Algorithms for the inverse modelling of transport and transformation of atmospheric pollutants

When studying air quality, a key parameter for assessment and forecast is information on emission sources. In applications, this information is not fully available and can be compensated by air quality monitoring data and inverse modelling algorithms. Because of the rapid development of satellite chemical monitoring systems, they are becoming more useful in air quality studies. Such systems provide measurements in the form of concentration field images. In this paper, we consider an inverse source problem and a corresponding data assimilation problem for a chemical transport model. The problem of assimilation of data given as images is considered as a sequence of linked inverse source problems. Each individual inverse problem solution is carried out by variational and Newton-Kantorovich type algorithms. In the numerical experiment presented, an emission source of a primary pollutant is reconstucted via the concetration field of a secondary pollutant. Both data assimilation and inverse problem solution algorithms are capable of approximating the unknown source.