Aerosol Optical Properties and Composition over a Table Top Complex Mining Area in a Monsoon Trough Region

Aerosol physiochemical properties over a varied mining plateau region at the eastern end of a monsoon trough are reported for the first time and analyzed at different time scales. Aerosol optical depth (single scattering albedo, SSA) is found to be 0.49 (0.9) in pre-monsoon, 0.4 (0.94) in monsoon, 0.46 (0.92) in post-monsoon, and 0.36 (0.89) in winter, with an annual mean of 0.43 (0.91). The volume-size distribution is tri-modal, with 0.02 (ultra-fine), 0.2 (accumulation) and 7 (coarse) μm, but with seasonal signatures. The angstrom exponent (AE) varies along with the AOD, especially in winter, although they are inversely related to each other during monsoons; the increase in size may be due to the effect of humidity. AODbc varies between 13.4%–4.7% of the total aerosols, with the highest contribution in March, when forest burning in the north east is at its peak. BC is the lowest in July, the mid monsoon month with the minimum biomass burning and brick-kiln activities. It is likely that the interactions of various minerals and intermittent rains help keep the aerosol size in a mixed state with regard to the relation between AE and AOD, although more work is needed to confirm this. The chemical composition of aerosols is derived from an aerosol chemical model based on the measured amount of black carbon and the assumed components. These components are selected based on back trajectories and earlier reports from the region. Their concentrations are adjusted by constraining the model output AOD and SSA to match (±2% @ 500 nm) that observed by a sun-sky radiometer. The chemical compositions of the winter and post-monsoon months are similar, while pre-monsoon period has more coarse mode minerals, and the monsoon period has more sea-salt (accu.). The component mass concentrations were grouped into various size bins based on their modal radii, and the results indicate that PM1 is at its maximum in winter whereas PM2.5 is highest in the post-monsoon period. Monsoons leads to the effective washout of 2.5–10 μm sized particles.


INTRODUCTION
Aerosols are studied through various considerations such as optical properties in terms of visibility (e.g., Silvia, 2002) or radiative forcing (e.g., Haywood and Ramaswamy, 2012); mass concentration for health related pollution (e.g., Dockery and Pope III, 1994).They are also studied for chemical properties in view of deposition and wash-out causing acid rain (e.g., Scudlark and Church, 1988) and cloud condensation nuclei connected to precipitation (e.g., Novokov and Penner, 1993).Aerosol size distribution modifies the cloud droplet size distribution spectra (e.g., Bréon et al., 2002).The radiative forcing by aerosols modulates surface fluxes (Ramanathan et al., 2001) hence also interferes with pre-monsoon period, northern hilly regions viz.Nainital, Dibrugarh, Darjeeling experience a high AOD during premonsoon period due to dust transported from Arabian deserts (Dumka et al., 2008;Chtatterjee et al., 2010;Gogoi et al., 2011).Kumar et al. (2011) report an increase in ozone, BC concentration and AOD (at 500 nm) by 34%, 145% and 150% respectively in the central Himalayas due to burning of agricultural residue in the pre-monsoon months.Further, their trajectory analyses show a deceleration of wind flow during the fire-impacted period indicating convection that helps transfer of aerosols present to higher levels.In contrast to other stations in IGB, the hill station, Darjeeling (Chatterjee et al., 2010(Chatterjee et al., , 2012) ) indicates an abundance of fine particles (~162% of annual mean) through pre-monsoon period as well, though variation in AOD is largely similar.Chemical composition and particulate matter analysis of Decesari et al. (2010) indicate that considerable PM 10 aerosols are transported to their high altitude aerosol observatory in Nepal all the yearlong from foothills of Himalayas through up-valley flows.It is interesting to note that coarse mode PM 10 aerosols are transported even in the significantly stable winter time as well.Wang et al. (2009) discussed anthropogenic absorbing aerosols as the cause of Indian summer monsoon's northward shift especially during the onset of season; in this context knowledge of trough region aerosol characteristics gains more importance since the summer monsoon is deterministic in India's economy and agricultural yield.
The current station under study Ranchi (85.3°E, 23.5°N, 650 m MSL) is situated at the south east end of IGB but is a plateau area, does not have any significant representation in terms of aerosol characterization.It is also a varied mining region (Fig. 1) implying that aerosols present there can be expected to have quite unique physical and chemical properties and mixing states.Table top mining related particulate pollution is an increasing concern as biological experiments by Knuckles et al. (2013) show that even 24 hours exposure caused considerable micro-vascular dysfunction and the strong valley flows from plateau easily transport pollutants downhill than uphill.Hence constraining and finding abatement measures on the plateau pollution is essential to save further degradation of adjacent valleys.Thus knowledge of the aerosol characteristics is essential as other lower altitude regions around also fall in the heavy mining belt of India where abatement measures are a concern since long (Ghose and Majee, 2001;Ghose, 2007).The wind reversals happening through 360° at this station also support existence of distinctive aerosol nature.
Through the current study we present the monthly and seasonal variation of aerosol optical and chemical properties including volume size distribution over the Ranchi plateau.Chemical composition of aerosols is derived from aerosol optical model by using measured black carbon mass concentration with other possible components adjusted so that the model derived AOD and SSA match with those obtained from sun-sky radiometer.An attempt is also made to inspect whether all the absorption is caused by BC alone or a mixture of BC or any other possible absorbing aerosols.

LOCATION DETAILS, EXPERIMENTAL SETUP AND CONCURRENT OTHER MEASUREMENTS
Fig. 1 presents a mineral map of India with Ranchi marked specifically to show the diversity in minerals and its plentiful availability.It has mines of coal and bauxite to the west; coal, asbestos and graphite to the north; iron, manganese, asbestos, coal and chromites to the east and iron, gypsum, copper, manganese and coal to the south.
The skyradiometer, model POM1-L (PREDE Inc, Japan make) that measures sky radiances directly and that of aureole to derive aerosol properties through inversion algorithm, SKYRAD.pack(used in SKYNET aerosol-radiation network; http://atmos.cr.chiba-u.ac.jp/aerosol/skynet), is used to make measurements of aerosols during daytime under clear sky conditions.Aoki and Fujiyoshi (2003) provide a detailed technical description of skyradiometer function.It measures both direct solar irradiance (every 1 min) and diffuse sky radiation at either 10 min time interval or 0.25 step in air mass at 7 wavelengths (0.315, 0.4, 0.5, 0.675, 0.87, 0.94, 1.02 µm); we opted 10 min interval for this study.Calibration methodology and data reduction procedures or quality checks for this instrument are followed as per Nakajima et al. (1996) and Boi et al. (1999).
The sky radiometer is installed on the terrace of Department of Applied Mathematics building at 85.3°E,600 m AMSL) taking care to avoid as much obstructions around as possible in the scanning pathway as BIT campus is thickly populated with tall Salwood trees and it was not permitted even to crop the branches of trees.Owing to the latitude (23.5°N) of the station, sky-radiometer did the vertical plane scan only during summer and some of the late evening data was lost due to presence of a tall tree to the west; being an eastern location, sunrise and sunset are always early with respect to local time.
Backward trajectories of wind (Figs.2(a), (b) and (c)) for 5 days for 2 nd week of every month are obtained through NOAA-Hysplit trajectory model (Draxler and Rolph, 2013) using GDAS data for surface, 1500 m and 3500 m altitudes and mixed layer depth (used in OPAC) is obtained from Giovanni (disc.sci.gsfc.nasa.gov/giovanni#maincontent)MERRA (Acker and Leptoukh, 2007).The meteorological observatory (LATAMOS) maintained by Birla Institute of Technology, Ranchi consists of a 32 meter tower mounted with meteorological sensors at different levels, another small post mounted with radiation (Table 1) sensors.Incoming solar radiation in short wave band of 0.3-3.0µm, SW in, is collected at 1Hz and averaged for 1min and logged.Apart from tower, it also has a conventional observatory (WMO/ IMD standards) with surface based meteorological sensors (Vaisala make) such as pan evaporimeter, sunshine hour recorder etc.Data from conventional observatory viz.wind, humidity, temperature, rainfall (Fig. 3(a)) provides meteorological background.Monthly-mean wind speed increases from winter to pre-monsoon, highest (3 m/s) in June before the onset of rains.Winds are relatively high and consistent in pre-monsoon period while they remain high in monsoon with lower wind maxima, revealing the influence of large scale flow.While humidity reduces steadily during pre-monsoon till March, thundershowers enhance it during April-June whereas post monsoon and winter continues to have better humidity levels and highly variable rainfall.After 2 months of almost no rainfall, January 2011 had a fair amount of rainfall.

SKYRADIOMETER DATA SCREENING AND METHODOLOGY
Aerosol optical properties are obtained by inverting the sun-sky radiances using the inversion algorithm viz.SKYRAD.packversion 4.2 (Table 1).Collected raw data are processed first to obtain calibration constant on monthly basis, checked for least deviation with the previous month and the calibration constant thus obtained is used for second level processing.The SKYRAD.pack uses less than 0.02 convergence error to obtain calibration constant through  improved Langley method and is capable of screening out the cloud contaminated data to a great extent.Further quality control is done by screening out the data that is above one standard deviation of daily mean for stringent cloud screening.Since the observation station has a tendency to get partly cloudy on most of the days of the year, we checked AOD data for cloud affected hours, verified with the pyranometer data and decided to filter out all AOD data above 1.0 (due to clouds) to avoid any cloud contamination as a further quality check.Solid view angles are checked periodically on clear days but settings are not changed as the results showed little deviation from the initial installation values; however a limited offset tuning together with cleaning of lens are done once a week.Concurrent to AOD, other parameters viz.single scattering albedo (SSA), refractive indices (RI) real and imaginary and volume-size distribution are also obtained.Black carbon (BC) concentration is also measured at the site during Jan-Dec 2011 using seven-channel aethalometer (model AE-47, Magee Scientific Company, USA; Table 1) in which BC mass concentration is obtained through 880 nm optical measurements.Operational and other details are available in the web page, http://www.mageesci.com/Aethalometer_book_2009.pdf.Data for September and December are not available due to instrumental problem and data for August is not used as there are only a few clear sky days and not statistically significant.
The chemical model OPAC (Optical Properties of Aerosols and Clouds; Hess et al., 1988) allows different types of chemical components with different number concentrations as input that can be adjusted to match with the observed AOD and SSA @ 500 nm.In the present study, we have used suitable combination of components based on back trajectories and possible potential sources for different seasons (Table 2).Observed mixed layer height and humidity are used in OPAC with possible 4 chemical components other than measured BC.Number concentrations of components are fine tuned constraining the total number concentration to approximate the 'Urban' type in OPAC.The model output consists of number, mass, volume concentrations and mixing ratios of different components.As described in Sreekanth et al. (2007), AOD bc is derived using OPAC by neglecting all other components but with only BC input.The volumesize distribution obtained from sky-radiometer observations is qualitatively verified with the volume concentration of OPAC components plotted against the modal-volume radii of respective components.Further sky-radiometer derived volume size distribution is converted to number size distribution to know approximate total number of PM 1 , PM 2.5 and PM 10 and compared the same with the mass concentration derived through OPAC in different bin sizes.Aerosols can be identified or grouped based on their refractive indices (Levoni et al., 1997).Inferring the aerosol types on the basis of their refractive indices based on their work suggest an abundance of water soluble aerosols mixed with soot (SSA < 0.85; high absorption) in February-March changing to more dust-like (SSA > 0.9; more scattering) but mixed with soot in pre (April-May) and post monsoon (October-December).RIR also indicate that aerosols prevailing in both pre and post monsoon also exhibit some  In March 2011 at Ranchi the aerosols exhibit an anomalous absorbing nature with the lowest SSA (0.84), though we see only a slight change in the RIR and almost no change RII.As per the detailed analysis and data set of fire counts/ sources by Venkataraman et al. (2006) March is the peak period of forest fire for NE India as well as central India.

Monthly and Seasonal Variations of Optical and Physical Properties
The wind trajectory fluctuates between northwest to north east during this month and it is capable of bringing in soot to higher levels in the atmosphere.Local emission around the month-end is also high with the festival "Holi" when large scale waste wood is burnt during "Holika Dahan/Dhuli Vandan".
Though the trajectory plots (Figs.2(a) and 2(b)) at the ground level show change of direction in April (NE), May (S) and June (SE), under the prevailing well mixed conditions, dust from Thar region that gets transported to upper layers of atmospheres over Ranchi is capable of maintaining the coarse mineral dust influx apart from the dust raised by local winds.In pre-monsoon month April, volume size distribution (Fig. 4(a)) indicates highest peak close to but less than 10 µm supports the transport of dust from W-NW direction.Change in surface wind direction and pre-monsoon showers reduce coarse particle concentration in May hence its coarse volume stays below that of April.In June coarse mode volume exhibits a double step at 1-6 µm and at around 11 µm, shows the effect of both coarse dust particles and humidity.The first half of June often experiences dusty winds and in the later half it changes over to rains as evidenced by the observer records and rainfall data.Growth of aerosols due to humidity impact (Tang, 1996) is also a deciding factor of the scattering property in this period as the peak in the monthly volume size distribution beyond 10 µm (Fig. 4(a)) seems to be an outcome related to high humidity.
During monsoon period sea salt influx seems to help sustain the high coarse mode distribution as explained by Chatterjee et al. (2010) in their chemical analysis of aerosol data at Darjeeling, a hill station further north to Ranchi.At Ranchi aerosols have the least optical extinction (AOD) in August apparently due to wash out.Though the aerosol concentration reduces (minimum AOD) presence of absorbing particles restricts SSA to low values.We are cautious to state that as the data length is particularly small in August to rely on.The accumulation mode peak is the maximum in October-November that is caused by various activities around that time.Minute inspections of these monthly volume-size distribution curves show a tri-modal distribution with a small peak in the nucleation mode particles in these months (Fig. 4(a); further illustrated in Fig. 5(b)).Burning activity that starts with Ramnavami in October to festival of lights, Deepawali in November, lots of crackers are burst that increases chemical/soot loading in the atmosphere.It is so significant that there are special studies conducted on aerosol loading in this period (e.g., Babu and Moorthy, 2001).Use of coal or firewood for warmers in winter further increases the soot aerosol loading of the atmosphere.
Angstrom Exponent, AE or α is the term indicating spectral dependency of AOD or aerosol size-AOD relations.α is calculated by the relation, α = -ln (τ1/τ2)/ln (λ1/λ2), where τ1 and τ2 are AODs at wave lengths λ1 and λ2 respectively.A best fit slope is obtained to find the AE between 400-1020 nm.Fig. 4(b) shows AE -AOD relation for different months with AE calculated for the entire wavelength range.Lee et al. (2010) proposed a classification of aerosol types depending on SSA and fine mode fraction and a class as "mixture" was suggested for aerosols having intermediate properties.The aerosols at Ranchi also may well be classified as "mixture" based on their AE-AOD variations and SSA values as the current results do not show demarcation of size based on AE values.The AE -AOD correlation coefficient falling below 0.6 (except for winter, 0.66) indicates uncertainty of dominant size (Fig. 4(b)).Thus it could be said that from December to February, increase in AOD is mainly caused by fine aerosols.In the dust dominated regions and during dust storms (Dey et al., 2004;Moorthy et al., 2007) AE vs. AOD show a clear, high negative correlation as coarse particles contribute to the increase in AOD.In Fig. 4(b) negative correlation is observed only during June-September In spite of the volume-size distribution (Fig. 4(a)) indicating the second peak (radius > 0.5 µm) much higher than the first peak (radius < 0.5 µm) in pre-monsoon as well as in monsoon months, an inverse relation between AE and AOD is absent in pre-monsoon.This behavior confirms that the aerosols fall into the category "mixture" (neither fine/absorbing nor coarse/scattering).Seasonal correlation of AE -AOD, shown in Fig. 4(b), is drastically low (0.03) in pre-monsoon season (March-May) having the highest data base of 76 days, proving the nonlocal and varied sources of aerosols while the same for winter (December, January, February) with a value of 0.66 indicates the dominant control of local sources which is corroborated by prevailing low wind speed.
Monthly averages are used to derive seasonal and annual optical properties (Fig. 5).Comparatively low AODs (-15% of annual mean) in winter associated with strongly absorbing (SSA is -4% of annual mean) nature changing over to larger AODs (+13% of annual mean) in pre-monsoon with higher scattering (SSA is -1% of annual mean), suggests that winter is characterized by relatively higher concentration of fine soot particles and reduced mixing height that gradually turns to increased influx of coarse dust particle along with higher mixing height in pre-monsoon which is consistent with the reports of other stations in the IGB north east (Chatterjee et al., 2012) and north west (e.g., Singh et al., 2004).Wind trajectories are also nearly similar.However, during monsoon and post monsoon months the generalization of trends in IGB ceases; while some stations like Kanpur to the west (Singh et al., 2004); monsoons are optically thick with high AOD.Darjeeling, located at the east (Chatterjee et al., 2012) experiences thinnest aerosol layer in monsoon but for Dibrugarh, a station further to the north-east, post monsoon reports to be of minimum AOD (Gogoi et al., 2011).AOD has an annual average of 0.43 that falls between AOD for monsoon (0.4) with the high standard deviation (SD) and AOD for post monsoon (0.46) with low SD.However, the highest AOD occurs in pre-monsoon (0.49).Unlike at Dibrugargh or Gandhi College (Gogoi et al., 2011;Srivastava et al., 2012) or Darjeeling (Chatterjee et al., 2012) winter is optically thinnest (lowest AOD) at Ranchi for 2011.Rainout in monsoon season certainly reduces aerosol concentration but high humidity aids the growth of aerosols and makes it more scattering irrespective of chemical composition thus resulting in enhanced AOD as well as SSA (Tang, 1996).The results show that SSA decreases seasonally from monsoon to winter thereafter increases a little during pre-monsoon.RII exactly follows SSA variations while RIR variation is opposite to that of SSA.Winter aerosols here are the most absorbing, and maintains the same nature throughout the season as indicated by minimum standard deviation (Fig. 5) reiterating the local signature.

Volume and Number Size Distribution
Seasonal variation of volume-size distribution further explains the behavior of AOD with the highest AOD corresponding to maximum area under the curve and the lowest corresponding to minimum area.Seasonal variation of number and volume size distributions are shown in Fig. 6.Number (Fig. 6(b)) and volume (Fig. 6(a)) size distributions indicate three peaks corresponding to radii of 0.02 µm, 0.2 µm and 7 µm representing ultra fine, fine and coarse modes respectively.The 3 modes are shown individually with dashed curves.The results are to be carefully interpreted as Nakajima (1996) estimated the error in volume-size distribution as 20% for the middle range and 35-100% at the edges.Yang and Wenig (2009) suggested that if the volume in the coarse (> 7 mm) or ultrafine (< 0.03 mm) exceeded the mean value of the whole spectrum, then they needed to be considered as anomaly and is to be discarded as they could be an outcome of cloud effect.However, in our case ultrafine range does not show such anomaly any time but the coarse mode is very high in June-August (Fig. 4(b)).Though they are considered, it results only in limited errors.A noticeable feature of this station (Fig. 6(a)) is that peak radii in fine and coarse modes for pre-monsoon are less than that for monsoon and postmonsoon but the area under the curve for pre-monsoon is much higher in comparison.In pre-monsoon, coarse mode volume concentration dominates as compared to other seasons and is attributable to transport of mineral dust from IGB and north east regions as indicated by back trajectories (Figs.2(a) and (b)).The increase in fine and ultrafine mode (peak radii 0.2 µm and 0.02 µm respectively) during post monsoon illustrates wind assisted fine dust loading of atmosphere after the rains.Increase of fine BC and ash due to re-start of biomass burning and other activities adds to it.Winter is the least in volume concentration for both fine and coarse modes.Number concentration is derived from volume concentration by dividing it by 4/3πr 3 .They maintain an almost inverse relation as the aerosols in the ultrafine mode having maximum number contribute to the minimal volume and vice versa in the coarse mode; for the same reason peak radii corresponding to peaks in number concentration are less than those for volume concentration.

Aerosol Optical Depth: Different Fractions
Variation of BC from October to May (other than monsoon period) is controlled by various types of biomass/fuel burning during festivities, generated either locally or transported by wind flow from elsewhere and day-to-day life.Apart from coagulation with other aerosol and resultant washout, temporary closing of brick kilns and absence of large scale biomass burning activities during monsoon explains the lowest observed BC in this period.
Following relations relate total AOD to absorption AOD which in turn is the sum of contributions from BC and non-BC absorbing aerosols.AOD total = AOD abs + AOD scat & AOD abs = AOD bc + AOD non-bc-abs. (1) In most of the months AOD bc is less than or equal to AOD abs (Fig. 7(a)) as expected.But in April and May AOD bc is more than AOD abs while dust particles are abundant in the air.Kim et al. (2004) studied properties of dust particles during ACE-Asia intensive observation period and found that there was an overestimation of coarse mode BC mass concentration by aethalometer during dusty period due to agglomeration of BC and dust particles.In addition to the possible over estimation error, scattering efficiency of coarse BC particles (mixture) could be more than fine BC particles (due to increase in size), then AAOD or AOD abs (= (1 -SSA) × AOD) (Russel et al., 2010) could be less than AOD bc which assumes all BC aerosols are with r mod N equal to 0.0118 µm (Hess et al., 1998) as per OPAC.This may be the reason of lower AOD abs in the above months.Scarnato et al. (2012) studied internal mixing of BC with NaCl and found that BC internally mixed with NaCl has better SSA than BC alone and increases with more embedding common salt.Having an influx of NaCl with the onward wind from sea during this period there exists a higher possibility of BC getting embedded on NaCl, explaining the lower AOD abs in July by the effective increase in SSA.AOD bc is 13.4% of total AOD in March and 4.7% in June; annual average, 8.8% that is less than 11% calculated for IGB (Ramanthan and Ramana, 2005).

Aerosol Chemical Composition from OPAC
OPAC is used in this study to determine possible chemical composition based on air mass source regions through back trajectories at surface, 1500 m and 3000 m altitudes (Figs.2(a), (b) and (c)) as well as earlier reports from the nearby regions through in-situ observations (pl.refer methodology section for settings of OPAC).The main constituents are water soluble, insoluble, soot, sea salt (acc.), and mineral accumulation/coarse and mineral transported above mixed layer.Fig. 7(b) shows monthly variation of AOD and SSA derived from OPAC and skyradiometer, is in agreement within 2% @ 500 nm.
Since AOD is mainly determined by aerosol volume, volume mixing ratios of different components are shown in Fig. 7(c).Soot, water-soluble and insoluble (Table 2) are the common components for all months but the 2 variable components are set as explained in the methodology section, but the same combination is maintained for each season.Thus mineral coarse mode is inducted during pre-monsoon and sea salt during monsoon.Transition months March (winter to pre-monsoon) and June (pre-monsoon to monsoon) have both characteristics; mineral accumulation and coarse for March (cold turning to dusty) and mineral coarse and sea salt for June (dusty turning to monsoon rain), makes it difficult to make correct prescription of components.Volume mixing ratio plot (Fig. 7(c)) portraits June with a high content of insoluble component as this also includes coarse mineral dust and in March it is the high mineral accumulation compensating some of mineral coarse mode.Sea salt accumulation mode particles are significantly high in July when air mass flows from the southern Bay of Bengal as indicated by back trajectories (Figs.2(b) and 2(c)).Postmonsoon months October and November are associated with low winds (with short trajectories at the surface); hence aerosols are likely to have local origin.Winter and postmonsoon have similar components and also have transport of minerals at higher level.The main difference between post-monsoon and winter is the humidity related growth of aerosols and lower fog scavenging at the surface in postmonsoon.Though through OPAC it is possible to get a fair idea of chemical nature of the components; its limitations such as limited components input, only 5 types at a time that may not be the reality, are to be kept in mind.Moreover, OPAC is highly sensitive to higher humidity and insensitive from 0-50%.Similarly mixing height is also crucial in deciding AOD.OPAC is generally set for underestimation of SSA, e.g., 0.82 for urban aerosols (Hess et al., 1988) that may be quite a low value looking at derived SSA from the in-situ observations (e.g., Srivastava et al., 2012).However, OPAC with all its limitations remains a handy tool for many aerosol researchers.

Particulate Matter Fractions
Study location being a varied mining region it is appropriate to know the possible chemical composition and health hazard each PM bins can have.A recent study by Perrone et al. (2010) reports about PM 1 having sulphate enriched with Al, Ar, Cu, Cr and Zn that can reduce cell viability to half in summer which is possible at this station as well.A study in China (Tao et al., 2012), with similar environment, for source apportionment of PM 1 showed secondary aerosol formation and biomass burning, diesel emissions, gasoline emissions and sea salt, and coal combustion were the greatest contributors to PM 1 in that order.Since the order of the PM 1 (derived from number concentration) is highest and cumulative number difference between PM 2.5 and PM 10 are orders low, PM 1 concentration and relative differences of PM 2.5 , PM 10 and TSP from PM 1 are plotted in Fig. 8(a).Aerosols above PM 10 are very low in winter as well as post monsoon and maximum submicron aerosols (PM 1 ) are in post-monsoon followed by monsoon.The significant reduction in PM 2.5 to PM 10 in monsoon apparently points at the most effective rain-washout in this size-range (Croft et al., 2009).Further, monsoon indicates an excess of aerosols above PM 10 .This probably is an artifact due to increase in humidity and this part of volume-size distribution is suspected as discussed earlier.
As seen in Fig. 8, a clear low in PM 1 and subsequent increase in higher sizes in pre-monsoon is associated with coarse mode aerosols transported during this period.
Since mass concentration is one of the outputs of OPAC, we have analyzed it using the respective component radii (Table 2) for quantifying pollution in the most common way (Fig. 8(b)).The station indicates poor air quality standard with PM 2.5 and PM 10 concentrations about 10 times higher than the guideline values set by WHO viz., 10 and 20 µg/m 3 respectively.The results show that though optically premonsoon is the most polluted (highest AOD) and winter in terms of atmospheric heating (lowest SSA) and low dispersion (lowest mixed layer height), post-monsoon has maximum number of TSP at this station.Due to variation  et al. (2012).The steady increase of PM 2.5 from the start of October to November at Delhi is remarkable which is similar to the OPAC output in this study.In winter, at Ranchi PM 2.5 are much less than in Delhi but in pre-monsoon it is more as Delhi has more coarse aerosols in pre-monsoon.

CONCLUSIONS
Aerosol optical properties and chemical composition over Ranchi are summarized as follows.1. AOD is the highest in Pre-monsoon, followed by postmonsoon and lowest in winter while SSA is the lowest in winter.Annual mean AOD and SSA are 0.43 and 0.91 respectively.2. Angstrom exponent falls between 1 and 2 indicating intermediate size aerosols and its linear increase with AOD in winter suggests dominance of fine particles (< 1 µm).3. OPAC derived aerosol chemical composition indicates significant volume concentration of sea salt accumulation in July, an active monsoon period, due to transport of marine aerosols as indicated by air mass back trajectories that originate in the Bay of Bengal.4. Pre-monsoon is associated with considerable amount of mineral coarse particles that can be attributed to transport of mineral dust from IGB as well as northeast regions as corroborated by back trajectories.5. OPAC derived AOD, SSA and fine mode radius (corresponding to the first peak in volume-size distribution) agree well with that measured by skyradiometer indicating the accuracy of aerosol chemical composition.However, in-situ measurements of chemical composition and internal/external mixing state are required for detailed understanding of aerosol properties over this complex mining plateau region as this could also affect pollution in the adjacent plains through valley flows.

Fig. 1 .
Fig. 1.Mineralogical map of India (source: Centre for Science and Environment, New Delhi) with the observation station, Ranchi shown in inset as Google map.

Fig. 2 .
Fig. 2. a) 7 day back trajectories derived from NOAA-Hysplit back trajectory analysis based on 14 th of every month at the location: January-April, 2011(right to left); b) Same as in Fig. 2(a) but for May-August; c) Same as in Fig. 2(a) but for September-December.

Fig
Fig. 3. a) Back ground meteorological parameters; wind speed, wind maximum, temperature, humidity, shortwave incoming radiation and rainfall for the year 2011; b) Monthly variation aerosol properties; AOD, SSA, RIR and RII.AOD from MODIS indicated by '*'.
Fig. 4. a) Monthly volume-size distribution; b) AE vs. AOD on seasonal scale; dots represent individual day average, correlation coefficient (number of days) are included.

Fig. 5 .
Fig. 5. Seasonal variation and annual mean of optical parameters.

Fig. 6 .
Fig. 6. a) Seasonal volume-size distribution; b) number-size distribution derived from it; three possible modes are shown with dotted lines.

Fig. 7 .
Fig. 7. a) Average day-BC concentration on monthly scale and a comparison of OPAC derived AOD bc and calculated AOD abs [(1 -SSA) × AOD]; b) Monthly plot of AOD and SSA (measured) vs. AOD and SSA (OPAC-derived); c) Volume mixing ratio of the different components.

Fig. 8 .
Fig. 8. a) number-size distribution of PM 1 (cumulative observed values, ref: Fig. 6(b) and difference in number from PM 1 to PM 2.5 , PM 10 and TSP; b) Monthly mass concentration in different PM size bins derived from OPAC considering modal radii.

Table 2 .
OPAC components assumed with details. ) conform to the properties of aerosols present at various times.They also point towards a possible influx of hematite from the iron ore mines to the south of the station with the wind blowing from south.