Cross-Sectional View of Atmospheric Aerosols over an Urban Location in Central India

Surface, column and vertically resolved variations of physical and optical properties of atmospheric aerosol over Hyderabad, a tropical urban location in central India are explored on the basis of ground based and satellite retrieved data. Annual mean aerosol optical depth (τ) observed with Microtops sun-photometer is 0.61 ± 0.07 and seasonally it varied from 0.71 ± 0.06 in pre-monsoon to 0.55 ± 0.05 in winter. Aerosol types are categorized based on Ångström exponent (α) and τ relations; revealed that the study region is dominated by mixed type (MT) aerosol followed by urban/industrial aerosols under high τ (HUI) category. A consistent diurnal variation of black carbon (BC) is observed irrespective of seasonal variation with annual BC mass concentration is found to be 9.7 ± 1.9 μg m. During Telangana Survey day, which was the least pollutant day showed a reduction of 75% BC concentration during day time in comparison to five years average values, indicating the influence of anthropogenic effect over Hyderabad. Vertical information’s on aerosol are analyzed using Cloud Aerosol Lidar Pathfinder Satellite Observations (CALIPSO) and ground based Lidar (LAMP) data. LAMP data analysis shows a significant elevated aerosol layer up to 4 km during pre-monsoon while aerosols are confined below 3 km during post-monsoon and winter. Long term CALIPSO observations revealed that during postmonsoon to winter, the study area is dominated (~60%) by ‘urban’ aerosol; while during pre-monsoon period ~75% of the aerosol type belongs to ‘dusty mix’ category. A decline in short wave flux at the top of the atmosphere (0.66 Wm yr) is observed, as revealed by long term Clouds and Earths Radiant Energy System (CERES) data analysis with higher decline rate observed in winter (1 Wm yr) followed by pre-monsoon (0.8 Wm yr).


INTRODUCTION
Aerosols plays an important role in global and regional climate change due to their potential in altering radiation budget by scattering and absorbing incoming solar radiation (Charlson et al., 1992) and to lesser extend outgoing terrestrial radiation.As per latest IPCC report (IPCC 2013) global estimation of Aerosol Direct Radiative Forcing (ADRF) is found to be varied from -0.85 to +0.15 Wm -2 with an uncertainty of 1 Wm -2 .Negative radiative forcing induces a cooling effect on the Earth's climate system, while a positive forcing can assist to warming along with greenhouse gases (GHGs).Although on a global scale, scattering aerosols dominate, but while down scaling to regional atmosphere especially in urban environments an enhancement in absorbing nature of aerosols are visible.Three primary aerosol species that have absorbing characteristics are iron oxides in coarse-mode dust aerosols (Sokolik and Toon, 1999), brown carbon (BrC), and black carbon (BC) in finemode aerosols (Yang et al., 2009).Among these, BC aerosols have gained considerable scientific attention in recent years, because there are overwhelming evidence that it has the potential to induce climate change and has also increased many folds due to rapid urbanisation and development (Streets et al., 2001).These aerosols because of their absorbing nature reduce the amount of solar radiation reaching the surface and increase lower-tropospheric heating (Meehl et al., 2008).Through direct radiative heating, aerosols can also indirectly influence cloud microphysics thereby alter the Earth's regional hydrological cycle and circulation (Ramanathan et al., 2001;Lau and Kim, 2006).Even though extensive studies have been made in understanding aerosol-cloud interactions, their underlying processes are far more complex and less understood (IPCC, 2013).Thus accurate quantification of aerosols effect on climate is possible only through long term analysis of synergetic observations from satellite, airborne, and surface-based observations, as well as from model simulations (Anderson et al., 2003).
Present study attempted to address the column, surface and vertically resolved variations of physical and optical properties of aerosol over Hyderabad, a tropical urban location in central India.Towards this, multi-layered aerosol observations from in-situ and satellite have been used.Importance of the present study is that it is first among the studies over this region where long term aerosol loading along surface, column and vertical profile are analysed together.The study provides insight to aerosol loading over the study area, its direct effect on short wave flux at TOA and primary source characterisation.

Site Description and Data Set Used
Aerosol sampling measurements are carried out at the premises of National Remote Sensing Centre (NRSC), situated at the heart of the city, Hyderabad, a fastest growing metropolitan cities in India, with a total area of 650 km 2 .It is the capital of newly formed state Telangana, with a population of 4.01 million (Census 2011).Main sources of aerosols are from vehicular emission, industrial emissions, local/long range transported dust and biomass burning aerosols (Jose et al., 2016b).The study site is situated on the Deccan plateau at a height of ~557 m above mean sea level.It has a hot semi-arid steppe climate with four dominant seasons; winter (Dec-Jan-Feb), pre-monsoon (Mar-Apr-May), monsoon (Jun-Jul-Aug-Sep) and post-monsoon (Oct-Nov).Long term meteorological data over Hyderabad (www.imd.gov.in)showed that temperature typically varied from 16 to 39°C, However during pre-monsoon, maximum temperature exceeds even 42°C for few days.Relative humidity typically ranges from 23% (pre-monsoon) to 75% (monsoon).Local winds are either easterly or westerly depending upon seasons and it oscillates between two typical directions viz., north and south.Long term annual mean  of rainfall over Hyderabad is ~827 mm.
Different ground based instruments, satellite/sensors and their data used in the present study are given in Table 1.

Columnar Aerosol Observations Aerosol Optical Depth and Ångström Exponent
Seasonal averaged spectral variation of Aerosol Optical Depth (τ) over Hyderabad during the period (2010-2014) is shown in Fig. 1(a).Vertical bar denotes standard deviation from the mean.Daily mean data on spectral τ have been used to obtain respective seasonal mean of τ at 500 nm (τ 500 ) during the study period (annual mean is 0.61 ± 0.07) and its values during different season viz., pre-monsoon, monsoon, post-monsoon and winter are 0.71 ± 0.06, 0.59 ± 0.04, 0.59 ± 0.05 and 0.55 ± 0.05, respectively.Spectral τ is one of the important input parameters in radiative transfer to determine its impact on Earths radiative balance.It can also reveal qualitatively seasonal distinctive feature of aerosol size distribution.In the figure, a high spectral τ values are observed during pre-monsoon in comparison with other seasons.Spectral dependence of τ is quantitatively expressed using Ångström coefficients which is obtained by performing an empirical regression analysis to Ångström relation (Ångström, 1964) as given in Eq. (1).
where, τ λ is aerosol optical depth at λ, β is Ångström turbidity coefficient which equals τ λ at λ = 1 µm and α is the Ångström exponent.From a comprehensive study on aerosol spectra from different environment it has been established that high alpha values (> 1) are normally associated with continental aerosols, mainly anthropogenic (Eck et al., 1999), while presence of natural aerosols viz., sea salt or dust can results in a flat AOD spectra and hence a low α value.The values of α (380-870 nm) during pre-monsoon, monsoon, post-monsoon and winter are found as 1.01 ± 0.2, 0.59 ± 0.18, 1.16 ± 0.04 and 1.08 ± 0.03, respectively.This variation in α values during different seasons clearly shows that during post-monsoon and winter columnar aerosol properties over Hyderabad behave as continental polluted site, while presence of coarse mode aerosol are observed in the columnar aerosol structure during premonsoon and monsoon seasons.Previous studies have shown that, this seasonal transformation of aerosol size distribution is a consistent feature over Indian landmass (Moorthy et al., 2007).Low values of τ during monsoon are associated with wet scavenging and also indicate the possible presence of sea salt aerosols.Possible hygroscopic growth of water soluble aerosols like NH 4 + , NO 3 -, SO 2 -4 etc., which mainly comes from industrial emissions and anthropogenic activities are also the reason behind observed lower values of α during pre-monsoon to monsoon.Further, frequency distributions of τ and α during different seasons are analysed.Frequency distribution can provide an accurate statistical characterization of τ and α and also reveals the distribution of these parameters.Statistical analysis of frequency distribution (O'Neill et al., 2000) of τ reveals a log-normal distribution.
Our analysis (Figs. 1(b) and 1(c)) reveals that over the study region ~57% of τ values falls between 0.3-0.7;~12% of τ is greater than 0.9 and about 7% are below 0.3.High values of τ (> 0.9) are found significantly higher proportion during pre-monsoon compared to post-monsoon and winter.Frequency distribution of α reveals that ~60% of its values are greater than 0.9; which clearly indicates that fine mode aerosols dominates the aerosol system over the study region.Presence of fine mode aerosol are found more (~75%) during winter and post-monsoon; while coarse mode particle are more frequent (~40%) during pre-monsoon.Frequency distribution of α during pre-monsoon reveals a mixture of coarse and accumulation mode particles.This could be due to presence of long range transported dust and biomass aerosols, in addition to local aerosols and are quiet frequent over Indian landmass during pre-monsoon (Gharai et al., 2013;Jose et al., 2015aJose et al., , 2016a)).

Dominant Species
Above analysis shows that there is a significant divergence in aerosol types over Hyderabad with significant seasonal changes.Divergence in aerosol types are due to varied sources and its emission mechanism.Hence an understanding of the aerosol types provides a better perspective while  estimating radiative effect due to varied aerosols.Most common and extensively used method in aerosol characterization is a scatter plot between τ and α (Eck et al., 2001).This plot gives a qualitative discrimination of the amount and dimensions of the observed aerosols.In the present study we characterized atmospheric aerosols using a scatter plot between τ and α, with threshold values suggested by Kaskaoutis et al. (2009).Dominant aerosol types over the region are classified into four types viz., (1) Marine Influenced aerosols (MI), (2) urban/industrial aerosols under High AOD (HUI), (3) desert dust particles under High AOD (HDD), (4) Mixed type (MT), which includes those aerosol which do not belong to the above mentioned classes or undetermined ones.Fig. 2 depicts the scatter plot between τ and α and percentage contribution of different aerosol types during the study period.Percentage contributions of MI, HUI, HDD and MT are 3%, 37%, 11% and 49%, respectively, suggests that aerosol belonging to the category of MT dominates the aerosols types over the study region followed by HUI category.Seasonal segregation revealed that during pre-monsoon, aerosol belonging to HDD category, contributes ~5% of the total (11%).Gharai et al. (2013) have indicated that dust storm frequency is quite high during pre-monsoon over Indian region with source of origin from Thar Desert and Persian Gulf regions.Backward trajectory analysis over Hyderabad (Jose et al., 2016b) also reveals that majority of wind trajectories are from these regions during pre-monsoon.In addition to this, local re-suspended soil dust also have significant role in columnar aerosol loading.Presence of MI aerosols is very less in all seasons, and is found maximum during monsoon.
Further, we compared in-situ (Microtops-II) and satellite (MODIS and MISR) retrieved τ 550 data over the study area.Satellite retrieved τ is compared with in-situ measurements within ± 30 min of satellite overpass.Microtops τ at respective satellite wavelength is obtained by using the Ångström power law (Eq.( 1)).About 140 synchronous colocated data points are used for the analysis.Analysis revealed correlation coefficients (r) of 0.69 and 0.71; root mean square error (RMSE) of 0.17 and 0.23 and biases of 0.26 and 0.05, respectively when compared Microtops data with MODIS and MISR data.To study the long term trend of aerosol loading over the study area, MODIS and MISR derived AOD over the period 2000 to 2014 is analysed.Analysis reveals a positive AOD trend; MODIS L2 AOD showed an increase of 0.015 AOD/year, while 0.007 AOD/year has been observed from MISR L3 AOD data.Both trend showed a good statistical significance with confidence level of 95%.

Surface Aerosol Observations Seasonal Variations of Black Carbon
Surface aerosol observation in this study is quantified in term of BC mass concentration (µg m -3 ).These absorbing aerosols have gained much importance recently from scientific community in the context of climate change because of their ability to heat the atmosphere; there by assisting warming caused by greenhouse gases.In addition to it, BC or soot measurement is also considered as an important benchmark to determine the air quality of a region.
Long term (2010-14) daily mean BC mass concentration (µg m -3 ) over Hyderabad is illustrated in Fig. 3(a).Besides the strength of emissions, variability is also associated with the variation of meteorological parameters like rainfall, surface temperature, mixing layer height, wind speed and direction etc. Influence of these meteorological parameters on BC variability is discussed in the following sections.One of the noticeable features in long term BC variability over the study area is that increase in BC mass concentration during the period 2011-12 followed by a slight decreasing trend.Since Hyderabad is a growing city such a change could not be attributed to change in source characteristics.It may be acknowledged that there are scientific reports where a decline in BC emissions has been observed due to improvements in combustion technology and changes in fuel composition (Kirchstetter et al., 2008;Wang et al., 2014).Monthly mean variation of BC and rainfall (mm) over the study site is shown in Fig. 3(b).It is interesting to note that during 2013-14 periods, numbers of rainy days are more compared to previous years.It is also seen from the figure that high values of BC during 2011-12 are associated with prolonged dry period with no rainfall.Sink mechanism due to rainfall could be one of the reasons for lowering the BC values during 2013-14 periods.Relatively higher BC values are also observed within the short non-rainy period of 2013-14.Though BC particles are not hygroscopic, rainfall can contribute to reduction of BC particles by wet scavenging and by increasing the probability of condensation of secondary inorganic aerosols on BC particle.Thereby it can make them hygroscopic and get scavenged by precipitation due to the presence of high humidity prevailing during Monsoon.Similar observation is reported over Bangalore city during the period 2003-07 (Vinoj, 2009).
Five year (2010-14) annual BC mass concentration over Hyderabad is 9.7 ± 1.9 µg m -3 .BC mass concentration observed over Hyderabad is comparable to those reported from other urban cites of the country.However BC mass concentration is quite high in comparison with rural local locations.Higher BC concentration in urban area compared to that in rural location indicates significant anthropogenic emissions to urban atmosphere (Begam et al., 2016).
Monthly mean variation of BC mass concentration in box plot is shown in Fig. 3(c).Range of box represents the values between 25-75 percentiles, the wisker represent standard deviation, mean and median values are described by straight line and square inside the box.In addition to inter-annual variation BC also exhibit significant seasonal variations with high values during pre-monsoon and winter months and minimum during monsoon.BC mass concentration during pre-monsoon, monsoon, post-monsoon and winter is found to be 10.42 ± 1.67, 7.7 ± 1.06, 11.46 ± 0.52, 10.38 ± 1.76 µg m -3 , respectively.Seasonal variation of surface aerosols concentrations depends on three major factor the emission (source), redistribution (both horizontal and vertical) and removal.Since the study area is an urban site, seasonal variation in emission source cannot be expected.Therefore the seasonal variation of BC depends on redistribution and removal mechanism.Redistribution is made effective by wind movement both horizontally as well as vertically.In addition, Atmospheric Boundary Layer (ABL) height determines the extent of aerosol mixing in the vertical scale and plays an important role in redistribution.Observed winter maximum could be due to boundary layer subsidence and resultant confinement of aerosols near surface.During pre-monsoon, a decline in values of aerosol optical properties are observed compared to winter at surface measurements, since dispersion of aerosols is high during pre-monsoon due to large ventilation coefficient.Minimum BC concentration during monsoon could be attributed to the wash-out effects due to precipitation.Thus seasonal variations of BC can be attributed to dynamics of monsoon circulation, ABL characteristics and the prevailing meteorological conditions (Kumar et al., 2015).

Diurnal Variation
Diurnal variation of BC provides an insight into the underlying process that controls their evolution.For better understanding of the diurnal behavior of BC, we used European Centre for Medium-range Weather Forecasting (ECMWF) Interim Reanalysis (ERA) ABL data set which gives the data for every three hours starting from 00:00 UTC with a resolution of 0.25 × 0.25° (http://data-portal.ecmwf.int).Fig. 4 portrays the diurnal evolution of BC during different season along with the evolution of ABL Height (m).Despite seasonal variation, BC showed a consistent and well defined diurnal variability with two peaks during morning (07:00-09:00) local time (LT) and night 20:00 LT onwards also with an afternoon minimum (11:00:17:00 LT).It is also seen from the figure that evening peaks during post-monsoon/winter are not as prominent as for other seasons.This could be due to presence of prolonged stable boundary layer during these seasons compare to premonsoon/monsoon.Morning peak arises due to combined influence of fumigation effect, (Stull, 2012) and morning buildup of local anthropogenic activities.Low value of BC concentration during afternoon hours has been attributed to dispersion of aerosols, due to increase in ABL height, in addition to low traffic density.In the evening hours, surface inversion begins and forms a shallow stable boundary layer causing aerosols to get trapped near the surface.Morning peak is observed to be higher than nocturnal peak in all seasons, except during monsoon.This suggests that aerosol mixing from residual layer (fumigation effect) is stronger or more significant during morning hours.In the Fig. 4, morning BC concentration peak is more significant during post-monsoon and winter than in pre-monsoon.Presence of shallow boundary layer due to relatively lower land surface temperature during morning hours of these seasons causes ABL height relative lower when compared with respective time period of pre-monsoon.Formation of surface based inversion layer and nocturnal boundary layer are not very strong during monsoon (Ganguly et al., 2006) and this could be one of the reasons for weak diurnal variations observed during monsoon.

A Case Study on the Influence of Anthropogenic Emissions on BC
To illustrate the influence of anthropogenic emissions on BC mass concentration, a case study is discussed.India's 29 th state "Telangana State (TS)" was born on 2June, 2014.Newly formed TS Government has decided to have a household survey on 19August, 2014 to ascertain socioeconomic conditions of more than 20 lakhs households on a single day.The government has declared survey day as a holiday for maximum participation in the survey.To participate in the survey, people temporarily staying in the city have moved to their native places prior to survey day and returned after the survey.On the survey day, Hyderabad resembled a ghost city with all shops and commercial establishments closed and the traffic movement also notably less.Thus the survey day became one of the rare days over Hyderabad to study the influence of anthropogenic emissions in modulating ambient air quality.Major contributors to air pollution in the urban areas are vehicular emissions followed by dust aerosols, biomass burning and industrial emissions.Fig. 5 illustrates the diurnal variations of BC on survey day in comparison with 5 years monthly (August) diurnal mean.The Aethalometer measurements revealed that BC concentration on 19August early morning mean has shoot up to 28.74 µg m -3 , which is three times higher than that was observed with five years monthly mean.This could be due to the movement of many numbers of vehicles on 19 th early morning as people were moving towards their households for the survey.During day time (10:00-18:00 hrs) average BC concentration has reduced to 1.57 ± 0.41 µg m -3 compared to 5 years monthly (August) mean concentration of 6.36 ± 2.42 µg m -3 .This house hold survey day on 19 August, 2014 was one of the unique days where BC emissions recorded the minimum concentrations over the recent years and can be considered as a reference data for the least polluting day.This case study clearly illustrates how anthropogenic emissions can modulate BC mass concentrations over the study region.

Influence of Prevailing Meteorology on BC
Prevailing meteorological condition is one of the dominant factors that affect BC mass concentration of any region.In order to understand the effect of meteorological parameters on BC, correlation analyses between measured BC and the corresponding meteorological parameters viz., temperature change, ΔT (difference between daily maximum and minimum temperatures) and wind speed are studied over Hyderabad.Scatter plot between monthly mean wind speed and BC mass showed an inverse correlation (r = 0.23) between wind speed and BC suggesting partly the proximity of BC sources at measurement site.It also suggests that regional/local transport also has a role in modulating aerosol systems over the study site.Previous studies (Jose et al., 2015) also reported influence of regional/long range transport on local BC mass concentration over Hyderabad.During pre-monsoon ~46% of the wind direction towards the study site are from south eastern side and ~30% are from north western side.While during post-monsoon and winter majority of wind advecting towards study site are from eastern side.It is also noteworthy to mention that most of the local industries are also located in the south eastern side of experimental site.Studies have reported that over Indian region during post-monsoon to winter season agriculture residue burning is prominent especially in the northern side of study region; while as the season shift to pre-monsoon contribution from forest fire is more.Among many other factors, with absorbing characteristics, BC can also play a part role in modulating ambient air temperature of a region.Scatter plot (not presented here) between monthly mean BC and ΔT showed a relatively low positive correlation with correlation coefficient 0.31.Similar observations between BC and these meteorological parameters have reported from other locations of the country like Ahemdabad, Kadapa, Trivandrum, Manora Peak and Pune etc. (Begam et al., 2016 and references therein).

Vertical Distribution of Aerosols over Hyderabad
Vertical distribution of aerosols in the troposphere have significant radiative impact.Intensity of shortwave radiative effect at top of the atmosphere (TOA) is strongly dependent on the vertical distribution of aerosols and is more sensitive for absorbing aerosols (Meloni et al., 2005).Thus it is important to understand the vertical distribution of aerosols in urban environments, where absorbing aerosols like BC are one of the dominating aerosols.Over Hyderabad, vertical profile of aerosol particles have been investigated by researchers using aircraft and balloons experimentation in campaign mode (Babu et al., 2011(Babu et al., , 2016)).Sinha et al. (2013) used one year in-situ lidar observations to study the aerosol vertical distributions.They found that the ABL aerosols contribute from a minimum of 43.0% to a maximum of 97.9% to the total AOD.In the present study, 3 year (2012-14) in-situ LiDAR for Atmospheric Measurement and Probing (LAMP) observations are analyzed to understand their seasonal variations.Comparison study is also performed with respect to available in-situ observations and co-located CALIPSO observations.

Vertical Distribution of Aerosols and Thermodynamic Structure
Vertical distribution of aerosol is closely linked to thermodynamic structure and variability of atmosphere and also can be used as a tracer to understand the stratification of atmospheric boundary layer (Tiwari et al., 2003).In general, aerosols that are lifted during daytime is due to presence of strong convective eddies while in absence of these strong eddies in night time, causes confinement of aerosols to lower levels.During night as the convective activity decreases, radiative cooling at the surface induces stable stratification near the ground, sometimes far above the nocturnal boundary layer and advection begins to play a more important role in determining aerosol concentration aloft (Tiwari et al., 2003).To study the vertical distribution of aerosol and associated thermodynamic structure of atmosphere over the study area a, case study (18 March, 2014) is analysed.Fig. 6(a) shows the temporal evolution of aerosol vertical profile in terms of the normalized range corrected signal (NRCS) on study days.NRCS is obtained by correcting the lidar raw signal for after-pulse behavior and energy normalization correction following the methodologies described by Kumar (2006).Fig. 6(b) shows time-averaged vertical profiles of Aerosol Backscatter Coefficient (ABC in Mm -1 ), potential temperature, θ (K), Relative Humidity (%), wind speed (m s -1 ) and direction.Respective vertical profiles of meteorological parameters are obtained from IMD radiosondae data.Nocturnal evolution of NRCS on 18 March, 2014 revealed stratified layers up-to ~3.5 km from surface.Analysis on 3 years of LAMP data also revealed that such stratified layers are more observed during premonsoon rather than winter.This can be due to the presence of long range transported aerosols and also due to enhanced convective activity during pre-monsoon.To examine the probable source of elevated aerosols, active fire locations (confidence > 60%) over Indian region from 14-19 March, 2014 are overlaid with five day air mass back trajectories ending over the study site, Hyderabad at three different altitudes 1, 2 and 3 km as shown in Fig. 6(c).The thick NRCS (Fig. 6(a)) observed around 1 km altitude during early morning hours can be inferred to the presence of long range transported aerosols.The mean profile of ABC (Mm -1 ) (calculated during 05:00-06:00 LT) also indicated the presence of stratified aerosol layers with two prominent peaks at ~0.9 km and ~2 km altitudes, respectively.The simultaneous profiles of θ and RH showed an inversion just above 1 km.The profile of wind speed and direction suggest that within 1 km moderately south easterly wind prevailed during the event day.Spatial distribution of fire counts also shows a number of potential fire location over the south eastern part of the study region and lower level wind trajectories (1 km and 2 km) indicates its possible transport to the study region (Fig. 6(c)).This clearly suggests that the inversion in θ at ~1 km altitude is caused by the presence of long range transported absorbing aerosol layer due to biomass burning.The increase in θ observed in lower atmosphere where aerosol loading due to diabatic heating will significantly inhibit atmospheric ventilation and thus prevent aerosol from vertical mixing.This may also lead to weakening of convection and may inhibit cloud formation.

Seasonal Variations of Aerosol Backscatter Coefficient (ABC)
Seasonal mean vertical profiles of aerosol backscatter coefficient (ABC, Mm -1 sr -1 ) during midnight hours (00:00-01:00 LT) obtained from LAMP is depicted in Fig. 7. Vertical profile of ABC calculated following an inversion technique (Klett, 1985) by assuming a constant lidar ratio.However we have assumed different lidar ratios for each season by following Sinha et al. (2013).Rayleigh vertical profile is calculated by using real time vertical profile observations of temperature and pressure from descending pass (~01:30 LT) of Atmospheric Infrared Sounder (AIRS) over the study location.Cloud contaminated data are removed from analysis by following Kim et al. (2007).Significant seasonal variability in vertical profile of ABC is observed during the study period which is similar to those reported from other parts of the country (Solanki and Singh, 2014 and references therein).The average maximum ABC (Mm -1 sr -1 ) values observed during pre-monsoon, post-monsoon and winter are 5.25 ± 2.37, 7.79 ± 6.97 and 7.06 ± 5.38 Mm -1 sr -1 , respectively.During pre-monsoon vertical extend of aerosol is observed up-to 4.2 km, while during post-monsoon and winter season aerosol are confined below 3 km.Long term CALIPSO data (2006-12) also showed a similar seasonal variations of aerosol extinction profiles as reported by Jose et al. (2016b).Winter time enhancement of aerosol loading near surface is mainly associated with prevailing meteorology supported by shallow boundary layer depth, weak ventilation and removal processes.Enhanced vertical extent of aerosol layers during pre-monsoon is due to high land surface temperature prevailing during these months, which in turn enhance the convection activities.One of the recent studies (Babu et al., 2016) on vertical distribution of atmospheric aerosols over the Indian mainland have reported a clear enhancement in aerosol loading and its absorbing nature at lower free troposphere levels (FT) over the entire mainland during pre-monsoon season compared to winter.In addition to this long range transported aerosols also play an important role in formation of elevated aerosol layers during premonsoon (Solanki and Singh, 2014).

Comparison of LAMP with CALIPSO Observation
A comparative study on vertical profile of ABC (km -1 sr -1 ) measured by space borne lidar (CALIPSO) and in-situ LiDAR (LAMP) has been performed over the study region.For the purpose of comparison, collocated profiles from both instruments have been used.CALIPSO profiles used for analysis are averaged within a spatial resolution of 1 × 1° over the study site and corresponding LAMP profiles are temporally averaged to ± 15 min of satellite overpass time.In the present study a total of seven cases have been studied, with 5 profiles collocated during winter and 2 during pre-monsoon.Synchronous CALIPSO retrieved vertical proles of temperature and pressure are used to remove molecular backscatter in the LAMP algorithm.Figs.8(a) and 8(b) illustrates averaged collocated profiles from CALIPSO and LAMP during pre-monsoon and winter, respectively.In order to quantify comparison, we calculated percentage bias (Mamouri et al., 2009) between LAMP and CALIPSO backscatter proles.It is quite obvious from figure that satellite underestimates ABC below 1.5 km and overestimates above.A similar observation is also reported over Manora Peak (Solanki and Singh, 2014).Percentage bias observed below 1.5 km during pre-monsoon and winter are -104 ± 71% and -21 ± 35%, while those above 1.5 km (upto 3.5 km) are 56 ± 31% and 51 ± 33%, respectively.However, mean bias for pre-monsoon and winter is found to be -3.6 ± 92% and 27 ± 48%, respectively.These seasonal mean biases are in reasonable agreement with such efforts observed over other parts of the world (Mamouri et al., 2009;Pappalardo et al., 2010).

Vertical Profiles of Aerosol Types
Long term data (2007-14) of CALIPSO over Hyderabad have been also used to analyze vertical profiles of aerosol types (Fig. 9).Aerosol types are classified by proving threshold values to CALIPSO retrieved profiles of Lidar Ratio, backscatter Colour Ratio and Particle Depolarization Ratio The respective threshold values are available in following literatures Tesche et al. (2009) and Burton et al. (2013).Although this classification is qualitative, it provides useful information on regional aerosol characteristics and air quality.It can be used to apportion by type and its vertical location in the column and understanding aerosol lifetime and transport (Burton et al., 2013).Our analysis reveals significant seasonal variations in vertical distribution of aerosol types over the study area.During post-monsoon to winter study site is dominated (~60%) by 'urban' aerosol types; while during pre-monsoon months about 75% of total aerosol types belongs to 'dusty mix' category.This observation is in quite agreement with columnar aerosol observations during day time and the classification scheme employed where aerosol belonging to HDD type dominates during pre-monsoon in comparison to post-monsoon and winter.Elevated polluted dust aerosol layers are also observed during pre-monsoon.These aerosols can pose significant climatic implications.Using satellite data and models, Vinoj et al. (2014) reported  that desert dust aerosol levels over the Arabian Sea, West Asia and the Arabian Peninsula are positively correlated with the intensity of the Indian summer monsoon.Presence of dust layer can redistributes aerosol radiative forcing vertically and increases about 60% of the radiative forcing and heating rate by 60 times at that altitude with respect to non-dusty cloud-free days (Das et al., 2013).Enhanced heating rate associated with this elevated aerosol layer can significantly reduce rainfall and effectively spins down hydrological cycle by decreasing monsoon intensity (Ramanathan et al., 2001).In contrast, because of their absorbing nature, elevated dust aerosols can also act as an 'Elevated Heat Pump' (Lau and Kim, 2006) and may increase the summer monsoon rainfall over northern India through forced ascend and enhanced convection during the early part of monsoon season.

Observational Evidence of Short Wave Flux Decline at TOA
Two major factors which influence the solar radiation are atmospheric aerosols and clouds, which can lead to either dimming or brightening phenomenon (Alpert and Kishcha, 2008).Estimating the magnitude of surface reaching solar radiation due to aerosol is complicated, largely because their vertical distribution in the atmospheric column also affects the impacts at the surface.In order to quantify the role of aerosol on the modulation of incoming solar radiation over Hyderabad, an attempt has been made to study the variation of long term  upwelling SW flux at TOA (SW-TOA) as observed by CERES (on board Terra platform) under clear sky condition (Terra MODIS cloud fraction, CF = 0).Time series of monthly mean variation of clear sky SW-TOA is shown in Fig. 10.Analysis reveals a reduction in SW-TOA (0.66 Wm -2 yr -1 ) with higher decline rate observed in winter (1 Wm -2 yr -1 ) followed by pre-monsoon (0.8 Wm -2 yr -1 ).Increase in aerosol loading over a period of time, over the study area might have played an important role in reduction of SW-TOA.This decreasing trend of SW-TOA implicitly indicates the predominance of absorbing aerosols over study area which enhances cooling at the surface and warming in the atmosphere.Similar observation is made by Badrinath et al. (2010) for downwelling solar radition over the study area using model data.Surface cooling combined with atmospheric heating may increase the stability of the boundary layer and reduce vertical mixing.This increase in atmospheric stability reduces natural removal processes for air pollutants, resulting in worse air pollution episodes.

CONCLUSIONS
Present study analyzed five years (2010-2014) surface, column and vertically resolved aerosol observations over Hyderabad, an urban location in Central India in conjunction with long term (2001-2014) satellite observations.Salient findings of the study are following: • Annual mean τ 500 using in-situ observations during study period is observed to be 0.61 ± 0.07 and its value during different season's viz., pre-monsoon, monsoon, post-monsoon and winter are 0.71 ± 0.06, 0.59 ± 0.04, 0.59 ± 0.05 and 0.55 ± 0.05 respectively.• Regional atmosphere is dominated by fine mode aerosols; however presence of coarse mode dust aerosol type (HDD) is observed during pre-monsoon.• Surface aerosol measurements quantified in terms of BC mass concentration also shows significant seasonal variations with winter maximum and minimum during premonsoon and monsoon.Annual BC mass concentration is found to be 9.7 ± 1.9 µg m -3 .A consistent diurnal variation is also observed irrespective of seasonal variation.• Vertical distribution of aerosol types as observed from CALIPSO reveals that during winter and post-monsoon, 'urban' aerosol types dominates (~60%) over the study area; while ~75% of aerosol types belongs to 'dusty mix' category during pre-monsoon.• A decline in short wave flux at TOA (0.66 Wm -2 yr -1 ) is observed by CERES data analysis with higher decline rate observed in winter (1 Wm -2 yr -1 ) followed by premonsoon (0.8 Wm -2 yr -1 ).

Fig. 1 .
Fig. 1.(a) Spectral variation of mean AOD during different seasons and frequency distributions (%) of (b) AOD and (c) Alpha over the study area.

Fig. 3 .
Fig. 3. (a) Daily variation of BC, (b) Monthly mean variation of BC and rainfall (mm) and (c) monthly variation (Box plot) of BC over the study period 2010-2014.

Fig. 4 .
Fig. 4. Diurnal variation of BC during different seasons along with the evolution of ABL height (m).

Fig. 5 .
Fig. 5.Diurnal variation of BC on TS household survey day in comparison with 5 years monthly (August) diurnal mean.

Fig. 8 .
Fig. 8. Collocated profiles from CALIPSO and LAMP during (a) pre-monsoon and (b) winter over the study area.

•
During Telangana Survey day, which was the least pollutant day showed a reduction of 75% BC concentration during day time in comparison to five years average values, indicating the influence of anthropogenic effect over Hyderabad.• Significant seasonality in vertical profile of Aerosol Backscatter Coefficient is observed.Elevated aerosol layers (~4 km) are observed during pre-monsoon while aerosols are confined below 3 km during other seasons.

Fig. 10 .
Fig. 10.Long term (2000-14) variation of shortwave radiation at TOA as observed by CERES under cloud free condition over the study area.

Table 1 .
Instrumentation used in the study.