Zebiak-Cane 模式中条件非线性最优扰动对 ENSO 春季预报障碍的影响

摘要

使用 Zebiak-Cane 模式和条件非线性最优扰动(CNOP)方法，研究初始误差和参数误差共同作用对ENSO春季预报障碍现象的影响。选取模式中的8个El Niño事件，包括4次强事件和4次弱事件，每个El Niño事件又分别从8个不同的起始时间做1 a的预报，这样一共64个预报实验。对每个实验分别计算CNOP误差(初始误差和参数误差同时存在时的最优误差)，通过分析误差增长，发现CNOP误差引起的1 a后的预报误差随着初始预报时间的不同有较大差异，并且不同强度的El Niño事件也会影响CNOP 误差的发展，增长位相中强事件的预报误差要比弱事件的预报误差大一些；而衰减位相中恰恰相反，弱事件的预报误差要比强事件的预报误差要大一些；同时也发现高频El Niño事件对误差增长率的影响较大。本结论有助于提高Zebiak-Cane模式预报ENSO的技巧。%In this paper, the impact of both the initial and parameter errors on the spring predictability barrier (SPB) of the Zebiak-Cane (ZC) model was investigated. We chose eight El Niño events in the ZC model, including four strong events and four weak ones, each with eight initial months for a total of 64 cases. Using the conditional nonlinear optimal perturbation (CNOP) approach, we calculated the CNOP errors (optimal errors when both initial and parameter errors were considered) for each event. By analyzing the error growth, we found that both the initial month and the intensity of the El Niño events can affect the one-year prediction errors of the CNOP errors. During the growing phase, the prediction errors of the strong events are larger than those of the weak events, while for the decaying phase, the weak events have larger prediction errors. In addition, the high-frequency events have a more noticeable impact on the seasonal error growth. This conclusion may help us improve the ENSO prediction using the ZC model.