Sulfur Dioxide (SO2) Monitoring Over Kirkuk City Using Remote Sensing Data

Air pollution mapping is now being an important issue to manage and enhance the environment of a city. The major problems of air pollution mapping is the data acquisition due to the high cost of instruments and the high spatial distribution requirements. This study aimed to monitor Sulfur Dioxide over Kirkuk city using Landsat-8 thermal bands to provide Department of Environment Kirkuk with low-cost Sulfur Dioxide concentration maps to better manage the city. The study used correlation analysis to find a relationship between Sulfur Dioxide ground-based measurements and satellite data. The ground-based measurements were collected from (17) stations distributed in Kirkuk city in January, 2014 using NOVA device to measure SO2 concentrations. The research showed a good correlation between groundbased measurements and satellite data with (R2=0.48 for band 11 and R2= 0.52 for band 10). Therefore, the study resulted that with band 10 of Landsat-8 data, better SO2 can be monitored than using band 11. It is recommended to other researchers to investigate the ability of free remote sensing data to monitor all elements that specify the air quality of a city. *Corresponding author: Ghada Hasan, Civil Engineering Department, Universiti Putra Malaysia, South City Academia A11-11, Selangor, Malaysia, Tel: +60-3-8946 6000; E-mail: maherrsgis@gmail.com Received August 30, 2014; Accepted September 19, 2014; Published September 26, 2014 Citation: Hasan G, Al Kubaisy MA, Nahhas FH, Ali AA, Othman N, et al. (2014) Sulfur Dioxide (SO2) Monitoring Over Kirkuk City Using Remote Sensing Data. J Civil Environ Eng 4: 155. doi:10.4172/2165-784X.1000155 Copyright: © 2014 Hasan G, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
Air pollution is a major concern in developing countries. The main sources of air pollution in developed cities are: emissions from vehicles (60-70%), coal based thermal power plants (10-15%), industrial units (10-15%) and domestic (5-10%) [1]. It causes a number of health problems and it has been linked with illnesses and deaths from heart or lung diseases [2]. Due to the high cost and limited number of air pollutant stations, it is difficult to provide a good spatial distribution of air pollutant measurements over a city. Although satellite measurements are a total column, not always correlated with surface air quality, satellite observations can help us in air pollution mapping with a high spatial distribution.
The objective of this study is to monitor Sulfur Dioxide pollutant over Kirkuk city, Iraq using free remote sensing data and in-situ measurements. The research uses Landsat 8 (TIRS) data to analyze the derived surface temperature and in-situ measurements of (17) ground stations around Kirkuk city. The relationship between satellite data and ground-based measurements will be investigated by correlation and regression analysis in GIS environment.

Related Works
The monitoring of aerosol concentrations becomes a high environmental priority particularly in urban areas [3]. Several local and international studies have shown the possible relationships between satellite data and air pollution.
A study by [4] investigated the applicability of measuring PM 2.5 concentrations using satellite data in a semi-empirical approach. The study compared daily PM 2.5 sampled on the ground surface over six stations and satellite data (MODIS). It showed good agreement (R 2 =0.68) [5,6]. Therefore, the research confirmed the potentiality of using satellite remote sensing observations for AQ monitoring.
Rohayu Haron Narashid [7] used images from Landsat (ETM+) and (8) continuous air quality monitoring stations to determine relationships between satellite thermal band data and air pollutant concentrations measured on the ground surface. The research resulted that the generated models by regression analysis can be used by environment staffs to monitor air quality in urban area.
Another research by Qihao Weng [8] investigated the relationship of air pollution pattern with urban thermal landscape using a GIS approach. The relationship between the spatial pattern of air pollution and land surface temperature were sought through GIS and correlation analysis.
The study resulted that there is a good relationship between satellite derived surface temperature and air pollutant concentrations. It has been concluded that thermal imagery could play a significant role in monitoring air quality.
Locally, a research by Al-Abdraba addressed monitoring and assessing of air pollution at several sites in North Oil Company in Kirkuk city. The study also tried to construct a mathematical equation for concentrations that discarded from sources close to ground surface. SO 2 concentration exceeded the international standards limitation in many studied sites. While Ali, studied Kirkuk Oil Refinery to assess the air, soil and plants contamination regarding the Kirkuk refinery system. The average concentration of SO 2 is recorded at 0.075 (ppm) in October, slightly higher than the allowable Iraqi National standards and (WHO), rather than lower concentration (0.006 ppm) is observed in March of that year. Furthermore, Salah Saleh [1] investigated the relationship between Land Surface Temperature (LST) derived from Landsat TM6 (thermal infrared) satellite data and ground-based air quality parameters measurements over Baghdad city, Iraq. A linear regression analysis was performed using least square method to model air quality of the city. The results showed significant agreements between ground-based measurements and Landsat satellite data. Also, the study indicated that satellite imagery is capable for air pollution mapping.

Sulfur Dioxide
Sulfur Dioxide (SO 2 ) is a significant indicator for volcanic movements [9]. Is produced mainly by combustion of fossil fuels or smelting of minerals, and is an intermediate product from organic sources in the ocean.
Global monitoring of sulfur dioxide gives information related to air quality, climate change, volcanology and volcanic hazards. Sulfate in air pollution is well known for production of acid rain and environmental degradation, as well as to reduce visibility and health effects.
It causes acidic pollution of lakes and streams and forms an aerosol that is important in climate change [10]. Other studies indicate that SO 2 causes nerve stimulation in the lining of the nose and throat. This can cause irritation, coughing and a feeling of chest tightness, which may cause the airways to narrow. People suffering from asthma are considered to be particularly sensitive to SO 2 concentrations.

Study Area
The study area of this research is Kirkuk city located in the north part of Iraq and it is a vital place as shown in Figure 1. Generally, this area consists of industrial, residential, oil fields places. The area of the region is approximately equal to (93.16 km 2 ) and the population of the city center currently is estimated to (784444) (Statistical Center of Kirkuk, 2011). Oil is the main source of its economic activities, it has the largest oil field in Iraq, natural gas and sulfur sources.
Geographically, it is located between the two great circles of (35 31 20 N) and (35 20 20 S), and two longitudinal circles of (44 26 10 E) and (44 16 30 W). It is 236 kilometers north of the capital Baghdad. Furthermore, the climate of Kirkuk city is generally continental, with hot and dry summers, and cool rainy winters.
The wind directions play an important role in the distribution of pollutants in air. The wind of the study area prevail north-western (Iraqi Meteorological Organization, 2011) and follows the climate of the Mediterranean Basin.

Data Sources
Ground-based measurements SO 2 pollutant concentrations were collected at (17) measuring stations around Kirkuk city in 9 th of January, 2014. The used device was NOVA instrument which designed to monitor different types of ambient air pollutants such as CO, CO 2 , NH 3 , NO 2 , SO 2 and H 2 S as shown in Figure 2.   Table 1 below shows the average measurements of SO 2 pollutants concentration measurements of 9 th of January, 2014 were higher than Iraqi and WHO standards air pollutants. So, air of Kirkuk city is polluted due to the traffic, personal and private electrical generators, industrial activities and poor vegetation cover in the city.

Satellite data
Two bands of thermal infrared region (10.4-12.5 µm) of Landsat8 image acquired on 9 th of January, 2014 for Kirkuk city was used for land surface temperature extraction. The Digital Number (DN) value for the thermal infrared of the corresponding ground base measurements are converted into radiance and apparent surface temperatures using NASA model [11][12][13].
Methodology DN conversion to surface temperature: In order to convert the raw data of Landsat thermal bands DN into surface temperature brightness [14], equation (1) Table 2 shows the used parameters.
Where: K1 and K2 are parameters of band-specific thermal conversion constant.
After we have got the TOAr values, then brightness temperature is calculated using equation (2) below [15].
After calculating temperature brightness in (Kelvin) unit, it is converted to Celsius and mapped in ArcGIS software as shown in Figure 3 [16].
Correlation analysis: Before investigating correlation analysis for ground-based and satellite data, the descriptive statistics are analysed for further understanding the nature of the collected data. Table 3 below shows important statistical parameters for two types of used data.
Relationship between Land Surface Temperatures (LST) [17] extracted from Landsat 8-TIRS satellite data with ground-based measurements of SO 2 concentrations around Kirkuk city were analyzed by correlation analysis.    between SO 2 pollutant concentrations and surface temperatures derived from satellite thermal bands. It also shows that the correlation between SO 2 ground-based concentrations and band 10 is stronger than band 11. The resulted correlation is accepted in terms of accuracy for monitoring purposes over large spatial distribution.

Regression analysis:
A non-linear regression analysis was performed to establish relationship between surface temperatures from Landsat data as independent variables against the in-situ SO 2 measurements concentrations data.
A third-order polynomial expression between SO 2 pollutant parameter and surface temperatures extracted from thermal bands of Landsat data were determined using regression analysis in Excel Software. Figures 4 and 5 shows the scatter plot of in-situ and satellite data and the fitting line.

Results
The regression analysis resulted an aerosol reflectance model for estimating SO 2 concentrations from raw DNs of thermal bands of Landsat 8 satellite data using the equations below: SO 2 (ppm) = 0.0117T 3 -0.3282T 2 + 2.837T -6.4733…. For band 10 SO 2 (ppm) = 0.0157T 3 -0.362T2 + 2.524 2 T -4.219….For band 11 Then, the two expressions above are used to calculate the SO 2 derived concentrations and it results as Table 5 below shows. The last    two columns are SO 2 concentrations in (ppm) unit that derived from the model which generated by regression analysis.
After that, the information that we have got from the generated model are used to map SO 2 pollutant dispersion over Kirkuk city using GIS tools. The below maps ( Figure 6) are produced using ArcGIS 10.2 software.
First of all, the SO 2 ground-based measurements are mapped to be compared with the derived SO 2 concentrations maps later and to assess the model applicability and accuracy for different purposes.
The highest concentration of SO 2 is found in the north of Kirkuk city and this is expected due to the oil fields and industries in the north part of the study area.
As we see from the Figure 7, the results of generated models using satellite thermal and in-situ data are reliable for large spatial distribution area. It can be seen that the highest concentration of SO 2 pollutant is emphasized in the West-North of the city in three maps. In addition, the resulted models will provide the Directory of Environment Kirkuk with low-cost and quickly produced maps for spatial decision making.

Air pollution Patterns & Landscape Temperature
Urban areas are associated with sources of a variety of air pollutants. In addition, cities are major contributors to global air pollution related to ozone depletion and carbon dioxide (CO 2 ) warming. Understanding air pollution patterns and thermal landscape patterns are important to better manage the cities [18].
Air pollution patterns in cities are mainly relate to the distribution of different land use and land cover categories. Land use of Kirkuk city are consists mainly of industrial, built-up, oil refinery and fields, bare areas with poor vegetated zones. It can be seen from Figure 7 that the high-concentrated of SO 2 is in the north and middle parts of the city which mostly are oil fields and built-up areas.
Investigating thermal landscape patterns that derived from satellite data of the study area gave us an imagination on how air pollutants are distributed in the city. Spatial pattern of air SO 2 pollutant is positively correlated with urban built-up density, oil refineries and with satellite derived land surface temperature in the study area.

Conclusion
This research showed good agreements in the relationship between Landsat-8 (TIRS) [19,20] data and in-situ SO 2 pollutant measurements in Kirkuk city. The correlation and regression analysis performed using "Analysis it" plug-in in Excel software and it showed the ability to analysis multi-variant data and also in plotting scatter diagrams, PCA analysis and curve fitting as well. While GIS technique is used in the study to map the collected and derived data. The IDW interpolation used in GIS environment to produce maps for SO 2 concentrations dispersion over Kirkuk city. The technique indicates to reliable results over the large spatial distribution when it compared with other researcher results.
The research resulted two models for SO 2 concentration over Kirkuk city, one for band 10 of Landsat TIRS data and the second is for band 11. However, the band 10 model is more accurate than band 11. Also the study generated three maps for SO 2 pollutant concentration over the study area using ground-based, values of the model of band 10 and 11. It showed a high correlation between three maps and by this we could conclude that remote sensing is a powerful tool for air quality mapping [21,22].
The results of this research indicate that air pollution can be mapped using satellite data to provide a larger area of coverage. Further research can be done to enhance the models using numerical and statistical analysis.