(ORDO NEWS) — Scientists have analyzed the nature of storms and strong waves in the Pacific Ocean. To do this, they used artificial intelligence and were able to train it to predict such phenomena. The researchers believe that the development will help to understand many natural phenomena and learn how to predict them.
For a long time, forecasting storms and waves depended on models based on physical equations. The problem is that the calculations turned out to be too complicated and cumbersome, since the specialists had to take into account many natural processes.
A research team from the Institute of Oceanology of the Chinese Academy of Sciences (IOCAS) has studied storms and large waves in the Pacific Ocean using artificial intelligence (AI). They created a deep learning satellite model that predicts changes in surface water temperature (SST) , which is associated with high waves and storms. The results are published in Science Advances.
“Artificial intelligence technology is a promising way to model complex ocean phenomena and circumvent the difficulties faced by traditional numerical models,” said Professor Li Xiaofeng, study leader.
The deep learning model uses satellite data to operate. On them, she studies what preceded the rise in temperature and what usually follows it, after which she makes a forecast. During nine years of testing (2010-2019), the model accurately and efficiently predicted changes in the surface temperature of the Pacific Ocean and subsequent phenomena. The algorithm is fast: it took the authors about a minute to complete the SST field prediction of the entire testing period on a typical desktop computer.
The model is based on actual observation data and makes accurate predictions using information from several physical parameters: for example, only SST satellite data. The developers believe that it can be modified: the model will use different data and is suitable for predicting other natural phenomena.
Machine learning models for predicting ocean phenomena have caught the eye of experts. Simultaneously with them, similar developments appeared, studying atmospheric processes. The study authors believe that in the future, with improved network architecture, complex phenomena, such as tropical cyclones, can be predicted quickly and with high accuracy.
The combination of new technology and old numerical method is also important. “Artificial intelligence, statistical and traditional numerical models can complement each other and provide a new perspective for studying complex ocean features,” said Professor Lee.
The development shows that the use of AI can be a reliable and promising way to model and predict complex ocean phenomena in the era of big data satellite remote sensing. The authors believe that it will facilitate interdisciplinary research in the field of oceanic and atmospheric phenomena.
Contact us: [email protected]