Wed. Nov 20th, 2019

Artificial intelligence able to foresee El Niño better than ever

Artificial intelligence is getting to know how it can forecast El Niño climate changes. It is expected that technology could be utilized to advance climate forecasts and allow policy-makers a greater amount of time to make appropriate preparations.

El Niño is capable of causing many severe natural disasters, and hence long-term damages. As warm waters in the tropical Pacific Ocean move eastward and trade winds fade, the weather design waves through the atmosphere, resulting and wildfires in some regions, and flooding in other regions. This is an example of Strong El Niño events.

Yoo-Geun Ham, a climate scientist at Chonnam National University in Gwangju, South Korea, along with his associates, developed an artificial intelligence capable of foretelling the advancement of an El Niño system, as many as 18 months prior to its occurrence.

The researchers fed the system with data between years 1871 and 1973, of international oceanic temperature. He AI was also given several thousand simulations, between years 1961 and 2005, of El Niño events, gathered by already present foretelling climate models. Data fed also included the ocean heat content, surface sea temperature, etc.

According to a report by the team in Nature, when Yoo-Geun Ham and his associates tested the AI against real data between years 1984 and 2017, they observed how it was more accurate than the best current model. Even though the program wasn’t perfect in how it was just 74% accurate in foretelling El Niño events, it is much better than the current model which is about 56% accurate in its predictions.

Yoo-Geun Ham stated how the AI system was corrected in predicting 24 out of a total of 34 climate occurrences as La Niña, El Niño, or usual 18 months before time. The current model got less than 20 correct.