(Text by ZHANG Xudong, zhangxd@qdio.ac.cn)
Xudong was processing a satellite image of internal solitary waves in the Andaman Sea. Credit: Institute of Oceanology, Chinese Academy of Sciences
The ocean is highly dynamic with multi-scale dynamical phenomena. Internal solitary waves (ISW) are generated and propagate in the ocean interior with little signatures on the sea surface. However, ISW has strong randomness, large amplitude (over 200 meters), and lone propagation distance (several hundred kilometers), which is a great threat to underwater navigation, and offshore engineering and significantly impacts ocean environments.
In the past two decades, satellite observations have been widely used to study ISWs in space. Great improvement has been achieved. I have used satellite images to study the spatial and temporal distributions, amplitude inversion, and forecast of ISWs for years. With the rapid development of satellite sensors, we have accumulated huge numbers of satellite observational data in different formats. It is almost impossible to retrieve valuable information with mere manpower. How to efficiently exploit the big data and mine ISW information hidden in the big data is a long-disturbing problem.
The booming development of artificial intelligence (AI) techniques shed light on this problem. For the past three years, we have applied AI algorithms coupling the ISW physical mechanism to realize ISW forecast and high-accuracy amplitude inversion. AI techniques work quite well and show their great potential in oceanographic studies. Traditional ISW forecast mainly relies on numerical simulation which is time costing and energy consuming, you need to run the numerical simulation on a server. With the help of AI techniques, we speed up the ISW forecast by at least two orders of magnitude with a personal computer.
As a fast-growing technology, AI can establish fast and direct mapping relationships in studying complex marine phenomena. It is an efficient tool and lightweight method to understand complex marine phenomena and mine useful information from big ocean data. We are very optimistic about the future application of AI techniques in oceanography studies.
(Editor: ZHANG Yiyi)