The climate phenomenon, El Niño, is dreaded by people living in many parts of the globe. El Niño causes the eastward shift of warm waters from the tropical Pacific Ocean and weakening of trade winds causing severe drought, wildfires, and flood in the South American continent. 

Presently, for climate scientists, it is almost impossible to predict El Niño events for more than one year in advance. However, a new study from Chonnam National University in South Korea claims that by using artificial intelligence, El Niño can be predicted 18 months in advance. 

Conventionally, scientists relied on a relatively small set of historical statistics and climate models to forecast El Niño. The research conducted by Yooo-Geun Ham and his team, however, uses a convolutional neural network, which is proficient at recognizing images. They trained them on global images of historical sea surface temperatures and deep ocean temperatures for forecasting El Niño events.

When tested against real data from 1984 to 2017, the program was able to predict El Niño events as early as 18 months. The forecasts had about 74% accuracy, which is a significant improvement over the best current model, which has only 56% accuracy.