1. Measurements on Atmospheric Aerosols and of the Aerosol Optical Depth during 2006 at Uccle, Belgium
Willy Maenhaut1, Wan Wang1, Xuguang Chi1, Nico Raes1, Jan Cafmeyer1,
Anne Cheymol2, Andy Delcloo2, Veerle De Bock2, and Hugo De Backer2
1 Ghent University (UGent), Dept. Anal. Chem., Inst. Nucl. Sci., Proeftuinstraat 86, BE-9000 Gent, Belgium
2 Dept. Observations, Royal Meteor. Inst. (RMI) of Belgium, Ringlaan 3, BE-1180 Uccle, Belgium
1. Introduction
Fig. 1
5. Aerosol chemical mass closure
During 2006, a study undertaken at Uccle (50º48’N, 4º21’E, 100 m asl), near Brussels, to examine the relationship between the vertical column-integrated Aerosol Optical Depth (AOD), the boundary layer aerosol characteristics, and the meteorological parameters, including air mass origin and mixing height
At 2006 meeting in Capetown preliminary results for the first 3 months of this study presented
Here, we present the full data set and examine interrelations between the various data
Mass closure calculations done for the separate PM2.5 and PM10 aerosol, and also for the coarse (PM10-2.5) size fraction, and this for each sampling
As gravimetric PM, data from the PM2.5(N) and PM10(N) collections used
For reconstituting PM: 8 aerosol types considered [Maenhaut et al., 2008]
organic matter (OM), estimated as 1.4 OC
elemental carbon (EC)
ammonium
nitrate
non-sea-salt (nss) sulphate
sea salt
crustal matter
other elements
OC and EC data obtained from the PM2.5(Q) and PM10(Q) samplers
data for the other 6 aerosol types obtained from PM2.5(N) and PM10(N) samplers
Average concentrations of the 8 aerosol types (components) in the PM10
aerosol for each of the four seasons shown in Fig. 3
Percentage contributions of the different aerosol types to the seasonally averaged
PM10 aerosol mass shown in Fig. 4
Secondary inorganic aerosols (SIA), which is the sum of ammonium, nitrate, and
non-sea-salt sulphate, was the major component in each season for both PM2.5
and PM10, in each case followed by OM
In PM2.5
SIA accounted for 59, 67, 46, and 42% of the mean aerosol mass in winter, spring, summer, and fall, respectively
the corresponding percentages for OM were 34, 20, 41, and 32%
2. AOD measurements and aerosol collections
Study site and study period
instruments and samplers set up on roof at RMI
at 17 meter above ground
measurements from 4 Jan. to 2 May and from 13 July to 30 Nov. 2006
AOD
obtained from Brewer ozone spectrophotometer
derived from direct sun observations at 5 isolated wavelengths in the UV-B (306.3, 310.1, 313.5, and nm) [Cheymol and De Backer, 2003]
99 days with at least 1 valid AOD measurement
up to 28 valid AOD data per day for these days; typically 13
Aerosol samplers (Fig. 1)
PM2.5(N): PM2.5 filter holder with 0.4 µm Nuclepore polycarbonate filter
PM10(N): PM10 filter holder with 0.4 µm Nuclepore polycarbonate filter
PM2.5(Q): PM2.5 filter holder with two Whatman QM-A quartz fibre filters
PM10(Q): PM10 filter holder, also with two Whatman QM-A filters
all filters had a diameter of 47 mm
the samplers operated at a flow rate of 17 L per min
the quartz filters had been pre-fired at 550C to remove organic contaminants
the purpose of the second quartz fibre filter in the PM2.5(Q) and PM10(Q) filter holders was to assess artifacts [Turpin et al., 2000]
Aerosol samplings
all samplers operated in parallel
samples collected during the daytime only
only on days with no or few clouds when 50% or more valid AOD data were to be expected
average collection time per sampling: 7.5 hours
109 days with aerosol samplings
Table 1 Overall and seasonal medians and ranges for the daily mean
AOD nm), the sample-averaged mixing height (in m), and the
PM2.5(N) mass (in µg.m–3).
AOD Mixing height PM2.5 mass
median (range) median (range) median (range)
Overall (0.052 – 1.08) (161 – 1840) (3.3 – 70)
Winter (0.122 – 0.60) (161 – 1150) (21 – 70)
Spring (0.084 – 0.90) (650 – 1840) (4.8 – 56)
Summer 0.42 (0.123 – 1.08) (900 – 1770) (4.9 – 23)
Fall (0.052 – 1.04) (300 – 1160) (3.3 – 37)
Fig. 3
Table 2 Overall median concentrations and interquartile ranges in the PM2.5 and PM10 aerosol. OC and EC were obtained from PM2.5(Q) and PM10(Q); ions and elements from PM2.5(N) and PM10(N). Data for the PM are in µg.m–3; all other data are in ng.m–3.
PM PM10
median (interq. range) median (interq. range)
PM (N) (8.7 – 23) (17.8 – 34)
PM (Q) (11.2 – 35) (18.8 – 45)
OC (Q) (690 – 4300) (2500 – 6200)
EC (Q) (350 – 1070) (570 – 1490)
Ammonium (690 – 4300) (740 – 4700)
Nitrate (640 – 5700) (1780 – 8000)
Sulphate (1670 – 5000) (2100 – 5500)
Na (IC) (94 – 380) (250 – 1350)
Mg (IC) (9.6 – 49) (57 – 192)
Al DL (DL – 34) (70 – 230)
Si (48 – 141) (300 – 700)
P DL (DL – 11.8) (13.4 – 30)
Cl (IC) (29 – 500) (185 – 1710)
K (53 – 162) (122 – 250)
K (IC) (33 – 107) (63 – 137)
Ca (48 – 112) (300 – 740)
Ca (IC) (58 – 134) (380 – 810)
Ti (DL – 5.5) (9.9 – 20)
V (DL – 3.1) (1.06 – 4.1)
Mn (2.1 – 8.7) (5.9 – 17.7)
Fe (69 – 200) (280 – 580)
Ni (0.80 – 2.5) (1.28 – 3.6)
Cu (2.8 – 7.1) (8.5 – 17.0)
Zn (15.6 – 78) (23 – 109)
Pb (5.2 – 23) (6.7 – 26)
3. Determination of particulate mass and
chemical analyses
All filters weighed before and after sampling with a microbalance to obtain the particulate mass (PM)
weighings done at 20C and 50% relative humidity
precision of net mass determination for the filter samples is estimated at
5 µg for the Nuclepore filters
30 µg for the Whatman QM-A quartz fibre filters
Quartz fibre filters analysed for organic and elemental carbon (OC and EC)
by a thermal-optical transmission (TOT) technique [Birch and Cary, 1996]
OC on the 2nd quartz fibre filter of PM2.5(Q) or PM10(Q) assumed to be mainly due to adsorption of VOCs (positive artifact), and subtracted from the OC on the 1st filter in order to arrive at artifact-free OC data
Nuclepore filters from the PM2.5(N) and PM10(N) samplers
analysed for 29 elements (from Na to Pb) by particle-induced X-ray emission (PIXE) [Maenhaut and Cafmeyer, 1998]
analysed for major anions and cations by ion chromatography (IC) [Maenhaut et al., 2002]
Fig. 4
4. AOD data, mixing heights, and mass
concentration data
The boundary layer mixing height was estimated for every 3 hours from ECMWF data and the AOD data and mixing heights were averaged over the duration of each filter sampling
The overall and seasonal medians and ranges for the daily mean AOD nm), the sample-averaged mixing height, and the PM2.5(N) mass are given in Table 1
The overall median of the daily mean AOD values nm) was 0.27 (range: 0.05 – 1.08)
the AOD values were clearly higher in spring and summer than in winter and fall
The overall median for the sample-averaged mixing height was 870 m (range: 160 – 1800 m)
it was substantially higher in spring and summer than in fall and lowest in winter
The overall median PM2.5 mass level was 14 μg/m3 (range: 3 – 70 μg/m3)
median much larger in winter than in the other three seasons
high winter levels were due to a strong air pollution episode with easterly winds, low wind speeds, and low mixing heights at the end of January – beginning of February
The overall medians (and interquartile ranges) in PM2.5 and PM10 for the PM and several aerosol species and elements given in Table 2
PM2.5/PM10 mass ratios for the PM: on average, 0.75 ± 0.08, 0.60 ± 0.14, 0.55 ± 0.15, and 0.54 ± 0.12 during winter, spring, summer, and fall, resp.
Time series for the PM10 PM and selected species shown in Fig. 2
Fig. 5
6. Relationship between AOD and boundary
layer PM2.5 mass
The relation between the daily mean AOD values and the PM2.5 and PM10 mass data was poor
The linear regressions lines (when forced through the origin) were
PM2.5 = 42 nm)
PM10 = 64 nm)
the coefficients are comparable to the data given by Schaap et al. [2009], who used AOD values at 550 nm
It was estimated to which extent our daily mean AOD data could be explained by the boundary layer aerosol
To this end, the PM2.5 aerosol mass data were multiplied by the sample-averaged boundary layer height and a mass extinction efficiency of 5 m2/g to obtain estimated AOD values
The scatter plot of these estimated AOD versus the vertical column-integrated daily mean AOD data nm) is shown in Fig. 5
boundary layer aerosols contributed, on average, for 23% to the vertical column-integrated AOD
the contribution for the individual days ranged from 6% to 63%
there was a slight tendency for a larger contribution of boundary layer aerosols during winter than during the other three seasons
Fig. 2
References
Birch, M. E., & Cary, R. A. (1996). Aerosol Sci. Technol. 25,
Cheymol, A., & De Backer, H. (2003). J. Geophys. Res. 108 (D24), 4800, doi: /2003JD
Maenhaut, W., & Cafmeyer, J. (1998). X-Ray Spectrom. 27,
Maenhaut, W., et al. (2002). Nucl. Instr. and Meth. B189,
Maenhaut, W., et al. (2008). X‑Ray Spectrom. 37,
Schaap, M., et al. (2009). Atmos. Chem. Phys. 9,
Turpin, B. J., et al. (2000). Atmos. Environ. 34,
Acknowledgements
We gratefully acknowledge the financial support from the Belgian Federal Science Policy Office (contract MO/34/014) and the Special Research Fund of Ghent University. We are also indebted to Sheila Dunphy for technical assistance.