On the use of US Epa Models-3 system for simulation of an extreme air pollution episode created by coal burning Thermal Power Plants

Stara Zagora is one of the biggest towns in Bulgaria located in the middle of the country. In the summer of 2004 two very high level SO2 pollution events happened there, leading to serious discontent among the population. Analogous events happened in 2005, too. An attempt for numerical study of one of these episodes —from 8 to 11 of July 2005 is the subject of the present paper. The most characteristic feature of the surface concentrations in the city during the period is that sharp SO2 peaks (about and over the alert threshold of 0.131 ppm) appear in the afternoon hours on a background of low concentration (under 10 ppb). As far as all this happened in summertime and the domestic heating can not be the reason for such pollution, the only sources can be the three thermal power plants disposed at 40 km southeast of the town. The 'Maritza-Iztok' TPPs are the main sulphur polluters not only in Bulgaria but in all SE Europe. The total SOx emission from the TPP is about 700 000 tones per year. The Models-3 system was applied for simulation and analysis of the episode with a sequential down scaling to a horizontal resolution of 1 km. The main impression from the analysis of time and space variation of the simulated wind fields during the period is that calm and non-oriented winds prevail. There is a very fast change of wind directions in the different points from the region and at different levels. As a whole, the calculated SO2 concentration fields have a reasonable behavior. The calculated concentrations however, in spite of the numerous runs with different parameters, do not agree well with the measurements. A suitable direction of the wind from TPPs to the town of Stara Zagora is simulated each afternoon. These flows form pollution spots in different places around the town, but not over it. Obviously the meandering of the plume is not simulated precisely enough in this episode with small and none-oriented winds. Several reasons for the ill-simulation are discussed in the study. The good result from these not so successful simulations is that they convinced the relevant authorities in the necessity of organization of operational air quality monitoring system for the region. Such a system is already created and operates practically automatically, supplying local authorities with two-days forecast for SO2, NOX and Dust concentrations in the region. Short description of the system is presented. (author)