Multipenalty and composed merit functions in Tikhonov-type regularization applied to atmospheric source identification problems

We propose a technique to simplify the minimization of the objective function arising in Tikhonov-type regularization when there are multiple observations, like multiphysics identification problems, and multiple penalty terms. It breaks the original problem into two or more minimizations that are performed sequentially and recursively. In this preliminary work, we consider the case of two minimization steps. We apply the estimation technique to the identification of pollutant emission sources in the atmosphere, considering the data from the Copenhagen experiment.