Minimization of SO2 Emissions at ADGAS (Das Island, UAE): II- Impact on Air Quality

In Part I of this work, two SO2 minimization schemes, namely, Fuel Gas Sweetening (FGS) and Seawater-Flue Gas Desulfurization (SW-FGD) schemes have been proposed to be implemented at the ADGAS plant (Das Island, UAE). The implementation of such schemes is expected to reduce the SO2 emissions by 77%. The FGS scheme is expected to reduce the H2S content in the fuel gas system by 94% and results in decreasing the total SO2 emissions due to fuel gas usage by 98%. The SW-FGD scheme is expected to reduce the SO2 emissions due to incomplete sulfur recovery by 99.5%.


Introduction
Das Island is an Emirati island in the Arabian Gulf. It is part of the emirate of Abu Dhabi, United Arab Emirates (UAE) but lies well offshore, about 160 km north-west of the mainland. It covers approximately 1.21 km by 2.4 km, and is almost rectangular in shape. It is characterized as a rocky-low island with about 10 m (highest point is 50 m) above sea level. The island is inhabited by oil and gas industry personnel. It exports crude oil and liquefied natural gas by tankers as far as Japan and Europe. The island was formerly a noted breeding site for turtles and seabirds. Despite the oil and gas production, turtles still feed safely in the area and the island is still remained an important landfall for migrant birds.
Throughout its history, the ADGAS Plant at Das Island suffers high rates of SO 2 emissions. The source of SO 2 emissions is mainly coming from the H 2 S containment in the feed natural gas. The SO 2 is the dominant air pollutant to the high pollution levels in the Western Region of the UAE. The high levels of SO 2 emissions there necessitate the need to research all possible means to combat the SO 2 impacts. Lewis et al. [1] described a Model Predictive Control (MPC) that can dramatically result in the minimization of flaring from fuel gas supply networks at the LNG facility at ADGAS. In our opinion, this will also contribute to the reduction of sulfur-containing gases emitted into the atmosphere.
Globally, the SO 2 emitted has the potential to travel in any direction for hundreds of kilometers depending on the meteorological conditions. SO 2 is relatively stable in the atmosphere and has the ability to travel as far as 1000 km [2]. The health and environmental impacts of SO 2 can be found elsewhere [3]. Jie [4] constructed a model to study the impact of SO 2 pollution on health over 78 Chinese counties and found that, after attaining the threshold (8 μg/m 2 ), continuous increase in industrial SO 2 emission density will raise the ratio of population suffering chronicle diseases, among which respiratory diseases occupy a significant proportion.
In general, the workers of ADGAS (and other inhabitants) are subject to be exposed to the outdoor air at all locations within the Das Island during their stay time. There is a lack of long-term studies on the effect of exposure on the health of the workers there. Such studies will help in determining the highest continuous exposure time an employee can withstand during his stay at the Island. In fact, the average exposure duration for the Das Island inhabitants to the various SO 2 GLCs for the various time averages cannot yet be easily determined.
The aim of Part I of this work was to explore feasible technologies that can be implemented at the ADGAS plant at Das Island (UAE) that will result in minimizing SO 2 emissions [5]. Two modifications on the SO 2 minimization schemes have been proposed and suggested to be applied on the current Fuel Gas Sweetening (FGS) and Flue Gas Desulfurization (FGD) systems. It is worth noting that the effect of implementing the SO 2 minimization schemes on plant operations through the FGS and FGD schemes aspects has been discussed elsewhere [5]. The total SO 2 emission rates from ADGAS three LNG trains under current and proposed SO 2 minimization schemes are presented in Table 1. The contribution of the various sources at the ADGAS plant toward total SO 2 emission rates at current and proposed conditions is shown in Figure 1.
Based on the results of part I of this work [5], the current contribution of fuel gas usage of Trains 1 & 2 is 37% of the total SO 2 emissions ( Figure  1). Upon implementation of the proposed FGS scheme, the total SO 2 emissions due to fuel gas usage will be reduced by 98.34% (Table 1) and result in reducing the contribution of fuel gas usage to only 3% of the total SO 2 emissions. On the other hand, the current contribution of the Sulfur Recovery Units (SRUs) to the total SO 2 emissions is 41% ( Figure 1) and upon implementation of the proposed FGD scheme the SO 2 emissions due to incomplete sulfur recovery in the SRUs will be reduced by 99.50% (Table 1) and it will become only 1% of the total SO 2 emissions upon the implementation of the proposed SO 2 minimization schemes.
Furthermore, the implementation of the proposed schemes is expected to result in reducing the total SO 2 emissions from the ADGAS plant by 76.88% (Table 1) and it will result in minimizing the SO 2 emissions from the fuel gas usage and incomplete sulfur recovery but leaves the SO 2 emissions from the Continuous Flaring of Flash Gas (CFFG) as the main contributor to the total SO 2 emissions. If the contribution of the flash gas flaring is excluded, the SO 2 minimization schemes will reduce the SO 2 emissions by 98.96% (from 21,391 to 224 ton/yr), which can be considered as the optimum minimization level [5].
The ultimate goal of this work was to investigate the impact of our proposed minimization schemes of SO 2 emissions [5] on the ambient air quality of the Das Island. The "BREEZE AERMOD GIS Pro" has been used in this work as the air quality model to establish and predict the baseline of the SO 2 GLC at the Das Island post the implementation of the proposed SO 2 minimization schemes. More about the steadystate air dispersion model used in this work can be found elsewhere [6][7][8][9].
Al-Nuaimi [10] studied the effect of upgrading of the SRUs at the ADGAS plant from conventional Claus ® to SuperClaus ® technology and used the Industrial Source Complex (ISC) model to only estimate the highest SO 2 GLCs at the Das Island. The study resulted in a maximum reduction of 33.9%, 0% and 11.4% for the 1-h, 24-h and 1-yr base, respectively. Deligiorgi et al. [11] have used AERMOD steady-state dispersion model to model air pollutant emission from a power plant and identified the dispersion patterns in complex topography in the Chania plain on the Crete Island (Greece). The meteorological assessment is based on a two year dataset (August 2004 -July 2006) from six automated surface meteorological stations. Case studies of the predicted GLCs of SO 2 are presented for days with commonly observed meteorological phenomena. Saqer et al. [12] used AERMOD to estimate the total emissions of SO 2 , non-methanated hydrocarbons (VOCs) and NO x from flares in two petroleum refineries and assessed the impact of these pollutants on air quality in industrial and suburban areas in Kuwait.
Currently the Das Island is lacking continuous monitoring of SO 2 GLCs, which is the indicator of effectiveness of any minimization scheme. However, starting from 2003, ADGAS hired a private company to carry out air quality monitoring at Das Island twice a year. Various pollutants, including SO 2 , were measured at specific locations for a period of 24-h. The used methodology in this work includes the following: • Establishing the design basis for the AERMOD GIS Pro software in order to predict the SO 2 ground-level concentrations under current conditions and predict the 1-h, 24-h and 1-year SO 2 levels on the Das Island. This will acquire site meteorological and air quality data, selection of receptors locations for SO 2 GLCs, and characteristics and rates of all SO 2 emission sources from the ADGAS plant.
• Using the AERMOD GIS Pro software to predict the SO 2 GLCs upon implementing the proposed fuel gas sweetening and flue gas desulfurization schemes by simulating the 1-h, 24-h and 1-year SO 2 levels over the Das Island.
• Exploring the compliance of the predicted SO 2 emission levels upon implementation of the proposed SO 2 minimization schemes to the United Arab Emirates Federal Environment Agency (UAE-FEA) standards.

Design Basis for the Modelling of the SO 2 Ground-Level Concentrations
The design basis for running the BREEZE AERMOD Pro software includes (1) physical characteristics of all SO 2 emission sources, (2)  locations of the selected receptor within the Das Island where the SO 2 GLCs are to be predicted, and (3) SO 2 emission rates from the various sources at the ADGAS plant under current conditions. These data, along with site meteorological data (obtained from AERMOD GIS Pro), are required to run the AERMOD air dispersion model to simulate the background SO 2 GLCs (measured by ADGAS) and predict them once the proposed SO 2 minimization schemes are implemented. More about simulation theory can be found in the literature [13][14][15][16].

Site meteorological data
The meteorological data of the site at hand has to be possessed, and if not available, well established meteorological data at the nearest station has to be used. Meteorological data of the Das Island are not available and the nearest meteorological station (i.e., Abu Dhabi International Airport) is about 160 km away from the Island. However, meteorological data for numerous locations around the world are available through "AERMOD GIS Pro". Acquisition of such data from the "AERMOD GIS Pro" supplier is considered the best choice because these data are reliable and accurate and the supplier provides the necessary data in a format that suits the software itself. Thus the meteorological data provided by AERMOD GIS Pro have been used in this work.

Receptors' locations for prediction of SO 2 ground level concentrations
An aerial view map of the Das Island along with the location of the receptors is shown in Figure 2. The rationale for the receptors' distribution is mainly to monitor the air quality in the residential and recreation areas on the Island. Three locations on the Island had been selected by ADGAS Company for monitoring of air quality and they were selected here as receptors prediction of the SO 2 GLCs over the Island. The physical characteristics of these receptors are given in Table  2.

Physical characteristics of the SO 2 emission sources
The physical characteristics of the SO 2 emission sources include the physical parameters and SO 2 emission rates from each emission source. The locations of the SO 2 emission sources have been obtained from the ADGAS plant records and placed on the Das Island Map. Each location was oriented through the determination of its longitude and latitude ordinates. The diameter and height of each emission source were also obtained. The available physical data were then collected from the data sheets of each stack at the ADGAS plant. See Table 3. It is worth mentioning here that each SO 2 emission source has been individually considered in this study and all the data from the various sources were fed into the AERMOD software. It is also assumed that the plant emissions are evenly distributed over the time.

Estimation of SO 2 emission rates from various sources at ADGAS
All the data necessary to estimate the SO 2 emission rates from all sources were collected. The SO 2 emission rate (kg/h) from each source was determined using the data available from the fuel gas usages (that are emitted from all non-flare sources such as boilers and gas turbines) and the data from flared gases in the continuous flaring of flashed gas sources. The data acquired are the flow rates (mol/h) and the H 2 S concentration (mol %) of the fuel gas usage and flare gas systems. The SO 2 emission rates were then calculated based on the following equations (1 and 2): Table 4 presents a summary of the estimated H 2 S and SO 2 emission rates under current conditions from the various sources at ADGAS. Table 5 summarizes the SO 2 emission rates at the current and proposed

Prediction of SO 2 Ground Level Concentrations
The main objective of air quality models is, in general, to simulate and/or predict the ambient level concentrations of any pollutant given the necessary emissions' source data (i.e., stack size, location and emission rate), and terrain description and meteorological data of the site of interest. Typical outputs of any air quality model include location and magnitude of the pollutant highest GLC and the magnitude of the pollutant GLC at specified receptor locations. Air quality models are also used to verify air quality standards' compliance of existing or proposed industrial facilities, and to assist in the design of effective control strategies to reduce emissions of harmful air pollutants [17].
In fact, the absence of continuous SO 2 GLC monitoring on the Das Island necessitates the use of air quality models in order to simulate the SO 2 GLCs under current conditions and reveal the effectiveness of the proposed SO 2 minimization schemes on the SO 2 GLCs, in particular, and air quality over the Island, in general. Table 6 shows a summary of the 1-h, 24-h and 1-year SO 2 GLC averages at the selected receptors for the years 2003 to 2007 under current conditions and proposed SO 2 minimization schemes. It also shows the corresponding SO 2 GLC standards for ambient air quality set by the UAE-FEA [18]. It is clear from Table 6 that the 1-h SO 2 GLC highest averages under the current conditions always exceed the UAE-FEA at the three receptors. However, the 24-h and the 1-yr SO 2 GLCs highest averages do not exceed the standards except for the 24-h SO 2 level at Al-Jimi receptor. The Al-Sahil and Al-Jimi receptors show SO 2 levels of almost 500 µg/m 3 or more for the 1-h average periods; this imposes serious health effects as per the WHO Air Quality Guidelines [19]. The 24-h SO 2 levels at the receptors frequently comply with UAE-FEA standards but are far exceeding the WHO 24-h set standards (20 µg/m 3 ). Exposure to such levels may exert serious health effects as indicated by the recorded WHO studies [19]. The 1-yr SO 2 levels at the three receptors always comply with UAE-FEA at the three selected receptors. In general, the ambient air quality at Das Island under the current conditions with respect to SO 2 GLCs is considered deteriorated and has the potential to impact the health of the residents of the Island as high SO 2 levels are experienced in the residential areas.

Measured vs. predicted SO 2 GLCs at the selected receptors
A comparison between the measured and AERMOD predicted 24-h highest SO 2 GLCs at the three receptors for the years 2003 to 2007 are shown in Table 7. In fact, AERMOD only predicts the highest   pollutant GLC for a given average. So for various receptors, AERMOD predicts the 24-h highest SO 2 GLCs for a given year. As seen in Table  7, the predicted SO 2 GLC is not occurring in the period when the actual SO 2 GLC measurements were made. Thus, the AERMOD model cannot be verified for the ADGAS SO 2 GLC database because the available data are not sufficient for validation purposes. In addition, the measured SO 2 GLCs are not indicative because there is a possibility that the meteorological conditions during the test period allow for the SO 2 GLCs to be lower than should be.

Temporal variations of highest SO 2 GLC distribution
The predicted temporal variations in the SO 2 highest GLC averages over the Das Island for the years 2003 to 2007 show similar trends for the 1-h, 24-h and 1-yr toward the distribution of the SO 2 GLC highest averages. This can be observed through examining the location of the highest SO 2 GLC averages, and the comparison of SO 2 GLCs at the selected receptors. (2) Comparison of the SO 2 GLCs at the Selected Receptors The average 1-h, 24-h and 1-yr, the mean and the Standard Deviation (SD) of the SO 2 levels at the selected receptors are presented in Table 8. The standard deviations for the 1-h, 24-h and 1-yr averages of the highest SO 2 level are considered low; hence this indicates that the predicted SO 2 levels at each receptor for the specified duration are close to each other. This in turn implies that the variations in SO 2 levels over

Simulation of the SO 2 ground level concentrations
Upon acquisition of the necessary data, the SO 2 GLCs on the Das Island have been predicted using the BREEZE AERMOD GIS Pro software. The 1-h, 24-h and 1-yr highest averages of SO 2 GLCs were predicted at the selected receptors' locations under current and modified SO 2 minimization schemes for the years 2003 to 2007. Upon these results, the impact of process modifications on the air quality has been predicted. Figure 3 shows a sample of the results for the 1-yr highest SO 2 GLCs at the three receptors under current and proposed conditions for the year 2007. Table 9 shows the predicted 1-h, 24-h and 1-yr highest SO 2 GLC averages at the same receptors under current and proposed conditions. Table 9 also shows the AERMOD predictions using the proposed SO 2 minimization schemes once the CFFG is eliminated. The elimination of the CFFG under the proposed SO 2 minimization schemes will result in reducing the SO 2 GLCs at the selected receptors by about 97%. In this case, the highest SO 2 GLC averages at the selected receptors will be about 4.6 µg/m 3 for the 1-hr basis and less than 1 µg/m 3 for the 24-h and 1-yr bases, which represent the minimum that can be achieved at the ADGAS plant.

Spatial distribution of highest SO 2 ground level concentrations
The BREEZE AERMOD GIS Pro software has been used in this work to generate the contour plots of the predicted SO 2 GLCs over the Das Island. The generated contour plots for the 1-yr highest average under current and proposed SO 2 minimization schemes for the year 2007 are shown in Figures 4 and 5, respectively, for comparison purposes. The complete set of the contour plots of the SO 2 GLC distribution over the Das Island is available elsewhere [20].
Based on these contour plots, the highest 1-h highest averages under current conditions occur in the central-west part of the Das Island (residential area). The 24-h highest averages occur in the middle and north-west parts of the Island while the 1-yr highest averages occur in the middle and north-east parts of the Island. . Table 10 shows the SO 2 highest GLCs under current and proposed SO 2 minimization schemes for the years 2003 to 2007.
As seen in Table 10, the current conditions highest SO 2 levels within Das Island, over the years 2003-2007, frequently exceed the standard limits set by the UAE-FEA. The 1-h highest SO 2 level is 1869 µg/m 3 (0.65 ppm). However, exposure to 0.15-0.25 ppm SO 2 (which is less than the current highest 1-h level at Das Island) has the potential to cause cardio respiratory effects to human body. The WHO limit is not to exceed 500 µg/m 3 SO 2 in 10-minutes periods as this imposes health risks on humans in the form of changes in pulmonary functions and respiratory symptoms [19]. On the other hand, the 24-h highest average SO 2 level is 507 µg/m 3 (or 0.18 ppm); the WHO 24-h average limit is 20 µg/m 3 . The effect of this is similar to that of the 1-h highest level. Moreover, the 1-yr highest average SO 2 level is 74 µg/m 3 (or 0.02 ppm); long time exposure to such concentration may have serious effects. Thus, the 1-h, 24-h and 1-yr highest SO 2 levels under the current conditions represent threat to the health of the Das Island residents. Table 10 also shows the highest SO 2 levels at Das Island, over the years 2003 to 2007, upon implementation of the proposed SO 2 minimization schemes. The 1-h highest SO 2 level is 636 µg/m 3 (or 0.22 ppm); the exposure to such level is associated with cardio respiratory response effect on human health. The 24-h and 1-yr highest SO 2 levels are 158 and 37 µg/ m 3 , respectively; the exposure to such levels has the potential to affect the health of the Island residents. However, none of these highest GLC levels exceed the UAE-FEA standards.
Lastly, it should be kept in mind that upon implementation of the proposed SO 2 minimization schemes, the highest 1-h, 24-h and 1-yr SO 2 levels will be shifted to the north and north-east parts of the Island. The north-east part of the Island is no more than an industrial area. See Figures 4 and 5. This shift in SO 2 levels to non-residential areas is justified by the elimination of the SO 2 emissions from the ADMA Gas Turbines (GTs) and the SRU incinerators of Trains 1, 2 and 3. In this case the only remaining contributor to the SO 2 emissions at the Island is the CFFG from the fuel gas sweetening units of Trains   as mentioned above, can be routed back to the ADGAS plant gas feed inlet and will result in the reduction of the SO 2 levels to much less than the UAE-FEA standard limits.

Compliance of the Predicted SO 2 Emission Levels with the UAE-FEA Standards Upon Implementation of the Proposed SO 2 Minimization Schemes
Currently, the SO 2 emission from the ADMA-GTs and the SRU incinerators of Trains 1, 2 and 3 are not complying with the UAE-FEA standard limits. Also, the current SO 2 emission rates from the Fuel Gas Users of Trains 1 & 2 are 89.54% of the UAE-FEA standard limits. This makes these sources susceptible to exceeding these limits in case of malfunction (sudden decrease in the UGAs removal efficiency as a result of process parameters changes). However, upon implementation of the proposed FGS scheme, the SO 2 emission rates (from the Fuel Gas Users of Trains 1 & 2) will be only 5.37% of the UAE-FEA limits (i.e., it will decrease the current SO 2 emission rates by 94%). Furthermore, Table  11 shows that the implementation of the proposed SO 2 minimization schemes has resulted in all SO 2 emission sources at the ADGAS plant to comply with the UAE-FEA limits and have the potential to challenge any future stringent limits imposed by the UAE-FEA with high level of confidence.  The compliance of the SO 2 levels with the UAE-FEA standards for the 1-h, 24-h and 1-yr at the Das Island have been plotted using the BREEZE AERMOD GIS Pro software. Figures 6(a) and 6(b) respectively represent the 1-yr compliance plot under current and under proposed SO 2 minimization schemes for the year 2007. The red color on these plots means the SO 2 GLC exceeds the limit set by the UAE-FEA. The complete set of the compliance plots is available elsewhere [20]. Table  11 shows the compliance results for the ADGAS plant SO 2 emission sources under current conditions and after implementation of the proposed SO 2 minimization schemes.
On the other hand, Table 12 summarizes the main observations of the spatial distribution compliances at the current and proposed SO 2 minimization schemes.
On the other hand, the contour plots indicate that the implementation of the proposed SO 2 minimization schemes will result in a greater area of the Das Island and most of the Island residential areas comply with the UAE-FEA air quality standards. Another aspect is that even though the northern part of the Island does not totally comply with the 1-h standard, its air quality will be improved. Also the highest predicted SO 2 level around the northern part of the Island will be about 650 µg/m 3 compared to the current 2000 µg/m 3 highest level. This means that under the proposed SO 2 minimization schemes all the Das Island area will comply with the 24-h and the 1-yr limits of the UAE-FEA. Lastly, the present study is expected to be of importance to modeling experts and managers of ADGAS plant at the Das Island. Also, the outcome of this study could further assist managers of ADGAS, and similar gas processing plants, on the methodology to reduce SO 2 emissions to meet air quality standards. Meanwhile, the general approach presented here may be of value in application to other similar plants and systems around the world.

Conclusions
Currently the SO 2 emission rates from the ADGAS plant at Das Island are as follows: • The SO 2 emission rates from the SRU incinerators of Trains 1, 2 & 3 and ADMA-GTs do not comply with the UAE-FEA standards. The SO 2 emission rates from Fuel Gas Users of Trains 1 and 2 are 89.54% of the UAE-FEA limits. This makes these sources susceptible to exceeding the standards' limits in case of malfunctions.   • The highest 1-h, 24-h and 1-yr SO 2 levels are 1869, 507 and 74 µg/m 3 , respectively.
• The locations of the highest SO 2 levels are more of Residential Areas which represent a threat to the health of the Das Island residents.
Upon implementation of the proposed SO 2 minimization schemes, the situation at the Das Island will be as follows: • The 1-h, 24-h and 1-yr SO 2 GLC highest averages at the specified receptors will comply with the UAE-FEA air quality standard. In addition, SO 2 levels > 500 µg/m 3 will not be experienced for short periods.
• The average of the 1-h highest SO 2 levels is 636.2 µg/m 3 . The averages of the 24-h and 1-yr highest SO 2 levels are 157.6 and 37.2 µg/m 3 , respectively. The highest SO 2 GLC will shift to the north-east part of the Island (i.e., the Industrial Area).
• • A greater area of the Das Island will comply with the UAE-FEA air quality standard. Also all the Island area will comply with the 24-h and 1-yr UAE-FEA standard.
• The implementation of the proposed SO 2 minimization schemes along with the elimination of the continuous flaring of flash gases will result in 99.2% reduction in the total SO 2 emissions at ADGAS. Once the flash gas flaring is eliminated, the SO 2 GLC 1-h highest averages at the selected receptors will be about 4 -5 µg/m 3 while the 24-h and 1-yr highest averages will be < 1 µg/m 3 .   On the other hand, Table 12 summarizes the main observations of the spatial distribution compliances at the current and proposed SO 2 minimization schemes.  • The routing of the Continuous Flaring of the Flash Gas (CFFG) back to the inlet feed gas of the plant trains will result in 99.19% reduction in the total SO 2 emissions at the ADGAS plant.
Lastly, based on the results of this work, it is recommended • To implement the proposed SO 2 minimization schemes presented in Part I of this work (i.e., fuel gas sweetening and flue gas desulfurization) in order maintain good air quality at ADGAS and in the Das Island. This also requires the replacement of the HYPAK packing of the UGAs of Trains 1 and 2 by IMTP packing.
• To have continuous monitoring of the air quality on the Island that should be easily accessed by the Island workers through brochures and/or broadcasting. This will help to worn the workers on occasions of high levels of SO 2 emissions.
• To carry out comprehensive studies on the occupational health of the workers and residence of the Island.