Abstract ID: 072
Summary of S2S Ensemble Sub-project
Lead Author: Yuhei Takaya
MRI, JMA, Japan
Keywords: Sub-project talk: Ensemble sub-project, ensemble prediction
Abstract: Ensemble techniques are a fundamental theme of the research and development of the sub-seasonal to seasonal (S2S) forecast for producing probabilistic forecasts. In a sub-seasonal time-range, the forecast uncertainty raises from both the uncertainty in initial conditions, and the stochasticity and uncertainty (errors) of forecast models. In addition, the dynamical system that describes the S2S variability encompasses various sub-components of the Earth System (atmosphere, ocean, land, etc.). Therefore, to produce reliable forecasts, the S2S forecast systems need to represent the inherent uncertainty in the complex coupled Earth System. Compared to extended-range forecasts (lead time of < 2 weeks) and seasonal forecasts (lead time of > 2 months), the ensemble techniques for the S2S forecast have been had not received much attention prior to the start of the project. In fact, it has been generally considered that ensemble prediction systems on S2S timescale lack sufficient spread among ensemble members (one of an uncertainty measure), and consequently, are overconfident. Therefore, efforts are needed to evaluate and improve methods for representing uncertainty in initial conditions (using methods such as singular vector, bred vector, and lagged average forecasting) and to develop methods to represent the uncertainty during model integration (such as stochastic parameterizations). Effects of the coupled system (e.g., the ocean) on the forecast uncertainty need to be explored. In addition to the ensemble techniques, the initialization strategy of S2S ensemble forecast systems (burst and lagged ensemble) is another practically important area in optimising the use of finite computer resources of operational forecasts. Understanding conditional predictability (uncertainty) and validating its expression in S2S ensemble systems will further enhance the usefulness of S2S forecast information, and the S2S Ensemble sub-project has studied the above aspects related to S2S ensemble forecasting. In this presentation, we will present the progress achieved in S2S Phase II on S2S ensemble forecasting and discuss future prospects.
Co-authors:
