[A-CG43] Earth & Environmental Sciences and Artificial Intelligence/Machine Learning

[ACG43-03] Recent trends in earth and environmental informatics

★Invited Papers

*Daisuke Matsuoka^1,2 (1.Japan Agency for Marine-Earth Science and Technology, 2.Japan Science and Technology Agency)

Keywords:Deep learning, Machine learning, AI, Big data, data-driven science

In recent years, the application of machine learning techniques, such as deep learning, has been actively promoted in the earth and environmental sciences. For example, the number of relevant papers published by the American Geophysical Union (search using machine learning as keywords) has increased rapidly as shown in the figure. Comparing the number of papers published in 2015 and 2020, there is a total increase of about seven times. Especially in the atmospheric and climate sciences, many studies have been published on pattern detection, parameterization, downscaling, bias correction, and future prediction. Some studies utilizing adversarial generative networks (GANs), which artificially generate data, have also been reported. In addition, benchmark data such as WeatherBench (Rasp et al., 2020; Rasp and Thuerey, 2021) is also available to the public with the aim of accelerating data-driven weather and climate prediction research. From these kinds of background, some research organizations such as NOAA and ECMWF have published roadmaps on the use of AI and data-driven science to provide a direction for future research and development. In this presentation, we will introduce some notable studies published in recent years, and present our personal opinions on the research trends and future directions.