Determination of The Permissible Distances in Vicinity of The Gas Station According to The Concentration of Benzene

Benzene is considered a toxic and hazardous pollutant in Tehran metropolis. The storage tanks of petroleum products and refueling in gas stations are among the main sources of benzene emissions. Using the software AERMOD and reviewing the benzene dispersion maps at different distances from 412 storage tanks at 148 gas stations, it was found the permissible distance of the emission source is dependent on various variables such as the number of loading times and the storage capacity. When, storage capacity in the range of 60,000 L to 96,000 L and the number of loading is in the range of 676 to 1,328 times a year, the concentration of benzene at a distance of 30 m of the emission source reaches the annual standard of 5 μg/m 3 . While, storage capacity in the range of 80,000 L to 128,000 L and the number of loading is in the range of 1,329 to 1,834 times a year, the concentration of benzene at a distance of 40 m of the emission source reaches the annual standard of 5 μg/m 3 . Also, based on the analysis of data and the linear regression equation, the permissible distance of the emission source can be predicted.


Introduction
Gasoline vapors, including benzene, adversely affect the environment, and large volumes of such emissions cause an increase in the greenhouse gases, which in turn leads to global warming, climate changes, acid rains, chemical reactions, smog, etc (Tsai 2016). Due to the high vapor pressure and high volatility, gasoline is easily evaporated, leading to formation of volatile organic compounds (VOCs). VOCs The dispersion map showed that the pollutants were mainly concentrated around emission sources. It was also found that mathematical modeling is a useful tool for studying dispersion of atmospheric pollutants (Macêdo and Ramos 2020).
Different programs are used to calculate emissions from organic liquids stored in petroleum tanks at gas stations. The TANKS 4.0.9d program can estimate VOCs and hazardous pollutants emitted from various types of storage tanks, including horizontal tanks. This software can be used to estimate emissions from tanks during storage and loading operations. The tanks studied of this project are horizontal and release vapors during storage, charge and discharge. Accordingly, in this study, the TANKS 4.09d program was used. The raw input data for this software include tank speci cations, location, and the properties of

Data And Methods
In this study, 148 gas stations in 22 districts in Tehran city ( Fig. 1) were examined.
First, the factors affecting the emission of gasoline vapors, including the number of gas stations, number of gasoline fuel storage tanks, maximum operating capacity of storage tanks, gasoline sales rate, number of loading and the number of active nozzles at gas station in Tehran was investigated. In the next step, to evaluate benzene emissions, data on the following parameters for 412 storage tanks were used as inputs to TANKS 4.0.9d Software: Physical speci cations of the storage tank (length and diameter), the number of loading times, the maximum operating capacity of the tank, weather parameters, the average daily and annual ambient temperatures, the maximum and minimum daily and annual temperatures, the average wind speed, speci cations of petroleum products stored in the tank, type of chemical compounds in the petroleum product, the weight percentage of compounds, pressure, and the vent vacuum of the storage tank. Based on information from the National Iranian oil products Distribution Company Fixed-roof horizontal underground tanks with a nominal capacity of 45,000 L and 24,000 L were considered. Due to the presence of gasoline vapors over the surface of the liquid stored in the tanks, the maximum operating capacity of the tanks is respectively 32,000 and 20,000 L.
Climatic information was obtained from the Meteorological Organization of Iran and related sites for MEHRABAD station in Tehran, which is the closest station to the study area for 12 months of 2018. AERMOD is capable of de ning point, surface and volume sources. To model benzene dispersion, data such as emission rate of volume sources (g/s), pollutant release height, volume source center relative to the ground surface (m), the initial side dimension of the volume source (m), the initial vertical dimension of the volume source (m), and x and y coordinates of the center of the volume source relative to the origin selected for the volume source were entered into the model. AERMOD is composed of a meteorological pre-processor called AERMET and a geological pre-processor known as AERMAP. For the pre-processor AERMET, the raw surface and upper air meteorological data (a total of 8,760 hourly data) including wind speed (FF), wind direction (DD), humidity (U), temperature (t) and cloud cover (n) for a 1-year (2018) were taken from Iran Meteorological Organization for MEHRABAD Station as the closest station to the study area. The collected data were then converted to format acceptable by AERMET.
For the pre-processor AERMAP, the digital elevation model (DEM) of Tehran was prepared to examine the effect of terrains on the concentration of pollutants and to process the topographical data of the region.
This pre-processor has been designed based on the USGS topographical maps and analyzes topographical information of the region. Using the results of these two pre-processors and additional data on the emission sources and the receiving grid, AERMOD performs calculations and outputs the nal results.
In order to evaluate the results of a dispersion model, statistical parameters are used (Haq et al 2019; Pandey and Sharan 2019). The AERMOD air dispersion model was investigated for its performance in predicting concentration of pollution emitted from petroleum re nery. A set of statistical parameters was employed to evaluate model performance. Results indicated that AERMOD can provide good results (Thepanondh et al 2016).
In this study to determine which variables are most in uential on benzene concentration and in order to nd the corresponding equation, data related to variables (number of loading times, tank capacity and permissible distance of the emission sources) were inserted into SPSS 20 software as primary data. Multivariate regression was used for the parameters of number of loading times and tank capacity as an independent variable and permissible distance as a dependent variable. The data obtained from modeling were analyzed using the above software and ANOVA test.
The ANOVA determines whether a regression model can signi cantly and appropriately predict the variations in dependent variables. To evaluate the statistical signi cance of the regression model, the last column of the table (sig) will be examined in the ANOVA test. If the value obtained in the (sig) part of the ANOVA was less than 0.05, it would mean that the regression model was signi cant.

Results & Discussion
3.1. Investigation of benzene concentration results and determination permissible distance of gas station from sensitive and residential areas.
According to the software output and calculations performed, emissions of benzene vapor emitted from 412 storage tanks in 148 gas stations in Tehran city was investigated. The lowest emission rate of benzene is about 0.0023 g/s and the maximum emission rate is 0.0314 g/s. In order to determine the permissible distance of emission source from sensitive and residential areas, using AERMOD dispersion model, the dispersion rate of benzene concentration with a grid distance of 20-30-40-50 meters was determined and their distribution in the area was simulated.
When the emission rate of benzene vapors is in the range of 0.0023 to 0.012 g/s, storage capacity in the range of 20,000 L to 64,000 L and the maximum number of loading is 675 times a year, the concentration of benzene at a distance of 20 m of the emission source reaches the annual standard of 5 µg/m 3 . Therefore, it can be concluded that in order to construct a gas station with similar conditions, it should be at least 20 meters away from residential and sensitive areas.
While the emission rate of benzene vapors is in the range of 0.013 to 0.0218 g/s, storage capacity in the range of 60,000 L to 96,000 L and the number of loading is in the range of 676 to 1,328 times a year, the concentration of benzene at a distance of 30 m of the emission source reaches the annual standard of 5 µg/m 3 . So, it can be concluded that in order to construct a gas station with similar conditions, it should be at least 30 meters away from residential and sensitive areas.
When the emission rate of benzene vapors is in the range of 0.0219 to 0.0314 g/s, storage capacity in the range of 80,000 L to 128,000 L and the number of loading is in the range of 1,329 to 1,834 times a year, the concentration of benzene at a distance of 40 m of the emission source reaches the annual standard of 5 µg/m 3 . Therefore, it can be concluded that in order to construct a gas station with similar conditions, it should be at least 40 meters away from residential and sensitive areas.
Based on the results, number of loading times and storage capacity are the main factors affecting benzene emissions from gas stations.  (Fig. 3a), a gas station with a storage capacity of 96,000 L and the loading number of 1,328 times per year at a grid distance of 20, 30, 40, and 50 m (Fig. 3b), a gas station with a storage capacity of 128,000 L and the loading number of 1,834 times per year at a grid distance of 20, 30, 40, and 50 m (Fig. 3c).
The permissible distance of the emission sources from each gas station is determined based on the factors mentioned in Table 1.  According to the above studies and considering benzene emissions and dispersion at different distances of the emission sources and comparing the results with the annual standard limit of benzene 5 µg/m 3 , it was determined that the permissible distance from the emission sources is dependent on various variables. The rst factor is the number of annual loading times and the second factor is the storage capacity of tanks in each gas station. As a result a xed number cannot be declared for the permissible distance. For a storage tank with a capacity of 64,000 L and the maximum loading of 675 times per year, the minimum permissible distance between the gas station and the residential areas equals 20 m. For a storage tank with a capacity of 96,000 L and the maximum loading of 1,328 times per year, the minimum permissible distance between the gas station and the residential areas equals 30 m. For a storage tank with a capacity of 128,000 L and the maximum loading of 1,834 times per year, the minimum permissible distance between the gas station and the residential areas equals 40 m. Accordingly, by increasing the capacity of storage tanks and the number of annual loads of the above declared values, permissible distance from residential areas more than 40 meters.  (Table 2).

Conclusions
Based on software output and calculations, the amount of benzene emitted from storage and the annual loading operations of underground tanks (charge and discharge) were obtained. Then, using the AERMOD program, benzene concentration in the environment surrounding the emission source was predicted. By comparing the predicted benzene concentration with the annual standard limit of benzene (5 µg/m 3 ), the minimum permissible distance of gas stations from residential and sensitive areas was determined.
On the other hand, In order to determine which variables (storage capacity and number of loading times) are most in uential on benzene concentration and to nd the equation for the permissible distance from the emission source, the parameter data (loading times, tank capacity, and permissible distance from the emission sources) were inserted into SPSS 20 as primary data. Multivariate regression was calculated for the parameters of loading times and tank capacity (independent variables) and permissible distance (dependent variable).
In the presence of additional factors affecting the emission of benzene vapors in the environment (such as vehicle refueling, liquid gasoline spills, and evaporation from leaking hoses and connections), that this will be an underestimation of total gasoline vapor emissions from gas stations. Therefore, setback distances might even need to be larger than suggested in this study.
According to the results of this study, construction of new gas stations should be supervised by the authorities and supervisory organizations to take into account the minimum permissible distance of residential areas and critical points based on the storage capacity and the number of loading times. Figure 1 The location of 148 gas stations in 22 districts in Tehran