项目名称: 空间依赖的参数优化对厄尔尼诺预测的影响研究
项目编号: No.41306006
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 天文学、地球科学
项目作者: 吴新荣
作者单位: 国家海洋信息中心
项目金额: 24万元
中文摘要: 厄尔尼诺事件对人类活动和地球气候系统有着深远的影响。研究表明,耦合数据同化和模式偏差校正能降低预报初始场的动力不一致性,提高厄尔尼诺的可预报性。目前,大多数业务化厄尔尼诺动力预测系统未同时进行耦合数据同化和模式偏差校正,而相关的研究工作则采用简单的耦合数据同化方法和模式偏差订正方法,这一定程度上限制了厄尔尼诺的预报技巧。本研究拟基于Zebiak-Cane模式,利用集合卡尔曼滤波方法建立耦合数据同化模型,并引入空间依赖的参数优化方案,同步进行状态估计和参数优化,从而有效降低预报初始场的误差以及由参数的不确定性导致的模式偏差。将耦合数据同化的分析场作为初始场,进行百年以上厄尔尼诺回报试验,借此深入探讨空间依赖的参数优化对厄尔尼诺预测的影响。本研究可以为业务化ENSO动力预测提供重要参考,因此具有重要的社会和经济价值以及广泛的应用前景。
中文关键词: 参数优化;集合卡尔曼滤波;ENSO;预报;空间依赖
英文摘要: El Ni?o events have extremely important influences on the human activities and the earth climate system. It has been realized that both coupled data assimilation (CDA) and the model bias-correction can reduce the initial shock and improve the predictability of El Ni?o. Most of the current operational systems for the prediction of El Ni?o do not instaneously perform CDA and correct model biases, and the related researth works adopt simple CDA and model bias-correction methods, which more or less limits the forecast skill of El Ni?o. Based on the Zebiak-Cane dynamical model, with an ensemble kalman filter, this study attempts to establish an ensemble coupled data assimilation (ECDA) model including the geographic-dependent parameter optimization (GPO) which can instaneously implement state estimation and parameter optimization, so as to effectively reduce the initial error of El Ni?o prediction, and the model bias induced by the uncertainty of model parameter. Started from the analysis results of ECDA, a long-term hindcast of El Ni?o will be performed so as to explore the impact of GPO on the prediction of El Ni?o. This study has a practical significance for the operational prediction of ENSO. Thus it has important social and economical values and broad application perspectives.
英文关键词: Parameter Opimtization;Ensemble Kalman Filter;El Ni?o–Southern Oscillation;Forecast;Geographic-Dependent