Atmospheric Research 92 (2009) 27–41

Contents lists available at ScienceDirect

Atmospheric Research j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / a t m o s

Seasonal and diurnal variations of black carbon aerosols over a Mediterranean coastal zone Auromeet Saha ⁎, Serge Despiau LSEET-LEPI, UMR 6017, Université du Sud Toulon-Var, La Garde, France

a r t i c l e

i n f o

Article history: Received 5 January 2008 Received in revised form 23 May 2008 Accepted 4 July 2008 Keywords: Black carbon Aethalometer Mediterranean Coast Diurnal variations

a b s t r a c t Measurements of Black Carbon (BC) aerosols were carried over Toulon (43.14°N, 6.01°E; 50 m above MSL), an urban coastal zone in the Mediterranean coast, Southeast of France. The data collected for 14 months during 2005–2006 has been used for this study. BC concentrations displays significant temporal variations, with monthly-mean concentrations varying between ~ 300 and ~ 1000 ng m− 3 during the study period. Daily-mean concentrations as high as ~ 1500 ng m− 3 were also observed on several occasions. Large concentrations occurred during the autumn and winter, followed by lower concentrations during spring and summer seasons. In addition to the seasonal variations, BC concentrations also showed strong diurnal variations, with a morning peak and a mid-afternoon minimum. The diurnal variations are found to be seasonally dependent, with the maximum diurnal amplitude occurring during the winter months. These diurnal and seasonal variations are found to be significantly influenced by local traffic sources, wind speed, and boundary-layer dynamics. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Aerosol Black carbon (BC) is the optically absorbing part of carbonaceous aerosols, primarily emitted from combustion. It is a major anthropogenic component of atmospheric aerosols, which has significantly different optical and radiative properties as compared to the other normal constituents. BC acts as an indicator of airmass affected by anthropogenic pollution (Penner, 1995). Absorption by BC lowers the aerosol single scattering albedo, increasing the amount of radiation absorbed in the atmosphere (Haywood and Shine, 1997). Due to its large absorption over a wide wavelength range, BC can significantly offset the whitehouse effect (Schwartz, 1996). Recent studies suggest that BC can alters the cloud lifetime (Ackerman et al., 2000), precipitation patterns (Menon et al., 2002), reflectivity and melting of snow and ice (Hansen and Nazarenko, 2003). Modelling studies by Wang (2007) also suggests that the direct radiative forcing of BC can cause a

⁎ Corresponding author. Now at: Department of Geography, University of Iowa, Iowa City, USA. Tel.: +1 319 353 2964; fax: +1 319 335 2725. E-mail addresses: [email protected], [email protected] (A. Saha). 0169-8095/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.atmosres.2008.07.007

significant change in the atmospheric circulation and tropical convective precipitation. Being present mostly in the submicron size range, it is easily inhaled and can be a health hazard (Horvath, 1993; Pope et al., 2002). Recently, BC has also been used as an indicator of exposure to diesel soot (Fruin et al., 2004). The two most important sources for atmospheric BC are fossil fuel combustion (for example automobile exhaust, industrial and power plant exhausts, aircraft emissions, etc) and biomass burning (burning of agricultural wastes, forest fires). While biomass burning may be the dominant BC source over tropical regions and most of the southern hemisphere, the role of fossil fuel combustion is usually more important in cities, especially over the northern hemisphere. Due to its environmental and climatic significances, as well as anthropogenic nature of its origin, characterisation of BC has attracted considerable interest in the recent years (Hansen et al., 2000). Knowledge of long-term changes in BC is important as it can help to assess the effect of BC emissions on regional and global climate. Several long term measurements of BC have been carried out in the remote and marine environments (Cooke et al., 1997; Sharma et al., 2004). However, very few

28

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

long-term measurements exist in the source regions. To fill this gap, several field experiments and campaigns have been carried out in the recent years such as Smoke, Clouds, Aerosols, Radiation-Brazil (SCAR-B), Tropospheric Aerosol Radiative Forcing Observational Experiment (TARFOX), Aerosol Characterisation Experiment — Asia (ACE-Asia), Indian Ocean Experiement (INDOEX), Experience sur Site pour Contraindre les Modeles de Pollution atmospherique et de Transport d'Emissions (ESCOMPTE), Mediterranean Intensive Oxidant Study (MINOS), etc. However, most of these experiments provide only a snapshot information and data is available for a limited period only. Additional studies are necessary to obtain a more complete picture of the global distribution of BC aerosol in the atmospheric boundary layer. With this perspective, BC measurements were initiated over Toulon (43.14°N, 6.01°E; 50 m above MSL), an urban coastal location. In this paper, we present the results on the seasonal and diurnal variations of BC and examine their association with anthropogenic activities and prevailing atmospheric processes. 2. Experimental details and database 2.1. Description of the measurement site The monitoring station is located in the Toulon University campus in La Valette. Fig. 1 shows the map of France and Provence-Alpes-Cote-D'Azur (PACA) region (top panel). The street map of Toulon is shown in the bottom panel of Fig. 1 with the station shown as black filled circle. The station is situated ~8 km away from the urban city of Toulon in the west and ~5 km from the Mediterranean coastline in the south. The district of Toulon consists of a total population of over half a million. The station is about 200 m away from a light traffic road and ~750 m away from the heavy traffic Highway A57 with a daily traffic density of N60,000 vehicles per day. Other than the traffic, there are no significant industrial and agricultural combustion activities for several kilometres around the station and hence the BC values are more representative of the region (at least in a meso-scale). 2.2. Instruments used 2.2.1. Black carbon mass concentrations Black carbon mass concentrations were continuously recorded using a 7-channel Aethalometer (Model AE-31, Magee Scientific). The Aethalometer measures the attenuation of light beam at seven different wavelengths viz. 370, 470, 520, 590, 660, 880 and 950 nm. Of these, the 6th Channel (centered at 880 nm) is considered as the standard channel for BC measurements, because BC is the principal absorber of light at this wavelength and other aerosol components have negligible absorption. The results reported in this contribution are mainly based on the measurements at 880 nm. Aethalometer measurements are based on light absorption properties of black carbon. The mass of black carbon accumulated on a quartz fibre tape is obtained by optical attenuation measurement and subsequent conversion using the calibration factor (specific attenuation cross-section, σ). More details on the instrument and the principle of operation are given elsewhere (Hansen et al., 1984; Weingartner et al.,

2003). Although the value of σ remains controversial, data obtained using Aethalometer satisfactorily describes concentration levels and trends in the urban atmosphere (Hansen and Novakov, 1990; Liousse and Cachier, 1992; Babu and Moorthy, 2002; Latha and Badarinath, 2005). In our present study, the value of σ was assumed to be 16.6 m2 g− 1 (as recommended by the manufacturer) for estimating the BC mass concentrations. As this value of σ is either comparable or higher than those used in different environments (Petzold et al., 1997; Sharma et al., 2002), the BC concentrations reported in this study represent minimum values. The other sources of uncertainty in BC mass concentrations using an aethalometer arise from instrumental noise, flow rate, filter spot area and detector response (Corrigan et al., 2006). Taking into account all these effects and the variations in σ, the overall uncertainty in the reported BC mass concentrations is estimated to be within ±10%. The aethalometer was installed in a small weather shelter and equipped with a sharp cut cyclone (Model SCC 1.829, BGI Inc.), which removed particles larger than 2.5 μm of particle aerodynamic diameter at a sample flow rate of 5 LPM. The inlet was located at a height of ~1.5 m above the ground level. In our present study, the aethalometer was configured to run at a flow rate of 4 LPM (which corresponds to an upper cut-off diameter of 3.24 μm), and the measurement cycle (time base) was set at 5 min. 2.2.2. Particle number concentration A Scanning Mobility Particle Sizer (SMPS Model 3936, TSI Inc.) consisting of a differential mobility analyzer (DMA Model 3080) and condensation particle counter (CPC Model 3010) was also being used at the measurement site. The SMPS provided measurements of near-surface aerosol number concentration and size distribution covering the size range between 0.01 and 0.41 μm. The data with the SMPS was acquired at 15 min intervals. 2.2.3. Meteorological parameters Local meteorological parameters such as air temperature, relative humidity, wind speed, wind direction, atmospheric pressure, global solar radiation, and rain intensity were measured at 1 min intervals using instruments and sensors onboard a tower at the measurement site co-located near the Aethalometer. The wind sensors were installed on the top of the tower (10 m above the ground level), whereas the other sensors were placed at ~ 1.5 m from the ground level. All these meteorological sensors and aerosol instruments were part of the monitoring station proposed under the Atmospheric Composition Change European Network Excellence (ACCENT) program (Despiau et al., 2005). 2.2.4. Traffic density Complementary information on the traffic density (number of motor vehicles per hour) at the Highway A57 (shown on the bottom panel of Fig. 1) was obtained to examine the influence of traffic on BC concentrations. It should however be noted that the traffic counts took both directions into consideration and did not differentiate the vehicles by type (truck, car, etc) or by the fuel they use (gasoline or diesel).

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

29

Fig. 1. The top panel shows the map of France and Provence-Alpes-Cote-D'Azur (PACA) region. The bottom panel shows the street map of Toulon with the observation site shown as black filled circle. The major highway A57 where regular measurements of traffic density are carried out is also shown.

2.3. Database The first measurements using Aethalometer were initially carried out on a campaign mode spanning about 20 days during June–July 2005. Beginning from the mid of

October 2005, continuous measurements were made until the end of October 2006. The data collected for 14 months during the period June 2005 to October 2006 (spanning more than 400 days) have been used for the present study.

30

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

3. Local meteorology Fig. 2 shows the meteorological conditions prevailing during the study period. The values obtained at hourly intervals are smoothed using a 24 h running mean. The frequency of occurrence of wind speed and wind direction for the more representative months is shown in Fig. 3. The winds are mostly weak (b3 m s− 1) and either from the east or west during most of the study period. However, in several occasions during the study period, the station was under the influence of mistral phenomenon. The mistral is characterised by severe winds that develops along the Rhone valley and influences the coast of south-eastern France and the western Mediterranean

climate (Drobinski et al., 2005). It brings clear skies, cold and dry continental air. As shown in Fig. 3, the mistral situations are generally characterised by very high wind speeds (N5 m s− 1) blowing from the westerly and north-westerly direction, followed by abrupt decrease in the relative humidity. Two typical mistral situations are indicated in Fig. 2. It was observed that during the study period, the frequency of occurrence of westerly winds (associated with the mistral phenomena) were higher during the months of March and August (Fig. 3). The relative humidity was relatively high (~60–85%) from September through April. However, dry conditions (RH b 65%) prevailed during the later period (May to August). There were isolated events of rainfall during the autumn and

Fig. 2. The panels from top to bottom show the temporal variation of air temperature, RH, wind speed, and wind direction prevailing at the observation site during the study period. The data obtained at hourly intervals are shown in grey and the smoothed data (24 h running mean) are shown as bold (dark) lines. No smoothing was performed for the wind direction. The vertical lines with arrows indicate two typical cases of mistral situations.

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

31

Fig. 3. Occurrence of wind direction and wind speed for the representative months of the study period.

winter months, whereas the summer (June to August) remained mostly dry. 4. Results and discussion 4.1. Temporal (annual) variations of BC Fig. 4 shows the daily-mean temporal variations of BC mass concentration during the entire study period. BC exhibits

considerable daily and annual variations. The daily data in a particular month are grouped together and the mean, standard deviation and standard error are estimated. The annual variations of BC concentrations are shown in Fig. 5. The monthly-mean BC concentrations varied between ~300 and ~1000 ng m− 3. Occurrences of high daily-mean BC concentrations (N1500 ng m− 3) were observed mostly during the winter months (November to February), followed by lower concentrations during spring, and summer months (April to August).

32

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

Highest monthly mean BC concentrations (1026.3 ng m− 3) were recorded in December and lowest concentration (303.6 ng m− 3) in August. An examination of the annual variation of monthly mean surface winds (Fig. 5) show that during the months of March and August, the average winds were higher compared to the other months. Hence the large decrease in the BC concentrations from February to March and during August can be very well attributed to the prevailing high wind speeds. Even though very high BC concentrations occurred during the winter months, on several occasions, very low concentrations (b300 ng m− 3) were also observed (Fig. 4). These can be attributed to the increase in the wind speeds associated with the mistral phenomenon, which was frequently encountered during the study period (Fig. 2). Hence, even though there is continuous BC emission in the atmosphere, the high wind speeds associated with the mistral phenomenon effectively disperse the aerosols, thereby causing a decrease in the BC concentration. More detailed results on the effect of wind speed on the BC concentration will be described later (Section 4.5). Being an urban location and since the observation site is in the close proximity of a Highway, traffic was expected to be a significant source of BC. Fig. 6 shows the variation of monthly mean traffic density. It can be seen that the traffic density also shows a seasonal dependence, with higher values being encountered during the summer months, and low values during winter months. This is in contrast to the seasonal variation of BC, which shows higher concentrations during winter and lower concentrations during summer months. Comparing Figs. 5 and 6, the high BC concentrations observed during the winter season cannot be attributed to traffic. Since traffic does not seem to explain the observed high BC concentrations in winter, other factors are explored. It can be seen from Fig. 2 (topmost panel) that with the arrival of the winter season (November onwards), the ambient tempera-

ture decreases gradually. It can be expected that the local population starts using space heating (which uses conventional fuels that emit BC). Hence, the higher BC concentrations encountered during winter season can be partly attributed to the enhanced BC emissions due to the usage of space heating. In Vienna, Hitzenberger et al. (1996) have also reported high BC concentrations during winter as compared to summer, and they attributed this partly to the increase in the BC emissions related to space heating. Therefore, the increase in the BC emissions (related to space heating) during winter seems to be one of the important factors in contributing to the high concentrations observed in the present study. The daily-averaged BC concentration for the entire study period (spanning over 400 days) was found to vary between 99.1 and 2522.6 ng m− 3, with a mean value of 642.02 ng m− 3 and standard deviation of 399.07 ng m− 3. These values are lower when compared to other coastal, semi-urban and urban locations (Table 1), such as Trivandrum (semi-urban, coastal location on the southwest tip of India; Babu and Moorthy, 2002), Goa (an semi-urban station in west coast of India; Leon et al., 2001), Voerde-Spellen (a rural site in Germany; Kuhlbusch et al., 2001), Uniontown (a semi-urban site in Pittsburgh, USA; Allen et al., 1999), and Fort Meade (a suburban location in Maryland, USA; Chen et al., 2001). The BC concentrations measured over Marseille (urban location over the French Mediterranean coast) during the ESCOMPTE campaign in June 2001 were found to be high at about 2720 ng m− 3 (Despiau et al., Personal communication). Based on the measurements using the thermal method during MINOS campaign, BC concentrations at Finokalia (marine boundary layer site located in the eastern Mediterranean in Greece) were found to be 1090 ± 360 ng m− 3 (Sciare et al., 2003). As seen from Table 1, the BC concentrations encountered in the present study are comparable to the values reported for La Reunion (an island station in Southern

Fig. 4. Day-to-day variation of daily-average BC mass concentration at Toulon during the study period. The vertical bars over the measurement points represent the standard error. The mean for the entire study period is also shown by the dashed line.

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

33

Fig. 5. Annual variations of BC concentration. The BC values for identical months in 2005 and 2006 are averaged. The annual variations of wind speed is also shown (right axis). The vertical bars over the measurement points represent the standard error. The dotted line indicates the average BC concentrations for the entire study period.

Hemisphere over Indian Ocean; Bhugwant et al., 2000) and the recently reported values for Nainital, a high altitude station in Central Himalayas by Pant et al. (2006). Table 1 does not provide a thorough comparison, though it only reflects the differences in magnitude of various BC sources and it should be used as an indication for the range of BC values encountered in the present study. It should be noted that the differences in the BC concentrations may result from different meteorological conditions, different times of the year, and differences in the sampling site characteristics, particle size

and sampling methods used and hence direct comparison of BC concentrations between various locations is quite difficult. 4.2. Diurnal variations of BC Besides the temporal (annual) variations, BC concentration also exhibits a pronounced diurnal variation. Typical diurnal variation of BC is shown in Fig. 7 (topmost panel) for the month of February 2006. The diurnal variation of fine particle (0.01–0.41 μm) concentration measured using SMPS is also

Fig. 6. Annual variation of traffic density obtained from Highway A57 (shown in Fig. 1). The vertical bars over the measurement points represent the standard error.

34

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

Table 1 Comparison of BC with other locations BC conc. (μg m− 3)

Location

Environment

Study period

Range

Mean

Toulon, France (43.14°N, 6.01°E)

Semi-urban, coastal

Jun. 2005–Oct. 2006

0.1–2.5

Marseille, France

Urban, coastal

Jun. 2001

1.1–4.0

0.95 (winter) 0.45 (summer) 2.72

Paris, France La Réunion, France (21.5°S, 55.5°E) Vienna, Austria (48.22°N, 16.35°E) Helsinki, Norway Voerde-Spellen, Germany Finokalia, Greece (35.32°N, 25.83°E) Uniontown, USA Trivandrum, India (8.55°N, 77.0°E)

Urban Semi-urban, island Urban Urban Rural Marine boundary layer Sub-urban Semi-urban, coastal

Aug.–Nov. 1997 Nov. 1996–Oct. 1998 Oct.–Dec. 1994 Nov. 1996–Jun. 1997 Sep.–Oct. 1997 Jul.–Aug. 2001 Summer 1995 Aug. 2000–Oct. 2001

1.1–49 0.27–0.65 – 1.2–1.5 – – 0.5–9.0 0.4–8.0

Goa, India (15.75°N, 73.13°E) Nainital, India (29.4°N, 79.5°E)

Semi-urban, coastal Remote, high altitude

Dec. 1998–Mar. 1999 Dec. 2004

1.0–5.0 0.2–2.7

shown (right hand axis in top panel of Fig. 7) for comparison. The monthly mean diurnal variation of traffic density and meteorological parameters (temperature, global solar radiation, wind speed and wind direction) are shown in the bottom panels. It can be very clearly seen that BC and fine particle concentration shows almost identical diurnal variations, which is expected, as BC is mostly comprised of fine particles. There is a gradual increase in BC mass and fine particle concentrations in the morning attaining a sharp peak at ~ 09:00 h local time (LT). The morning build up of local anthropogenic activities associated with the go-to-work traffic rush (Fig. 7) is responsible for this peak. In addition to this, the morning peak might also be associated with the fumigation effect in the boundary layer, which brings aerosols from the nocturnal residual layer shortly after the sunrise (Stull, 1988). As the day advances, the BC concentration continuously decreases and reaches the diurnal minimum around 13:00 h LT, even though the traffic density does not show a significant decrease. As the day advances, increased solar heating leads to increased turbulent effects and a deeper boundary layer, leading to faster dispersion of aerosols and hence a dilution of BC concentration occurs near the surface. An examination of meteorological data shows that the diurnal variations of the wind speed reaches a peak during the mid-afternoon (5th panel in Fig. 7). The well-developed boundary layer and increase in the wind speed is likely responsible for the observed low concentrations in the afternoon hours. Here it is to be noted that BC is not totally lost from the atmosphere, it is only re-distributed over a large spatial extent by the boundary layer dynamics. The BC concentration continues to be lower until 17:00 h and thereafter it slowly tends to increase to reach the secondary maximum at ~20:00 h LT. This increase in the evening time BC concentrations can be partly attributed to the evening traffic rush. Additionally, shallower nocturnal boundary layer and lower wind speeds during night (bottommost panel in Fig. 7) leads to a rapid reduction in the ventilation effects and consequently confine the aerosols causing the secondary peak. As night advances, there is a progressive and strong reduction in the traffic density (2nd panel in Fig. 7) and BC generation, while the existing ones closer to the surface are partly lost by sedimentation. Thus the concentration gradu-

14.3 – 8.8 1.38 2.0 1.09 2.0 1.5 (wet) 5.0 (dry) – 1.4

Reference

Present study ESCOMPTE, Despiau et al. (Personal communication) Ruellan and Cachier, 2001 Bhugwant et al., 2000 Hitzenberger et al., 1996 Pakkanen et al., 2000 Kuhlbusch et al., 2001 Sciare et al., 2003 Allen et al., 1999 Babu and Moorthy, 2002 Leon et al., 2001 Pant et al., 2006

ally decreases towards the early morning (between 02:00 and 06:00 h LT). The amplitude of the diurnal variations is typically ~4, and is of the order of magnitude observed at other continental locations. Similar diurnal variations of BC are also observed at other continental locations (Allen et al., 1999; Bhugwant et al., 2000; Ruellan and Cachier, 2001; Babu and Moorthy, 2002; Latha and Badarinath, 2005). However, a different pattern with a midday peak was observed at Nainital, a high altitude location (~ 2 km above MSL) in Central Himalayas in India (Pant et al., 2006) and Look Rock, a background complex terrain site (550 m above MSL) in US (Tanner et al., 2005), and was ascribed to the upward transport of BC from the nearby plains, due to the elevation in the day-time boundary layer height. Based on the Aethalometer and Lidar measurements in a field campaign at Mexico City during April–May 2003, Barnard et al. (2005) have also observed pronounced diurnal variations in the BC mass concentrations, with very high concentrations in the morning and reduced concentrations in the afternoon. They attributed the diurnal variations mainly to the emission rate of BC and the mixing of BC upwards by the development of the convective boundary layer during the day, as inferred from the lidar backscatter data. Making measurements at a high flow road in Paris during August– November 1997, Ruellan and Cachier (2001) also observed a diurnal variation in BC concentrations. They attribute the variations to the traffic intensity and regime, as well as to the boundary layer height. 4.3. Seasonal variations of BC diurnal cycle In order to examine the seasonal changes on the BC diurnal evolution, the BC concentrations are examined for the month of July 2006 (representing the summer season) in Fig. 8, along with the traffic density and meteorological parameters. Comparing Figs. 7 and 8, it can be clearly seen that there are significant changes in the diurnal variations of BC from February to July. As compared to February, the morning peak is less pronounced in July and the evening peak is completely absent. It is interesting to note that there is slight enhancement in the BC concentrations during the mid-night hours

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

35

Fig. 7. The panels from top to bottom shows the diurnal variation of (1) BC and fine particle concentration (topmost panel), (2) traffic density (2nd Panel), (3) global solar radiation (3rd panel), (4) surface air temperature (4th panel), (5) wind speed (5th panel), and (6) wind direction (bottommost panel), for the month of February 2006. The vertical bars through the measurement points in the various panels represent the standard error. Dashed vertical lines in the top panel show the local sunrise and sunset times for 15th of the month. In all the panels, the diurnal variation of BC is also shown for comparison.

in July. This increase can be attributed to the increase in the mid-night traffic density in July, as compared to February (top panel of Fig. 9). In order to examine this aspect in more detail, the relative deviation in the traffic density (∇) is further estimated as  j¼

T:DJul −T:DFeb T:DJul

  100

ð1Þ

where T.D is the traffic density. Fig. 9 (bottom panel) shows the diurnal variation of ∇. It can be seen from the figure that there is a significant increase (as high as ~ 70%) in the mid-night traffic density in July, as compared to February. This increase in the night-time traffic is expected for summer months, as more people are involved in outdoor leisure activities. Any increase in the traffic density (leading to the increase in BC emissions) in the night becomes important, as there is little dispersion

36

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

owing to low wind speeds and shallow boundary layer, thereby causing an enhancement in the BC concentrations as noticed in Fig. 8. In order to examine the seasonal variations of BC diurnal cycle in further detail, the hourly data for one entire month are grouped together. Further, the data for the identical months in 2005 and 2006 are also combined together. The average, standard deviation and standard error are estimated for all the monthly data (January to December). Fig. 10 shows

the monthly mean diurnal variation of BC mass concentration for the various months (January to December). There is marked seasonal dependence on the diurnal variations of BC. It can be seen that the BC concentrations during the winter season (November to February) show stronger diurnal variations, compared to the spring and summer season. This can be mainly attributed to the seasonal changes in the boundary layer dynamics. The winter-time boundary layer is shallower compared to its summer counterparts, resulting in

Fig. 8. Same as for Fig. 7, but for the month of July 2006.

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

the increased confinement of aerosols, causing an increase in the BC concentrations in winter. It is further observed that the secondary (evening) peak is prominent only during the winter season (November through February), gradually diminishes during spring and autumn season, and finally it almost disappears during the summer months (June to August). The secondary (evening) peak is a characteristic feature of the winter months. There may be several factors responsible for it. One of the facts is that the day-time boundary layer evolution is confined to shorter time scales in winter (as can be inferred from the diurnal variation of temperature and global solar radiation shown in Fig. 7), as compared to summer months, when the boundary layer remains evolved for much longer duration of the day and becomes shallower much later in the night (Fig. 8). Furthermore, by comparing Figs. 7 and 8, it can also be seen that the regimes of high wind speeds (diurnal maximum) are limited to shorter time span in February, unlike in July when high wind speeds are experienced for extended periods of the day. In addition to this, as discussed earlier, it can also be expected that during the winter months, there can be additional BC emissions due

37

to the usage of space heating, particularly during the nighttime when the temperatures are very low. All these factors (shallow boundary layer, low winds, use of space heating) can be held responsible for the occurrence of secondary (evening) peak observed during the winter season in the present study. It is also interesting to note that, there is a shift in the morning peak from ~ 10:00 h LT in December to ~08:00 h LT in August. This seasonal shift in the morning peak can be associated with the change in the local sunrise time from winter (~ 06:00 h LT) to summer (07:30 h LT). As described earlier, the morning peak in BC concentration over continental locations is associated with fumigation effect in the boundary layer and time of its occurrence is associated with local sunrise (Stull, 1988). It also can be noted from Fig. 10 that the morning rise in the BC concentrations occurs almost simultaneously, irrespective of the season. This can be due to the fact the morning increase in the traffic density occurs almost simultaneously and doesn't show any significant changes from month-tomonth. However, the seasonal difference in the morning peak BC concentrations can be attributed due to the fact that the atmospheric processes (boundary layer height and winds)

Fig. 9. The top panel shows the diurnal variation in traffic density for the months of February (continuous line) and July 2006 (dotted line). On the bottom panel is shown the relative deviation in the traffic density.

38

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

Fig. 10. Monthly mean diurnal variation of BC mass concentration from January to December. Each point represents the average concentration for that hour for the entire month. BC values for identical months in 2005 and 2006 are averaged.

Fig. 11. Diurnal variation of BC mass concentration (top panels) and traffic density (bottom panels) for weekdays, Saturdays, and Sundays for two different periods: (1) November to February (left panels), and (2) March to October (right panels). Each point represents the average concentration for that hour for the respective days (Weekdays, Saturdays, and Sundays). The vertical bars over the measurement points represent the standard error.

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

rather starts influencing the BC concentrations at different time of the day, depending on the season. Based on the Aethalometer measurements at Hyderabad, a tropical-urban location in India, Latha and Badarinath (2005) have observed seasonal variations of BC with high concentrations during dry season and low concentrations during the monsoon season. Based on the continuous BC measurements using Aethalometer, Babu and Moorthy (2002) have also observed significant seasonal changes in the diurnal BC concentrations at Trivandrum, a tropical coastal location in India. They attributed the observed variations to the seasonal changes in the synoptic meteorology and boundary layer dynamics. 4.4. Weekday/weekend effect Since traffic is a major source of BC, we decided to closely examine the average BC concentrations with respect to the traffic density on various days of the week. It was observed that the BC concentrations during weekdays (Monday through Friday), Saturdays and Sundays showed similar variations and therefore the entire data were grouped

39

separately as Weekdays, Saturdays and Sundays. Further, we have already seen in our earlier discussion that the diurnal variations of BC showed changes from month to month. The BC diurnal variations showed presence of strong evening peak only during the winter months of November to February, whereas for the rest of the period (March to October), the evening peak was either insignificant or almost absent. Taking these aspects into consideration, we examined the weekday/ weekend effect separately for these two periods. Fig. 11 shows the diurnal variation of hourly average BC concentration (top panels) and diurnal variation of Traffic density (bottom panels), separately for weekdays, Saturdays, and Sundays for two different cases: (1) November to February (left two panels), and (2) March to October (right two panels). It is clearly seen that the traffic density during Saturday and Sunday displays quite different patterns, as compared to the weekday pattern for both the periods. There is a sharp increase in the morning traffic density to reach the peak at ~ 08:00 h LT during the weekdays. However, the increase is gradual for Saturdays and slower for Sundays. Associated with this, the peak in traffic density is slightly shifted from ~ 08:00 h LT (during weekdays) to ~ 12:00 h LT

Fig. 12. Variation of daily average BC concentration with wind speed, shown separately for two different periods: (a) November to March (bottom panel) and, (b) March to October (top panel). The ordinate is given in logarithmic scale. The straight line is the linear fit to the measurement points.

40

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41

(Saturdays and Sundays). This is irrespective of the season, except that the traffic density is slightly higher during the spring and summer months (March to October) during all the days (weekdays, Saturdays and Sundays). The diurnal variations of BC concentrations (shown in the top two panels of Fig. 11) shows that the BC during weekdays depicts higher concentrations and stronger diurnal variations, compared to the concentrations observed during Saturdays and Sundays. The morning BC peak is highly pronounced during the weekdays and, it becomes less prominent during Saturdays and becomes almost insignificant during Sundays. This is irrespective of the season, except that the BC concentrations are higher during winter, compared to summer. Furthermore, it can also be observed that the late night BC concentrations during weekends (Saturdays and Sundays) are relatively higher for the period March to October (top right panel in Fig. 11). As discussed earlier (Section 4.3), this can again be attributed to the increase in the night-time traffic density during the weekends, as more people stay outdoors for leisure activities during late nights in weekends (particularly Friday and Saturday evenings) during the summer. Except for the midnight, the BC concentrations on Sundays are the lowest compared to the weekday and Saturday values. 4.5. Effect of wind speed As we have already seen in our earlier discussion, wind speed is an important factor in affecting the diurnal variations of BC. In Fig. 12, we examine the relationship between the daily averaged BC concentrations and the prevailing wind speed for the two respective periods: (1) November to February, and (2) March to October, following the criteria described in Section 4.4. Notwithstanding a fair amount of scatter, BC concentration decreases with increase in the wind speed in both the cases. The correlation is better for the winter data (R2 N 0.8), as compared to the later period (0.67). Further the slope and intercept for the winter data was −0.159 and 2028.3 respectively and for the later period, it was −0.09 and 908.2 respectively. The steeper slope and the higher intercept value for the winter data can be mainly attributed to the high BC concentrations observed, which is typical to the season. The increase in wind speed causes an increase in the ventilation effects, thereby dispersing the aerosols in the ambient air and consequently causes a decrease in the observed BC concentrations. Comparing the daily mean elemental carbon concentration and wind speed at Birmingham during 1995, Harrison et al. (1997) have found quite a similar correlation. Making Aethalometer measurements at Helsinki during 1996–1997, Pakkanen et al. (2000) have also observed a strong dependence between the daytime BC concentration and average wind speeds up to 2.5 m s− 1. However, they found no clear dependence of wind speed on BC concentration for wind speeds N2.5 m s− 1. 5. Summary and conclusions Long-term measurements of BC mass concentration made for about 14 months during 2005–2006 from an urban, coastal station Toulon were used to characterize its seasonal and diurnal variation and its association to local traffic sources and prevailing meteorology.

1. The average BC concentration at Toulon is less than the values reported for the other coastal, semi-urban and urban locations and is comparable to that for a remote location. 2. BC shows significant temporal variations, both diurnal as well as annual. Diurnal variations show a morning peak and an afternoon decrease. The diurnal variations are more pronounced during winter season, with the presence of secondary evening peak. These diurnal variations are found to be associated with local traffic density (even though its impact was not directly related one-to-one), with the wind speed and boundary layer dynamics showing more significant influence. 3. The annual variations are significant (amplitude N3) and strongly linked to the intra-seasonal changes in the boundary layer dynamics. High concentrations are seen from October through February and lower values during March to September. 4. Local traffic sources, wind speed and boundary layer dynamics are the most important parameters influencing the BC concentration over Toulon. Acknowledgements The present work is supported by French CNRS, INSU, and PNTS Projects. The authors would like to acknowledge the staff of Direction départementale de l'équipement (DDE) Toulon for providing the traffic density data. One of the authors (AS) would also like to acknowledge the French Ministry for Education & Research and the Université du Sud Toulon-Var for the financial support. References Ackerman, A.S., Toon, O.B., Stevens, D.E., Heymsfield, A.J., Ramanathan, V., Welton, E.J., 2000. Reduction of tropical cloudiness by soot. Science 288, 1042–1047. Allen, G.A., Lawrence, J., Koutrakis, P., 1999. Field validation of a semicontinuous method for aerosol black carbon (Aethalometer) and temporal patterns of summertime hourly black carbon measurements in southwestern PA. Atmospheric Environment 33, 817–823. Babu, S.S., Moorthy, K.K., 2002. Aerosol black carbon over a tropical coastal station in India. Geophysical Research Letters 29, 2098. doi:10.1029/ 2002GL015662. Barnard, J.C., Kassianov, E.I., Ackerman, T.P., Frey, S., Johnson, K., Zuberi, B., Molina, L.T., Molina, M.J., Gaffney, J.S., Marley, N.A., 2005. Measurements of black carbon specific absorption in the Mexico City metropolitan area during the MCMA 2003 field campaign. Atmospheric Chemistry and Physics Discussion 5, 4083–4113. Bhugwant, C., Cachier, H., Bessafi, M., Leveau, J., 2000. Impact of traffic on black carbon aerosol concentration at la Réunion Island (Southern Indian Ocean). Atmospheric Environment 34, 3463–3473. Chen, L.W.A., Doddridge, B.G., Dickerson, R.R., Chow, J.C., Mueller, P.K., Quinn, J., Butler, W.A., 2001. Seasonal variations in elemental carbon aerosols, carbon monoxide and sulfur dioxide: implications for sources. Geophysical Research Letters 28, 1711–1714. doi:10.1029/2000GL012354. Cooke, W.F., Jennings, S.G., Spain, T.G.,1997. Black carbon measurements at Mace Head, 1989–1996. Journal of Geophysical Research 102, 25339–25346. Corrigan, C.E., Ramanathan, V., Schauer, J.J., 2006. Impact of monsoon transitions on the physical and optical properties of aerosols. Journal of Geophysical Research 111, D18208. doi:10.1029/2005JD006370. Despiau, S., Piazzola, J., Missamou, T., 2005. A French Mediterranean site for atmospheric studies and surveillance. First ACCENT Symposium, Urbino, Italy. Drobinski, P., Bastin, S., Guenard, V., Caccia, J.-L., Dabas, A.M., Delville, P., Protat, A., Reitebuch, O., Werner, C., 2005. Summer mistral at the exit of the Rhone valley. Quarterly Journal of Royal Meteorological Society 131, 353–375. Fruin, S.A., Winer, A.M., Rodes, C.E., 2004. Black carbon concentrations in California vehicles and estimation of in-vehicle diesel exhaust particulate matter exposures. Atmospheric Environment 38, 4123–4133. Hansen, A.D.A., Novakov, T., 1990. Real time measurements of aerosol black carbon during the carbonaceous species methods comparison study. Aerosol Science and Technology 12, 194–199.

A. Saha, S. Despiau / Atmospheric Research 92 (2009) 27–41 Hansen, J., Nazarenko, L., 2003. Soot climate forcing via snow and ice albedos. Proceedings of the National Academy of Science 101, 423–428. Hansen, A.D.A., Rosen, H., Novakov, T., 1984. The aethalometer — an instrument for the real-time measurement of optical absorption by aerosol particles. Science of the Total Environment 36, 191–196. Hansen, J., Sato, M., Reudy, R., Lacis, A., Oinas, V., 2000. Global warming in the 21st Century: an alternative scenarios. Proceedings of the National Academy of Sciences 97, 9875–9880. Harrison, R.M., Deacon, A.R., Jones, M.R., Appleby, R.S., 1997. Sources and processes affecting concentrations of PM10 and PM2.5 particulate matter in Birmingham (UK). Atmospheric Environment 31, 4103–4117. Haywood, J.M., Shine, K.P., 1997. Multi-spectral calculations of the radiative forcing of tropospheric sulphate and soot aerosols using a column model. Quarterly Journal of Royal Meteorological Society 123, 1907–1930. Hitzenberger, R., Fohler-Norek, C., Dusek, U., Galambos, Z., Sidla, S., 1996. Comparison of recent (1994) black carbon (BC) data with those obtained in 1985 and 1986 in the urban area of Vienna, Austria: implications for future environmental policy decisions. Science of the Total Environment 189/190, 275–280. Horvath, H., 1993. Atmospheric light absorption — a review. Atmospheric Environment 27A, 293–317. Kuhlbusch, T.A.J., John, A.C., Fissan, H., 2001. Diurnal variations of aerosol characteristics at a rural measuring site close to the Ruhr-Area, Germany. Atmospheric Environment 35, S13–S21. Latha, K.M., Badarinath, K.V.S., 2005. Seasonal variations of black carbon aerosols and total aerosol mass concentrations over urban environment in India. Atmospheric Environment 39, 4129–4141. Leon, J.-F., Chazette, P., Dulac, F., Pelon, J., Flamant, C., Bonazzola, M., Foret, G., Alfaro, S.C., Cachier, H., Cautenet, S., Hamonou, E., Gaudichet, A., Gomes, L., Rajot, J.-L., Lavenu, F., Inamdar, S.R., Sarode, P.R., Kadadevarmath, J.S., 2001. Large-scale advection of continental aerosols during INDOEX. Journal of Geophysical Research 106 (D22), 28427–28440. doi:10.1029/ 2001JD900023. Liousse, C., Cachier, H., 1992. Measurement of black carbon aerosols in the atmosphere of two different source regions: real-time data for the Paris region and a savanna site of the Ivory Coast. Environmental Technology 13, 959–967. Menon, S., Hansen, J., Nazarenko, L., Luo, Y.F., 2002. Climate effects of black carbon aerosols in China and India. Science 297, 2250–2253. Pakkanen, T.A., Kerminen, V.-M., Ojanen, C.H., Hillamo, R.E., Aarnio, P., Koskentalo, T., 2000. Atmospheric black carbon in Helsinki. Atmospheric Environment 34, 1497–1506.

41

Pant, P., Hegde, P., Dumka, U.C., Sagar, R., Satheesh, S.K., Moorthy, K.K., Saha, A., Srivastava, M.K., 2006. Aerosol characteristics at a high-altitude location in central Himalayas: optical properties and radiative forcing. Journal of Geophysical Research 111, D17206. doi:10.1029/2005JD006768. Penner, J.E., 1995. Carbonaceous aerosols influencing atmospheric radiation: black and organic carbon. In: Charlson, R.J., Heintzenberg, J. (Eds.), Aerosol Forcing of Climate. John Wiley and Sons, Chichester, pp. 91–108. Petzold, A., Kopp, C., Niessner, R., 1997. The dependence of the specific attenuation cross section on black carbon mass fraction and particle size. Atmospheric Environment 31, 661–672. Pope, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K., Thurston, G.D., 2002. Lung cancer, cardiopulmonary mortality and long-term exposure to fine particulate air pollution. Journal of the American Medical Association 287, 1132–1141. Ruellan, S., Cachier, H., 2001. Characterization of fresh particular exhausts near a Paris high flow road. Atmospheric Environment 35, 453–458. Schwartz, S.E., 1996. The white house effect-Short wave radiative forcing of climate by anthropogenic aerosols — an overview. Journal of Aerosol Science 27, 359–382. Sciare, J., Cachier, H., Oikonomou, K., Ausset, P., Sarda-Esteve, R., Mihalopoulos, N., 2003. Characterization of carbonaceous aerosols during the MINOS campaign in Crete, July–August 2001: a multi-analytical approach. Atmospheric Chemistry and Physics 3, 1743–1757. Sharma, S., Brook, J.R., Cachier, H., Chow, J., Gaudenzi, A., Lu, G., 2002. Light absorption and thermal measurements of black carbon in different regions of Canada. Journal of Geophysical Research 107 (D24), 4771. doi:10.1029/ 2002JD002496. Sharma, S., Lavoué, D., Cachier, H., Barrie, L.A., Gong, S.L., 2004. Long-term trends of the black carbon concentrations in the Canadian Arctic. Journal of Geophysical Research 109, D15203. doi:10.1029/2003JD004331. Stull, R.B., 1988. An introduction to boundary layer meteorology. Kluwer Academic Publishers, Dordrecht. 680 pp. Tanner, R.L., Bairai, S.T., Olszyna, K.J., Valente, M.L., Valente, R.J., 2005. Diurnal patterns in PM2.5 mass and composition at a background, complex terrain site. Atmospheric Environment 39, 3865–3875. Wang, C., 2007. Impact of direct radiative forcing of black carbon aerosols on tropical convective precipitation. Geophysical Research Letters 34, L05709. doi:10.1029/2006GL028416. Weingartner, E., Saathoff, H., Schnaiter, M., Streit, N., Bitnar, B., Baltensperger, U., 2003. Absorption of light by soot particles: determination of the absorption coefficient by means of aethalometers. Journal of Aerosol Science 34, 1445–1463.