THE EXAMPLE OF THE VERIFICATION OF THE SO 2 DISPERSION MODEL FROM INDUSTRIAL SOURCES AND SMALL HOUSE STOWS IN THE CITY OF ZENICA

dispersion patterns from industrial plants and home furnaces in the area of the City of Zenica. The validation results show a good correlation of the concentration of SO 2 concentration obtained by modeling and measurement at the station for air quality measurement "Crkvice".

The quality of air in the Zenica area has been disturbed for many years, which causes justified concern for the citizens of Zenica. Monitoring shows that air is more polluted, that polluters are ever more numerous and that their impact on the health of the population and the environment is increasingly pronounced. Since monitoring of air quality is a costly and demanding job, dispersion modeling is increasingly used to determine the distribution of pollution or concentration at all points of the observed domain. Validation and calibration of the dispersion model is performed using air quality monitoring stations. Once the model is verified the number of air quality monitoring stations in the area covered by the model can be reduced to the minimum number of smart stations used only for calibration of the model. Therefore, the models can save considerable resources, because they allow reduction of the number of stations for measuring air quality. The paper describes the verification of SO 2 dispersion patterns from industrial plants and home furnaces in the area of the City of Zenica. The validation results show a good correlation of the concentration of SO 2 concentration obtained by modeling and measurement at the station for air quality measurement "Crkvice".

…………………………………………………………………………………………………….... Introduction:-
Sulphur dioxide can affect both health and the environment. Exposures to SO 2 can harm the human respiratory system and obstruct breathing, and children and elderly are particularly sensitive to effects of SO 2 . Sulphur oxides (SO x ), can react with other compounds in the atmosphere to form small particles, thus contribute to particulate matter (PM) pollution causing additional health problems. At high concentrations, gaseous SO x can contribute to acid rain that can harm sensitive ecosystems [1]. In the Zenica Valley, almost 12.874 tons of SO 2 is emitted from industrial plants alone [2]. The contribution of small house stoves is still unknown, and contribution of traffic on the SO 2 concentration is not significant. This model of SO 2 dispersion is made as an effort to determine the impact of boilers on air quality in the Zenica Valley.
The paper deals with an example of verification of sulfur dioxide dispersion model from 20 stationary sources in the City of Zenica and from three industrial sources (primary point sources of metallurgical plants of the Arcelor Mittal Zenica). The height of the chimneys of stationary sources is less than 80 m high, and in period of appearance of the inverse layer, they significantly pollute the ground layer of the atmosphere [3]. The modeling was carried out for 1092 2010 due to the existence of all necessary input data such as air quality data, relevant emission measurements of pollutants from all sources, meteorological data of the atmosphere and fuel consumption data for each pollution source. The measurements of SO 2 emissions from 20 stationary sources was carried out by the Institute "Kemal Kapetanović" Zenica [4]. The data for three industrial sources are taken over from the Federal Ministry of Environment and Tourism. Table 1 gives the characteristics of the air pollution sources, the amount of fuel consumed and the type of fuel and the total SO 2 emissions. Annual production was also given for industrial sources. The position of the source in the Zenica Basin and domain within which modeling was performedis given in Figure  1. In addition to this data, the meteorological parameters which were obtained on the basis of the measured values at the meteorological station "Brist", which define atmospheric conditions, are also entered into the model. Validation of the dispersion model was performed at the stationary source in the area of "Cantonal Hospital of Zenica" because it is the only source that works continuously throughout the year and for which exist reliable data needed for verification of the model. In addition, near this stationary source there is air quality monitoring station "Crkvice" for continuous monitoring of sulfur dioxide concentrations in the air. Verification was performed by comparing the results of the model with the measured values obtained from previously mentioned monitoring station, for the summer and winter measurement periods. The summer period was chosen because during the summer there are no 1093 other active pollution sources in the area of the stationary source "Cantonal Hospital". Verification during the winter period was performed in order to see the contribution of small house stoves and traffic to the observed site, because to the lack of reliable data on emission of polluting substances for these sources.
The verification of the model gives the right way for estimating the influence of the pollutant source on air quality in the specific geo-urban and industrial conditions of the Zenica valley. Applying such a verified mathematical model to the emission simulation of sulfur dioxide in the given space can significantly reduce costs that require the assessment of air quality by continuous measurement. In addition, modeling can give much more data than those gained by the network of air quality monitoring stations.

Methodology:-
The measurement of SO 2 concentration in flue gases was performed by the portable TESTO 350 XL flue gas analyzer. The TESTO 350 XL is device for analysis of the composition of gases generated by fuel combustion, pressure measurement, gas flow rate and the efficiency of the combustion process. The measurement is performed with built-in electrochemical cells. Dry flue gases flow through cells from which was previously removed moisture in the dryer that is an integral part of the device. The measurements also included the total flow rate through which the sulfur dioxide emission into the air was calculated.
Dispersion modeling was performed using the AERMOD program package. The model is applicable in rural and urban areas, plain and complex field, for point and surface sources. Using relatively easy approach AERMOD combines current concepts of flow and dispersion in the complex field. Where it is appropriate, it can be modeled that the plume fly or collide with the terrain or to track it. This approach is designed to be physically realistic thus avoiding the need to pre-define the terrain types. The measurement of sulfur dioxide concentration in the ambient air for the observed period was carried out at the air quality monitoring station "Crkvice".
When determining the boundaries of the model in a specific geo-urban area with respect to the configuration of the terrain and wind rose, a uniform network of receptors has been defined in order to obtain a map of ground-level concentration of pollutants.  1094 The required input data for the modeling process are obtained on the basis of the measurements of sulfur dioxide concentrations in waste flue gases and flue gas flow. The data of the measured emissions of SO 2 emitted into the atmosphere through the chimney of the observed stationary sources for the given time period are entered in the AERMOD dispersion modeling software. The meteorological parameters, which define the atmospheric conditions, obtained based on measured values at the meteorological station were also entered in the AERMOD. The data about terrain configuration are entered in the form of digitized maps, referring to the shape of the terrain (altitude), the type or purpose of the land and other characteristics of the terrain. Output of the modeling process can be presented in different forms.
It should be mentioned that the obtained ground level sulfur dioxide concentrations are not the total concentration, but only the result of concentrations from the emissions from the chimney of the observed stationary sources. Thus, the presented dispersion modeling results do not contain the influence of any other sources of polluting or the natural concentration of pollutants in the atmosphere. Figures 2 and 3 show the results of the dispersion modeling of annual emissions of sulfur dioxide from the observed sources for the year 2010.  As can be seen from the results of dispersion modeling of sulfur dioxide concentrations over the observed period, air quality limit values were exceeded due to the influence of pollutants emissions from stationary sources (according to the Rulebook on air quality limit values "Official Gazette of FBiH" No. 05 /12)

Verification of the model:
In order to make the data obtained as a result of modeling comparable to the measured concentrations of pollutants at the control points, the calculation of average daily sulfur dioxide concentrations for the selected summer and winter seasons, as a dominant influence on the air quality within the model boundaries, was performed.
It is especially important to analyze sulfur dioxide concentrations from the dispersion modeling process from the aspect of the hospital's chimney impact on the environment, as it is located near the air quality station "Crkvice" and is the only large source of pollution active throughout the year. Therefore, the emphasis in further analyzes is on monitoring changes in ambient sulfur dioxide concentrations from the "Crkvice" measuring site.
By analyzing the matching of the modeling and measurement results, it is possible to evaluate the applicability of the SO 2 dispersion model used. For the validation process, it is necessary to find a set of measurement data from the field and to extract the pollutants that can be connected to a specific source of pollution in the modeled space in a certain part of the modeled period. 1096 The choice of the representative period for validation of the model was done on the basis of the following criteria: 1. Select a period for which there are sufficient quality data of air quality measurement (more than 60% of quality measurements), 2. Select a period in which there are no significant, dominant impacts of other pollutants (during the observed period there is no in function a larger number of small house stoves that would contribute more to the increase of ground concentrations in the immediate vicinity of the control metering point where the air quality monitoring station is located) . 3. Select period with stable atmospheric conditions affecting reduced horizontal dispersion and lack of convective air mixing, 4. The time period must be long enough (3-10 days) to allow sufficient time to transport polluting substances to all edges of the defined area or network of receptors, and thus to the location of the metering stations Verification of the model for the summer period: As the most optimal period of the year for the validation of the dispersion model, the summer period was taken, due to the absence of other low height polluting sources, that is, domestic fireplaces which are not in use during the summer period. The most suitable stationary source of contamination was the chimney of the Zenica Cantonal Hospitals boiler, because boiler works continuously throughout the year. The boiler in the winter period due to the heating period is in higher capacity that was accounted for during modeling. Validation of the model was done with regard to sulfur dioxide due to its uniform dispersion and relatively known behavior in the atmosphere, i.e. stability and distances of dispersion. Table 2 provides the basic input data of the stationary source "Cantonal Hospital of Zenica" used for validation. For the verification of the model the period from 01. June until 31. August 2010 was taken from the summer season of the year 2010, for which most of the criteria mentioned above were met. During the observed period, 98% of the measured values, which were relevant to the modeling process and validation of the obtained results, were correct and usable. This refers to data on the measurement of sulfur dioxide concentrations from the "Crkvice" measuring site, the measurements of the emissions from the Zenica Cantonal Hospital Boiler and other data required for modeling. Graphic representation of daily averages of sulfur dioxide concentrations for the selected summer period of the model verification is given a Figure 4. As it can be seen in the Figure 4 certain parts of the selected summer season have increased concentrations of the sulfur dioxide. This means that certain pollutants have an impact on air quality throughout the year i.e. high concentrations of the sulfur dioxide in the periods when the influence of local pollution sources is reduced to a minimum.
1097 From the selected summer period, two shorter periods (episodes) with low and elevated concentrations of sulfur dioxide were selected. First period with low concentrations from 18. June to 24. June and second period with elevated concentrations from 11. August to 17. August. Daily average sulfur dioxide concentrations for selected periods are given In Table 3, and graphical interpretation of the data from table 3. is given in Figure 5.  For both summer verification periods, Average daily values calculated using the dispersion model, and these values are given in Table 4. The table shows the measured and modeled values and percentage of matches of modeled and measured values. Graphical representation of data from table 4 is given in Figure 6a. Figure 6b shows wind speed and direction diagrams for the first summer period of model verification.   As it can be seen from the figure 6a) during the first episode of lower concentration of pollutants, results of the model follow the trend of the measured concentrations, but the percentage of correlation of these results is low. The model results are significantly smaller than the measured values. From the Figure 6b it is obvious that wind direction cannot be dominant influence factor for this deviation.

I summer period Date
Modeled and measured values of SO 2 concentrations for the II summer verification period are given in Table 5. The table also shows percentage match of modeled and measured values. Graphical representation of data from table 5 is given in Figure 7a. Figure7 shows wind speed and direction diagrams for the second summer period of model verification.

Analysis of the Verification Results for Summer Period:
Analysis of the data for the summer verification period has shown that the application of the given dispersion model for this kind of pollutant sources and pollutants into the air is adequate, with the provision of a sufficient data set for the purpose of quality verification and the design of the model itself. Matching trends of measured and modeled values pollutants in the air is one of the first indicators of the correctness of the applied dispersion model. In the case mismatching trends of measured and modeled data, it is necessary to carry out additional checks of the wind roses and polluters and check the times of data we have at our disposal. In some cases, it might come to delay in the comparison of measured data with emission data from emission sources, which may be due to terrain configuration, ground winds and distances between sources of pollutants and measurement stations. During the process of designing this model, the trends of the model and the quality of air quality have matched which means that model was properly set. However, there was a deviation of the modeled and measured values at the end of the verification intervals. Such errors are most often due to insufficient precision of the data averaging periods. In the case of this model, average annual energy consumption data, average daily air quality values and hourly wind data were used, hence the deviations that appeared were expected. Deviation of the value at the end of verification intervals does not have a significant impact on the final results when yearly average values are modeled. In that case, deviations that occur during the verification process should be taken with the caution, and considered as expected.

Verification of the model for the winter period:
High concentrations of sulfur dioxide in the ambient air were recorded at Crkvice measurement station during the whole observed period. In addition to the chimney of the cantonal hospital, small house stoves are a significant source of pollution in this period. But, this source of pollution is not taken into account since its influence is unknown and cannot be estimated.
Verification of the model for the summer period showed a very good correlation with high concentrations at the Crkvice measurement station in periods when atmospheric conditions were favorable for such transport of pollutants from the emission source of the Cantonal Hospital. Concerning this, it is possible to estimate the influence of other sources in the Zenica valley, on an annual basis, by comparing the influence of the Cantonal Hospital on the Crkvice measurement station in the summer period and influence of all sources in the winter period. Data used for SO 2 dispersion model verification in the winter period is showed in table 6. For the purpose of model verification, the period from 01. December to 31. December of the year 2010 was taken for the winter season. This period during the winter season is selected because sulfur dioxide concentrations are much more pronounced in this period, and it is assumed that the sources in the surrounding of the measurement station have greatest influence on the air quality. Graphical representation of daily averages of sulfur dioxide concentrations for the selected winter period of the model verification is given a Figure 8. Daily average sulfur dioxide concentrations for the periods of low and high sulfur dioxide concentration periods are shown in Table 7. Graphical representation of data from table 7 is given in Figure 8.  Modeled and measured values of SO 2 concentrations (daily averages) for the first winter verification period are given in Table 8. The table also shows match percentage between modeled and measured values. Graphical representation of data from table 8 is given in Figure 10a. Figure 10b shows wind speed and direction diagrams for the first winter period of model verification. a) b) Figure 10:-Daily averages of sulfur dioxide obtained by measuring and modeling (a) and (b) the wind direction and wind speed diagram for the first winter period of the model verification.
As it can be seen from the figure 10a) during the first winter episode of high concentration of pollutants, results of the model do not follow the trend of the measured concentrations. Percentage of correlation of these results is very low ( Table 8). The model results are significantly smaller than the measured values.

1102
Modeled and measured values of SO 2 concentrations for the II summer verification period are given in Table 9. The table also shows match percentage of modeled and measured values. Graphical representation of data from table 9 is given in Figure 11a. Figure 11b shows wind speed and direction diagrams for the second winter period of model verification.  Figure 11:-Daily averages of sulfur dioxide obtained by measuring and modeling (a) and (b) the wind direction and wind speed diagram for the second winter period of the model verification.
As it can be seen from the figure 10a) during the second winter episode of lower concentration of pollutants, results of the model follow the trend of the measured concentrations. Percentage of correlation of these results is low (Table  9).

Analysis of the Verification Results for the Winter Period:
On the base of presented data for the winter period of model verification, it is clear that the verification of the model cannot be carried out during the winter period and that it is necessary to select the period of the year when there are no other sources of pollution. Only then can the real impact of the emission source on air quality be assessed. In order to verify the model and in the winter period it is necessary to include other pollution sources in the environment into the model and to consider the overall influence of the surrounding sources on the air quality condition. This kind of verification cannot be carried out without the good register of pollution sources, greater the data volume of the pollution sources, long-term data of the condition of atmosphere and correct data about ground winds at micro location.
The apparent presence of other sources of pollution, which could not be accounted for, could be the reason for huge mismatch between modeled and measured values in the winter period of the model verification. The wind direction