Real-time forecast of air pollution episodes in the venetian region. Part 1: the advection-diffusion model

Abstract

The paper describes a mathematical model of pollution dispersion in an airshed. The model is of the advection-diffusion type and is of rather general form, in spite of some simplifying assumptions concerning the meteorological inputs and the diffusion coefficients. The paper suggests a numerical solution algorithm of the advection-diffusion equation, which proves to be accurate, unconditionally stable and computationally efficient. In particular, such an algorithm (which is of the Carlson-Crank-Nicolson type) allows one to use a non-uniform grid. This characteristic leads to a relevant computational saving in the presence of non-uniformly distributed emission sources (pollution from an industrial area). The model and the solution scheme are applied to the Venetian lagoon sulphur dioxide pollution case. The model performance is good in normal pollution situations, but unsatisfactory in the presence of episodes (high concentrations) because of the input inaccuracies (mainly lack of information about actual emission schedulings). This is the reason why the model has been corrected through the ‘stochastic embedding’ described in Part 2.