Scientists team up with Yandex Cloud to develop a neuronet to predict El Niño

El Niño occurs when warm water builds up along the equator in the eastern Pacific. The warm ocean surface warms the atmosphere, which allows moisture-rich air to rise and develop into rainstorms.

A team of specialists from the Higher School of Economics and the Yandex School of Data Analysis, teamed up with Yandex Cloud to develop a neural network to predict the El Niño climatic phenomenon. The new algorithm helps to more accurately predict changes in the average temperature of oceanic surface waters, which can cause natural disasters in certain regions of the world.

Currently, the model already predicts El Niño 1.5 years in advance, and in the future, scientists plan to extend the forecast period to 2 years.

The university research team trained neural networks on an array of thousands of temperature maps with synthetic and real data collected from 1800 to the present. Apart from standard machine learning methods for predicting such phenomena, ML specialists test the Autoformer architecture in training. This will allow to process the sequence of temperature maps with high quality. The scientists used the Yandex DataSphere ML development, which has all the necessary tools and dynamically scalable cloud resources for the full cycle of machine learning development.

“Cloud technologies help to conduct more efficient experiments in a scientific environment. In projects such as El Niño research, fast and flexible access to services for testing different machine learning models is important. Each such test with a new architecture helps to predict the phenomenon as early as possible and more accurately,” said Anna Lemyakina, National and Strategic Projects Director at Yandex Cloud.