近年来，英国气象局通过改善气候模式的描述能力和机理研究提高了气候模式的预测能力。然而，数值模式对大气对流过程的模拟依旧不尽如人意，其症结在哪儿，如何改进对流参数化方案？英国气象局业务预报模式对西太平洋热带气旋的天气预报和季节预测水平如何，其预报技巧关键来源是什么？为了让大家了解数值预报方面的最新研究进展，特邀特邀英国雷丁大学Hilary Weller副教授、Xiangbo Feng研究员来校访问交流，欢迎大家踊跃参加！
报告一题目：Multi-fluid Modelling of Atmospheric Convection
报告二题目：How well do the UK Met Office operational global models predict West Pacific tropical cyclones: weather and seasonal forecasts?
报告时间：2019年4月16日（周二）上午9：30
报告地点：气象楼423会议室
主 持 人：张文君 教授
欢迎广大师生前来参加！
报告一摘要：
Atmospheric convection is one of the biggest challenges of weather and climate modelling. Convection is the driver of atmospheric circulation but most clouds are smaller than the grid size. As a consequence, all but the highest resolution models use convection parameterisations. These parameterisations have become sophisticated, estimating the mass that is transported upwards and how this will influence the momentum, temperature, moisture and precipitation. However, there are still problems with these parameterisations; they tend to produce unrealistic hot columns of air and the main regions of convection in the tropics are usually misplaced. These are large heat sources for the global atmosphere and so errors in these locations have knock-on effects across the globe.
There are two assumptions made in convection parameterisations that could be to blame for their poor performance. Firstly there is no memory of the properties of convection from one time-step to the next. The convection properties are calculated each time step from scratch as if there had been no convection in the previous time step. Secondly, even convection schemes called "mass flux" do not actually transport air upwards. They transport the heat, moisture and momentum but the distribution of mass in the vertical is not changed by the convection scheme. Removing these assumptions is tricky. It involves solving same equations of motion in clouds and outside clouds. This is the multi-fluid approach. Separate equations for velocity, temperature, moisture and volume fraction are solved for the air in clouds and the environmental air, outside clouds. As the clouds and the environment are interwoven, we assume that they share the same pressure.
I am working on making this approach work as part of the NERC/Met Office Paracon project to develop big changes to the way that convection is parameterised. I will describe some of the progress and difficulties in solving multi-fluid equations for convection.
报告二摘要：
Prediction skills of the UK Met Officeoperational global models forwestern Pacific tropical cyclones (TCs) are vigorously evaluated, thanks to the WCSSP Southeast Asia project- FASCINATE through the Newton Fund. For the weather forecasts (out to seven days ahead), TC prediction skill of the Met Office global numerical weather prediction (NWP) model has been assessed from the past 10 years of operational forecasts. Error statistics for TC tracks and intensities show a steady improvement in Western Pacific, with increasing model resolution and other model upgrades, such that for location there is about two days’ additional skill in the most recent forecast (2017) compared with the earliest (2008-2010). An emphasis is put on the recent development of including ocean coupling in NWP. The merit of air-sea coupling for TC predictions is examined by re-forecasting the 2016 TC season using different model configurations. For the seasonal forecasts (out to four months ahead), predictability of the Met Office global seasonal forecasting system (GloSea5) for TCs are presented. GloSea5 has skillful TC track predictions in most areas of Western Pacific, except in the east of Japan. The simulated ENSO teleconnection is the main source for such skill. Notable biases of TC track density are also found in GloSea5 associated with biases in the large-scale environment.
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
2019年4月8日
