Development of new 4-dimnesional variational data assimilation technique using Monte-Carlo method.

提出了一种新的数据同化方案，称为“集合四维数值方法”。该方法可以得到成本函数的宏观梯度，而经典的伴随法得到的是微观梯度。结果表明，该方法适用于强非线性系统。在本研究中，新方案的有效性得到了海洋生态系统模型同化系统的证实。此外，该方法还应用于大气-海洋耦合模式。

The new data assimilation scheme is developed called as "ensemble 4-dimensional ariational method". This method can obtain the macroscopic gradients of cost function, hile the classical adjoint method obtains microscopic gradients. As a result, this new cheme can apply for strong nonlinear system. In this study, the efficiency of the new cheme is confirmed by the assimilation system for marine ecosystem model. In ddition, the new method is applied for atmosphere-ocean coupled models.