Background error covariance is an important part of variational dataassimilationsystem, which is used to spread the observation information to other grid points and vertical levels of the model.
Furthermore, the conventional observations are done in the MM5 model system, and a four-dimensional variational dataassimilation test is made based on observed data.
最后对MM5伴随模式系统进行了梯度检验,并利用实际资料进行四维变分资料同化试验。
3
Considering that the model is traditionally supposed to be exact and only initial data fields are amended in Adjoint AssimilationSystem (AAS).