报告题目：Learning model parameters from observations by combining data assimilation and machine learning
报告专家：Prof. Dr. Tijana Janjic
报告时间：2023年10月09日（周一）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：
Prof. Dr. Tijana Janjic is the Heisenberg Progessor of Data Assimilation at the Catholic University of Eichstätt-Ingolstadt (KU) in Germany. She has been the Head of data assimilation branch of Hans Ertel Center for Weather Research at the University of Munich until 2018. She is the associate editor of Journal of Advances in Modeling Earth Systems and the associate editor of Quarterly Journal of the Royal Meteorological Society. She obtained Heisenberg Award for 2019 and The Quarterly Journal Editor’s award for 2020.
报告摘要：Parametrization of microphysics and processes in the surface and boundary layers typically contain several tunable parameters. Parameters can be estimated from observations by the augmented state approach during the data assimilation. However, when parameters are estimated by this approach, stochastic model for the parameters needs to be pre-specified to keep the spread in parameters. Alternatively, it is poissible to use data assimilation for the state estimation while using ML for parameter estimation in order to overcome this problem. We train two types of artificial neural networks as a function of the observations or analysis of the atmospheric state. The test case uses perfect model experiments with the one-dimensional modified shallow-water model. Through perfect model experiments we show that Bayesian neural networks (BNNs) and ensemble of point estimate neural networks (NNs) are able to estimate model parameters and their relevant statistics. The estimation of parameters combined with data assimilation for the state decreases the initial state errors even when assimilating sparse and noisy observations.
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年10月6日
