报告题目：Using machine learning and data assimilation to allow all-sky all-surface assimilation of microwave radiances in weather forecasting

报告专家：Dr. Alan Geer

报告时间：2023年12月13日（周三）16:00

会议形式：Webex-线上会议（会议ID: 27441291796）

会议链接：https://yuefeizeng.my.webex.com/meet/pr003558 (电脑浏览器直接进入，无需密码)

主 持 人：曾跃飞 教授

专家简介:

Dr. Alan Geer is the Principal Scientist at ECMWF. He has strong interests in radiative transfer modelling and has helped improve the representation of cloud overlap, cloud microphysical characteristics, and now surface radiative transfer. His current aim is to extend the use of radiances at ECMWF to full all-sky and all-surface conditions. The recent explosion of machine learning has given him a chance to explore how to combine machine learning with data assimilation, especially focused on using observations to help fill the gaps in, and improve, existing physical modelling frameworks. His publications have been cited more than 40,000 times.

报告摘要：

The use of satellite radiances to derive initial conditions for weather forecasting has progressed from using the data in clear sky to now all-sky assimilation. The final big challenge is the earth surface. If radiance observations could be used more completely, in 'all-sky all-surface' conditions, it would provide not just new information on surface variables but also help provide new information on the lower atmosphere above those difficult surfaces. One way forward is to combined machine learning and data assimilation to learn the poorly known surface characteristics and the poorly known models at the same time. A first example is the simultaneous learning of the sea ice concentration, the sea ice state and a model for the sea ice surface emissivity. The model learnt in this work will be used in the next upgrade to the ECMWF forecasting system to allow assimilation of radiances over sea ice surfaces.

欢迎广大师生踊跃参加！

气象灾害教育部重点实验室

气象灾害预报预警与评估协同创新中心

大气科学学院

2023.12.11