Spatio-Temporal Analyses of Atmospheric Sulfur Dioxide Column Densities Over Pakistan by Using SCIAMACHY Data

Spatio-Temporal Analyses of Atmospheric Sulfur Dioxide Column Densities over Pakistan by Using SCIAMACHY Data

Palwasha Khattak1, Muhammad Fahim Khokhar1,2*, Naila Yasmin1,3

1 Institute of environmental sciences and engineering, National University of Sciences and Technology Islamabad, Pakistan 2 Visiting scientist, Satellite Group Max-Planck Institute for Chemistry Mainz, Germany 3 R2V GIS consultancy Rawalpindi, Pakistan

ABSTRACT

This paper presents the results of atmospheric Sulfur dioxide (SO2) column densities obtained from satellite observation over Pakistan during the time period of 2004–2012. The level-2 data product of the Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY) instrument on-board ENVISAT-1 is used. SO2 column densities are retrieved by applying differential optical absorption spectroscopy (DOAS) technique. The spatio-temporal distribution of SO2 column densities along with the seasonal variation over main cities and regions of Pakistan are discussed. The Nabro volcano eruption in the year 2011 caused high SO2 levels over East Africa, the Middle East and South and Southeast Asia. Satellite images exhibited several episodes of volcanic SO2 column densities transported to Pakistan. The temporal trend in SO2 column densities was calculated, and the significance of the data set was tested with statistical analysis. An overall increase of about 70% (8.7% per year) in SO2 column densities over Pakistan during the time period of 2004–2012 has been calculated. This study presents the first spatial and temporal analyses of SO2 column densities over Pakistan by using satellite observation.

Keywords: Satellite data; DOAS; Column densities; SO2; Pakistan; ArcGIS.

INTRODUCTION around the world (Ferrari and Salisbury, 1999; Finlayson- Pitts and Pitts, 2000; Platt and Stutz, 2008). Atmospheric Sulfur dioxide (SO2) results from both SO2 life time varies from few days to several weeks (e.g., natural and anthropogenic sources (Krueger et al., 1995; Finlayson-Pitts and Pitts, 2000; Platt and Stutz, 2008). In the Krotkov et al., 1997; Khokhar et al., 2006; Lee et al., troposphere due to its reaction with hydroxyl (OH) mainly 2008) including the oxidation processes occurring in the and other oxidizing agents, it has a limited lifetime of just soil and over oceans, volcanic eruptions, biomass burning a few days (Finlayson-Pitts and Pitts, 2000; Platt and metal smelters and combustion of fossil fuels (Eisinger and Stutz, 2004). However, in the lower stratosphere its lifetime Burrows, 1998; Finlayson-Pitts and Pitts, 2000). Sulfur (S) increases from several weeks to year (Eisinger and Burrows, is found mostly in oil and coal and when burnt, sulfur 1998; Platt and Stutz, 2008). Lee et al. (2009) calculated combines with oxygen (O2) and SO2 is produced (Eisinger the global SO2 life time during different seasons. They and Burrows, 1998; Finlayson-Pitts and Pitts, 2000; Mirza performed atmospheric chemistry transport model - GEOS- et al., 2013). Chem (Bey et al., 2001) simulations by taking into account SO2 is very reactive and rapidly converted into other gas-phase oxidation, dry deposition, and aqueous-phase compounds such as sulfuric acid (H2SO4), sulfurous acid oxidation (in clouds) of atmospheric SO2. Pakistan is –2 (H2SO3) and sulfate particles (SO4 ). These compounds situated between 24°–37°N and 62°–78°E and SO2 lifetime are very harmful to the environment as well as to the over these latitude ranges from 17 h in summer to 40 h in human health (Finlayson-Pitts and Pitts, 2000; Ejaz et al., winter. The longer SO2 lifetime in winter reflects reduced 2009a). SO2 is also one of the key precursors of acid rain; oxidation by both aqueous (i.e., H2O2) and gas-phase (i.e., hazardous for forests and many fresh water ecosystems OH) processes as compare to summer time with relatively larger OH production and monsoon season. Traditionally, monitoring of atmospheric gases including SO2 has been carried out via air quality monitoring stations * Corresponding author. located mainly in the cities. These stations along with Tel.: +92-51-90854308 tropospheric O3 also provide SO2 measurements and other E-mail address: [email protected] criteria pollutants (e.g., Zerefos et al., 1986; Fioletov et al.,

1884 Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014

1998). Such monitoring stations are located sparsely and The SO2 slant column densities are retrieved in the monitoring on a global scale through these stations is not spectral window of 315–326 nm by using DOAS technique possible. Since last few decades the availability of satellite (Platt and Perner, 1979; Khokhar et al., 2005; Richter et al., instruments, the atmospheric SO2 monitoring on global 2006; Platt and Stutz, 2008). After the radiometric correction scales has become possible (Eisinger and Burrows, 1998; and application of m-factor to levle1 SCIAMACHY data, Khokhar et al., 2005; 2006; Carn et al., 2008). Space borne the retrieval of SO2 columns involve three main steps: (1) instruments such as Global Ozone Monitoring Experiment determining total SO2 slant column densities (SCD) by (GOME - since 1996, Khokhar et al., 2005; Thomas et al., spectral retrieval (Khokhar et al., 2005; Richter et al., 2005), Ozone Monitoring Instrument (OMI - since 2004; 2006; Krotkov et al., 2006, 2008; Lee et al., 2008), (2) by Krotkov et al., 2006), the Global Ozone Monitoring eliminating the latitude-dependent offset by normalizing Experiment-2 (GOME - 2 since 2008) in addition to Scanning SO2 SCD over pacific ocean region with relatively Imaging Absorption Spectrometer for Atmospheric no/negligible atmospheric SO2 (for details see, Khokhar et Chartography (SCIAMACHY - since 2002, Bovensmann al., 2005, 2006; Lee et al., 2009) and (3) by applying an et al., 1999; Lee et al., 2008) have opened new corridors appropriate air mass factor (AMF) to convert slant columns by providing continuous monitoring of various trace and into vertical columns (e.g., Khokhar et al., 2005; Thomas greenhouse gases globally. et al., 2005; Khokhar et al., 2008; Lee et al., 2009). Pakistan has no continuous air quality monitoring facility. The retrieval of the SO2 slant column- integrated along Pakistan Environmental Protection Agency (Pak-EPA) and the light path, is affected largely by stronger ozone (O3) Pakistan Council of Scientific and Industrial Research absorption, therefore, both SO2 and O3 should be taken into (PCSIR) have the facility of mobile laboratories. According account in the retrieval process to avoid/minimize spectral to report from economic survey of Pakistan 2012–2013, all interference. This "interference" when concentrations of of the air quality monitoring stations are non-functional SO2 are low, may produce negative SO2 slant column values, since year 2010 (ESoP, 2013). Thus, satellite instruments with an error of the same magnitude. This then represents the play a pivotal role in bridging such gaps in countries like SO2 background level, i.e., the apparent SO2 absorption in Pakistan. the absence of emissions of SO2. This is further dependent This study has first time emphasized on the long-term on solar zenith angle (SZA) and more prominent for higher record of SO2 present in the atmosphere of Pakistan by SZA due to increased slant light path and larger horizontal using satellite observations. Focus is mainly on spatial and ozone concentrations. For low SZA, i.e., at low and mid temporal variations in atmospheric SO2 column densities latitudes, the interference between the SO2 and O3 absorption over Pakistan during the time period of 2004–2012. For results in less negative SCD (low back ground level) and this purpose differential optical absorption spectroscopy errors. Therefore, emissions of SO2 (by pollution or volcano (DOAS) technique (Platt and Perner, 1979; van Roozendael, eruptions) will be detected against this background. This 1999; Khokhar et al., 2005) was applied to retrieve SO2 effect is corrected by applying background correction column densities from nadir observations of SCIAMACHY method used by various studies (e.g., Khokhar et al., 2005; (Bovensmann et al., 1999; Khokhar et al., 2006; Lee et al., Lee et al., 2009; Theys et al., 2013). 2008; Rix et al., 2012) instrument. Retrieval of data used in this study involved the absorption cross-sections of SO2 and ozone at two METHODOLOGY temperatures (223°K and 243°K) along with two Ring spectra to correct for the Raman scattering (one to account SCIAMACHY (Bovensmann et al., 1999) is a multi- for the general Ring effect of the Fraunhofer and one for channel grating spectrometer on board Environmental Satellite the SZA dependent contribution of ozone) were fitted in (ENVISAT-1) by European Space Agency (ESA) and was the wavelength range 315–326 nm. In addition to correction launched on March 1, 2002. It has ground pixel size of 60 for polarization effects, a wavelength calibration using a  30 km2 in standard scan mode. Several studies have been high-resolution solar spectrum is performed. Further details conducted on the degradation of SCIAMACHY instrument about retrieval algorithms can be obtained from various and its impact on level 1 and derived level 2 products (e.g., studies (e.g., Afe et al., 2004; Khokhar, 2006; Richter et Noël et al., 2007; Bramstedt et al., 2009; Tilstra et al., al., 2006; Lee et al., 2008, 2009, 2011). 2012). An official correction for instrument degradation, An air mass factor (AMF) is mandatory for the called m-factor correction, was available to the users of conversion of the retrieved slant columns into vertical SCIAMACHY data since 2007 (Bramstedt et al., 2009). By column densities (VCD). The SO2 AMF depends strongly implementation of these correction factors, the trend/artifact on surface albedo, clouds, aerosols optical properties, the caused by instrument degradation in SCIAMACHY data was vertical distribution of SO2, and the ozone column (Khokhar successfully removed (Tilstra et al., 2012). SCIAMACHY et al., 2005; Thomas et al., 2005; Khokhar et al., 2008; level-2 data for year 2004–2012 from the Tropospheric Krotkov et al., 2008; Lee et al., 2009). For data used in this Emission Monitoring Internet Service (TEMIS) European study, if no cloud cover information is available for a given Research Project (http://temis.nl/) was used to get a ground pixel, the AMF and VCD of that ground pixel were quantitative analysis of SO2 column densities over Pakistan. not computed and the data files will get the "no data" value/flag. Only the AMF for the clear-sky part were given SO2 Retrieval Analysis a value that can be computed always.

Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014 1885

In case of volcanic eruption, SO2 AMF calculation is from monthly mean values and negative SO2 VCDs caused strongly dependent on the SO2 plume height. In general, by spectral interference, instrument degradation and/or there is no information available on the altitude of the SO2 imperfect removal of ring effect were not considered. plume, so an assumption must be made. Therefore, VCDs As SCIAMACHY takes six days for global coverage were computed for three different assumed plume heights 2 due to its dual mode of limb and Nadir viewing geometry, km for Low AMF, 6 km for Middle AMF and 14 km for therefore, it limits the quantification of background SO2 High AMF, with an assumed plume thickness of 1 km (For level in addition to violent dust storms and monsoon further details see documents available at TEMIS website season (e.g., Chaudhary et al., 2009; Ahmed et al., 2012). http://sacs.aeronomie.be/info/vcdamf.php). HYSPLIT trajectory However, anthropogenic SO2 analyses performed for various model was used to have first guess of SO2 cloud height and cities of Pakistan indicate background SO2 levels of 0.135 then VCD were computed by using the respective AMF ± 0.02 Dobson units (3.6 ± 0.5 × 1015 molecules/cm2) (middle or High) during days of volcanic activity. Finally, during the time period of 2004–2011. monthly mean average VCD were extracted over Pakistan. SO2 columns are mostly higher over major cities of The data files contain the value of the SCD and VCD in Northern Areas, Punjab, Sindh, Khyber Pakhtunkhwa and in Dobson Units and an estimate for the retrieval error on the the western part of Baluchistan provinces. It is comparatively SCD for each individual pixel. Possible errors in the AMF low in interior Sindh and Southern Baluchistan because of calculation are therefore not taken into account. Several less populated and industrial activities in these regions. It’s studies have reported AMF error of 10–15% by Thomas et worth mentioning that few small places like Nok Kundi al. (2005), about 10% error by Khokhar et al. (2008) and (western Baluchistan) Dadu (Sindh) and Okara (Cenral 35–70% in polluted regions (e.g., East Asia and the eastern Punjab) exhibited relatively large SO2 columns. Nok Kundi US) and less than 10% over clear ocean regions by Lee et is situated on N40-highway of Pakistan which connects al. (2009). The total error in the SO2 retrieval from satellite Quetta, Pakistan to the border cities of Iran. Nok Kundi is data is below 50% (Khokhar et al., 2005; Thomas et al., 221 km from Iranian city of Zahedan (the capital of Sistan 2005; Khokhar, 2006; Lee et al., 2011; Theys et al., 2013) and Baluchestan Province, Iran with population of 552, 706). SCIAMACHY observations with back ground correction, It has a hot desert climate with extremely hot summers, mild cloud fraction less than 10% , SZA less than 88 degree and winters and gets just 35 millimeters (1.4 in) of rainfall over fit error (Chi sq < 5 × 10–5) are used only in this study. the year. There is no larger industry in this region and is Further details about the data products, error analysis, also less populated. The observed SO2 over Nok Kundi cloud correction and AMF calculations are available through region is mainly due to traffic on N40-Highway and trans- a document titled “Sulphur Dioxide Monitoring within boundary air pollution either anthropogenic SO2 coming TEMIS” available at webpage (http://sacs.aeronomie.be/ from Zahedan, Iran or due to regional/global volcanic info/docs.php, last accessed on February 21, 2014). eruptions. Monthly mean SO2 data, gridded to a latitude-longitude This is further confirmed by mean wind fields (2004–2011) grid of 0.25 by 0.25 degrees is used. For each grid cell the at 1000 hPa levels over Pakistan and neighboring countries average of all measurement crossing that grid cell (using presented in Fig. 2, prepared by using the NCEP Reanalysis forward scans only) is computed from all orbit files for the Derived data provided by the NOAA/OAR/ESRL PSD, given month. Monthly mean spatial maps of SO2 were Boulder, Colorado, USA, from their Web site at generated over Pakistan and time series analysis was carried http://www.esrl.noaa.gov/psd/. Wind vectors are indicating out. This helped in identification of regions with major SO2 the mean flow of air masses over Pakistan and neighboring column amounts over Pakistan during the selected time countries at 1000 hPa during the study period. In general, period. the Nok Kundi is impacted by the air masses coming from Zahedan, Iran. Furthermore, wind fields also indicated that RESULTS AND DISCUSSION many cities (Sialkot, Lahore, Okara, Faisalabad, Sargodha, Multan and Bahawalpur) of Pakistan are also influenced by Spatial Distribution the air masses coming from Indian region which is densely Pakistan’s climate is mainly arid to semi-arid with populated and famous for extensive industrial activities exception of slopes of Hindukush and Karakorum mountain (western Gangetic plain). Therefore, the enhanced level of ranges, where rain fall ranges from 700 to 2000 millimetre SO2 over various cities of Pakistan is also contributed by (mm) per year (Chaudhary et al., 2009; Ahmed et al., 2012). trans-boundary anthropogenic SO2 pollution coming from Pakistan receives summer rainfall due to monsoon system India, Iran and other neighboring countries. while winter precipitation is due to western systems (e.g., In case of Dadu, Sindh; it is middle level urbanized Chaudhary et al., 2009; Wang et al., 2011 and references district. However, the observed level of SO2 are mainly due therein). to oil and gas exploration/drilling activities at Zamzama and Spatial distribution of SO2 column densities in DU (1 Bhit Badra oil and gas fields in the region. Similarly, DU = 2.69 × 1016 molecules/cm2) over Pakistan is presented Okara (central Punjab) is not highly urbanized region but in “Fig. 1” during the time period of 2004–2011. Black still there exist several textile mills, thermal power circles are indicating the location of main cities of Pakistan. generation and cottage industries that may cause such high Grey lines in each map are representing the district level of SO2 over this region in addition to trans-boundary boundaries. Yearly mean over each grid point was calculated air pollution.

1886 Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014

Fig. 1. Spatial distribution of annual mean SO2 column densities in Dobson units (DU) over Pakistan during time period of 2004–2011. Temporal and spatial variation from 2004–2011 in SO2 columns densities over Pakistan can be identified. Circles are indicating the location of major cities. Annual maps are prepared by taking average of monthly mean SO2 VCDs over each grid point. SO2 VCDs less than zero (mainly caused by spectral interference etc. - see the text) were not included while calculating the annual mean.

Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014 1887

Fig. 2. Wind fields over Pakistan and neighbouring region averaged over 2004 to 2011. Wind fields are generated from NCEP Reanalysis Derived data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their Web site at http://www.esrl.noaa.gov/psd/. Arrows (wind vectors) are indicating the wind direction and magnitude of each arrow is subject to respective wind speed (larger arrows for higher wind speed and smaller arrows for calm winds).

In general, “Fig. 1” exhibited a temporal increase in SO2 the reasons for such high SO2 columns could be due to the column densities from year 2004 to 2011. Annual maps albedo effect of underlying snow during winter months (for clearly indicate that higher SO2 column densities are details see Khokhar et al., 2006), which enhances the observed over the areas of KPK, Punjab, Sindh and instrument sensitivity and also causes imperfect correction Baluchistan with larger population and traffic densities, of ring effect. agricultural practices, and industrial activities. There are Additionally, in these areas about 70–80% of household some exceptions; (1) enhanced SO2 levels are observed uses biomass for the purpose of cooking and heating over larger regions due to tans-boundary volcanic SO2 (NASSD, 2003). This consumption increases in winter for caused by Jabel at Tair, Dalaffilla, and Nabro volcanoes space heating and has slightly enhanced the SO2 column from African region during the years 2007, 2008 and 2011, densities. Moreover the non-availability of compressed respectively. (2) Over northern areas of Pakistan enhanced natural gas (CNG) and mainly consumption of conventional levels of SO2 are partially caused by the albedo effect fuel (diesel and other petroleum fuel) in motor vehicles caused by ground snow/ice during winter months (Khokhar might have resulted in increased SO2 emissions. Another et al., 2005, 2006). These discrepancies are further discussed reason is the trans-boundary SO2 pollution resulting from in the following sections of this study. global volcanic eruptions such as: Dalaffilla (13.79°N, 40.55°E), Ethiopia and Nabro (13.37°N, 41.70°E), Eritrea Atmospheric SO2 over Major Cities of Pakistan volcanic activity in East African region has caused high It is interesting to note that Northern Areas of Pakistan SO2 columns over Pakistan during November 2008 and (Skardu, Gligit and Chitral) exhibited larger SO2 levels June 2011, respectively. In order to avoid the influence of during the selected time period (2004–2012) in spite of that trans-boundary SO2 resulting from global and/or regional it is considered as less polluted region with less population volcanic activities two steps were taken as under: (1) for and industrial activities. However, spatial maps of Pakistan analyses of anthropogenic SO2 over major cities presented exhibited larger SO2 column densities especially during in Fig. 3 and for identification of seasonal cycle presented winter months over this region. in Fig. 4, SO2 data for the months affected with trans- This region of the country has cold winters (temperature boundary volcanic SO2 were not considered. (2) for the is well below zero, Chaudhary and Rasul, 2004), and snow conversion of anthropogenic SO2 SCD into VCD, AMF covered areas from December to April. Therefore, one of was calculated by considering SO2 vertical distribution

1888 Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014

Fig. 3. Annual variation of anthropogenic SO2 column densities measured over major cities of Pakistan during the time period of 2004–2011. Months affected by trans-boundary volcanic SO2 were not include in this time series.

Fig. 4. SO2 seasonal cycle observed over different regions of Pakistan (provinces-solid lines) and over Pakistan (dashed line) devised from the SCIAMACHY observations over the period of 2004–2011. Months affected by trans-boundary volcanic SO2 were not include in this time series.

within first 2 km from the earth surface. largely the SO2 slant columns retrieval and radiative Frequent high SO2 columns are observed over many cities transfer modelling for AMF calculation by inducing both of Pakistan such as Quetta, Dera Ismail Khan, Peshawar, shielding and albedo effects (Khokhar et al., 2008; Wagner Bahawalpur, Faisalabad, Lahore, Sargodha, Skardu, Chitral, et al., 2008) as mentioned in the previous sections. Karachi and Gilgit. Table 1 presents the database of annually “Fig. 3” shows a consistent behaviour of steadily averaged SO2 column densities over these cities during the increasing SO2 for all cities till the year 2010 (with one time period of 2004–2011. South Asian region (India, exception of Skardu with maximum SO2 column densities Bangladesh and Pakistan) is experiencing worst air quality during the year 2011). The highlighted values (bold) in due to rapid urbanization and industrialization. Decent Table 1 are the highest SO2 columns observed over the growth in their GDPs has resulted in more cars, factories, selected cities during the time period of 2004–2011. and more people to consume coal and other petroleum Anomalous behaviour of low SO2 columns over Karachi products. According to (Ramanathan et al., 2001), brown (the largest city with population of 23.5 million and haze is a severe problem over the tropics. It is mainly due industrial hub of Pakistan) was observed during the year to increased emissions of aerosols and precursor gases like 2004 to 2008 as compared to other cities. From year 2009 SO2 that have increased over South Asia by a factor of 3 to onward, Karachi exhibited highest SO2 column densities. 4 since 1970. Furthermore, these aerosols may affect According to reports from Pak-EPA, EURO II emission

Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014 1889

16 2 Table 1. Annually averaged SO2 columns in DU (1DU = 2.68 × 10 molecules/cm ) over major cities of Pakistan during 2004–2011. Cities 2004 2005 2006 2007 2008 2009 2010 2011 Bahawalpur 0.104 0.173 0.187 0.166 0.298 0.062 0.367 0.282 Chitral 0.129 0.146 0.177 0.186 0.185 0.238 0.426 0.151 D.I. Khan 0.149 0.147 0.16 0.203 0.298 0.193 0.371 0.261 Lahore 0.147 0.131 0.153 0.148 0.242 0.226 0.355 0.291 Faisalabad 0.161 0.171 0.18 0.228 0.241 0.233 0.399 0.151 Gilgit 0.186 0.191 0.285 0.203 0.212 0.244 0.434 0.405 Peshawar 0.111 0.149 0.116 0.195 0.171 0.215 0.479 0.145 Quetta 0.132 0.124 0.107 0.105 0.065 0.124 0.318 0.191 Sargodha 0.156 0.163 0.197 0.163 0.254 0.172 0.436 0.419 Skardu 0.16 0.239 0.295 0.205 0.168 0.23 0.393 0.452 Karachi 0.039 0.106 0.155 0.108 0.129 0.288 0.541 0.386

standards for motor vehicles were implemented in January exhibits the seasonal cycle in atmospheric SO2 over the four 2012 and hardly met by most of the vehicle in Pakistan. provinces, Northern Areas and all Pakistan. Climatological- This and increased emissions due to rapid urbanization, SO2 column densities for each month of the year were trans-boundary fraction of SO2 and excessive use of calculated by taking average over whole time period of furnace oil for thermal power generation could be the main 2004–2012 for respective month. As mentioned in the reasons of observed temporal increase of SO2 column previous section, Pakistan was affected by the various densities over main cities of Pakistan. regional volcanic activities such as Jebal at Tair in October “Fig. 3” exhibits the average annual behaviour of SO2 2007 and Nabro volcano eruption in June 2011 etc. and pollution and does not reflect the seasonal cycle over these was very well captured by SCIAMACHY observations as cities. An analysis about seasonal variation in SO2 column indicated by the very high SO2 column (see “Fig. 5”). In densities over major regions of Pakistan is presented in order to avoid such anomalous high SO2 column densities, next section. data for the months affected by observed regional volcanic events for all regions was excluded from this analysis. Seasonal Variation Over Pakistan SO2 seasonal cycle is largely affected by The hydroxyl radical (OH) is very important especially the monsoon season (high rainy season generally prevails in the tropospheric chemistry of sulfur species (Finlayson- from July to September, Jhajharia et al., 2009). During Pitts and Pitts, 2000). The oxidation process of almost all these months the SO2 amount is low compared to the post sulfur compounds (including reduced sulphur species) are monsoon months (September to October). This is due to initiated by OH (Finlayson-Pitts and Pitts, 2000; Hassan et the washout effect that lowers the SO2 amount in the al., 2013). In general low SO2 amounts are observed in atmosphere during rainy season by wet deposition and via summer (due to high OH productivity) compared to high its aqueous phase conversion into sulfate particles. In amounts in winter (less OH productivity). SO2 conversion general, “Fig. 4” exhibits similar seasonal cycles of SO2 into sulfate particles and sulfuric acid (H2SO4) involves its column densities over all provinces of Pakistan, although, reaction with O3, hydrogen peroxide (H2O2), OH and they differ in various aspects like having different traffic hydroperoxyl (HO2). Therefore, OH is the key parameter density, population density and industrial activities. Probably, which determines the life time and seasonal cycle in the the washout during the monsoon season and high OH SO2 concentrations in addition to some other minor and (Finlayson-Pitts and Pitts, 2000; Platt and Stutz, 2004) local phenomena (such as monsoon rains etc.). The SO2 concentrations during summer months are the dominant seasonal cycle may also be slightly dependent on both factors in determining the seasonality in the SO2 column anthropogenic (coal, petroleum products and biomass densities over Pakistan. burning) and natural (monsoon, snow) activities on regional basis. For instance during winter season, Northern Temporal Analysis Areas of Pakistan are covered with snow and people burn The temporal record of SO2 over Pakistan was analysed biomass fuel for heating and cooking purposes which by plotting the monthly averaged SO2 column from 2004 causes relatively high SO2 concentrations (NASSD, 2003). till 2012, as presented in “Fig. 5”. The mean SO2 columns The underlying snow causes the high surface albedo and were calculated by taking the mean of SO2 column enhances the instrument sensitivity (albedo effect, for densities over all cities of Pakistan. The bars show the details see Khokhar, 2006). While anthropogenic activities spatial variation (standard deviation) for each month over in other regions of Pakistan may not reflect such variation all cities. In general, high SO2 column densities are in SO2 emissions. observed over cities with high population/traffic density The SO2 distribution varied for each year throughout the and industrial activities (see Fig. 1). In June, 2011 unexpected time period of 2004–2011 with large SO2 column densities high SO2 columns were observed, caused by the Nabro during winter and lower during the rest of the year. “Fig. 4” volcanic eruption. It was very explosive and sulfur dioxide

1890 Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014

Fig. 5. Temporal record of SO2 column densities observed by SCIAMACHY instrument over Pakistan during the time period of 2004–2012. The arrows indicating the month influenced by trans-boundary SO2 pollution caused by respective volcanic eruptions. Dashed lines are the linear fits implemented to calculate the temporal increase in SO2 VCDs with (Blue) and without (red) trans-boundary volcanic eruptions. rich volcanic eruption (Fee et al., 2011; Carn et al., 2012). did not have overpass over the region on January 27, 2011 For instance Clarisse et al. (2012) reported total SO2 mass because of switching between limb and nadir viewing of 1.5 Tg retrieved from daily IASI observations and Carn modes (Bovensmann et al., 1999). Therefore, SO2 emitted et al. (2012) estimated 1 to 2 Tg SO2 based on OMI and from Tor Zawar volcanic eruption was not recorded by the Atmospheric Infrared Sounder (AIRS) data. It has caused SCIAMACHY observations. not only the stratospheric aerosol load (e.g., Bourassa et Pakistan is currently experiencing a severe energy/power al., 2012) but also large amounts of SO2 pollution was crisis and consequent 10–12 hours of electricity blackout in detected over surrounding regions. It has influenced addition to natural gas outages. Pakistan’s coal, oil and Pakistan’s atmosphere with maximum SO2 of 7 DU detected natural gas reserves are running out leaving the economy on June 17, 2011 from daily observations of SCIAMACHY to rely on imported fossil fuel. The situation has gradually instrument. Other peaks in the time series were also worsened overtime forcing industrialists to use poor investigated and it was found that they were also caused by quality fossil fuels (especially furnace oil etc.) to meet their various volcanic activities in neighbouring regions such as industrial needs of electricity. Consequently, it is deteriorating in October, 2007 (Jabal al-Tair, Yemen), November 2008 the environment of Pakistan. According to a study Qudrat- (Dalaffilla, Ethiopia) and January, 2010 (Tor Zawar, Ullah and Karakul (2007), has indicated a recent increasing Pakistan). The monthly averages affected by these volcanoes trend in oil- and gas-based power generation while a are indicated by the red data points in “Fig. 4”. Additionally, decreasing trend in the hydro-based power generation in some other peaks in SO2 columns (especially, Jan–May Pakistan. It also mentioned that as a result the environment 2010) of amplitude comparable to the SO2 peaks caused by is exposed to the additional 167 million tons of CO2 in trans-boundary volcanic eruptions can be identified. The addition to other air pollutants such as SO2 and NO2 etc. A reasons behind such anomalous high SO2 are not clear yet, temporal increase is observed in the thermal power generation however, it might be due to various factors such as observed data of Pakistan presented in the Table 2. For instance the seasonal cycle (high during winter months), instrumental thermal power generation was increased during year 2008 degradation (as discussed in the methodology and data to 2011 with maximum during the year 2010. Among section) and regional/global volcanic activities. According thermal power generation, the use of coal and furnace oil to reports from global volcanisms program (GVP, 2010), Tor was larger as compared to use of gas as fuel. Similar Zawar (30.48°N, 67.49°E) from Pakistan, Erta Ale volcano behavior of increasing SO2 columns was reflected by (13.60°N, 40.67°E) from Ethiopia and Eyjafjallajökull satellite observations over Pakistan with maximum during (63.63°N, 19.62°W) volcano from Iceland and Niyragongo 2010 followed by a small decrease in 2011. This is further (1.52°S, 29.25°E) from Congo have been active during this confirmed by statistics analysis presented in Table 2. The time of the year 2010. However, satellite instrument and correlation between annul mean SO2 columns densities Hybrid Single Particle Lagrangian Integrated Trajectory over Pakistan and power generation in Pakistan by using Model-4 (HYSPLIT4 - Draxler et al., 2013) analyses could different types of fuel was calculated. Resulting correlation 2 not track SO2 plume over Pakistan caused by aforementioned coefficients (r ) and slope are also given in Table 2. volcanic activities. Therefore, it is difficult to speculate Maximum r2 = 0.79 is observed for thermal power generation 2 about the origin of such enhanced SO2 over Pakistan and to (+ve slope) followed by use of gas (r = 0.77 with –ve 2 2 correlate it with trans-boundary SO2 without any scientific slope), furnace oil (r = 0.68 with +ve slope) and coal (r = evidence. In case of Tor Zawar volcano, SCIAMACHY 0.18 with –ve slope) used for power generation. It is worth

Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014 1891

Table 2. Share of coal, oil and gas consumption by power sector of Pakistan together with total Hydro and Thermal power generation.

Power Generation in Pakistan Power generation in Pakistan Annual SO2 by using fuelsa,b by sectorc Column Year Densities Oil Gas Coal Hydro Thermal Over Pakistan (GWh) (GWh) (000 metric tonne) (GWh) (GWh) (DU) 2008 39.18 33.7 162.2 28667 57602 0.172 2009 42.27 31.8 112.5 27763 56614 0.161 2010 46.07 28.7 125.5 28492 60746 0.335 2011 43.09 27.2 96.5 32259 58316 0.298 r2 0.68 0.77 0.18 0.24 0.79 1 slope 0.026 –0.026 –0.013 0.021 0.44 1 GWh: Giga watt hours 2 r & slope are calculated by different fuel type verses annual SO2 columns densities a Source: Ministry of Petroleum and Natural Resources Pakistan b Source: Hydrocarbon Development Institute of Pakistan c Source: Pakistan Electric Power Company (Pvt) Limited (PEPCO), National Transmission & Distribution Company Limited (NTDC)

to note that annual SO2 columns and gas used for power analysis presented in “Fig. 6” indicates the consumption of generation are anti-correlated. Such anti-correlation was compressed natural gas (CNG) in Pakistan along with the expected and has been stated in various studies (e.g., number of motor vehicles and SO2 column densities during Goyal, 2003; Chelani and Devotta, 2007). Therefore, the 2004–2011. According to the Pakistan Economic Survey observed temporal increase in SO2 columns during the time 2011–2012 (MoF, 2012), the number of motor vehicle on period of 2004–2011 is mainly contributed by trans-boundary road has increased from 4.5 million in year 2000 to 11 SO2 pollution caused by both regional volcanic eruptions million in the year 2011. These statistics indicate the huge (Khattak et al., 2014) and anthropogenic activities (“Fig. 2”), growth in the transport sector of Pakistan, which is directly and use of furnace oil and coal for thermal power generation related to air quality degradation (Ejaz et al., 2009a) and in Pakistan to meet the industrial needs (especially, onward health effects especially process of wound healing (Ejaz et from year 2008) and to overcome the power outages. al., 2009b). The growth in motor vehicles on roads of In order to estimate the temporal increase in SO2 columns Pakistan and annual SO2 column densities averaged over over Pakistan, regression analysis was performed by applying main cities of Pakistan did not reflect the similar trend. As linear fit to both data sets. The red dashed line represents the there is gradual increase in SO2 emissions until year 2010 trend in SO2 columns over Pakistan including the volcanic followed by a decrease in 2011 (contrary to trends in motor eruptions, whereas the blue dashed line represents the trend vehicles). Similar discrepancy was exhibited by correlation excluding the volcanic events. Analysis showed 120% between total number of motor vehicles and observed SO2 2 increase in SO2 columns with volcanic eruptions and 70% columns with r = 0.50 presented in “Fig. 7” (green triangles). 2 increase at the rate of 8.7% per year in SO2 columns The correlation factor is not very high and r indicates that without volcanic eruptions over the time period of 2004 to the increase in the number of vehicles cannot fully explain 2012. As Pakistan is a developing country, the cause of this the observed trend in the SO2 columns. It was further increase in SO2 columns can be associated with various investigated by taking into account the type of fuel consumed factors include increase in motor vehicles, new industries by the transport sector of Pakistan during the selected time (Ramanathan et al., 2001; Ejaz et al., 2009a; MoF, 2012), period. Table 3 presents the statistics of share of fuel used use of low quality and sulfur containing fuel, burning of by the transport sector of Pakistan, growth in compressed low grade coal and biomass for domestic and industrial natural gas (CNG) as fuel and number of vehicles converted purposes (especially thermal power generation), and open from conventional fuel to CNG fuel and SO2 column burning of solid waste. Additionally, the emission control densities during 2004–2011. According to the International technologies, strategies, environmental legislation and air Association of Natural Gas Vehicles (IANGV) there are quality standards are hardly adopted and met in the approximately 4 million Natural Gas Vehicles (NGVs) in country. For example, even the ‘Euro II’ Emission Standards use worldwide, of which 1.6 million are in Argentina and for motor vehicles are not met. 1.5 million in Pakistan. Additionally, due to its affordability (CNG is cheaper than gasoline and diesel in Pakistan) Source Apportionment during the last few years, a large number of motor vehicles Four major air pollution sources that should be immediately has been shifted from conventional fuel (petroleum and addressed by Pakistan are: vehicular emissions, industrial diesel) to CNG. Therefore, the use of CNG as fuel has emissions, emissions from solid waste burning and Natural resulted in fewer emissions of SO2 and other pollutant dust or trans-boundary pollution (PCAP, 2008). The from motor vehicles (Goyal, 2003; Chelani and Devotta,

1892 Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014

Fig. 6. Compressed natural gas (CNG) consumption in billion cubic feet (y-axis, source: Ministry of Transport and Communication (NTRC), yearly mean SO2 column densities in DU (on secondary y-axis) over Pakistan and total no of motor vehicles (y-axis source: Ministry of Transport and Communication (NTRC) in Pakistan during the time period of 2004–2011).

Fig. 7. Correlation plot among SO2 column densities, total number of motor vehicles in Pakistan during the year 2004- 2011 (green triangles), number of vehicles converted to CNG from conventional fuel (blue diamond) and amount of CNG fuel consumed by transport sector in Pakistan. Dashed lines are showing the linear fits for respective datasets.

2007). The change in fuel type is probably the main cause by transport sector was also decreased during year 2011. behind the poor correlation between temporal behaviour of SO2 columns and total number of motor vehicles in Pakistan. Regression Analysis It is further supported by the correlations calculated among Partial Autocorrelation Function (PACF) was applied on SO2 columns and CNG consumption by transport sector in the SO2 data for the time period 2004–2012, it suggests an Pakistan (r2 = 0.7, dark brown circles in “Fig. 7”) and autoregressive nature of the time series. The descriptive number of motor vehicles converted to CNG fuel from analysis shows that data is right skewed with high peak, conventional fuel (r2 = 0.7, blue diamonds in “Fig. 7”). indicating non normality (see Table 4A - presents the KPSS This explains the fact that SO2 emissions decreased in year test (Kwiatkowski et al., 1992) statistics). The Autoregressive 2011. Although, the number of motor vehicles in Pakistan model is employed to capture the time series structure. As were still increasing in year 2011 but on the other side the the series is stationary; ordinary least square method provide number of vehicles converted to CNG from conventional consistent estimator. A trend variable ‘T’ is included in fuel was maximum during year 2011. Similar trend can be regressors to capture the temporal trend in the series. The seen in Table 3 that share of petroleum products consumed intercept in the model represent monthly mean column

Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014 1893

Table 3. Share of Oil and CNG consumption by the transport sector of Pakistan. Share of Oil/petroleum Share of CNG Yearly mean SO No of vehicle Total No 2 products Consumption Consumption by Column Densities Year converted to CNG of vehicle by Transport sector Transport sector Over Pakistan (million)c (million)d (%)a (million cubic feet)b (DU) 2004 63.07 1.5 0.45 2.9635 0.114 2005 61.50 2.1 0.66 3.1464 0.138 2006 55.77 3.2 1.05 3.8688 0.136 2007 47.38 4.6 1.30 4.5429 0.145 2008 51.90 5.6 1.40 5.1263 0.172 2009 49.34 7.0 1.70 5.4564 0.161 2010 46.32 7.8 2.00 5.5012 0.335 2011 47.08 9.3 2.74 5.5586 0.298 a Source: Hydrocarbon Development Institute of Pakistan (CNG and oil consumption by Transport sector) b Source: Hydrocarbon Development Institute of Pakistan (CNG and oil consumption by Transport sector) c Source: OGRA, Ministry of Petroleum & Natural Resources (CNG fitted vehicles) d Source: National Transport Research Centre (total number of vehicles)

Table 4A. KPSS test and descriptive statistics of SO2 data. retrieved from SCIAMACHY observations were used to Parameter Value study the spatial and temporal variation over Pakistan for Mean 0.1797 the time period of 2004–2012. The main findings are: SD 0.1321 SO2 columns vary throughout the country. The annual Skewness 2.4903 average distribution of SO2 showed that SO2 is typically Excess Kurtosis 9.1523 higher in Northern part of the country, mostly because of KPSS with trend 0.0714 trans-boundary volcanic SO2, albedo effect, increased use of bio-fuel (for cooking and space heating) and usage of

low quality fuel in motor vehicles. It is also high in various Asymptotic critical values of KPSS major cities of Punjab, Sindh, Khyber Pakhtunkhwa and 1% 5% 10% Baluchistan due to anthropogenic activities. It was 0.216 0.146 0.119 comparatively low in less populated areas of interior Sindh and Southern Baluchistan. Table 4B. Regression Analysis Parameters. Meteorological analysis indicated that many cities of the Punjab and Baluchistan Provinces are influenced by polluted Standard Coefficient t-value t-prob air masses coming from India and Iran, respectively. Error Temporal analysis showed that SO has increased since SO –1 0.3387 0.1007 3.36 0.0011 2 2 2004 and this increase is statistically significant. Constant 0.06852 0.0267 2.57 0.0117 Anthropogenic SO increased 70% during the time period Trend 0.001075 0.0004876 2.21 0.0299 2 of 2004 to 2011 at rate of 8.7 percent per year (120% by including the trans-boundary SO2 pollution due to major densities. The estimated model is given below in Eq. (1). volcanic eruptions). This increase in SO2 columns over Pakistan is mainly SO2t = 0.0685 + 0.001T + 0.3384SO2t-1 (1) due to rapid urbanization, use of poor quality petroleum 2 products in vehicles and for thermal power generation (r = where SO2t and SO2t-1 are mean SO2 column densities for 0.8), use of bio fuel (char coal) for cooking and space the current and previous month, respectively. heating, open solid waste burnings and no implementation Regression analysis presented in Table 4B indicates that of strict emission standards for industry and motor vehicles all of three regressors are significant at 5% significant level. (EURO II emission standards are not implemented yet in The current concentration depends upon the last month’s its true spirit). Although, there is no strong correlation concentration about 34% with 0.1% per mensem increasing observed between the increase in number of motor vehicles trend in SO2 column densities. Especially, t-values greater in Pakistan and SO2 column densities but still it may be than 2 (see Table 4B) confirm that calculated trend in SO2 partially attributed to extensive increase in vehicular emission column densities is statistically significant. and related activities in Pakistan during the time period of 2004–2011. The observed poor correlation is mainly due to CONCLUSION extensive use of CNG as a fuel rather than a conventional fuel during last few years. This study has emphasised on the amount of SO2 present The Seasonal variation showed high SO2 in winter and in the atmosphere of Pakistan by using satellite data during comparatively low values in summer months which is the time period of 2004–2012. SO2 column densities mainly driven by OH concentration and monsoon season

1894 Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014

(local/regional effect). Carn, S.A., Yang, K., Wang, J. and Park, S.P.F. (2012). The study also revealed that some major volcanic Satellite Measurements and Modeling of the 2011 Nabro eruptions in the African region have affected SO2 column (Eritrea) Volcanic Clouds. American Geophysical Union densities over Pakistan through trans-boundary SO2 pollution Chapman Conference Selfoss, Iceland. Abstract. RS-5. (Khattak et al., 2014). Among those eruptions, the Nabro Iceland 10–15 June 2012 Available at http://www.agu volcano event during June 2011 was explosive; sulfur .org/meetings/chapman/2012/bcall/pdf/Final_Program.p dioxide enriched and has major effects on atmospheric df, last accessed on 17.07.2013. composition of Pakistan by contributing substantial amount Chaudhary, Q.Z., Mehmood, A., Rasul, G. and Afzaal, M. of SO2 column densities. (2009). Climate Change Indicators of Pakistan, Pakistan Meteorological Department, Technical Report August ACKNOWLEDGMENT 2009. Chelani, A.B. and Devotta, S. (2007). Air Quality Assessment Authors gratefully acknowledge the TEMIS project for in Delhi: Before and after CNG as Fuel. Environ. Monit. providing level-2 satellite data of SO2 measurements and Assess. 125: 257–263. details about the retrieval algorithm and error analyses. We Dalaffilla Volcano. Volcanolive.com. Available at: http://w are thank full to Mr. Zeeshan Shahid from Institute of ww.volcanolive.com/dalaffilla.html (Last Accessed on Atmospheric Physics (IAP) Chinese Academy of Sciences May 1, 2013). (CAS) for meteorological analysis and to NOAA/OAR/ De Smedt, I., Müller, J.F., Stavrakou, T., Van der A.R., ESRL PSD, Boulder, Colorado, USA, for winds data from Eskes, H. and Van Roozendael, M. (2008). Twelve Years their Web site at http://www.esrl.noaa.gov/psd/NUST, Engr. of Global Observations of Formaldehyde in the M. Zaheer and Mr. Obaid from IGIS-NUST, for providing Troposphere Using GOME and SCIAMACHY Sensors. guidance about ArcGIS software. Our special gratitude Atmos. Chem. Phys. 8: 4947–4963. goes to NUST for providing financial support to conduct De Smedt, I., Van Roozendael, M., Stavrakou, T., Müller, this study. J.F., Lerot, C., Theys, N., Valks, P., Hao, N. and Van der A.R. (2012). Improved Retrieval of Global Tropospheric REFERENCES Formaldehyde Columns from GOME-2/Metop-A Addressing Noise Reduction and Instrumental Degradation Afe, O.T., Richter, A., Sierk, B., Wittrock, F. and Burrows, Issues. Atmos. Meas. Tech. 5: 2933–2949. J.P. (2004). Bro Emissions from Volcanoes: A Survey Draxler, R., Stunder, B., Rolph, G., Stein, A. and Taylor, Using GOME and SCIAMACHY Measurements. A. (2013). HYSPLIT4 User's Guide Overview (http://re Geophys. Res. Lett. 31: L24113. ady. arl.noaa.gov/HYSPLIT.php). Ahmed, S.S., Abbasi, Q., Jabeen, R. and Shah, M.T. Eisinger, M. and Burrows, J.P. (1998). Tropospheric Sulfur (2012). Decline of Conifer Forest Cover in Pakistan: A Dioxide Observed By the ERS-2 GOME Instrument. J. GIS Approach. Pak. J. Bot. 44: 511–514. Geophys. Res. 25: 4177–4180. Bey, I., Jacob, D.J., Yantosca, RM., Logan, J.A., Field, Ejaz, S., Chekarova, I., Ahmad, M., Nasir, A., Ashraf, M. R.D., Fiore, A.M., Li, Q., Liu, R.Y., Mickley, L. and and Lim, C.W. (2009b). Pollution Dilemma in Asian Schultz, M.G. (2001). Global Modeling of Tropospheric Population: CNG and Wound Healing. Environ. Toxicol. Chemistry with Assimilated Meteorology: Model Pharmacol. 28: 323–332. Description and Evaluation. J. Geophys. Res. 106: Ejaz, S., Ejaz, A., Sohail, A., Ahmed, M., Nasir, A. and 23073–23095. Lim, C.W. (2009a). Exposure of Smoke Solutions from Bourassa, A.E., Robock, A., Randel, W.J., Deshler, T., CNG-Powered Four-Stroke Auto-Rickshaws Induces Rieger, L.A., Lloyd, N.D., Llewellyn, E.J. and Distressed Embryonic Movements, Embryonic Degenstein, D.A. (2012). Large Volcanic Aerosol Load Hemorrhaging and Ectopia Cordis. Food Chem. Toxicol. in the Stratosphere Linked to Asian Monsoon Transport. 47: 1442–1452. Science 337: 78–81. ESoP - Economic Survey of Pakistan (2013). Economic Bovensmann, H., Burrows, J.P., Buchwitz, M., Frerick, J., Survey of Pakistan 2012–2013, Ministry of Finance, Noe, S., Rozanov, V.V., Chance, K.V. and Goede, Pakistan. A.P.H. (1999). SCIAMACHY: Mission Objectives and EUMETSAT (2005). GOME-2 Products Guide, http://www. Measurement Modes. J. Atmos. Sci. 56: 127–150. eumetsat.int/en/area4/eps/product_guides/GO ME-2/GO Bramstedt, K., Noël, S., Bovensmann, H., Burrows, J.P., ME2-PG.pdf. Lerot, C., Tilstra, L.G., Lichtenberg, G., Dehn, A. and Fehr, Fee, D., Carn, S.A. and Prata, F. (2011). Infrasound and T. (2009). SCIAMACHY Monitoring Factors: Observation SO2 Observations of the 2011 Explosive Eruption of and End-to-End Correction of Instrument Performance Nabro Volcano, Eritrea, American Geophysical Union, Degradation, Proceedings of the 2009 Atmospheric Fall Meeting 2011, Abstract: V41E-07. Science Conference. Ferrari, L. and Salisbury, J. (1999). Sulfur Dioxide, National Carn, S.A., Krueger, A.J., Arellano, S., Krotkov, N.A. and Environmental Health Forum, Australia. ISBN: 0 642 Yang, K. (2008). Daily Monitoring of Ecuadorian 95982 X, National Library of Australia Cataloguing-in- Volcanic Degassing from Space. J. Volcanol. Geotherm. Publication, Available at (http://www.health.gov.au/inte Res. 176: 141–150. rnet/main/publishing.nsf/Content/7AD4C69916960162

Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014 1895

CA257BF00019394C/$File/suldiox.pdf). Ozone Mapping Spectrometer (TOMS) Instruments. J. Finlayson-Pitts, B.J.J. and James, N.P.J. (2000). Chemistry Geophys. Res. 100: 14057–14076. of the Upper and Lower Atmosphere, Academic Press, Kwiatkowski, D., Phillips, P.C.B., Schmidt, P. and Shin, Oval rd. London. Y. (1992). Testing the Null Hypothesis of Stationary Fioletov, V.E., Griffioen, E., Kerr, J.B., Wardle, D.I. and Against the Alternative of a Unit Root. J. Econom. 54: Uchino, O. (1998). Influence of Volcanic Sulfur Dioxide 159–178 on Spectral UV Irradiance as Measured By Brewer Lee, C., Richter, A., Weber, M. and Burrows, J.P. (2008). Spectrophotometers. J. Geophys. Res. 25: 1665–1668. SO2 Retrieval from SCIAMACHY Using the Weighting Goyal, P. (2003). Present Scenario of Air Quality in Delhi: Function DOAS (WFDOAS) Technique: Comparison A Case Study of CNG Implementation. Atmos. Environ. with Standard DOAS Retrieval. Atmos. Chem. Phys. 8: 37: 5423–5431. 6137–6145. GVP: Global Volcanism Program (2010). Available at: http: Lee, C., Martin, R.V., van Donkelaar, A., O'Byrne, G., //www.volcano.si.edu/index.cfm (Last Accessed: March, Krotkov, N., Richter, A., Huey, L.G. and Holloway, J.S. 2014). (2009). Retrieval of Vertical Columns of Sulfur Dioxide GVP: Global Volcanism Program. Available at: http://ww from SCIAMACHY and OMI: Air Mass Factor Algorithm w.volcano.si.edu/index.cfm (Accessed: April, 2013). Development, Validation, and Error Analysis. J. Geophys. Hassan K., El-Abssawy, A.A. and Khoder, M.I. (2013). Res.114: D22303. Characteristics of Gas-Phase Nitric Acid and Lee, C., Martin, R.V., van Donkelaar, A., Lee, H., Dickerson, Ammonium-Nitrate-Sulfate Aerosol, and Their Gas- R.R., Hains, J.C., Krotkov, N., Richter, A., Vinnikov, K. Phase Precursors in a Suburban Area in Cairo, Egypt. and Schwab, J. (2011). SO2 Emissions and Lifetimes: Atmos. Pollut. Res. 4: 117–129. Estimates From Inverse Modeling Using in Situ and http://earthobservatory.nasa.gov/IOTD/view.php?id=36889 Global, Space-Based (SCIAMACHY And OMI) (Last Accessed on May 1, 2013). Observations. J. Geophys. Res. 116: D06304. http://www.volcano.si.edu/world/volcano.cfm?vnum=0201 Marbach, T., Beirle, S., Platt, U., Hoor, P., Wittrock, P., -07=&volpage=erupt (Last Accessed on May 1, 2013). Richter, A., Vrekoussis, M., Grzegorski, M., Burrows, Jhajharia, D, Shrivastava S.K., Sarkar, D. and Sarkar, S. J.P. and Wagner, T. (2009). Satellite Measurements of (2009). Temporal Characteristics of Pan Evaporation Formaldehyde from Shipping Emissions. Atmos. Chem. Trends under the Humid Conditions of Northeast India. Phys. 9: 8223–82234. Agric. For. Meteorol. 149: 763–770. Mirza, A.I., Kazmi, J.H. and Shirazi, S.A. (2013). Spatio- Khattak, P., Khokhar, M.F. and Khan, S.A. (2014). Trans- Periodic Monitoring and Analysis of Atmospheric Sulphur Boundary Volcanic SO2 Detected over Pakistan from Dioxide (SO2) in the Metropolitan Areas of Lahore- Satellite Observations during the Time Period 2004– Pakistan. Pak. J. Sci. 65: 83–89. 2012. Aerosol Air Qual. Res. 14: 1543–1557. MoF: Ministry of Finance (2012). Pakistan Economic Survey Khokhar, M.F., Frankenberg, C., Van Roozendael, M., 2011–12, Transport and Communications, Ministry of Beirle, S., Kuhl, S., Richter, A., Platt, U. and Wagner, T. Finance, Government of Pakistan. (2005). Satellite Observations of Atmospheric SO2 from NASSD-Northern Areas Strategy for Sustainable Volcanic Eruptions during the Time-Period of 1996– Development (2003). Background Paper, Health and 2002. Adv. Space Res. 36: 879–887. Environment International Union for Conservation of Khokhar, M.F. (2006). Retrieval and Interpretation of Pakistan (IUCN), Pakistan program. Tropospheric SO2 from UV/VIS Satellites Instruments. Noël, S., Bramstedt, K., Bovensmann, H., Burrows, J.P., Faculty of Physics and Geosciences, University of Gottwald, M. and Krieg, E. (2007). SCIAMACHY Leipzig, Germany, PhD Thesis. Degradation Monitoring Results, Proc. ‘Envisat Khokhar, M.F., Platt, U., and Wagner, T. (2008). Temporal Symposium 2007’, ESA SP-636, Montreux, Switzerland. Trends of Anthropogenic SO2 Emitted by Non-Ferrous PCAP: Pakistan Clean Air Program (2008). Government of Metal Smelters in Peru and Russia Estimated from Pakistan, Pakistan Environmental Protection Agency. Satellite Observations. Atmos. Chem. Phys. Discuss. 8: Ministry of Environment. 17393–17422. Platt, U. and Perner, D. (1979). Detection of Nitrous Acid Krotkov, N.A., Carn, S.A., Krueger, A.J., Bhartia, P.K. and in the Atmosphere by Differential Optical Absorption. Yang, K. (2006). Band Residual Difference Algorithm Geophys. Res. Lett. 6: 917–920. for Retrieval of SO2 from the Aura Ozone Monitoring Platt, U. and Stutz, J. (2008). Differential Optical Absorption Instrument (OMI). IEEE Trans. Geosci. Remote Sens. Spectroscopy (DOAS), Principle and Applications, 44: 1259–1266. Springer Verlag, Heidelberg. Krotkov, N.A., Krueger, A.J. and Bhartia, P.K. (1997). Qudrat-Ullah, H. and Karakul, M. (2007). Modelling for Ultraviolet Optical Model of Volcanic Clouds for Policy Assessment in the Electricity Supply Sector of Remote Sensing of Ash and Sulfur Dioxide. J. Geophys. Pakistan. Int. J. Energy Sect. Manage. 1: 240–256, doi: Res. 102: 21891–21904. 10.1108/17506220710821125. Krueger, A.J., Walter, L.S., Bhartia, P.K., Schnetzler, C.C., Ramanathan, V., Crutzen, P.J., Lelieveld, J., Mitra, A.P., Krotkov, N.A., Sprod, I. and Bluth, G.J.S. (1995). Cass, G.R., Althausen, D., Anderson, J., Andreae, M.O., Volcanic Sulfur Dioxide Measurements from the Total Cantrell, W., Chung, C.E., Clarke, A.D., Coakley, J.A.,

1896 Khattak et al., Aerosol and Air Quality Research, 14: 1883–1896, 2014

Collins, W.D., Conant, W.C., Dulac, F., Heintzenberg, (2005). On the Retrieval of Volcanic Sulfur Dioxide J., Heymsfield, A.J., Holben, B., Howell, S., Hudson, J., Emissions from GOME Backscatter Measurements. J. Jayaraman, A., Kiehl, J.T., Krishnamurti, T.N., Lubin, Atmos. Chem. 50: 295–320. D., McFarquhar, G., Novakov, T., Ogren, J.A., Podgorny, Tilstra, L.G., De Graaf, M., Aben, I. and Stammes, P. I.A., Prather, K., Priestley, K., Prospero, J.M., Quinn, (2012). In-Flight Degradation Correction of SCIAMACHY P.K., Rajeev, K., Rasch, P., Rupert, S., Sadourny, R., UV Reflectance and Absorbing Aerosol Index. J. Satheesh, S.K., Shaw, G.E., Sheridan, P. and Valero, Geophys. Res. 117: D06209. F.P.J. (2001). Indian Ocean Experiment: An Integrated Van Roozendael, M., Fayt, C., Lambert, J.C., Pundt, I., Analysis of the Climate Forcing and Effects of the Great Wagner, T., Richter, A. and Chance, K. (1999). Indo-Asian Haze. J. Geophys. Res. 106: 28371–28398. Development of a Bromine Oxide Product from GOME, Richter, A., Burrows, J.P., Nuss, H., Granier, C. and Proc. ESAMS'99-European Symposium on Atmospheric Niemeier, U. (2005). Increase in Tropospheric Nitrogen Measurements from Space, ESTEC, Noordwijk, The Dioxide over China Observed from Space. Nature 437: Netherlands, 18–22 January 1999, ESA WPP-161, p. 129–132. 543–547. Richter, A., Wittrock, F. and Burrows, J.P. (2006). SO2 Wagner, T., Beirle, S., Deutschmann, T., Eigemeier, E., Measurements with SCIAMACHY, Proceeding of Frankenberg, C., Grzegorski, M., Liu, C., Marbach, T., Atmospheric Science Conference, SP-628, ESA/ESRIN, Platt, U. and Penning de Vries, M. (2008). Monitoring of Noordwijk, Netherlands. Atmospheric Trace Gases, Clouds, Aerosols and Surface Rix, M., Valks, P., Hao, N., Loyola, D., Schlager, H., Properties from UV/vis/NIR Satellite Instruments. J. Huntrieser, H., Flemming, J., Koehler, U., Schumann, U. Opt. A: Pure Appl. Opt. 10: 104019, doi:10.1088/1464- and Inness, A. (2012). Volcanic SO2, Bro and Plume 4258/10/10/104019. Height Estimations Using GOME-2 Satellite Wang, S.Y., Davies, R.E., Huang, W.R. and Gillies, R.R. Measurements during the Eruption of Eyjafjallajökull in (2011). Pakistan’s Two-Stage Monsoon and Links with May 2010. J. Geophys. Res. 117, doi: 10.1029/2011JD the Recent Climate Change. J. Geophys. Res. 116, doi: 016718. 10.1029/2011JD015760. Theys, N., Campion, R., Clarisse, L., Brenot, H., Van Gent, Zerefos, C.S., Mantis, H.T., Bais, A.F., Ziomas, I.C. and J., Dils, B., Corradini, S., Merucci, L., Coheur, P.F., Van Zoumakis, N. (1986). Solar Ultraviolet Absorption by Roozendael, M., Hurtmans, D., Clerbaux, C., Tait, S. Sulphur Dioxide in Thessaloniki, Greece. Atmos. Ocean and Ferrucci, F. (2013). Volcanic SO2 Fluxes Derived 24: 292–300. from Satellite Data: A Survey Using OMI, GOME-2, IASI and MODIS. Atmos. Chem. Phys. 13: 5945–5968. Thomas, W., Erbertseder, T., Ruppert, T., Van Roozendael Received for review, December 10, 2013 M., Verdebout, J., Balis, D., Meleti, C. and Zerefos, C. Accepted, May 27, 2014