  In my experience, I have used the Nino 3.4 index – this is a way of characterizing El Nino/La Nina events solely by using sea surface temperature. If sea surface temperature increases over a standard threshold of 0.5 degrees Celsius, then it is considered El Nino conditions. Similarly, if sea surface temperature decreases by 0.5 degrees Celsius, then it is considered La Nina conditions. Anything less than a 0.5 degree change in either direction is considered Neutral conditions.

There are two types of models that are most commonly used in ENSO forecasting: dynamical and statistical models. The dynamical models are models that use physics and our understanding of the climate and atmospheric conditions to simulate what’s going to happen. The use of the physics is very precise and we have a pretty good understanding of how our atmosphere works. Statistical models, on the other hand, involve absolutely no prior understanding of how the climate or the oceans work. Instead, they use statistical analyses based solely on past years of data to predict what’s going to happen. Most of the models in ENSO forecasting are dynamical models.

What do the forecasts look like?

The forecasts look like a cube (or, rectangular prism). If you imagine a three dimensional matrix (or block), along the bottom side, we have time (in months, as we get new predictions every month), along the height side we have lead time. Lead time tells us how far in advance we’re making the prediction. On any given month, when a prediction is made, forecasters will predict the sea surface temperature up to a year in advance. So, if they’re making a prediction now, in August 2015, they will make predictions for August, September, October,… all the way up to August 2016. As you can probably gather, the predictions for August and September 2015 are really good, because we have a much clearer understanding of what’s going on. Predictions with a lead time of 12 months (a year in advance) aren’t as good.

Finally, the last dimension of our cube is ensemble members. When a model predicts things, some part of that prediction involves uncertainty. We don’t know exactly how the atmosphere will behave or exactly what the temperature will be like, so to be safe, we run the model a few times and store that information in different members. Thus, we can have an understanding of the variation in this model. Each model has somewhere on the order of ten members.

That is all the data that’s found when the data from a model is compiled. Then, forecasters use that data to come up with a forecast for the next 12 months.  The creation of the models is difficult, but the issuing of the forecast is also complicated because the forecaster has to sort through the cube of data and separate the noise (random or seasonal climate variations) from the signal (temperature fluctuations due to an impending El Nino or La Nina).

How good are these forecasts?

Well, that depends on your definition of “good”. These are all highly tuned models and, for the most part, they do a good job of forecasting an increase in temperature when an increase happens and vice versa. However, keep in mind that for the last few years we have been predicting an El Nino of a certain magnitude when one failed to appear. Now, that doesn’t necessarily mean that will happen this year (especially because the intensity of the El Nino that is being predicted now is much, much higher than in past years), but these models aren’t perfect.

Something to keep in mind while we look at the forecasts for this year: forecasts that occur in the spring and early summer are not as good as forecasts that occur in the rest of the year. This is known as the spring predictability barrier, and its causes are still being debated. Once we get past that time, forecasts get much more accurate.

So, what’s going on with THIS El Nino?

I mentioned earlier that a standard threshold for an event being considered an “El Nino” is a sea surface temperature increase of 0.5 degrees. Here is the forecast for this year:

Image from: http://iri.columbia.edu/our-expertise/climate/forecasts/enso/current/

As you can see, we’ve gone above and beyond that 0.5 degree threshold – some models are even going above the 3.0 degree mark, which does not happen very often. So, for these models, pretty much all are in agreement that we’re going to have an El Nino. They disagree to an extent about the intensity of this El Nino, but most are predicting one far above the average.

What are the effects if this El Nino really is that big? Yes, probably California gets the rain that it has desperately needed. We can also predict much more rainfall in other parts of the United States as well. However, according to NOAA, these impacts are not guaranteed – there is a higher probability of increased rain, but it’s not a sure thing. Since ENSO is a Pacific phenomenon, it does not affect the United States as much as tropical countries and those in South America and parts of Asia and Australia. However, either way, this year will be fascinating one for weather and ENSO – we will keep an eye out on the forecasts as we get closer to December!

If you’re interested in learning more, NOAA has a pretty cool ENSO blog that has explanations, discussions, and frequent updates about the forecasts: https://www.climate.gov/news-features/department/8443/all

 

Works Cited

https://www.ncdc.noaa.gov/teleconnections/enso/enso-tech.php

http://www.elnino.noaa.gov/lanina_new_faq.html

http://www.bom.gov.au/climate/enso/history/ln-2010-12/ENSO-when.shtml

http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/impacts/warm_impacts.shtml

https://landshape.wordpress.com/2012/06/12/dynamical-vs-statistical-models-battle-over-enso/

http://web.lasg.ac.cn/UpLoadFiles/File/papers/2010/2010AOSL-wc-dws.pdf

https://www.climate.gov/news-features/blogs/enso/august-2015-el-ni%C3%B1o-update-supercalifragilisticexpealidocious

http://iri.columbia.edu/our-expertise/climate/forecasts/enso/current/

http://www.cpc.ncep.noaa.gov/products/analysis_monitoring/lanina/enso_evolution-status-fcsts-web.pdf