Technology 3 min read

Researchers use Neural Network to Predict El Niño in Advance

thechatat /

thechatat /

Researchers have figured out how to predict El Niño events 18 months in advance using artificial intelligence.

El Niño, which means The Little Boy, or Christ Child in Spanish, was first recognized by fishermen off the coast of South America in the 1600s. The term refers to a climatic cycle in the Pacific Ocean which has a global impact on weather patterns.

It begins with warm water in the western tropical Pacific Ocean that shifts eastward, toward the coast of South America. Instead of pooling near Indonesia and the Philippines, the Pacific’s warmest surface waters collect offshore of northwestern South America during El Niño season.

As you can imagine, this leads to a series of event, including wildfires in South America and drought in southern Africa, India, and Asia. It also causes flooding in North America’s Pacific coast.

Unfortunately, predicting the dreaded event can be challenging.

Scientists have to rely on a relatively small set of historical data, such as ocean temperature to predict El Niño. Although a few forecast use climate models, they didn’t provide a detailed picture of the ocean required for a long-range forecast.

As a result, climate scientists have never been able to forecast El Niño events one year in advance. But that’s about to change.

Thanks to a type of AI called convolutional neural network (CNN), researchers can now make a long-range forecast – about 18 months in advance. That means threatened regions can now make adequate preparations for drought or flood.

So, how does the model work?

Using Convolutional Neural Network to Predict El Niño

Convolutional neural networks are proficient at recognizing images.

For example, you can train the neural network to recognize a catfish in photos by identifying specific characteristics. These include barbels around the mouth, scaleless skins, rayless posterior fins among others.

Similarly, researchers trained the neural network using images of historical sea surface temperatures and deep ocean temperatures. These include data from El Niño events spanning from 1871 to 1973, including several thousands of simulations of the same data by the climate models.

They taught the AI to learn how these data correspond to the future emergence of El Niño events. And when compared with real data from 1984 to 2017, it was able to predict the event as far as out as 1.5 years.

With that said, the program was far from perfect. For one, it could only predict El Niño events 18 months into the future with 74 percent accuracy.

But, since the current model could only offer 56 percent accuracy for the same time frame, the CNN model is still a step ahead. It was also able to pinpoint the part of the pacific that would experience the most heat.

In a paper published in the journal Nature, lead author of the study, Yoo-Geun Ham wrote:

“The CNN model is also better at predicting the detailed zonal distribution of sea surface temperatures, overcoming a weakness of dynamical forecast models. A heat map analysis indicates that the CNN model predicts ENSO events using physically reasonable precursors.”

The researchers have already issued forecasts that extend into 2021. Now, Ham and his team are tweaking the model to extend the forecast even further.

Read More: New Study Reveals Record Of El Niño Events In The Past 400 Years

First AI Web Content Optimization Platform Just for Writers

Found this article interesting?

Let Sumbo Bello know how much you appreciate this article by clicking the heart icon and by sharing this article on social media.

Profile Image

Sumbo Bello

Sumbo Bello is a creative writer who enjoys creating data-driven content for news sites. In his spare time, he plays basketball and listens to Coldplay.

Comments (0)
Most Recent most recent
share Scroll to top

Link Copied Successfully

Sign in

Sign in to access your personalized homepage, follow authors and topics you love, and clap for stories that matter to you.

Sign in with Google Sign in with Facebook

By using our site you agree to our privacy policy.