“…This can be done by considering ak = aCt) as the vector of concentrations of a pollutant at the grid points of a grid dispersion model (Bankoff and Hanzevack, 1975;Melli et aI., 1981) or at the pollutant monitoring points (Sawaragi et aI., 1976) of the study area. The state vector aCt) might also be extended to include some additional adaptive parameters (Bankoff and Hanzevack, 1975), but we will not discuss this extension here. Then the transition matrix F becomes either the matrix of the spatial discretization (K-model) of the transport and diffusion equation (Bankoff and Hanzevack, 1975;Melli et aI., 1981) (including time-dependent emission and meteorological inputs) or a multiple regression matrix (Sawaragi et aI., 1976).…”