| Session Details | |
| Section | OS - Ocean Sciences AS - Atmospheric Sciences |
| Session Title | Advances In Data Assimilation And Ensemble Forecast: Applications To Studies And Predictability Of Atmosphere-ocean Variability |
| Main Convener | Dr. Fei Zheng (Chinese Academy of Sciences, China) |
| Co-convener(s) | Dr. Noel Keenlyside (University of Bergen, Norway) Prof. Jin-Yi Yu (University of California, Irvine, United States) Dr. Stéphane Vannitsem (Royal Meteorological Institute of Belgium, Belgium) Prof. Wansuo Duan (Chinese Academy of Sciences, China) |
| Session Description | Uncertainties in predicting ocean-atmospheric variability on a range of temporal and spatial scales (e.g., thunderstorms, torrential rains, tropical cyclones, MJO, ENSO, IOD, Pacific and Atlantic decadal variability, and etc.) arise from the initial conditions and the formulation of the forecast models themselves. Successful applications of advanced data assimilation techniques to weather-climate forecasting systems in the past decade have led to an increased interest in their use for the initialization of numerical models. Ensemble forecast can be much better to estimate the probabilistic distribution of prediction uncertainties caused by not only initial errors but also model errors, even combination of them. Both data assimilation and ensemble forecast still encounter great difficulties. Related ocean and atmosphere motions display very different time scales but they interact with each other and strongly influence weather and climate. Consequently, both ensemble forecast and data assimilation have to deal with the issues on the different scales that ocean and atmosphere presented in initialization. This session invites researchers to share their novel data assimilation and ensemble forecast approaches for improving predictions in weather and climate events, as well as for gaining insight into the causes and mechanisms for atmosphere-ocean variability. We also welcome talks aimed at increasing understanding of the fundamental limits of predictability. Such talks could include analyses of the relevant initial error dynamics and model error physics and methods that attempt to quantify the predictability. |