Hi all,
I understand that for data assimilation, the objective is to minimize the error covariance related J function. For adjoint sensitivity analysis, are we minimizing the indicator function? During that process, all the observation-related inputs are not required, right? Thank you very much.