[Pollution with sulfur dioxide in an inhabited industrial area. Methodologic problems in monitoring].

The area around Rho and Pero in the north west of the province of Milan is densely populated and highly industrialised. It is also characterised by high levels of sulphur dioxide pollution. During the 1980s it was decided to check the sources of production of the pollutant and its distribution in the area. The main area for improvement was the refinery at Rho where it was aimed to disperse gases at a higher level by raising the chimneys and to use fuel gas in those burners which were connected to lower chimneys. For domestic systems it was decided to diffuse the use of methane as much as possible and limit oil-fired central heating, by statute from the Mayor. In order to rationalise the programme of action the pollution was to be determined for each single area by subdivision of the territory into a grid. By means of a census of all the production sources of sulphur dioxide appropriate data were collected to estimate each area's contribution to pollution. The theoretical hourly quantity of sulphur dioxide was calculated on the basis of the combustion equation. The data were then analysed, in total and divided into component parts, i.e. the type of fuel used, the season of the year and industrial or domestic usage. With Gifford and Hanna's ATDL mathematical model the data thus obtained, together with meteorological variables, were used to simulate the expected effect on the distribution of pollution after carrying out the programmed actions. The simulations foresee a reduction of up to 80% of sulphur dioxide concentrations in the areas of greatest use of liquid fuels. In the territory in question the monitoring of the pollutant was assured by the presence of five fixed positions of continuous survey. These turned out to be useful above all for analysis of the phenomenon over a period of time. In order to study spatial distribution better 374 findings were gathered in the winter of 1985-6 from various points of the territory by means of a mobile instantaneous detector. From these data the average aerial concentration was calculated on a grid, with sides of 500 metres, first for all the observations and then separately for the direction of the wind. Taking into account meteorological factors and the actual hour (during daytime) when the concentration of the pollutant was recorded, it was possible to estimate concentrations by means of linear models in which meteorological parameters were taken as independent variables.(ABSTRACT TRUNCATED AT 400 WORDS)