Integration of Chemical Transport Model and Artificial Neural Network for PM10 Concentration Forecasting

A hybrid model for forecasting PM10 surface concentrations in the Moscow region has been developed and tested. The hybrid model consists of a chemical transport model and an artificial neural network (ANN). The ANN has been trained to predict PM10 concentrations based on numerical forecasts of concentrations and meteorological parameters on a 2 km grid, with the use of PM10 automatic measurements data as target values. The results of the ANN testing with the use of independent samples that included the episodes of high PM10 pollution are discussed.