Using a Simulation Tool to Model the Ground Level Concentrations of Green House Gases Emitted by Flaring in Petroleum Production in Kuwait Oilfields

Air pollution and its effects on the ecosystem has been a source of concern for many environmental pollution organizations in the world. In particular climatologists who are not directly involved in petroleum industry sometimes express concerns about the environmental impacts of gas emissions from flaring at well heads. For environmental and resource conservation reasons, flaring should always be minimized as much as practicable and consistent with safety considerations. However, any level of flaring has a local environmental impact, as well as producing emissions which have the potential to contribute to the global warming. In the present research the Industrial Source Complex (ISCST3) Dispersion Model is used to calculate the ground level concentrations of two selected primary pollutants (i.e. methane and non-methane hydrocarbons) emitted due to flaring in all of Kuwait Oilfields. In additional, the performance of the ISCST3 model is assessed, by comparing the model prediction with the observed concentration of methane and non-methane hydrocarbons obtained from the monitoring sites. The described model evaluation is based on the comparison of 50 highest daily measured and predicted concentrations of methane and non-methane hydrocarbons. The overall conclusion of this comparison is that the model predictions are in good agreement with the observed data (accuracy range of 60-95%) from the monitoring stations maintained by the Kuwait Environmental Public Authority (EPA). A specific important conclusion of this study is that, there is a need for a proper emission inventory strategy for Kuwait Oil Company (KOC) as means of monitoring and minimizing the impact of methane and non-methane hydrocarbons released because of flaring activities.


INTRODUCTION
Kuwait is a major oil exporting country and its economy, growth and prosperity is heavily dependent on oil production. KOC is at the heart of the petroleum production in Kuwait. The oilfields involve various types of industrial operations and activities, such as drilling, production of crude oil, fuel combustion and flaring of gases which all result in gas emission into atmosphere. In practice, all other sources of emissions are small compared with emissions from flaring. Consequently, a wide range of air pollutant emissions is generated on various sites. Such emissions include carbon dioxide, nitrogen and sulfur oxide gases, methane and non-methane hydrocarbons and suspended particulates.
A comprehensive impact assessment study has been previously published [1] which provides an account and estimates of all emissions of primary pollutants associated from flaring activities in the Kuwait Oilfields. This inventory records the annual emissions of air pollutants: NO X , SO 2 , CO, CO 2 , methane and non-methane hydrocarbons. The emissions are generated from various point sources and aggregated to obtain total pollutants load of ambient air. The emissions of pollutants from the flaring associated with all types of operations in the oilfields, Gathering Centers (GC), Booster Stations (BS), tank areas and other oil production related activities.
In the present research the previously published data are used as the necessary input for the ISCST3 model. Obviously methane and non-methane hydrocarbons are not the only green house gasses which result from flaring activities.
However these gases provide a typical sample which can be used as an input for the ISCST3 model to investigate of the effects of gases emitted from flaring in all of Kuwait Oilfields.

EPA MONITORING STATIONS IN THE STATE OF KUWAIT
Kuwait EPA has established a number of fixed monitoring stations to collect air quality data in the urban areas. These stations continuously measure the levels of pollutants such as SO 2 , NO 2 , CO, NO, CO 2 , H 2 S, O 3 and TSP (total suspended particles) in the air. In additional, the hourly air pollutants concentrations are measured continuously by fixed ambient air stations located over the State of Kuwait.
It is important to note that, in general, all of the monitoring stations are considered as urban stations distributed within the residential areas except for Um Al-Aish station, which is located in the northern part of the country far away from the residential areas. Fig. 1 shows the area map and the locations of Kuwait-EPA air quality monitoring sites. These monitoring stations are equipped with an automatic analyzer and meteorological sensors.
In order to assess the air quality in Kuwait, measurement of the concentrations of pollutants are collected from the Kuwait-EPA air quality-monitoring network. These concentration values of methane and non-methane hydrocarbons are analyzed and compared with the specified limits and guidelines published by the EPA and the model predictions for the ground level concentrations of methane and non-methane hydrocarbons from flaring in Kuwait Oilfields.

Geography of Kuwait:
Kuwait has a small area of about 17,818 km 2 . At its most distant points, it is about 200 km north to south and 170 km east to west. Kuwait is shaped roughly like a triangle, surrounded by land on its northern, western and southern sides and sea on its eastern side, with 195 km of coastlines. The bulk of the Kuwaiti populations live in the coastal capital city of Kuwait. Smaller populations inhabit the nearby city of Al-Jahrah. Kuwait's land is mostly flat and arid with little or no ground water.

Meteorological conditions in Kuwait:
Kuwait has a typical desert climate, hot and dry most of the time. Rainfall varies from seventy five to 150 mm a year across the country, however, rainfall ranging from twenty-five mm a year to as much as 325 mm have also been recorded.
In summer, average daily temperatures range from 42-46°C, the highest recorded temperature has been 51.5°C. The summers are relentlessly long, punctuated mainly by dramatic dust storms in June and July when northwesterly winds cover the cities in sand. In late summer, which is more humid, there are occasional sharp, brief thunderstorms.
Starting from November colder winter weather sets in temperatures dropping as low as 0°C at nights; daytime temperature remains in the 15-20°C range. Frost rarely occurs; rain is more common and falls mostly in the winter and spring.
Winters (November through February) are cool with some precipitation and average temperatures around 13°C (56°F) with extremes from 2-27°C. The spring season (March) is warm and pleasant with occasional thunderstorms. Surface coastal water temperatures range from 15°C (59°F) in February to 35°C (95°F) in August. The winter months are often pleasant, featuring some of the region's coolest weather, with daytime temperatures hovering around 18°C (64°F) and nights being genuinely chilly. Sandstorms occur throughout the year but are particularly common in spring. real meteorological conditions measured and recorded so that a clear picture can be drown about the climate in the state of Kuwait and its affect on the air pollution.
In order to use the meteorological data as input in the ISCST3 model a pre-processing program based on the US EPA. PCRAMMET is utilized to convert the Kuwait data into a suitable format.
A one year hourly record of the surface and upper air meteorological data for year 2006 obtained from the Kuwait International Airport (KIA) weather station is used in the present study for simulation of the dispersion of methane and non-methane hydrocarbons emitted from flaring in all Kuwait Oilfields areas (NK, SEK, WK).
One of the main meteorological factors that can affect the behavior of the pollutants trends during a day is the mixing height and depth of the mixing layer. US-EPA [2,3,4] stated that the estimation of mixing heights from upper air meteorological data is a critical parameter for understanding the formation, dispersion and transfer of ozone and precursors during pollution episodes. The upper air meteorological data were obtained from routine measurements at the KIA weather station for the year 2006. These data were used to calculate the mixing heights (Fig. 2.) and investigate the effects of upper air meteorological data in the diurnal behaviors of ozone and its precursors.
The morning and afternoon mixing height estimates are determined based on the method described by Holzworth [5] and Hanna [6] . Mixing height is estimated by plotting maximum surface temperature and drawing a line parallel to the dry adiabatic lapse rate from the point of maximum surface temperature to point at which the line intersects the ambient lapse rate early morning as shown in Fig. 2.
The prevailing wind direction in Kuwait is along the north westerly quadrant throughout the year, but it is more so in summer. Figure 3a and 3b show detailed wind rose plots for the year 2006 and the main two seasons in Kuwait.  Figure 3a shows the wind rose plot for the winter (November-March) where most of the time the prevailing wind direction is from the North West with calm conditions about 19.11% of the total time and an average wind speed of 4.35 m sec −1 . Figure 3b provides the wind rose plot for summer (April-October) and shows that the prevailing wind direction is also from the North West. This indicates that there no marked seasonal variation in the wind direction throughout the year. The prevailing wind direction in summer is more established than winter season with calm wind 10.92% of the total time and an average wind speed of 4.92 m sec −1 . Moreover, there is no significant diurnal variation in the prevailing wind direction during the day and night times. This tends to minimize the effects of any land or sea breeze in the dispersion of the pollutants in the urban areas of Kuwait.
The most important meteorological element that can control the level of the atmospheric pollution is the wind. The wind in the state of Kuwait results from the influence of the pressure systems, which dominate the area during each season.
In Fig. 4, the frequency distribution of the winds is illustrated. There are about 32.2% of wind speed record is in between 3.6-5.7 m sec −1 and about 17.5% of wind speed record in between 2.1-3.6 m s −1 . As shows in detailed wind rose plots (Fig. 3, a and b), the main prevailing wind direction in NW is more frequent than other directions (i.e., N, NNW and W). In addition, it can be noted that the North West wind direction coincides with high wind speeds than other directions.
The effect of the wind speed is a very important parameter in the dispersion of pollutants as the relationships between the wind speed and the concentration of pollutants downwind of a source is of inversely proportional. This means that when the wind speed reaches its highest level it actually helps in reducing the concentration of any air pollution, thus reducing its hazardous effects on the residential area. On the other hand, slow winds allow for high concentration of pollutants moving slowly over residential areas. Table 1 shows the Mean Monthly Wind Speed (MMWS) and the Mean Monthly Ambient Temperature (MMAT) for 2006. These mean monthly meteorological data were computed from the hourly records during each day of 2006. The annual mean wind speed in 2006 is low being only 4.04 m sec −1 , while MMWS reaches its highest in June (5.23 m sec −1 ) and in July (6.07 m sec −1 ) and its lowest in January   Figure 5 shows the MMAT, maximum and minimum temperatures recorded for each month. The maximum temperature in summer ranges from 40-51°C. Table 2 shows the frequency distribution count for the wind direction under a specify winds speed class in 2006. The frequency of the calm winds was 14.3% of the 8736 hourly record data. In meteorology, the wind direction considered as the direction from which the wind is bellows therefore, a North West (NW) wind will move pollutants to the South East (SE) of the source. Hence, this consideration was taken into account in a construction of Table 1 and the wind rose plot shown in Fig. 3 to make the analysis of the wind    data more consistencies with the modeling results. However, it is very important to note that the ISCST3 model considers the wind direction as the direction towards which the wind is blowing.

MATHEMATICAL MODEL
Industrial Source Complex (ISCST3) dispersion model modified by the US EPA in 1999 is used in the present study. The ISCST3 algorithm is based on a Gaussian plume dispersion model (i.e., it solves the steady-state Gaussian plume equation) and calculates short-term pollutant concentrations from multiple point sources at a specified receptor grid on a level or gently sloping terrain. The ISCST3 model includes a wide range of options for modeling air quality impacts of pollution sources, making it a popular choice for the modeling community is a variety of applications.
Since the ISCST3 model is specially designed to support the US EPA's regulatory model programs [3,4] , the regulatory modeling options, as specified in the revised guidelines for air quality models (USA-EPA, 1999), are the default mode of operation for the models. These options include the use of stack-tip downwash, buoyancy-induced dispersion, final plume rise, a routine for processing averages when calm winds prevail, default values for wind profile exponents and for the vertical potential temperature gradients and the use of upper bound estimates for super squat buildings having an influence on the lateral dispersion of the plume.
The model is capable of handling multiple sources, including point, volume, area and open pit source types. Line sources may also be modeled as a string of volume sources or as elongated area sources. Several source groups may be specified in a single run, with the source contributions combined for each group.
The ISCST3 model implementation requires three main inputs data as follows: Source information: The source parameters required for the ISCST3 numerical model are pollutant emission rate (g sec −1 ), location coordinates (UTM), source height (m), exit inner diameter (m), exit gas speed (m sec −1 ) and exit gas temperature (°C). All the required information on the location coordinates, the emission rates and heights of stacks, diameters, speeds and temperatures of the gas flow at the exits of the stacks were collected from flaring activities from Kuwait oil field as stated in previous published.

Receptor information:
The ISCST3 model have considerable flexibility in the specification of receptor locations, has the capability of specifying multiple receptor networks in a single run and may also mix Cartesian grid receptor networks and polar grid receptor networks in the same run.
Two different kinds of Cartesian coordinate receptors were used as an input to the ISCST3 model, these are: • The uniform grid system of 441 receptors which cover approximately 55 by 53 km. The grid base elements is a square with side length of 1 km. Figure 6 describes one plot figure of the grid under study • Discrete Receptors points corresponding to the location of the major pollution centers and the existing monitoring stations in the State of Kuwait. This means that concentrations in each point in the grid, which is 1km, will be estimated in addition to the discrete point of the existing monitoring stations. The matrix of concentrations can be plotted as a contour map for the selected meteorological data file Indeed, the uniform grid receptors are not need for the model evaluation, neither for investigation of the efficiency of the monitoring sites, but it is a way to have a general view of the pollutants dispersion over the study area.
These receptors are selected based on actual sites in UTM location coordinate of Kuwait map as shown in Fig. 6. Meteorological information: The meteorological data required are anemometer height (m) wind speed (v), wind direction (degree) clockwise from the north, air temperature, total and opaque cloud cover (%), stability class at the hour of measurement (dimensionless) and mixing height (m). The anemometer height, wind speed, wind direction, air temperature and cloud cover are usually obtained from direct measurements.
The hourly stability class and the hourly mixing height are estimated using PCRAMMET. PCRAMMET is a meteorological pre-processor for preparing National Weather Service (NWS) data for use in the ISCST3 US-EPA. The routine measurements of the surface and upper air meteorological data obtained from KIA for the year 2006 is used to run the PCRAMMET to generate an hourly ASCII input meteorological file containing the meteorological information parameters needed for the running of the ISCST3 model. The anemometer height of this station is 10 m. The average hourly meteorological data for 2006 were assumed to be valid for the periods investigated The stability class was defined on the basis of Pasquill categories, which are mainly a function of the hour of measurement, wind speed and sky cover (i.e., the amount of clouds). Based on temperature profile measurements, the mixing height was estimated by the model.

DESCRIPTION OF THE STUDY AREA
The subject area of the present study covers all of the Kuwait's oil producing zones which are located in three selections in the state of Kuwait (Fig. 7). Figure 7 shows the Kuwait map with the location of the three oil producing areas (SEK, WK and NK) and the location of the residential areas.
The distance between the farthest northeast and southeast points of the state boundaries is about 200 Km and from farthest east to west is about 170 Km. Therefore, the total length of the border line is about 685 Km. To cover all of this area, modeling is divided into three individual tasks to calculate the ground level concentrations of methane and non-methane hydrocarbons. The modeling tasks are:

RESULTS AND DISCUSSIONS
ISCST3 model was set up to simulate the ground level concentrations of methane and non-methane hydrocarbons emitted from flaring activities in KOC at all points covered by the receptors information.
Modeling was then carried out by summing the steady state concentration contributions from each source at each receptor point in the study area. The calculations were done based on the model input parameters as described in the previous sections. The simulated results of the emission scenarios using the ISCST3 are on an hourly basis for the predicted concentrations of methane and non-methane hydrocarbons.
The hourly, daily and annual average maximum ground level concentrations of methane and nonmethane hydrocarbons were predicted and output results were compared with Kuwait Ambient Air Quality Standards (KAAQS) at all of the grid point receptors under the study area (443 receptors) as shown in Fig. 6. The maximum hourly, daily and annual Allowable levels of pollutants specified by KAAQS are shown in Table 3. The maximum hourly level as indicated by KAAQS can be exceeded only once a month during a year in the same location. However, the daily and annually allowable limits are not to be exceeded.

North Kuwait oilfield area results
Methane emission: Table 4a-c show the modeling results for the 50 highest hourly, 50 highest daily and the 50 highest annual maximum ground level   in UTM (m) CONC.
- concentrations of methane, respectively calculated at the uniform grid receptors described previously. Isopleths plots (contours) were generated, as shown in Fig. 8a-c. These present the maximum hourly, daily and annual ground level concentration of methane in µg m −3 calculated at the specified uniform grid receptors. The data presented in Table 4a-c and Fig. 8a-c reveal that predicted ground level concentrations of methane for the specified time exceeds the KAAQS ambient air quality standard over the study area.
As shown in Table 4a the predicted maximum hourly average ground level concentration of methane in the study areas exceeds KAAQS by as much as 2761.5 µg m −3 , hour 02:00, 19th January 2006 at the receptor coordinate of X = 774643.69, Y = 3305361.25. As shown in Fig. 8a, this receptor is located nearly 54 km in the NNW direction from the centre of NK Oilfields and not far from the residential areas.
The predicted maximum daily average ground level concentration of methane in the study areas in  -   Table 4c and Fig. 8c show that the highest annual maximum concentration of methane is 6 µg m −3 . Table 5a-c show the modeling results for the 50 highest hourly, 50 highest daily and the 50 highest annual maximum ground level concentrations of non-methane hydrocarbons, respectively calculated at the uniform grid receptors described previously. Isopleths plots (contours) were generated, as shown in Fig. 9a-c. These present the maximum hourly, daily and annual ground level concentration of non-methane hydrocarbons in µg m −3 calculated at the specified uniform grid receptors.

Non-methane hydrocarbon emissions:
The described data reveal that predicted ground level concentrations of non-methane for the specified time exceeds the KAAQS ambient air quality standard over the study area.
As shown in Table 5a and Fig. 9a, the predicted maximum hourly average ground level concentration of non-methane hydrocarbons in the study area exceeds KAAQS by as much as 34442.8 µg m −3 , hour 02:00, 19th January 2006 at the a receptor coordinate of X = 774643.69, Y = 3305361.25.
The predicted maximum daily average ground level concentration of non-methane hydrocarbons in the study area given in Table 5b    -    Table 5c and Fig. 9c show that the highest annual maximum concentration of non-methane hydrocarbons is 200.17 µg m −3 . There were some unexpected problems in NK Oilfields in the year 2005 and the amount of gas flared, as a percentage of production, was almost double of the amount recorded from 2004. The above results reflect this, as well as, the increase in flaring in January 2006, due to regular shut down of Condensate Recovery Unit (CRU's) in NK Oilfields and the prevailing wind direction in Kuwait. Considering Table 4a-c, 5a-c and Fig. 8a-c, 9a-c together, it can be concluded the weather pattern in Kuwait in January 2006, especially the mean prevailing wind direction, significantly contributed to high concentrations of methane and non-methane hydrocarbons at ground level in residential areas located downwind of the NK Oilfields. Table 6a-c show the modeling results for the 50 highest hourly, 50 highest daily and the 50 highest annual maximum ground level concentrations of non-methane hydrocarbons, respectively calculated at the uniform grid receptors described previously. Isopleths plots (contours) were generated, as show in Fig. 10a-c. These present the maximum hourly, daily and annual ground level concentration of non-methane hydrocarbons in µg m −3 calculated at the specified uniform grid receptors.

Non-methane hydrocarbons emission:
As shown in Table 6a the predicted maximum hourly average ground level concentration of nonmethane hydrocarbons in the study area is 5363 µg m −3 , hour 2:00, 28th January 2006 at the receptor coordinate of X = 790158.13, Y = 3203288.25 The predicted maximum daily average ground level concentration of non-methane hydrocarbons in the study area in Table 6b is 473.15 µg m −3 , hour 24:00, 16th January 2006 at the receptor coordinate of X = 790158.13, Y = 3203288.25. For the same location, Table 6c and Fig. 10c show that the highest annual maximum concentration of non-methane hydrocarbons is 17.943 µg m −3 .   -    receptors described previously. Isopleths plots (contours) were generated, as shown in Fig. 11a-c. These present the maximum hourly, daily and annual ground level concentration of methane in µg m −3 calculated at the specified uniform grid receptors.  -     -     As shown in Table 7a the predicted maximum hourly average ground level concentration of methane in the study area is 655.48 µg m −3 , hour 2:00, 28th January 2006 at the receptor coordinate of X = 790158.13, Y = 3203288.25.

Methane emission:
The predicted maximum daily average ground level concentration of methane in the study area in Table 7b is 57.647 µg m −3 , hour 24:00, 16th January 2006 at the a receptor coordinate of X = 790158.13, Y = 3203288.25. For the same location, Table 7c and Fig. 11c show that the highest annual maximum concentration of methane is 2.125 µg m −3 .
The main reasons for high levels of methane and non-methane hydrocarbons encountered in the above results is the increased amount of flaring in January 2006 resulting from frequent shutdowns of CRU's, shortage of gas compression facilities and malfunction of the BS's in SEK Oilfields. Again, these data given strong indication regarding the significant influence prevailing wind direction on the ground level concentrations of methane and non-methane hydrocarbons.

West Kuwait oilfield area results
Non-methane hydrocarbons emission: Table 8a-c show the modeling results for the 50 highest hourly, 50 highest daily and the 50 highest annual maximum ground level concentrations of non-methane hydrocarbons, respectively calculated at the uniform grid receptors described previously. Isopleths plots (contours) were generated, as show in Fig. 12a-c. These present the maximum hourly, daily and annual ground level concentration of non-methane hydrocarbons in µg m −3 calculated at the specified uniform grid receptors.
As shown in Table 8a the predicted maximum hourly average ground level concentration of nonmethane hydrocarbons in the study area is 3485.2 µg m −3 , hour 15:00, 25th August 2006 at the receptor coordinate of X = 766258.06, Y = 3192914.25.
The predicted maximum daily average ground level concentration of non-methane hydrocarbons in Table 8a: ISCST3 output data modeling results for the 50 highest hourly average concentrations of non-methane Table 8b and Fig. 12c show that the highest annual maximum concentration of non-methane hydrocarbons is 45.639 µg m −3 . Table 8c: ISCST3 output data modeling results for the 10th highest  annual average  concentrations of  non-methane  hydrocarbons Location coordinate in UTM (m) CONC.
-   Table 9a-c show the modeling results for the 50 highest hourly, 50 highest daily and the 50 highest annual maximum ground level concentrations of methane, respectively calculated at the uniform grid receptors described previously. Isopleths plots (contours) were generated, as show in Fig. 13a-13c. These present the maximum hourly, daily and annual ground level concentration of methane in µg m −3 calculated at the specified uniform grid receptors. As shown in Table 9a the predicted maximum hourly average ground level concentration of methane in the study area is 221.02 µg m −3 , hour 15:00, 25th August 2006 at the receptor coordinate of X = 766258.06, Y = 3192914.25.
The predicted maximum daily average ground level concentration of methane in the study area in Table 9b is 31.075 µg m −3 , hour 24:00, 25th August 2006 at the receptor coordinate of X = 766258.06, Y = 3192914.25. For the same location, Table 9c and Fig. 13c show that the highest annual maximum concentration of methane is 2.764 µg m −1 . The above data reflect the increase of emissions as a result of increase in flaring in August 2006 due to regular shut down of Shuaiba AGRP and CRU's of WK gathering centers. In addition to this complete shutdown of the two main GC's in WK Oilfields for survey, have contributed to the increase of flaring.
After the comparison between the simulated results for emission scenarios in the North, Southeast and West Kuwait Oilfields it can be concluded the following: • NK Oilfields have generated a high ground level concentration of methane and non-methane hydrocarbons emissions than SEK and WK Oilfields. This is because of the unexpected problems in NK Oilfields. The amount of gas flared in these fluids as a percentage of production in January 2006, was about double that of the previous year -  • Methane and non-methane hydrocarbons are not the only green house gasses which result from flaring activities. The flaring of excess gas is the largest single source of atmospheric emissions arising from KOC operations. However, flaring produces carbon dioxide , oxides of sulphur and nitrogen (NOx) and other chemical species that are produced due to incomplete combustion, such as carbon monoxide, aldehydes, ketones and other organic compounds known as VOCs (Volatile Table 9c: ISCST3 output data modeling results for the 10th highest annual average concentrations of methane Location coordinate in UTM (m) CONC.
-----------------------------------------Rank (  Organic Compounds). However the methane and non-methane hydrocarbons gases provide typical samples which can be used as an input for the ISCST3 model to investigate of the effects of emission from flaring in all Kuwait Oilfields • There is a need for an emission inventory strategy for KOC to minimize the impact of methane and non-methane hydrocarbons released from flaring activities

MODEL PERFORMANCE
The performance of the model is evaluated based on the comparison of 50 highest daily measured and predicted concentrations of methane and non-methane hydrocarbons from KOC flaring at each monitoring station. The overall conclusion of this comparison is that the model predictions are in good agreement with the observed data with accuracy of 60-94% at the monitoring stations used by Kuwait EPA.

CONCLUSION
The simulation of real hourly air quality in the State of Kuwait for the year 2006, inserting the source emission data for that year into the ISCST3 software indicates that the levels of methane and non-methane hydrocarbons from flaring activities in NK Oilfields exceed the allowable daily ambient air standard set by Kuwait EPA.
The model prediction show that these green house gas levels are as much as 248.49 and 3099.8 µg m −3 above the accepted KAAQS for methane and nonmethane hydrocarbons, respectively.
Overall, the statistical comparison between the 50 highest daily measured and predicted concentrations at Kuwait existing monitoring sites shows that the model is in good agreement with the observed data.
This study can be extended to include other pollutants such as NO X , SO 2 , CO, CO 2 and the organic components. Therefore, there is a need for a proper emission inventory strategy for KOC to minimize the impact of NO X , SO 2 , CO, CO 2, methane and nonmethane hydrocarbons released from flaring activities.