FORECASTING OF HIGH-LEVEL AIR POLLUTION IN URBAN-INDUSTRIAL AGGLOMERATION BY MEANS OF NUMERICAL WEATHER FORECASTING

The purpose of the work was performing Air Quality (AQ) forecast model. The idea has made a prediction of AQ in urban-industrial agglomeration basing on weather forecasts. Mezoscale meteorological forecasting UM model (Unified Model) developed by the UKMO, artificial neural networks and regression method was used. Two-step approaches were used to AQ forecasting. The first step was based on classification of meteorological conditions and air pollutants and the second step was based on training of the network in situations responsible for high concentrations. The obtained good fit between forecast air pollution levels (PM10, SO2) and real-time observed air quality. However, the significance advantage this works is enforcement results mezoscale numerical weather forecast to pollution prediction.