数值天气预报同化系统的构造及试验研究

Construction of the Data Assimilation System for Numerical Weather Prediction and Its Experiments

采用伴随码方法构造了数值天气预报的伴随模式系统,并对其正确性进行了检验,发现伴随码推导伴随模式是很有价值的工具。用理想场和实际资料进行的试验发现:权重函数和尺度因子选取的不同及不同的网格距对伴随模式同化系统会产生不同的影响。试验表明伴随模式系统有较强的同化能力。

In this paper,the adjoint code method is employed for the construction of adjoint model system for numerical weather prediction,verifications are made for the accuracy of the system,and results indicate that the adjoint code method is a very effective tool.Experimental evidence with ideal and real data shows that different weighting and scaling and grid spacing have different influence on the adjoint assimilation system,and that the system have good ability of assimilation.