During Sep-Oct 2009, when the winds were mostly from the South-West direction (as shown in Figure 2), with the 7 day back trajectories originating mostly from Arabian sea where SO₂ concentration is a minimum in the GEOS-Chem generated plots (Figure 1), the SO₂ time series showed a near constant continuum with very few spikes (Figure 3). The mean SO₂ concentration for this sampling period is 4.3 ppbv. The spikes in the spectrum seen on the 2^nd and 3^rd of Oct 2009 correlate very well with a change in wind trajectory pattern (Figure 2) when an air parcel from the high SO₂ source region surrounding 30˚N, 75˚E reached the sampling site.

The SO₂ profile had minimal diurnal variability (Figure 4) which is explained as due to the low PBL heights as well as high atmospheric OH radical concentration during these months over the sampling region (as will be discussed in section 3.4). The PBL height for this month over the sampling region is amongst the lowest of the year and the sampling site Gurushikhar is more or less within the free troposphere, nullifying the PBL height variation effects and hence the observed minimum diurnal variation for the SO₂ levels. Also the high OH radical concentration due to a higher water vapour content and incoming solar radiation (insolation) rate—the two parameters necessary for the production of OH radical in the atmosphere via photo chemistry—possibly makes sure sufficient oxidation of SO₂ even when the concentration is high to minimise the diurnal variation in the SO₂ oxidation efficiency. The very negligible contribution from long range transport during this sampling period—as almost all the back trajectories originated in the Arabian sea where the SO₂ concentration is a minimum in the GEOS-Chem generated plots, resulting in non dependence of the measured SO₂ levels on possible diurnal variability of wind pattern—can be another possible reason for the minimal diurnal variation in the SO₂ levels. The GEOS-Chem generated plots for the sampling period (shown in Figure 3) however showed some diurnal variation in SO₂ levels which may be attributed to the difference in the altitudes of the sampling site and the co-ordinates used in the model calculations (due to the constraint that at 4˚ × 5˚ resolution, the nearest possible grid point available with the model was 26˚N, 75˚E, where the terrain height may be different).

During Nov 2009 when the wind back trajectories (Figure 5) extended mostly to North-East, the SO₂ profile showed short duration spikes superimposed on the back ground values. The mean of the SO₂ concentrations recorded (Figure 6) for this sampling period is 3.4 ppbv. For this month, the wind back trajectories showed occasional long range transport from the high SO₂ region (surrounding 30˚N, 75˚E) along with those from the low SO₂ regions (e.g., regions lying in the North-East of the sampling site) in the GEOS-Chem generated plots.

Because of the wind pattern specific for this month, only short duration random air parcels arrive at the sampling site at Mt. Abu from the high SO₂ region surrounding 30˚N, 75˚E—as shown by the back trajectory analysis (Figure 5)—explaining the observed short duration spikes in the SO₂ profile.

A diurnal pattern is observed (Figure 7) in the measured SO₂ spectra, which may be attributed to the combined effects put in by the insolation dependent variation in OH radical levels and PBL height as well as the diurnal variation in the wind pattern carrying polluted air parcel to the sampling site from long distances.

The spikes as well as the concentration levels decreased from 12^th Nov 2009, as is the case with the GEOS—Chem generated SO₂ time series shown also in Figure 6—which has association with the wind trajectory change from high SO₂ regions surrounding 30˚N, 75˚E to relatively low SO₂ regions lying North-East of the sampling site.

Detection of SO₂ Plume during a Major Oil Fire Incident A major fire had erupted at an oil depot of Indian Oil Corporation Limited (IOCL) at Jaipur, Rajasthan (26.92˚N, 75.82˚E) on 29th October 2009 at around 7:30 PM (IST). Approximately 60,000 kilolitres of oil were burned in this fire incident which lasted for almost a week till 6^th Nov 2009, resulting in a significant rise in the air pollution levels over the region.

Studying the possible long range transport of SO₂ containing plumes to the high altitude sampling site at Mt. Abu from the disaster site, located only a few co-ordinates away was another objective of the sampling for this month. Three short duration high intensity spikes were recorded (Figure 8) each on 2^nd (around 0125 hrs) 4^th

(around 2230 hrs) and 5^th (around 2120 hrs) of Nov 2009 which was found associated with this major oil fire from back trajectory analysis—which showed air parcels reaching the sampling site from the disaster site (26.92˚N, 75.82˚E), during the above days (Figure 5).

During Dec 2009, the winds—as seen in the back trajectory plots shown in Figure 9—were mostly from North and North-East of the sampling site and occasionally carried high SO₂ from the source region surrounding

30˚N, 75˚E to produce many spikes in the measured spectra (Figure 10). The mean SO₂ concentration for this sampling period is 3.5 ppbv. The observed shift in concentration from higher during the initial sampling days to lower towards the end of the sampling period is explained by the observed shift in wind back trajectories from high SO₂ source regions (surrounding 30˚N, 75˚E) to low SO₂ regions (lying in the eastern side of the sampling site).

A diurnal pattern is observed during this sampling period (Figure 11) with the minimum around 1000 hrs IST, which is explained as due to the combined effects put in by the diurnal variability in the PBL height and OH radical levels. A possible periodic variation in the wind pattern carrying polluted air to the sampling site from long distances also could contribute to the observed diurnal variability.

The GEOS-Chem generated SO₂ time series (seen in Figure 10) also reproduced the trend observed for this month including the hump in the concentration values between 10^th and 15^th of this month.  

The GEOS-Chem based calculations showed (as will be discussed in section 3.5) that it is in the month of December every year, the contribution of long range transported

SO₂ over Mt. Abu attains its peak.

The sampling period in Feb 2010 was characterized by high SO₂ concentrations (Figure 12) and clear diurnal variability. The trend observed is high SO₂ concentrations during the initial days of the sampling followed by a dip and the concentration again rose towards the end of the sampling period. This is well explained by comparing backward wind trajectories (Figure 13) with GEOSChem generated SO₂ profiles (Figure 1). During the initial days of sampling all three components of the wind trajectories (plotted for altitudes 50 m, 500 m and 3000 m AGL) extended to the high SO₂ regions (surrounding 30˚N, 75˚E) in the GEOS-Chem generated plots. But during the time when dip in SO₂ values were observed, the only trajectory at 50 m AGL carried polluted air parcel from the high SO₂ region (30˚N, 75˚E) whereas the higher altitude trajectories (500 m and 3000 m AGL) came from the West Asian region where no major hot spots of SO₂ prevail in the GEOS-Chem generated plots. The fact that the Mt. Abu site enters the free tropospheric

zone during the winter time explains the non-influence of pollutant carrying winds in the lower layers of the atmosphere on the ambient SO₂ levels for this month.

The mean SO₂ concentration for this sampling period is 7.7 ppbv. The strong diurnal pattern observed for this sampling period (Figure 14) is attributed to the large diurnal PBL height variation as seen in the GEOS-Chem generated PBL values for this month (discussed in section 3.4).

For this sampling period, the GEOS-Chem generated SO₂ profile (seen in Figure 12) deviated in magnitude from experimental observations, suggesting the prominent roles of local sources as well as parameters not incorporated in the model in controlling the ambient SO₂ levels during this month.

During the sampling period in Mar 2010 also, relatively high SO₂ concentrations were recorded (Figure 15). Occasional long range transport of SO₂ to the sampling site is indicated by the back trajectory analysis which showed (Figure 16) arrival of air parcel from both the high SO₂ region surrounding 30˚N, 75˚E and low SO₂ regions such as the Arabian sea and West Asia. The mean SO₂ concentration for this sampling period is 9.2 ppbv. A well defined diurnal variability (Figure 17) is a characteristic feature for this period, which is attributed to the very large diurnal variability in the PBL heights as seen in the GEOS-Chem generated time series PBL values (discussed in section 3.4). The diurnal pattern resembled the one observed in Feb 2010 with the whole graph shifted by ~3 hrs to the left along the time axis. This shift is explained as due to the increase in the insolation rate from February to March, to cause the PBL rise to take place in the morning hours itself compared to February. Possible diurnal variation in the wind pattern carrying polluted air to the sampling site also could contribute to the observed diurnal variability.

It is evident from the above discussion that interplay between seasonal PBL height and the OH concentration variability can modulate the diurnal pattern of the SO₂ concentrations. The fact that the rates at which these two parameters vary—both seasonal as well as diurnal are different, leads to a seasonal dependence for the diurnal variability of SO₂ levels.

For example, during the months Feb, Mar and Apr the effect put in by the sudden change in PBL height in the morning hours leads to the sudden decrease in the SO₂ values between 0600 hrs and 0800 hrs IST, where the OH radical variation is not sufficient to cause much of a change in the SO₂ variability pattern. Whereas, the sudden changes in the OH concentrations occurring during Jun-Sep is the major player controlling the SO₂ diurnal pattern during these months.

To visualize the above mentioned relative contributions clearly, 3-dimensional graphs were plotted with the GEOS-Chem generated OH concentration along X-axis, PBL height along Y-axis and the SO₂ concentration along the Z-axis for different months of the year 2009 (Figure 21). During all the months, the SO₂ concentration decreased with PBL height increase as well as the OH radical concentration, but with varying magnitude. During Jan 2009, a PBL height variation from the lowest of the month to ~600 m takes the SO₂ value to a constant for all the OH values lying above ~2e⁺⁶. Similarly even with the highest OH concentration available with the month (~7e⁺⁶), the SO₂ concentration is not pulled down for PBL values below ~400 m. The random peaks observed in the SO₂ values for Jan has probably got to do with the influence of other parameters randomly controlling the SO₂ concentration over the region, and the transient long range transport is one such possible parameter.

During Feb 2009, for all values of OH concentration above ~4e⁺⁶, an increase in PBL height above ~700 m is sufficient to pull the SO₂ values to the minimum for that month. Below an OH level of ~4e⁺⁶, a combined effect of OH radical as well as PBL height variation is found to influence the SO₂ concentration. Similarly for OH concentrations below ~2e⁺⁶, even with the highest PBL height recorded for the month, the SO₂ concentration remained high. Compared to Jan 2009, the graph plotted for Feb 2009 has fewer uncorrelated humps for the SO₂ values, indicating the possible shift in the transport pattern.

The month, Mar 2009 saw a clear interplay between the PBL and OH values in deciding the diurnal pattern of ambient SO₂ over this region. This is one such month where both PBL height and OH levels are significant. As a result, for OH values between ~1e⁺⁶ and 4e⁺⁶, the increase in PBL height from the minimum of the month to ~1500 m, reduces the SO₂ concentration at a steady rate to reach the minimum of the month. For all OH values above ~4e⁺⁶, the SO₂ concentration drops to the minimum, for PBL heights above ~700 m. Interference from

other players such as transient long range transport is a minimum in this month as evidenced by the absence of any random humps in the SO₂ values.

The behaviour observed in Apr, May and Jun were somewhat similar to the one in Mar. The scenario started changing from Jul and became clearly visible by Aug, when the atmospheric OH concentration over the sampling region reached its peak values of the year. In Aug, the OH concentration was sufficient to pull down the SO₂ values to the minimum of the month for all PBL values above ~150 m, showing the major role the OH radical play in the diurnal pattern of SO₂ for this month. The non observation of any humps in the SO₂ values correlate very much with the minimum transient transport expected

for these months (as will be discussed in section 3.5), underlining the argument that the third major player controlling the ambient SO₂ levels over this region is the long range transport.

During Sep 2009 and Oct 2009, when the boundary layer heights are on the lower side, the OH radical and the transport are the only major players, resulting in many humps in the SO₂ values in addition to the contribution from heterogeneous phase oxidation on water droplets. During Nov and Dec, an overlap of low PBL heights with low OH concentration along with higher transport events (as discussed below) keep the SO₂ values on the higher side during most of the time.