Vertical profiles of aerosol absorption coefficient from

STOTEN-12587; No of Pages 14 Science of the Total Environment xxx (2011) xxx–xxx

Contents lists available at ScienceDirect

Science of the Total Environment 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 / s c i t o t e n v

Vertical profiles of aerosol absorption coefficient from micro-Aethalometer data and Mie calculation over Milan L. Ferrero a,⁎, G. Mocnik b, B.S. Ferrini a, M.G. Perrone a, G. Sangiorgi a, E. Bolzacchini a a b

POLARIS Research Center, Department of Environmental Sciences, University of Milan-Bicocca, Piazza della Scienza 1, 20126, Milan, Italy Aerosol d.o.o., Kamniska 41, SI-1000 Ljubljana, Slovenia

a r t i c l e

i n f o

Article history: Received 29 August 2010 Received in revised form 27 January 2011 Accepted 11 April 2011 Available online xxxx Keywords: Black carbon Vertical profile Absorption coefficient Aethalometer Optical particle counter Particulate matter Air pollution

a b s t r a c t Vertical profiles of aerosol number–size distribution and black carbon (BC) concentration were measured between ground-level and 500 m AGL over Milan. A tethered balloon was fitted with an instrumentation package consisting of the newly-developed micro-Aethalometer (microAeth® Model AE51, Magee Scientific, USA), an optical particle counter, and a portable meteorological station. At the same time, PM2.5 samples were collected both at ground-level and at a high altitude sampling site, enabling particle chemical composition to be determined. Vertical profiles and PM2.5 data were collected both within and above the mixing layer. Absorption coefficient (babs) profiles were calculated from the Aethalometer data: in order to do so, an optical enhancement factor (C), accounting for multiple light-scattering within the filter of the new microAeth® Model AE51, was determined for the first time. The value of this parameter C (2.05 ± 0.03 at λ = 880 nm) was calculated by comparing the Aethalometer attenuation coefficient and aerosol optical properties determined from OPC data along vertical profiles. Mie calculations were applied to the OPC number–size distribution data, and the aerosol refractive index was calculated using the effective medium approximation applied to aerosol chemical composition. The results compare well with AERONET data. The BC and babs profiles showed a sharp decrease at the mixing height (MH), and fairly constant values of babs and BC were found above the MH, representing 17 ± 2% of those values measured within the mixing layer. The BC fraction of aerosol volume was found to be lower above the MH: 48 ± 8% of the corresponding ground-level values. A statistical mean profile was calculated, both for BC and babs, to better describe their behaviour; the model enabled us to compute their average behaviour as a function of height, thus laying the foundations for valid parametrizations of vertical profile data which can be useful in both remote sensing and climatic studies. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Aerosols affect the climate due to their ability to scatter and absorb sunlight (direct effect), and to act as cloud condensation nuclei (CCN), thus modifying the lifetime of clouds, droplet size and precipitation rate (indirect effect) (Ramanathan and Feng, 2009; IPCC, 2007; Koren et al., 2004, 2008; Kaufman et al., 2002; Ramanathan et al., 2001; Penner et al., 2001; Ackerman et al., 2000). Different aerosol species (black carbon, sulfate, organics and dust) contribute to surface dimming (Ramanathan and Carmichael, 2008; IPCC, 2007); the absorbing species (i.e. black carbon, dust) can absorb atmospheric sunlight thus “masking” (and cooling) the surface whilst warming the atmosphere in the process (Ramanathan and Carmichael, 2008). This can affect

⁎ Corresponding author. Tel.: + 39 0264482814; fax: + 39 0264482839. E-mail address: [email protected] (L. Ferrero).

atmospheric thermal structure and regional circulation systems such as monsoons (Ramanathan and Feng, 2009). These processes depend strongly on the vertical distribution of aerosols throughout the whole atmospheric column. Consequently, measurements of vertical profiles are required in order to better understand the effect of aerosols on climate (Corrigan et al., 2008; Ramana et al., 2007; Podgorny and Ramanathan, 2001). These data can be obtained by direct and indirect methods such as tethered balloons (Ferrero et al., 2007, 2010; McKendry et al., 2004; Stratmann et al., 2003; Maletto et al., 2003), aircraft (Taubman et al., 2006), unmanned aerial vehicles (UAVs) (Corrigan et al., 2008; Ramana et al., 2007), lidars (Kim et al., 2007; Amiridis et al., 2007; Eresmaa et al., 2006) or sunphotometers (Schuster et al., 2005). Direct methods enable the aerosol's physical–chemical and optical (scattering and absorption) properties to be measured at the same time, and of such methods, only balloons and UAVs can be used to collect long-term measurements at a reasonable cost (Ferrero et al., 2010; Corrigan et al., 2008). However, the payload limitations of these platforms require a

0048-9697/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2011.04.022

Please cite this article as: Ferrero L, et al, Vertical profiles of aerosol absorption coefficient from micro-Aethalometer data and Mie calculation over Milan, Sci Total Environ (2011), doi:10.1016/j.scitotenv.2011.04.022

2

L. Ferrero et al. / Science of the Total Environment xxx (2011) xxx–xxx

new generation of light-weight, battery-powered miniaturized instruments (Ferrero et al., 2010). Until recently, it was necessary to adapt commercial instruments (e.g. Aethalometers) that were heavy and primarily designed for installation in ground stations (Corrigan et al., 2008). However, a new portable, lightweight, battery-powered microAethalometer (microAeth® Model AE51, Magee Scientific, USA) has been developed. Its primary intended application is to measure personal exposure to emissions such as diesel exhaust, but its small size and light weight suggest its use for vertical profile measurements. In this context, special attention has to be paid to the interpretation of its data in terms of calculating an absorption coefficient. Past studies of filter-based methods have clearly shown that calculation of the absorption coefficient (babs) requires the application of various different kind of compensatory factors to the light attenuation signal measured by filter based absorption photometers, due to the multiple light-scattering effect of the filter fibers, and to aerosol loading on the sampling filter itself (Schmid et al., 2006; Arnott et al., 2005; Weingartner et al., 2003; Bond et al., 1999). This procedure, and the optical enhancement factor values, were basically calculated for Aethalometers using synthetic, laboratory-generated aerosols (Arnott et al., 2005; Weingartner et al., 2003) and from a limited number of ambient aerosol measurement campaigns conducted at ground-level (Schmid et al., 2006; Arnott et al., 2005; Weingartner et al., 2003); as a result, the estimated uncertainty in the retrieved absorption coefficients may range from 5% to 40% (Corrigan et al., 2008). Vertical profile application introduces another uncertainty factor, due to the sampling, on the same filter, of different kinds of aerosols collected at different heights in different atmospheric layers. To investigate these effects, we measured vertical profiles of aerosol properties above Milan, one of the most polluted areas of Europe. We also derive an optical enhancement factor suitable for use along vertical profiles, calculated for the new microAeth® Model AE51. We also discuss vertical profiles of BC and absorption coefficient measured in this location.

Fig. 1. Tethered balloon (a) fitted with: the OPC 1.108 “Dustcheck” (Grimm Aerosol Technick), the meteorological station (LSI-Lastem) and the microAeth® Model AE51 (Magee Scientific) (b).

atmosphere, and rendering the standardized results directly comparable with those obtained at different locations, heights, dates and times; UTC time is used throughout this article.

2. Experimental 2.1. Aerosol characterization 2.1.1. Vertical aerosol profiles Measurements of vertical aerosol and BC profiles were carried out at the Torre Sarca site in Milan (University of Milan-Bicocca; 45°31′ 19″N, 9°12′46″E), situated in the Po Valley basin, on the 2nd and 3rd of December 2008 (11 profiles). The instrument package consisted of: 1) an optical particle counter (OPC, 1.108 “Dustcheck” Grimm, 15 class-sizes ranging from 0.3 μm to 20 μm); 2) the new microAethalometer microAeth® Model AE51 (Magee Scientific); 3) a meteorological station (BABUC-ABC, LSI-Lastem: pressure, temperature and relative humidity). The sampling platform was carried aloft by a helium-filled tethered balloon (diameter 4 m, volume 33.5 m3, payload 15 kg), both shown in Fig. 1. An electric winch controlled the ascent and descent rate, which was set at a fixed value of 30.0±0.1 m/min; a measurement time resolution of 6 s was chosen for each instrument, giving 3.0 m of vertical resolution for each measurement. The maximum height reached during each launch depended on atmospheric conditions; for the majority of profiles, maximum height was 510 m AGL. Tethered balloon soundings have already been successfully used in the Po Valley to measure aerosol property changes (number size distribution and chemical composition) as a function of height in the low troposphere over Milan; further details of the experimental approach can be found in Ferrero et al., 2010. All vertical profile data were normalized to standard conditions (0 °C, 1013 hPa), as suggested by Hänel (1998), making measurements independent of the actual thermodynamic state of the

2.1.2. Aerosol sampling and chemical characterization During the course of vertical profile measurements, PM2.5 samples were also collected at ground-level (at the Torre Sarca site) and at a remote, high-altitude site (Alpe S. Colombano, 2280 m ASL, 46°27′17″N, 10°18′53″E). The Alpe S. Colombano site lies on the Italian side of the Alps, facing northern Italy. During winter, it is located in the free troposphere; during summer, the site is influenced by regional transport, a higher mixing height and valley breezes bringing polluted air masses from the Po Valley up to the site (Ferrero et al., 2005; Belis et al., 2006). The Torre Sarca and Alpe S. Colombano sites have been active since 2005; PM2.5 is constantly monitored (in accordance with EN-14907 standards), and was sampled at both sites using the dual channel FAIHydra gravimetric system (PM2.5 sampling head, flow 2.3 m3/h; WHATMAN pre-fired quartz fiber filters, Ø = 47 mm). PM2.5 chemical composition was assessed at both sites. PM2.5 samples were analyzed using an ion chromatography coupled system (Dionex ICS-90–ICS-2000), in order to determine the ionic inorganic fraction (Ferrero et al., 2010; Perrone et al., 2010), and using the Thermal Optical Transmission method (TOT, Sunset Laboratory inc.; NIOSH 5040 procedure, http://www.cdc.gov/niosh/nmam/pdfs/ 5040f3.pdf) to quantify elemental and organic carbon (EC and OC) content (Gualtieri et al., 2009). The ionic fraction is denoted here as the water soluble (WS) fraction of the aerosol. The organic matter (OM) fraction was estimated from OC using different coefficients to account for the presence of hetero-atoms (H, O, N, etc.). Following the work of Turpin and Lim (2001), the chosen factors were 1.6 for the urban Torre Sarca site, and 2.1 for the remote site of Alpe S.



Colombano. Table 1 shows the particle chemical composition measured at the two sites. 2.2. Micro-Aethalometer The new microAeth® Model AE51 was developed to measure personal exposure to BC. Due to its light weight (250 g), small size (117 × 66 × 38 mm3) and long-life battery (~ 24 h), it can be used in vertical profile measurements on tethered balloons or UAVs. The measured signal is the attenuation of light (ATN) from a LED source at 880 nm, transmitted through a PTFE-coated borosilicate glass fiber filter (Fiberfilm™ Filters, T60 material, Pall Corporation) while being continuously loaded by the aerosols. The ATN is given by: ATN = 100⁎lnðI0 = IÞ

ð1Þ

where I0 and I are the light intensities transmitted throughout a reference blank spot, and throughout the aerosol-laden 3 mm diameter sample spot of the filter respectively. These data enable us to estimate the absorption coefficient (babs) (Schmid et al., 2006; Arnott et al., 2005; Weingartner et al., 2003) from: babs =

A⋅Δ ATN 1 ⋅ 100Q ⋅Δt C ⋅Rð ATN Þ

ð2Þ

1 σATN

⋅

wavelength) operating in a test chamber with different BC concentrations at low attenuation values. The comparison was then repeated using ambient air. Eqs. (2) and (3) allow one to estimate babs, given BC and σATN; in Eqs. (2) and (3), the expression (A ⁎ ΔATN)/(100Q ⁎ Δt) is also referred to as the attenuation coefficient (bATN). C and R(ATN) constitute the investigated parameters. Briefly, C is a constant optical enhancement factor (≥1) which compensates for the enhanced optical path through the filter caused by multiple scattering induced by the filter fibers themselves; R(ATN), on the other hand, rectifies the “shadowing” effect due to the enhanced absorption of scattered light caused by an increase in aerosol loading over time, which in turn results in a reduction in the optical path (Schmid et al., 2006; Weingartner et al., 2003). Due to the different filter material used in the microAeth® Model AE51, compared with that of other Aethalometers, a new C value has to be calculated. A full mathematical treatment of the aforesaid optical enhancement procedure is given in Schmid et al. (2006), Arnott et al. (2005) and Weingartner et al. (2003). In the aforementioned studies, optical enhancement factors are calculated for other types of Aethalometer, and as reported in Schmid et al. (2006), Eq. (2) is suggested for the purpose of calibrating the Aethalometer's response. Hence, Eq. (2) is used in this present study to calculate a C value for the new microAeth® Model AE51. 2.3. Mie calculation of absorption coefficient along vertical profiles

and the BC concentration from: BC =

3

A⋅ΔATN b = ATN 100Q ⋅Δt σATN

ð3Þ

where A is the sample spot area (7.1.10− 6 m2), Q is the volumetric flow rate (1.7 10− 6 m3/s), ΔATN indicates the ATN variation during the time period Δt (6 s), C is the multiple scattering optical enhancement factor, and R(ATN) is the aerosol loading factor, which depends on light attenuation. σATN (12.5 m2/g) is the apparent mass attenuation cross-section for the black carbon collected on the PTFEcoated borosilicate glass fiber filter, considering the optical components of the instrument; σATN is provided by the manufacturer and was obtained by comparing the BC values measured with the microAeth® Model AE51, with an AE31 Aethalometer (880 nm

Aerosol optical properties were calculated using a Mie code, based on the work of Bohren and Huffman (1983), along vertical profiles from OPC data. As Guyon et al. (2003) and Howell et al. (2006) have reported, OPC data can be used to estimate the aerosol optical properties, and thus Mie theory can be applied to aerosol vertical profile data. The scattering efficiency is calculated on the basis of the integration of the scattered power over all directions, and the extinction efficiency results from the extinction theorem (Mätzler, 2002; Seinfeld and Pandis, 1998; Bohren and Huffman, 1983; Van de Hulst, 1981; Ishimaru, 1978). The absorption efficiency (Qabs) is calculated as the difference between the extinction and scattering efficiencies. The absorption coefficient (babs) at 880 nm is calculated

Table 1 Aerosol properties considered in this study: aerosol mass fraction measured at the Milan and Alpe S. Colombano (ASC) sites (WS: water soluble; OM: organic matter; EC: elemental carbon), density of each component (ρ) and the complex refractive index (m). The figures marked ts are those chosen for the purposes of this study, while the figures in brackets are the ones most commonly found in studies of densities; those figures found in existing literature are marked (by letters) as follows. ρ (g cm− 3)

Aerosol component

Mass fraction (%) Milan (ground)

ASC (2300 m asl)

WS OM EC Missing mass

36.06 35.86 11.83 16.25

23.60 49.81 1.37 25.22

1.75ts (1.7a, 1.72b, 1.77b,c) 1.45ts (1.4c, 1.5a) 2.0ts,d,e (1.5f,g, 1.8c, 2d,e, 2.05h-2.25h) 1.0ts,i (H2O), 2.6ts,l (Dust)

m (n + ik)

m (n + ik)

λ = 880 nm

λ = 780 nm

1.520 + i0.012ts,m,n,o 1.520 + i0.012ts,m,n,o 2.09 + i0.60ts,a,p,q 1.324 + i4.05e− 7 ts,r (H2O), 1.520 + i0.008ts,m,n,q (Dust)

1.525 + i0.010ts,m,n,o 1.525 + i0.010ts,m,n,o 2.07 + i0.61ts,a,p,q 1.326 + i1.41e− 7 ts,r (H2O), 1.525 + i0.008ts,m,n,q (Dust)

a

Chazette and Liousse (2001). Fierz-Schmidhauser et al. (2010). c Schmid et al. (2009). d Schuster et al. (2005). e Hand and Kreidenweis (2002). f Horvat (1993). g Janzen (1979). h Hess and Herd (1993). i Pesava et al. (2001). l Dong et al. (2010). m Shettle and Fenn (1979, 1976). n D'Almeida et al. (1991). o World Climate Programme (1986). p Ackerman and Toon (1981). q Raut and Chazette (2008). r Velazco-Roa and Thennadil (2007). b


4


from the integration of Qabs over the whole number-size distribution (Seinfeld and Pandis, 1998): Dpmax

babs = ∫ 0

πD2p Qabs ðm; αÞn Dp dDp 4

ð4Þ

where m and α are the aerosol complex refractive index and the size parameter at 880 nm respectively, and n(Dp) represents the numbersize distribution. The aerosol refractive index was calculated from measured aerosol chemical composition (Section 2.3.1). The aerosol size distribution function was obtained from the log-normal interpolation of aerosol number-size distribution data measured by the OPC. For this, the OPC size channels were corrected in agreement with the ambient aerosol refractive index. The size correction procedure and the log-normal interpolation are discussed in Section 2.3.2. 2.3.1. Aerosol refractive index Mie calculations require the aerosol complex refractive index (m = n+ik) as an input parameter. In this study, coarse (dp N 1 μm) and fine (dp ≤ 1 μm) particles have different refractive indexes. Coarse particles (dp N 1 μm) are assumed to be composed of dust, while the mean aerosol complex refractive index m of fine particles is calculated from the aerosol chemical composition measured and reported in Table 1. The calculation of the mean aerosol refractive index m of the fine particles, using chemical composition data, must be performed using a mixing rule. Many authors (Fierz-Schmidhauser et al., 2010; Liu and Daum, 2008; Stier et al., 2006, 2007; Riemer et al., 2003; Guyon et al., 2003; Ebert et al., 2002; Pesava et al., 2001; Chazette and Liousse, 2001; Horvath, 1998; Levoni et al., 1997) have used a linear volume-average mixing rule or linear mass-average mixing rule (Raut and Chazette, 2008) in their studies. However, it is well known that a linear volume (mass)-average mixing rule can overestimate the imaginary part (k) of m in the presence of highly absorbing inclusions (i.e. BC) in a nonabsorbing medium (i.e. NH4NO3 and (NH4)2SO4) (Stier et al., 2007; Lesins et al., 2002; Chýlek et al., 1995). A more correct approach would be to consider all possible positions of the inclusions relative to the host medium. Aspnes (1982) formulated such a mixing rule as follows: n εeff −εh ε −εh = ∑ fi i εeff + 2εh ε i=1 i + 2εh

ð5Þ

where εeff is the complex effective dielectric constant of the mixture pffiffiffiffiffiffiffi (meff ≈ εeff ), εh represents the dielectric function of the host medium, εi and fi are the complex dielectric constant, and the volume fraction, of the i-th component respectively. Depending on the choice of host medium, we may obtain three different mixing rules: 1) Maxwell–Garnett (MG) if the host medium is one of the components (εi = εh) (Stier et al., 2007; Schuster et al., 2005; Bohren and Huffman, 1983; Aspnes, 1982; Heller, 1965); 2) Lorentz–Lorenz (LL) if the host medium is the vacuum (εh = 1) (Liu and Daum, 2008; Aspnes, 1982; Heller, 1965); 3) Bruggeman (BR) if no choice of host medium is made, and inclusions are considered embedded in the effective medium itself (Stier et al., 2007; Aspnes, 1982; Heller, 1965; Bruggeman, 1935): this approach is also known as “effective medium approximation” (EMA). Stier et al. (2007) and Aspnes (1982) point out that the BR mixing rule overcomes the dilemma of the choice of host medium among the various aerosol components. From this point of view, the BR mixing rule considers all possible positions of each component (BC, dust, water soluble materials…) in an aerosol particle, thus simulating the real complexity of aerosols and making the BR mixing rule suitable for use in calculating the aerosol meff. For this reason, the BR mixing rule

has been chosen here to calculate the aerosol effective complex refractive index. The refractive index within the mixing layer has been estimated starting from ground-level particle chemical composition; above the mixing layer, the refractive index is calculated from the particle chemical composition considered as an average of that measured at ground-level in Milan, and that measured in the free troposphere at Alpe S. Colombano. The reliability of this assumption will be discussed in Section 3.2: briefly, the BC content in aerosol volume, as well as the ionic fraction, decreased above the mixing layer, approaching the average value of Milan and Alpe S. Colombano. The refractive indexes of pure aerosol components used in the calculation (λ = 880 nm), as well as the pure densities (ρ) for each aerosol component, are reported in Table 1 (together with the respective references). Pure densities were used to estimate the volume fraction of each aerosol component from the aerosol's chemical composition, as reported in the literature (Fierz-Schmidhauser et al., 2010; Pesava et al., 2001; Chazette and Liousse, 2001; Heller, 1965). The value of BC refractive index m (2.09 + i0.60) was chosen because it lies at the mid-point of the published data, and is consistent with the value estimated in a smog chamber at 700 nm (Riemer et al., 2003). The value of BC density ρ is the same as used by Schuster et al. (2005) in AERONET data retrieval, and once again it is to be found in the mid-range of the data reported in Table 1. The OM refractive index was assumed to be the same as WS (the total aerosol ionic content). This assumption derives from the observation of those values reported in the literature, which are close to WS ones in the green region of the visible spectrum (Raut and Chazette, 2008; Randriamiarisoa et al., 2006; Schmid et al., 2009); furthermore, as Moosmüller et al. (2009) point out, the OM is not absorbing in the infrared part of the spectrum, as brown carbon absorption mainly takes place in the blue and ultraviolet parts of the spectrum. Many scholars presumed the missing mass was essentially water adsorbed on particles (Pesava et al., 2001; Marley et al., 2001; Gebhart and Malm, 1994), while others assigned the refractive index of dust to the missing mass (Raut and Chazette, 2008). Since it has been shown that a certain amount of water is chemically bound to particles (Subramanian et al., 2007; Hueglin et al., 2005; Rees et al., 2004), and also that this amount is comparable to dust in winter (Hueglin et al., 2005; Rees et al., 2004), we have assigned the missing mass to both water and dust. Finally, the aerosol refractive index was modulated point by point along vertical profiles, considering the hygroscopic growth of the aerosol: rw = r ⋅

1−RH 1−RHref

−ε

ð6Þ

where RH is ambient relative humidity, r is the aerosol radius at RH = RHref (dry conditions) and rw at ambient RH; ε is the coefficient controlling the aerosol's hygroscopic growth. The value of coefficient ε was 0.26, as reported in both Randriamiarisoa et al. (2006) and in Raut and Chazette (2008); this value was chosen because the ground-level chemical composition of dry aerosol measured in Milan (Table 1), is similar to that reported in the aforementioned studies. RHref was 65%, and it was estimated from the aerosol chemical composition following the work of Potukuchi and Wexler (1995a, 1995b); this value is also in accordance with other mutual deliquescence relative humidity figures reported in literature (Badger et al., 2006; Randriamiarisoa et al., 2006). From hygroscopic growth, the BR mixing rule was applied point by point along height to calculate the final aerosol refractive index every 3.0 m for each profile. Finally, the calculated aerosol refractive indexes were used to calculate the absorption coefficient profiles (Eq. (4)): the results thus obtained were compared to the Aethalometer data in order to



estimate the optical enhancement factor C (Section 3.3). The accuracy of m estimation is discussed in Section 3.1.1. 2.3.2. Aerosol size distribution Mie calculations also require the aerosol number size distribution as an input parameter. The size distribution is measured using an OPC (Grimm 1.108), whose size channels are calibrated using polystyrene latex spheres (PLS, m = 1.59) which are non absorbing and whose refractive index has a larger real part compared to ambient aerosol (see Section 3.1.2). This may result in an “undersizing” of atmospheric aerosols (Guyon et al., 2003; Liu and Daum, 2000; Schumann, 1990). Moreover, the OPC has a PLS equivalent size range between 0.3 μm and 20 μm, which renders the coarse mode clearly defined, while the accumulation mode is only partially measured: no measurements are available for Aitken mode particles. As many authors report (Randriamiarisoa et al., 2006; Guyon et al., 2003; Liu and Daum, 2000), if we neglect the Aitken mode (dp b 0.1 μm) this may give rise to a ~2–4% error in the calculation of aerosol optical properties, while the “truncation effect” of the OPC in the accumulation mode (~0.3 μm, ~half of the accumulation mode) can not be neglected, because particles in the accumulation mode are highly efficient in absorbing light. Thus, to avoid the first problem (“undersizing”), the OPC size channels were corrected to account for the ambient aerosol refractive index m. The size correction procedure requires the calculation of the OPC response function (S) which describes the intensity of light scattered into a given angular interval normalized to the incident radiation; S represents the partial light scattering cross section of the particle (cm2) related to the specific optical design of the OPC. For a plane, linearly polarized, monochromatic wave λ, the response function S can be computed as follow (Baron and Willeke, 2005; Heyder and Gebhart, 1979):

5

size corrected channels: these results agree with literature studies (Liu and Daum, 2000; Schumann, 1990). Table 2 shows that correction resulted in an increase of the lower cutoff diameter, from 0.30 μm up to 0.33 μm; thus, if the size correction solves the OPC “undersizing” problem, the “truncation effect” is accentuated by the same correction. The complete aerosol size distribution function n(Dp) was therefore obtained from the log-normal interpolation of aerosol number-size distribution data measured by the OPC. This procedure makes the calculated babs (Eq. (2)) closely dependent on the reliability of the particle number-size distribution interpolation procedure. However, the log-normal interpolation of OPC data has already been successfully used by Deshler et al. (2003), and extinction profiles calculated from balloon borne OPC have been successfully compared with extinction profiles estimated using an automated lidar ceilometer (Vaisala LD40, λ = 855 nm) installed at the Torre Sarca site at the same time as our balloon campaign (Angelini et al., 2009). We discuss the accuracy of the log-normal interpolation of the accumulation mode in Section 3.1.2 Moreover, to achieve the effect of size correction on the calculated babs, the log-normal interpolation was conducted on both the two size distributions: uncorrected (original OPC sizes); and corrected (for SAMB). The results thus obtained were used to calculate uncorrected and corrected babs that were compared to the Aethalometer data, allowing to estimate uncorrected and corrected optical enhancement factors C, namely CUSD (USD = “uncorrected size distribution”) and CCSD (CSD = “corrected size distribution”) respectively (Section 3.3). The final optical enhancement factor C for the microAeth® Model AE51 is given in Section 3.3.

3. Results and discussion 3.1. Data quality

2

λ ∬ iðθ; ϕ; x; mÞ sinθdθdϕ 4π2 ΔΩ

ð7Þ

where θ0 represents the mean scattering angle of the optical arrangement, ΔΩ the receiver aperture, x the dimensionless size parameter, m the refractive index and i(θ,ϕ,x,m) the Mie scattering function composed by the perpendicular and parallel components: i1(θ,x,m) and i2(θ,x,m) respectively. The OPC 1.108 uses 780 nm polarized laser light to illuminate the aerosols in the airflow, and the PLS calibration curve is calculated by the manufacturer under the plane wave approximation (we gratefully acknowledge the manufacturer for this information). The optical arrangement of the OPC 1.108 is the same as that of the Grimm 1.109 OPC described in Heim et al. (2008), and consists of: 1) a wide angle parabolic mirror (121°, from 29.5° to 150.5°, θ0 = 90°) that focuses scattered light on the photodetector located on the opposite side; 2) 18° of direct collected scattered light on the photodetector (from 81° to 99°, θ0 = 90°). Knowledge of the specific optical design of the OPC 1.108 allows us to calculate the response function S (Eq. (7)) following the methodology reported in Heyder and Gebhart (1979); the response function was calculated both for PLS (SPLS) and for ambient aerosol (SAMB) along each vertical profile, within and above the mixing layer. The refractive indexes of ambient aerosol used in SAMB calculations were calculated as reported in Section 2.3.1: 1) dust refractive index for coarse particles (dp N 1 μm), 2) m for fine particles calculated applying the EMA (Section 2.3.1) to the aerosol chemical composition along vertical profiles, at the OPC laser wavelength (780 nm; Table 1 also reports m of pure aerosol components at 780 nm). Calculations of SPLS and SAMB allow us to correct the size channels of the OPC according to the ambient m. Fig. 2 shows the response curves SPLS and SAMB for the average refractive index of the whole campaign (m = 1.480 + 0.034i at 780 nm). Table 2 shows the corresponding new

3.1.1. Aerosol optical properties The reliability of the calculated aerosol optical properties can be evaluated by considering some parameters such as the refractive 1.0E-04 S_PLS 1.0E-05

S_AMB

1.0E-06

1.0E-07

S (cm2)

Sðθ0 ; ΔΩ; x; mÞ =

1.0E-08

1.0E-09

1.0E-10

1.0E-11

1.0E-12

1.0E-13 0.10

1.00

10.00

100.00

Fig. 2. Response curves for 1.108 Grimm OPC in terms of partial light scattering cross sections for PLS (SPLS, m = 1.59) and ambient aerosol (SAMB, m = 1.480 + i0.034 for fine fraction and m = 1.525 + i0.008 for coarse fraction).


6


index and the single scattering albedo (SSA) which, once averaged along vertical profiles, can be compared with the same data measured by the Aerosol Robotic Network (AERONET; Table 3). The closest AERONET station is the Ispra site (E 8°37′36″,N 45°48′11″, 57 km from the Torre Sarca site); AERONET data for the 2nd and 3rd of December were not available, and we used the averaged values for December as a reference in order to assess the accuracy of our optical estimation. The average refractive index calculated in this study along vertical profiles was 1.476(±0.032) + i0.034(±0.008), which is in keeping with the columnar AERONET estimation of 1.457(±0.087) + i0.034 (±0.011). Considering the SSA, the average values calculated in this study along vertical profiles for the CSD and USD cases were SSACSD = 0.803 ± 0.040 and SSAUSD = 0.728 ± 0.074 respectively; these were in keeping with the columnar AERONET estimation of 0.721 ± 0.069.

Table 2 Size channels of OPC 1.108 Grimm for PLS (m = 1.59) and ambient aerosol (m = 1.480 + i0.034 for fine fraction: dp ≤ 1 μm; m = 1.525 + i0.008 for coarse fraction: dp N 1 μm). OPC Channel

PLS (μm)

Ambient Aerosol (μm)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

0.30 0.40 0.50 0.65 0.80 1.00 1.60 2.00 3.00 4.00 5.00 7.50 10.00 15.00 20.00

0.33 0.46 0.62 0.91 1.15 1.38 1.86 2.60 3.85 5.50 7.00 13.18 20.89 35.89 49.55

Table 3 Aerosol optical and size distribution properties calculated in this study at Torre Sarca site (TS) in Milan for both the USD (uncorrected size distribution) and CSD (corrected size distribution) cases. Ispra site is located ~ 57 km far from Torre Sarca, while MB (Milano-Bresso) site is is located ~ 2 km far from Torre Sarca. n and k are the real and immaginary part of the complex index of refraction. SSA indicates the Single Scattering Albedo, while Dg and σg the geometric mean diameter and the geometric standard deviation respectively. ML indicates the mixing layer. Parameter

This study — TS USD

n columnar k columnar SSA columnar Dg (μm) within ML σg within ML Dg (μm) columnar σg columnar

1.476(± 0.032) 0.034(± 0.008) 0.728(± 0.074) 0.099(± 0.074) 1.802(± 0.102) 0.110(± 0.040) 1.810(± 0.156)

AERONET

Van Dingenen et al. (2004)

CSD

Ispra

MB morning

MB afternoon

MB night

0.803(± 0.040) 0.166(± 0.030) 1.894(± 0.257) 0.201(± 0.095) 1.851(± 0.351)

1.457(± 0.087) 0.034(± 0.011) 0.721(± 0.069) – – 0.152(± 0.036) 1.766(± 0.160)

– – – 0.114 1.83 – –

– – – 0.096 2.00 – –

– – – 0.123 1.74 – –

Fitted particle conc. (cm -3)

1100 1000

n(Dp)_USD

900

n(Dp)_CSD

800 700 600 500

y = 1.091x - 7.760 R² = 0.998

400 300

y = 1.105x - 11.741 R² = 0.997

200 100 0 0

100

200

300

400

500

600

700

Measured particle conc.

800

900

1000

1100

(cm -3)

Fig. 3. Linear regression between measured and fitted particle number concentrations in the OPC size range (N0.3 μm). Blue circles indicate average particle number concentration measured and fitted within and above the mixing layer for the corrected size distribution (CSD); red diamonds indicate average particle number concentration measured and fitted within and above the mixing layer for the uncorrected size distribution (USD). Error bars indicate the variability of particle number concentration (as standard deviation) around corresponding average concentrations, for measured and fitted concentrations. An F test was conducted, which revealed the absence of any statistical difference between variances associated to measured and fitted concentrations both within the mixing layer and above it. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)



2–3 December 2008 December 2008

mean σ mean σ

p (hPa)

t (°C)

RH (%)

ws (m/s)

1001.3 1.1 1008.8 10.3

4.0 1.1 3.7 2.7

77.4 7.6 74.6 14.1

1.5 0.5 1.6 0.8

3.1.2. Log-normal interpolation The estimation of the absorption coefficient (babs) from OPC data depends very much on the reliability of the particle number-size distribution interpolation procedure. Several factors have to be taken into account here, and the accuracy of n(Dp)USD and n(Dp)CSD have to be considered. The reliability of the OPC data interpolation procedure is firstly guaranteed by a comparison of the interpolated particle number concentrations with the measured ones within the OPC size ranges (uncorrected and size corrected). Considering the whole profile dataset, interpolated data showed a good fit with measured ones both for n(Dp)USD and n(Dp)CSD (R2 0.997 and 0.998 for n(Dp)USD and n(Dp)CSD respectively; Fig. 3); the average difference between interpolated and measured particle number concentration was found to be only + 6.56 ± 0.83% for n(Dp)USD and + 6.18 ± 0.71% for n(Dp)CSD: these are small differences compared to the ones reported in other works (~ 8–15% in Schmid et al., 2006). Additionally, the geometric mean diameter (Dg) and the geometric standard deviation (σg) of the accumulation mode were calculated, for both n(Dp)USD and n(Dp)CSD, as mixing layer and columnar averaged values, and were compared each other (Table 3): as expected, n(Dp)CSD shows a larger Dg compared with n(Dp)USD due to size correction of OPC channels (Table 2). Finally, these values were compared with AERONET (Ispra site) and literature data (Van Dingenen et al., 2004; Milan-Bresso site: ~ 2 km from the Torre Sarca site, during year 2004); Table 3 reports all the data. Size distribution data (both n(Dp)USD and n(Dp)CSD) agree well with AERONET data. For n(Dp)CSD the agreement is better if the Dg value within the mixing layer is also considered: the optical signal originated mainly within the mixing layer. Conversely, n(Dp)USDDg agree better with literature data (Van Dingenen et al., 2004). Since AERONET data were collected during the same period of balloon launches, and the size correction procedure reported in Section 2.3.2 is more appropriate from a physical point of view, we consider the n(Dp)CSD size distribution most reliable to calculate aerosol optical properties along vertical profiles. The reported values, as well as the agreement between the calculated aerosol optical properties and AERONET data, underline the accuracy of optical estimations from OPC data, and their appropriateness for the calculation of the Aethalometer optical enhancement factor (Section 3.3).

a

600 BC Particle conc.

500

Height (m AGL)

Date

very close to that measured along vertical profiles: 76.8 ± 6.0% (range: 61–88%); as a result, the columnar average value of the calculated hygroscopic growth factor, applied in the refractive index determination (Section 2.3.1, Eq. (6)), was 1.128 ± 0.082 (range considering each single 3.0 m measuring point along vertical profiles: 1.000– 1.318). The average ground-level wind speed was very low (less then 2 m/s) and favoured the accumulation of pollution within the mixing layer. Ground-level meteorological conditions during the vertical profile measurements (2–3 December) are summarized in Table 4 and

400

300

200

100

0 0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

BC (µg/m3) 0.0

100.0 200.0 300.0 400.0 500.0 600.0 700.0

Particle conc. (cm-3)

b

600 Potential temperature Relative humidity 500

Height (m AGL)

Table 4 Ground meteorological data measured in Milan both on 2nd–3rd December 2008 and along the whole month of December 2008; σ respresent the standard deviation.

7

400

300

200

100

3.2. Vertical profiles 0

3.2.1. Meteorology and mixing height determination We present the first BC vertical profile measurements performed in Italy; the data show a vertical distribution of BC (the main absorbing aerosol species) over Milan, in the middle of the Po Valley, which represents Italy's largest urban and industrial area. Vertical profiles of BC, aerosol parameters and meteorological measurements were made on the 2nd and 3rd of December 2008, from the morning to the afternoon. Measurements were conducted under a synoptic low pressure system (average pressure was 1001.3 ± 1.1 hPa), thus causing an alternation of cloudy and clear sky during measurements. Under these conditions, the relative humidity average value at ground on the 2nd and 3rd of December was 77.4 ± 7.6% (range: 57–89%)

280.6 280.8 281.0 281.2 281.4 281.6 281.8 282.0 282.2 282.4

T (K) 62.7

64.2

65.7

67.2

68.7

70.2

71.7

73.2

RH (%) Fig. 4. (a). Vertical BC (blue line) and aerosol (light blue line) profiles measured on the 3rd of December 2008 (9:41–9:58 UTC). The main horizontal axis shows BC concentrations (μg/m3), while the secondary x-axis shows total particle number concentration (cm− 3). (b) Potential temperature (T, red line) and relative humidity (RH, green line) vertical profiles for the same balloon launch. The main horizontal axis shows potential temperature (K), while the secondary x-axis shows relative humidity (%). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


8


Table 5 a–b. BC concentrations and absorption coefficients averaged every 50 m along height for each profile measured both on the 2nd of December 2008 (Table 5a) and on the 3rd of December 2008 (Table 5b); particle, BC, T and RH derived mixing heights (p-MH, BC-MH, T-MH and RH-MH) are also reported for each profile. 2/12/2008 h 9.53–10.20 Parameter

BC (μg/m3)

2/12/2008 h 10.25–10.42 babs (Mm−1)

BC (μg/m3)

2/12/2008 h 10.57–11.14 babs (Mm−1)

BC (μg/m3)

2/12/2008 h 14.30–14.47 babs (Mm−1)

BC (μg/m3)

2/12/2008 h 14.50–15.07 babs (Mm−1)

BC (μg/m3)

babs (Mm−1)

Height (m a.g.l.)

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

0–50 50–100 100–150 150–200 200-250 250–300 300–350 350-400 400–450 450–500 p-MH (m AGL) BC-MH (m AGL) T-MH (m AGL) RH-MH (m AGL)

13.2 11.3 10.8 8.9 2.9 1.4 1.0 0.9 1.0 1.2 196 199 227 224

1.2 0.2 0.3 0.7 0.8 0.2 0.1 0.1 0.1 0.1

80.4 68.7 65.6 54.2 17.7 8.3 6.3 5.6 6.2 7.2

7.5 1.1 2.0 4.2 4.9 1.0 0.5 0.1 0.2 0.8

12.2 11.8 10.0 4.0 1.1 0.9 0.9 1.0 1.0 1.2 157 149 147 140

0.6 1.1 1.1 1.8 0.3 0.1 0.1 0.1 0.1 0.1

74.5 72.2 60.9 24.4 6.8 5.5 5.5 5.9 6.2 7.2

3.9 6.6 6.4 10.8 1.6 0.4 0.2 0.3 0.5 0.2

15.9 13.0 12.4 10.5 6.2 2.2 1.1 219 222 216 199

3.0 0.2 0.1 1.4 1.1 1.1 0.1 – -

96.8 79.5 75.7 64.1 38.0 13.3 6.5 – -

18.1 1.4 0.6 8.2 6.7 6.6 0.3 – -

12.8 9.6 9.5 10.0 9.7 9.3 9.7 9.0 5.4 2.7 396 396 412 406

1.9 0.1 0.2 0.4 0.2 0.2 0.1 0.9 0.4 1.1

78.1 58.5 58.0 61.1 59.1 56.8 59.4 55.0 32.9 16.6

11.8 0.8 1.0 2.4 1.0 1.1 0.9 5.3 2.6 6.6

11.0 8.4 8.3 8.7 9.5 9.8 8.2 5.9 4.4 1.9 443 440 431 452

1.4 0.3 0.1 0.2 0.4 0.6 0.4 0.8 0.9 0.2

67.0 51.2 50.4 53.0 57.8 59.8 50.0 35.9 26.8 11.8

8.3 1.6 0.4 1.3 2.2 3.6 2.7 4.7 5.5 1.5

a

compared with that of the whole December month: from a meteorological point of view, the vertical profile measurements were conducted during typical winter days, and can be considered representative of this period. Fig. 4a shows, as a case study, BC and aerosol profiles measured on the 3rd of December 2008 (9:41–9:58 UTC); 50 m averaged BC concentrations as a function of height, are summarized for the whole campaign in Table 5a–b. On the 3rd of December 2008 (9:41–9:58 UTC), aerosol and BC pollution loading was restricted to the first 264 m in the lower troposphere (Fig. 4a–b). At 264 m the aerosol concentration and BC concentration, as well as potential temperature (T) and relative humidity (RH) (Fig. 4b), showed a clearly defined mixing height boundary (MH). The MH was inferred using the gradient method

200 180 160

bATN (Mm-1)

140 120 100 y = 2.221x R² = 0.970

80 60 40 20 0 0

10

20

30

40

50

60

70

80

babs_USD(Mm-1)

b

600 Abs from BC Abs calculated

180 500

y = 2.050x R² = 0.985

160

Height (m AGL)

bATN (Mm-1)

140 120 100 80 60 40

400

300

200

20 100

0 0

10

20

30

40

50

60

70

80

babs_CSD(Mm-1)

0 0.0

Fig. 5. Linear regression between bATN calculated from the microAeth® Model AE51 measurements, and the aerosol babs calculated on the basis of Mie theory as described in Section 2.3 for: (a) uncorrected size distribution (babs_USD); (b) corrected size distribution, (babs_CSD). Full dots and diamonds indicate averages within the mixing layer, while open circles and diamonds indicate averages above the mixing layer. Error bars indicate data variability (as standard deviation) regarding the corresponding averages both within and above the mixing layer.

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

Mm-1 Fig. 6. Vertical profiles of the absorption coefficient calculated from aerosol (light blue line) and BC (blue line) profiles measured on the 3rd of December 2008 (9:41–9:58 UTC). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)



2/12/2008 h 15.18–15.35 BC (μg/m3)

babs (Mm− 1)

2/12/2008 h 15.37–15.54

3/12/2008 h 8.46–9.03

BC (μg/m3)

BC (μg/m3)

babs (Mm− 1)

3/12/2008 h 9.10–9.27 babs (Mm− 1)

BC (μg/m3)

9

3/12/2008 h 9.41-9.58 babs (Mm− 1)

BC (μg/m3)

3/12/2008 h 10.05–10.22 babs (Mm− 1)

BC (μg/m3)

babs (Mm− 1)

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

m

σ

9.8 10.7 7.8 7.7 8.1 7.2 6.2 3.9 1.7 1.8 374 377 413 419

0.9 0.7 0.6 0.2 0.2 0.2 0.5 1.0 0.4 0.3

60.0 65.3 47.6 46.8 49.3 43.9 38.0 23.9 10.4 11.2

5.7 4.3 3.7 1.0 1.2 1.4 2.8 6.1 2.2 1.9

7.3 7.3 7.2 5.9 3.7 2.8 3.7 2.5 325 322 340 297

0.2 0.2 0.1 0.6 0.9 0.2 0.3 0.5

44.7 44.3 44.0 36.0 22.4 17.4 22.3 15.5

1.2 1.2 0.7 3.4 5.7 1.5 1.8 3.2

7.4 9.7 8.3 7.9 5.2 2.2 1.7 1.5 1.1 1.0 213 217 162 241

0.5 0.7 0.2 0.3 1.4 0.2 0.1 0.2 0.1 0.1

45.1 59.3 50.7 48.2 31.6 13.2 10.6 8.9 6.9 5.9

3.3 4.0 1.0 1.7 8.7 1.3 0.9 1.0 0.3 0.3

7.8 4.8 3.3 3.8 2.7 1.6 1.4 1.2 1.2 1.0 221 221 235 244

1.1 0.3 0.2 0.2 0.3 0.3 0.1 0.1 0.1 0.1

47.7 29.5 20.3 22.9 16.5 9.8 8.2 7.3 7.1 6.3

6.7 1.8 1.3 1.0 2.0 1.7 0.2 0.3 0.1 0.5

6.1 4.8 3.9 3.5 4.5 3.1 1.0 1.4 0.9 0.6 264 264 287 284

0.5 0.4 0.5 0.4 0.3 1.7 0.1 0.3 0.1 0.1

37.0 29.0 23.9 21.2 27.7 19.1 6.4 8.4 5.4 3.6

3.2 2.3 2.9 2.6 1.6 10.2 0.8 2.1 0.6 0.3

7.6 6.2 6.1 5.8 5.6 4.8 3.1 4.4 1.1 0.7 404 401 407 400

1.0 0.1 0.1 0.1 0.4 0.7 0.3 1.1 0.3 0.1

46.6 37.8 37.1 35.5 34.2 29.0 18.7 26.9 6.5 4.1

6.1 0.8 0.3 0.9 2.5 4.6 1.6 6.6 1.7 0.6

(Ferrero et al., 2010) from particle (p-MH), BC (BC-MH), T (T-MH) and RH (RH-MH) profiles; Table 5a–b reports values of p-MHs, BCMHs, T-MHs and RH-MHs. These values coincide for all measured profiles, and the MHs calculated from different parameters, either meteorological and pollution tracers, showed a marked correlation each other, together with very low root mean squared errors (RMSE) and with linear best fits close to the ideal ones (BC-MH= 0.996*p-MH+ 0.774, R 2 = 0.999, RMSE = 3.5 m; T-MH = 1.049*p-MH-8.468, R 2 = 0.948, RMSE = 24.4 m; RH-MH = 1.014*p-MH + 4.470, R2 = 0.950, RMSE= 23.7 m): these data are in agreement with that reported in Ferrero et al. (2010) and underline the reliability and physical inter-consistence of using particle, BC, T and RH soundings when determining MH. Angelini et al. (2009) also showed that balloon derived MH was in keeping with MH estimation performed using an automated lidar ceilometer (Vaisala LD40, λ = 855 nm) installed at the Torre Sarca site at the same time as the balloon launching. All of these results underline the accuracy of BC-MH and p-MH in estimating the mixing height. Hereinafter we are going to refer to the mixing height as that derived from BC concentration gradient, indicated simply as MH.

3.2.2. Black carbon profiles Fig. 4a and the results shown in Section 3.2.1 indicate that aerosol and BC profiles were shaped in the same way by atmospheric turbulence (either thermal or mechanical forces), and that the majority of BC lies within the MH, as would be expected over a large urban area. To a smaller degree, BC profiles within the mixing layer revealed a layer of high concentration close to the ground, probably due to proximity to combustion sources (traffic, domestic heating, industry). This ground-level layer (~50 m) is evident in most profiles (Fig. 7) with concentrations + 24 ± 4% higher than the average BC concentration measured within the whole mixing layer. These BC concentrations near ground-level are of particular importance due to the health effects of BC (Hesterberg et al., 2006; Mar et al., 2000; Vedal, 1997). Above the MH, both BC and aerosol concentrations were fairly constant from the MH to the highest elevation of the balloon (Fig. 4). Over the course of the entire campaign, the BC concentrations ranged between 0.82 and 2.06 μg/m3 (average value: 1.39 ± 0.14 μg/m3) above the MH. These values represent 17 ± 2% of BC concentrations measured within the mixing layer (8.40 ± 0.95 μg/m3), and indicate the presence of stable BC concentrations above the MH, at least during the measuring campaign period. At the same time, the average

particle concentration above the MH was 206 ± 40 cm− 3, representing 36 ± 5% of the total particle number concentration measured within the mixing layer (588 ± 67 cm− 3); these data confirm the winter averages (Dec., Jan. and Feb.) reported in Ferrero et al. (2010), indicating that, also from the pollution point of view, the campaign was conducted during days which, on average, were representative of the mean seasonal conditions (compare Section 3.2.1 and Table 4). Another sign that vertical profile measurements were conducted during typical winter days is given by the PM2.5 ground-level concentrations measured at Torre Sarca on the 2nd and 3rd of December 2008: 52 and 50 μg/m3 , respectively, and are thus in agreement with the December average of 53 ± 1 μg/m3. Regardless of the mechanism responsible for the BC profiles, BC experienced a higher concentration drop at the MH than did the aerosol number concentration. This means that the BC content of the aerosol was not constant along the vertical profiles, but sharply decreased above the MH; the aerosol chemical composition is different above the MH than below it. The BC fraction of total aerosol volume was found on average to be 48± 8% lower above MH than within the mixing layer. Ferrero et al. (2010) showed, after measuring vertical profiles for a period of three years, that during winter the inorganic ionic aerosol content also behaves similarly, and is 26± 6% lower than it is when measured near the ground. These data are in accordance with the average aerosol chemical composition measured in the free troposphere at the Alpe S. Colombano site (Table 1), which showed lower BC and inorganic aerosol mass contents than those measured at the same time at groundlevel in Milan (at the Torre Sarca site, see Table 1), while organic matter content was found to be higher in the more remote site. Because BC particles are emitted from primary sources, this difference in aerosol content could be due both to the presence of these sources within the mixing layer, and to secondary particles formed above the MH. This would be in agreement with the results reported by many authors (Morgan et al.;, 2009; Schneider et al., 2006; Hueglin et al., 2005), which pointed out a lower vertical gradient of organic species than of inorganic ones; the organic matter appeared to be more important above the MH, as has also been reported by Sun et al. (2009). Because the aerosol volume content of a chemical species is a proxy of the aerosol mass fraction of the same species (without considering particle density), we applied the fraction of BC reduction above the MH (48 ± 8%) to the BC mass fraction reported in Table 1 for


10


600 2-12-2008-h 9:53 -10:20 2-12-2008-h 10:25 -10:42 2-12-2008-h 10:57 -11:14

500

2-12-2008-h 14:30 -14:47 2-12-2008-h 14:50 -15:07

Height (m AGL)

400

2-12-2008-h 15:18 -15:35 2-12-2008-h 15:37 -15:54 3-12-2008-h 8:46 -9:03

300

3-12-2008-h 9:10 -9:27 3-12-2008-h 9:41 -9:58 3-12-2008-h 10:05 -10:22

200

100

0 0

5

10

20

15

BC µg/m3 0

25

50

75

100

125

150

babs(Mm-1) Fig. 7. Vertical BC profiles measured over Milan on the 2nd and 3rd of December 2008 (11 profiles altogether) with BC concentrations (μg/m3 main x-axis), and the corresponding absorption coefficients (Mm− 1, secondary x-axis) calculated as described in Section 3.2.

Milan (Torre Sarca PM2.5 chemical composition). By adopting this procedure, we were able to estimate the BC mass fraction above the MH on the 2nd and 3rd of December 2008: the estimated BC mass fraction above the MH was 6.2%, which is very close to the value of 6.6% calculated as the average of the Milan and Alpe S. Colombano EC data reported in Table 1. This result substantiates the assumption made in Section 2.3.1, whereby the aerosol refractive index above the MH is calculated from the averaged chemical composition between that at ground-level in Milan, and that in the free troposphere at Alpe S. Colombano. We also used PM2.5 chemical composition at ground-level in order to estimate a refractive index for the whole mixing layer (Section 2.3.1). This was verified by comparing the BC profile data measured within the mixing layer, with EC concentrations measured in ground-level PM2.5 samples. We averaged the BC concentrations measured within the mixing layer for each profile for the 2nd and 3rd of December 2008, and estimated the BC content in ground-level PM2.5 samples. The averages were calculated within the mixing layer to account for the atmospheric turbulence that continuously mixes the aerosol from ground-level to the MH; even if vertical BC profiles are discontinuous during the day, these averages are a first estimation of daily BC concentrations calculated on the basis of readings of the microAeth® Model AE51. The estimated BC mass fraction in the PM2.5 samples was 14.9 ± 1.2%, that is, slightly higher than the EC mass fraction (11.8 ± 0.9%)

measured using the TOT method (see Section 2.1.2 and Table 1); the difference of ~ 3% could be due to the fact that the BC measurements were conducted during the daytime only, when primary emission sources are more active than they are at night. 3.3. Aethalometer optical enhancement factor and absorption coefficient profiles The Aethalometer optical enhancement factor C was calculated for the microAeth® Model AE51 in order to derive an estimation of babs along vertical profiles over Milan. The optical enhancement factor C was not previously estimated for the microAeth® Model AE51, and its determination is of crucial importance since the C value depends on the filter material (PTFE-coated borosilicate glass fiber) and instrument specifications, which are completely new in this particular case. The experimental design of vertical profiles does not require any estimation of the aerosol loading factor R(ATN): all vertical BC profiles were conducted by changing the filter ticket after each profile. Every Aethalometer measurement cycle (ascent and descent of the balloon) took less than 40–50 min. As a result, ATN never achieved values higher than 20 during all profiles, meaning that the bATN measurements were not affected by the “shadowing” effect due to filter loading. The average ATN measured along vertical profiles was 5 ± 1, hence there was no need to use R(ATN) in the determination of the



optical enhancement factor C (Schmid et al., 2006; Arnott et al., 2005; Weingartner et al., 2003). In order to calculate the C factor, first bATN was calculated from BC profiles by means of (3); the reference absorption coefficients (here denoted babs_USD and babs_CSD) were calculated by applying Mie theory to aerosol number size distribution data, n(Dp)USD and n(Dp)CSD respectively. Coefficients bATN, babs_USD and babs_CSD were averaged within and above the mixing layer, for every balloon profile, and were compared to each other. A similar averaging procedure had been used previously (Corrigan et al., 2008; Schmid et al., 2006; Arnott et al., 2005; Weingartner et al., 2003) to smooth out short-term signal variations. The linear regression between averaged bATN and babs_USD and babs_CSD is shown in Fig. 5a–b, and shows a high degree of correlation (R2 = 0.970 and 0.985 for babs_USD and babs_CSD respectively). The estimated optical enhancement coefficients CUSD and CCSD were 2.22 ± 0.06 (95% confidence limits: 2.10–2.34) and 2.05 ± 0.03 (95% confidence limits: 1.98–2.12) respectively. CCSD is ~8% higher than CUSD; this reflects the effect of the OPC size correction (Section 2.3.2), which results in an increase of the geometric mean diameter (Dg, Section 3.1.2, Table 3), and consequently in an absorption coefficient babs_CSD on average ~ 13% higher than babs_USD. The OPC size correction does not significantly affect the result and 95% confidence limits of CCSD and CUSD overlap; this is because the size distribution is interpolated to prevent the “truncation effect” which, as reported in Section 2.3.2, can not be neglected considering Dg values reported in Table 3 (for both n (Dp)USD and n(Dp)CSD) compared to the OPC lower cutoff (~0.3 μm). Thus the interpolation procedure compensates for the highest source of possible bias (the “truncation effect”), while the size correction allows to improve the whole algorithm performance (from interpolation to the optical properties calculations); this turns in a higher R2 for CCSD, in a narrower standard deviation and confidence width. Moreover, CCSD calculation is more appropriate from a physical point of view, thus we consider CCSD the most reliable estimation, and hereinafter C = CCSD = 2.05 ± 0.03. This C factor is specific to the new micro-Aethalometer microAeth® Model AE51. We wanted to compare the new optical enhancement factor to those reported in the literature; we took account of the fact that the new microAeth® Model AE51 has a mass attenuation cross section of 12.5 m2/g, rather than the 16.6 m2/g, reported for other Aethalometers at the 880 nm wavelength (e.g. AE22, AE31, Magee Scientific). In order to compare the C value calculated for the microAeth® Model AE51 with that of other Aethalometers reported in the literature, C was recalculated according to a mass attenuation crosssection of 16.6 m2/g, in order to compare it with the data currently available in other studies; the recalculated C factor (Cr) amounted to 2.73, which agrees with those reported by Weingartner et al. (2003), who found optical enhancement factors in the 2.13–3.90 range (from pure soot particles to internally mixed coated particles) at 660 nm, and with those given by Schmid et al. (2006), who calculated the wavelength dependence of optical enhancement factors to give a C value of 3.14 at 880 nm. Our Cr value (2.73 at 880 nm) was found to be close to those values reported in other studies (Schmid et al., 2006; Arnott et al., 2005; Weingartner et al., 2003), when soot particles were coated with secondary compounds. This fits well with the aerosol chemical composition measured in Milan during balloon launches: 36 ± 3% of the mass fraction of PM2.5 was due to NH4NO3 and (NH4)2SO4, while a further 36 ± 2% was due to organic matter (OM). 3.3.1. Absorption coefficient profiles By using the mass attenuation cross-section σATN (12.5 m2/g) and the optical enhancement factor C (2.05), the micro-Aethalometer BC data can be used to calculate aerosol absorption coefficient profiles.

11

The reliability of this form of estimation using short-term measurements (6 s) conducted along vertical profiles, is clearly shown in Fig. 6, where the absorption coefficient profiles resulting from the Mie calculation and from the micro-Aethalometer measurements are compared to each other. Fig. 6 reports the babs for the same aerosol and BC profiles shown in Fig. 4a. All BC and absorption profiles measured using the microAeth® Model AE51 during the field campaign are reported in Fig. 7, while Table 5a–b also contains 50 m averaged babs values as a function of height. The values of babs reported in Table 5a–b and Fig. 7, reflect the BC behaviour with respect to height previously discussed in Section 3.2.2. Obviously, babs profiles are shaped in the same way as BC profiles, and the main atmospheric absorbing layer was found within the MH, with an average absorption coefficient of 51.2 ± 5.8 Mm− 1 (range: 26.3– 81.3 Mm− 1). From the MH to the top of the profile, babs remained fairly constant (see Fig. 7), on average 8.5 ± 0.8 Mm− 1 (range: 5.0– 12.5 Mm− 1), representing 17 ± 2% of those values measured within the mixing layer (as for BC, Section 3.2). The shape of the babs (and BC) profiles, characterized by a convergence value above the MH, led us to

a

1.00 0.80

Measurements

0.60

Interpolation

0.40 0.20

Hs 0.00 -0.20 -0.40 -0.60 -0.80 -1.00 0

20

40

60

80

100

120

Abs % (or BC%)

b

9.00 8.00

Measurements

7.00

Interpolation

6.00

Alpe S. Colombano (2280 m asl)

5.00

Hs 4.00 3.00 2.00 1.00 0.00 -1.00 0

20

40

60

80

100

120

Abs % (or BC%) Fig. 8. (a) the statistical mean profile of babs (or BC) along standardized height Hs, as a percentage in relation to the value of babs (or BC) measured at ground-level (denoted as Abs% or BC%); (b) the statistical mean profile of Abs% (or BC%) extrapolated up to 2280 m ASL near the Alpe S. Colombano site. Blue dots indicate points of the statistical mean profile averaged from measured babs (or BC) profiles; the red line indicates the behaviour of the interpolant function; while the asterisk (*) indicates the percentage of EC measured at Alpe S. Colombano compared to that calculated at ground-level in Milan. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


12


calculate a statistical mean profile in order to account for this behaviour. A simple average of the data collected at the same height is not the proper metric, as such an average does not take into account the difference in MH at different times, and can only produce smoothed profiles not clearly shaped by atmospheric turbulence, as is the case in Morgan et al. (2009). One way of avoiding this problem is to average vertical profile data for a standardized height (here denoted as Hs) which assumes a value of 0 at the MH, and values of −1 and 1 at ground-level and at twice the MH, respectively (Ferrero et al., 2010). The standardized height (Hs) enables us to average out data in relation to their position at the MH, and can be defined as: Hs =

z−MH MH

ð8Þ

where z is the absolute height above ground. Averaging babs (or BC) data along Hs enables us to compute the statistical mean profile by taking the daily evolution of MH into account. Fig. 8a shows the behaviour of babs (or BC) along Hs: this behaviour is reported as a percentage, and refers to the value of σabs (or BC) measured at ground-level, denoted as Abs% (or BC%); this enables us to draw the shape of a generic profile regardless of the initial ground value of babs (or BC). Fig. 8a shows that the mean profile of Abs% is characterized by: 1) a sharp decrease at the point where Hs = 0 (which corresponds to the MH); 2) higher values near ground-level compared to the whole mixing layer (Section 3.2.2); 3) a lower variance above the MH, pointing to the presence of a relatively constant value of babs (or BC), as already mentioned in this section and in Section 3.2.2. The presence of a clearly defined mean profile enabled us to interpolate it in order to find a generic function to describe the behaviour of Abs% (or BC%) along the new standardized height Hs. This is a realistic approach considering that, as was mentioned in Sections 3.2.1 and 3.2.2, meteorological parameters and aerosol number concentration along vertical profiles on the 2nd and 3rd of December, were very close to the seasonal mean over Milan, and so these two days could be considered to be representative of winter over the city of Milan. One way of building a generic function for Abs% (or BC%) is to consider a weighted sum of powers of Hs (up to four), in order to describe the non-linear relationship between Abs% (or BC%) and the standardized height (Ferrero et al., 2010). The statistical mean profile of Abs% (or BC%) can be described using the following equation: 2

ð2−iÞ

∑ p2−i Hs

Abs% =

i=0 4

ð9Þ ð4−jÞ

∑ q4−j Hs

j=0

where p2 − i and q4 − j are specific parameters designed to represent the shape of Abs% (or BC%) (Table 6). The rational of the two polynomials in Eq. (9) enables us to shape the profile by taking the MH into account, and also to estimate a ~ 1/H2s dependence of Abs% (or BC%) when Hs is NN1 (going towards the free troposphere). The effectiveness of Eq. (9) in parametrizing Abs% (or BC%) is shown in Fig. 8a, by comparing it with the experimental mean profile (R2 = 0.993, RMSE = 3.17%): Eq. (8) clearly describes the aforementioned characteristics 1), 2) and 3) of the babs (or BC) mean profile.

Moreover, we tested Eq. (9) by trying to predict BC% at 2280 m ASL at the Alpe S. Colombano site, in the free troposphere; in this process, we considered the median value of the MH estimated during the campaign (264 m). The result, portrayed in Fig. 8b, clearly shows how the parametrization of BC% (predicted BC% = 1.13%) is close to the percentage of EC measured at the Alpe S. Colombano site compared with the ground-level EC measured in Milan (EC% = 1.10 ± 0.10%). A second test was conducted by predicting the absorption Aerosol Optical Depth (AODabs) integrating the Abs% along the atmosphere up to 5000 m; AODabs was 0.020, in keeping with the estimate from AERONET data: 0.015. Therefore, Eq. (9) can be used to describe the behaviour of black carbon and the absorption coefficient as a function of height, at least over Milan, once the MH and the absolute ground-values of babs (or BC) are known. This lays the basis for the development of valid parameterizations of vertical profile data, which are useful both in remote sensing and in climatic studies; this is of vital importance over highly built-up areas such as the Po Valley, where BC aerosol is one of the main aerosol components, as Putaud et al. (2004) and Baltensperger et al. (2002) have shown. The parameterization that emerges is also based on results derived from robust, inexpensive techniques, which are very useful in collecting statistically significant long-term data series to be used to calculate the main atmospheric aerosol behaviour (Ferrero et al., 2010).

4. Conclusions Vertical profiles of black carbon and aerosol number-size distribution were measured on the 2nd and 3rd of December 2008, using a tethered balloon fitted with the newly-developed micro-Aethalometer (microAeth® Model AE51) and an OPC (Grimm 1.108). BC profiles clearly identified the mixing layer boundary (MH), which was characterized by a strong vertical concentration gradient. BC profiles also showed a shallow layer of increased concentrations close to the ground, due to the proximity of combustion sources. This ground-layer was +24 ± 4% higher than the average BC concentrations measured within the whole mixing layer. Fairly constant concentrations of BC were found above the MH, representing 17 ± 2% of those BC concentrations measured within the mixing layer. The BC fraction of aerosol volume fell to 48 ± 8% above the MH, compared to ground-level data. This caused a change in the optical absorption properties of the aerosol at different heights. This result was confirmed by separate analyses of the chemical composition of particles taken from PM2.5 samples, collected at the same time at both ground-level and at a high altitude site. Profiles of the absorption coefficient (babs) were calculated from the Aethalometer measurements. In order to do so, an optical enhancement factor (C) for the new microAeth® Model AE51 was calculated for the first time. Mie calculations were applied to the OPC to correct the number-size distribution data, and to calculate the aerosol optical properties as a function of height. Aerosol chemical composition data was used to calculate the aerosol refractive index by means of the Bruggeman mixing rule. The comparison between the Aethalometer attenuation coefficient and the aerosol optical properties estimated from OPC enabled us to calculate the optical enhancement factor (C = 2.05 ± 0.03) for the new microAeth® Model AE51 at 880 nm, and thus to calculate the absorption coefficient profiles using this factor.

Table 6 Estimated parameters for the parameterization of Abs% and BC% vertical profiles over Milan. Model parameters

p0

p1

p2

q0

q1

q2

q3

q4

Estimated values Model performance

9.95 RSS = 120.70

18.59

105.40 R2 = 0.993

0.25

0.58 R2adj = 0.990

4.35

3.71 RMSE = 3.171

1.00



A statistical mean profile was established for babs and BC data, in order to better describe their behaviour. The statistical mean profile referred to a standardized height (Hs). A simple model, based on the same standardized height, was successfully created in order to simulate the vertical behaviour of babs and BC. This model was able to represent the main characteristics of the mean profile, namely: 1) a sharp decrease at the MH, 2) higher values of babs (and BC) near the ground than in the entire mixing layer, 3) a lower degree of variance above the MH, thus suggesting the presence of fairly constant values of babs (and BC) 4) in the free troposphere: the model estimated a ~1/H2s dependence of babs (and BC). The validity of this parametrization was confirmed by comparing the predicted BC% at 2280 m asl at the Alpe S. Colombano site in the free troposphere (predicted BC% = 1.13%) to the percentage of EC measured at Alpe S. Colombano compared to the ground-level EC readings taken in Milan (EC% = 1.10 ± 0.10%). This formed the basis for the development of parametrizations of vertical profile data to be used in remote sensing and climatic studies.

Acknowledgements This paper presents some results from the Italian SATMAP (Mapping particulate matter from satellite) project. Grisa Mocnik would like to thank the Slovenian Ministry of Higher Education, Science and Technology for the financial support (grant 3211-09000304). We thank Giuseppe Zibordi for his effort in establishing and maintaining the AERONET Ispra site. We acknowledge Friedhelm Schnider for information concerning the optical and laser parameters of the Grimm 1.108 OPC.

References Ackerman AS, Toon OB, Stevens DE, Heymsfield AJ, Ramanathan V, Welton EJ. Reduction of tropical cloudiness by soot. Science 2000;288:1042–7. Ackerman TP, Toon OB. Absorption of visible radiation in atmosphere containing mixtures of absorbing and nonabsorbing particles. Applied Optics 1981;20(20): 3661–8. Amiridis V, Melas D, Balis DS, Papayannis A, Founda D, Katragkou E, et al. Aerosol Lidar observations and model calculations of the Planetary Boundary Layer evolution over Greece, during the March 2006 total solar eclipse. Atmos Chem Phys 2007;7: 6181–9. Angelini F, Barnaba F, Landi TC, Caporaso L, Gobbi GP. Study of atmospheric aerosols and mixing layer by lidar. Radiat Prot Dosimetry 2009;137(3–4):275–9. Arnott WP, Hamasha K, Moosmüller H, Sheridan PJ, Ogren JA. Towards aerosol lightabsorption measurements with a 7-wavelength aethalometer: evaluation with a photoacoustic instrument and 3-wavelength nephelometer. Aerosol Sci Technol 2005;39:17–29. Aspnes DE. Local-field effects and effective medium theory: a microscopic perspective. Am J Phys 1982;50(8):704–9. Badger CL, George I, Griffiths PT, Braban CF, Cox RA, Abbatt JPD. Phase transitions and hygroscopic growth of aerosol particles containing humic acid and mixtures of humic acid and ammonium sulphate. Atmos Chem Phys 2006;6:755–68. Baltensperger U, Streit N, Weingartner E, Nyeki S, Prévôt ASH, Van Dingenen R, et al. Urban and rural aerosol characterization of summer smog events during the PIPAPO field campaign in Milan, Italy. J Geophys Res 2002;107(D22):8193. doi: 10.1029/2001JD001292. Baron PA, Willeke K. Aerosol measurements. Principles, techniques and applications. Second edition. Wiley-Interscience; 2005. Belis CA, Vannini P, Moraschetti G, Fermo P, Piazzalunga A, Mognaschi G, et al. PM mass concentration and chemical composition in the alpine remote site Bormio-San Colombano (N. Italy; 2,200 m a.s.l.). Chem Eng Trans 2006;10:65–70. Bohren CF, Huffman DR. Absorption and scattering of light by small particles. New York, NY: John Wiley; 1983. Bond TC, Anderson TL, Campbell D. Calibration and intercomparison of filter-based measurements of visible light absorption by aerosols. Aerosol Sci Technol 1999;30: 582–600. Bruggeman D. Calculation of various physics constants in heterogenous substances. I. Dielectricity constants and conductivity of mixed bodies from isotropic substances. Ann Phys 1935;24:636–64. Chazette P, Liousse C. A case study of optical and chemical ground apportionment for urban aerosols in Thessaloniki. Atmos Environ 2001;35:2497–506. Chýlek P, Videen G, Ngo D, Pinnick R, Klett J. Effect of black carbon on the opticalproperties and climate forcing of sulfate aerosols. J Geophys Res 1995;100: 16325–32.

13

Corrigan CE, Roberts GC, Ramana MV, Kim D, Ramanathan V. Capturing vertical profiles of aerosols and black carbon over the Indian Ocean using autonomous unmanned aerial vehicles. Atmos Chem Phys 2008;8:737–47. D'Almeida GA, Koepke P, Shettle EP. Atmospheric aerosols global climatology and radiative characteristics. A. Deepak Publishing: Hampton; 1991. Deshler T, Hervig ME, Hofmann DJ, Rosen JM, Liley JB. Thirty years of in situ stratospheric aerosol size distribution measurements from Laramie, Wyoming (41_N), using balloon-borne instruments. J Geophys Res 2003;108(D5):4167. doi: 10.1029/2002JD002514. Dong Z, Li Z, Xiao C, Wang F, Zhang M. Characteristics of aerosol dust in fresh snow in the Asian dust and non-dust periods at Urumqi glacier no. 1 of eastern Tian Shan, China. Environ Earth Sci 2010;60:1361–8. Ebert M, Weinbruch S, Rausch A, Gorzawski G, Hoffmann P, Wex H, et al. Complex refractive index of aerosols during LACE 98 as derived from the analysis of individual particles. J Geophys Res 2002;107(D21):8121. doi:10.1029/2000JD000195. Eresmaa N, Karpinnen A, Joffre SM, Räsänen J, Talvitie H. Mixing height determination by ceilometer. Atmos Chem Phys 2006;6:1485–93. Ferrero L, Bolzacchini E, Perrone MG, Del Nevo E, Belis C, Gianelle V. Transport and deposition of particle-bound PAHs at an high altitude alpine site (Oga-Bormio m.2280, Italy). European Aerosol Conference; 2005. abstract book, 187. Ferrero L, Bolzacchini E, Petraccone S, Perrone MG, Sangiorgi G, Lo Porto C, et al. Vertical profiles of particulate matter over Milan during winter 2005/2006. Fresenius Environ Bull 2007;16(6):697–700. Ferrero L, Perrone MG, Petraccone S, Sangiorgi G, Ferrini BS, Lo Porto C, et al. Verticallyresolved particle size distribution within and above the mixing layer over the Milan metropolitan area. Atmos Chem Phys 2010;10:3915–32. Fierz-Schmidhauser R, Zieger P, Gysel M, Kammermann L, DeCarlo PF, Baltensperger U, et al. Measured and predicted aerosol light scattering enhancement factors at the high alpine site Jungfraujoch. Atmos Chem Phys 2010;10:2319–33. Gebhart KA, Malm WC. Examination of the effects of sulfate acidity and relative humidity on light scattering at Shenandoah National Park. Atmos Environ 1994;28 (5):841–9. Gualtieri M, Mantecca P, Corvaja V, Longhin E, Perrone MG, Bolzacchini E, et al. Winter fine particulate matter from Milan induces morphological and functional alterations in human pulmonary epithelial cells (A549). Toxicol Lett 2009;188:52–62. Guyon P, Boucher O, Graham B, Beck J, Mayol-Bracero OL, Roberts GC, et al. Refractive indexof aerosol particles over the Amazon tropical forest during LBA-EUSTACH 1999. J Aerosol Sci 2003;34:883–907. Hand JL, Kreidenweis SM. A new method for retrieving particle refractive index and effective density from aerosol size distribution data. Aerosol Sci Technol 2002;36: 1012–26. Hänel G. Vertical profiles of the scattering coefficient of dry atmospheric particles over Europe normalized to air at standard temperature and pressure. Atmos Environ 1998;32:1743–55. Heim M, Mullins BJ, Umhauer H, Kasper G. Performance evaluation of three optical particle counters with an efficient “multimodal” calibration method. J Aerosol Sci 2008;39:1019–31. Heller W. Remarks on refractive index mixture rules. J Phys Chem 1965;69(4):1123–9. Hess WH, Herd CR. Microstructure, morphology, and general physical properties, in Carbon Black. In: Donnet J, Bansal R, Wang M, editors. Boca Raton. Fla: CRC Press; 1993. p. 89-173. Hesterberg T, Bunn III WB, Chase GR, Valberg PA, Slavin TJ, Lapin CA, et al. A critical assessment of studies on the carcinogenic potential of diesel exhaust. Crit Rev Toxicol 2006;36:727–76. Heyder J, Gebhart J. Optimization of response functions of light scattering instruments for size evaluation of aerosol particles. Appl Opt 1979;18(5):705–11. Horvath H. Atmospheric light absorption—A review. Atmos Environ 1993;27A(3): 293–317. Horvath H. Influence of atmospheric aerosols upon the global radiation balance. In: Harrison RM, Van Grieken R, editors. Atmospheric particles. New York: Wiley; 1998. p. 543. Howell SG, Clarke AD, Shinozuka Y, Kapustin V, McNaughton CS, Huebert BJ, et al. Influence of relative humidity upon pollution and dust during ACE-Asia: size distributions and implications for optical properties. J Geophys Res 2006;111: D06205. doi:10.1029/2004jd005759. Hueglin C, Gehrig R, Baltensperger U, Gyselc M, Monnd C, Vonmonta H. Chemical characterisation of PM2.5, PM10 and coarse particles at urban, near-city and rural sites in Switzerland. Atmos Environ 2005;38:3305–18. Intergovernmental Panel on Climate Change (IPCC). Climate change 2007: the scientific basis. In: Solomon S, editor. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. New York: Cambridge Univ. Press; 2007. p. 2007. Ishimaru A. Wave propagation and scattering in random media, Vol. 1. Orlando, FL: Academic Press; 1978. Janzen J. The refractive index of colloidal carbon. J Colloid Interface Sci 1979;69(3): 436–47. Kaufman YJ, Tanré D, Boucher O. A satellite view of aerosols in the climate system. Nature 2002;419:215–23. Kim SW, Yoon SC, Won JG, Choi SC. Ground-based remote sensing measurements of aerosol and ozone in an urban area: a case study of mixing height evolution and its effect on ground-level ozone concentrations. Atmos Environ 2007;41: 7069–81. Koren I, Kaufman YJ, Remer LA, Martins JV. Measurements of the effect of Amazon smoke on inhibition of cloud formation. Scienze 2004;303:1342–5. Koren I, Martins JV, Remer LA, Afargan H. Smoke invigoration versus inhibition of clouds over the Amazon. Science 2008;321:946–9.


14


Lesins G, Chýlek P, Lohmann U. A study of internal and external mixing scenarios and its effect on aerosol optical properties and direct radiative forcing. J Geophys Res 2002;107:4094. doi:10.1029/2001JD000973. Levoni C, Cervino M, Guzzi R, Torricella F. Atmospheric aerosol optical properties: a database of radiative characteristics for different components and classes. Appl Opt 1997;36(30):8031–41. Liu Y, Daum PH. The effect of refractive index on size distributions and light scattering coefficients derived from optical particle counters. J Aerosol Sci 2000;31(8):945–57. Liu Y, Daum PH. Relationship of refractive index to mass density and self-consistency of mixing rules for multicomponent mixtures like ambient aerosols. J Aerosol Sci 2008;39:974–86. Maletto A, McKendry, Strawbridge KB. Profiles of particulate matter size distributions using a balloon-borne lightweight aerosol spectrometer in the planetary boundary layer. Atmos Environ 2003;37:661–70. Mar TF, Norris GA, Koenig JQ, Larson TV. Association between air pollution and mortality in phoenix, 1995–1997. Environ Health Perspect 2000;108:347–53. Marley NA, Gaffney JS, Baird JC, Blazer CA, Drayton PJ, Frederick JE. An empirical method for the determination of the complex refractive index of size-fractionated atmospheric aerosols for radiative transfer calculations. Aerosol Sci Technol 2001;34:535–49. Mätzler C. MATLAB functions for mie scattering and absorption. Institut für Angewandte Physik research report; 2002. No. 02–08, June. McKendry IG, Sturman AP, Vergeiner J. Vertical profiles of particulate matter size distributions during winter domestic burning in Christchurch, New Zealand. Atmos Environ 2004;38:4805–13. Moosmüller H, Chakrabarty RK, Arnott WP. Aerosol light absorption and its measurement: a review. J Quant Spectrosc Radiative Transf 2009;110:844–78. Morgan WT, Allan JD, Bower KN, Capes G, Crosier J, Williams PI, et al. Vertical distribution of sub-micron aerosol chemical composition from North-Western Europe and the North-East Atlantic. Atmos Chem Phys 2009;9:5389–401. Penner JE, Andreae M, Annegarn H, Barrie L, Feichter J, Hegg D, et al. Aerosols, their direct and indirect effects. Climate change 2001: the scientific basis. Cambridge, UK: Cambridge University Press; 2001. Perrone MG, Gualtieri M, Ferrero L, Lo Porto C, Udisti R, Bolzacchini E, et al. Seasonal variations in chemical composition and in vitro biological effects of fine PM from Milan. Chemosphere 2010;78:1368–77. Pesava P, Horvath H, Kasahara M. A local optical closure experiment in Vienna. J Aerosol Sci 2001;32:1249–67. Podgorny IA, Ramanathan V. A modeling study of the direct effect of aerosols over the tropical Indian Ocean. J Geophys Res 2001;106(D20):24097–105. Potukuchi S, Wexler AS. Identifying solid-aqueous phase transitions in atmospheric aerosols-I. Neutral-acidity solutions. Atmos Environ 1995a;29(14):1663–76. Potukuchi S, Wexler AS. Identifying solid-aqueous phase transitions in atmospheric aerosols-II. Acidic solutions. Atmos Environ 1995b;29(22):3357–64. Putaud JP, Raes F, Van Dingenen R, Brüggemann E, Facchini MC, Decesari S, et al. A European aerosol phenomenology—2:chemical characteristics of particulate matter at kerbside, urban, rural and background sites in Europe. Atmos Environ 2004;38:2579–95. Ramana MV, Ramanathan V, Kim D, Roberts GC, Corrigan CE. Albedo, atmospheric solar absorption and heating rate measurements with stacked UAVs. Q J R Meteorol Soc 2007;133:1913–31. Ramanathan V, Carmichael G. Global and regional climate changes due to black carbon. Nat Geosci 2008;1:221–7. Ramanathan V, Feng Y. Air pollution, greenhouse gases and climate change: global and regional perspectives. Atmos Environ 2009;43:37–50. Ramanathan V, Crutzen PJ, Kiehl JT, Rosenfeld D. Aerosols, climate, and the hydrological cycle. Science 2001;294:2119–24. Randriamiarisoa H, Chazette P, Couvert P, Sanak J, Mégie G. Relative humidity impact on aerosol parameters in a Paris suburban area. Atmos Chem Phys 2006;6: 1389–407. Raut JC, Chazette P. Vertical profiles of urban aerosol complex refractive index in the frame of ESQUIF airborne measurements. Atmos Chem Phys 2008;8:901–19. Rees SL, Robinson AL, Khlystov A, Stanier CO, Pandis SN. Mass balance closure and the Federal Reference Method for PM2.5 in Pittsburgh, Pennsylvania. Atmos Environ 2004;38:3305–18.

Riemer N, Vogel H, Vogel B, Fiedler F. Modeling aerosols on the mesoscale-γ: treatment of soot aerosol and its radiative effects. J Geophys Res 2003;108(D19):4601. doi: 10.1029/2003JD003448. Schmid O, Artaxo P, Arnott WP, Chand D, Gatti LV, Frank GP, et al. Spectral light absorption by ambient aerosols influenced by biomass burning in the Amazon Basin. I: Comparison and field calibration of absorption measurement techniques. Atmos Chem Phys 2006;6:3443–62. Schmid O, Chand D, Karg E, Guyon P, Frank GP, Swietlicki E, et al. Derivation of the density and refractive index of organic matter and elemental carbon from closure between physical and chemical aerosol properties. Environ Sci Technol 2009;43: 1166–72. Schneider J, Hings SS, Hock BN, Weimer S, Borrmann S, Fiebig M, et al. Aircraft-based operation of an aerosol mass spectrometer: measurements of tropospheric aerosol composition. J Aerosol Sci 2006;37:839–57. Schumann T. On the use of a modified clean-room optical particle counter for atmospheric aerosols at high relative humidity. Atmos Res 1990;25:499–520. Schuster GL, Dubovik O, Holben BN, Clothiaux EE. Inferring black carbon content and specific absorption from Aerosol Robotic Network (AERONET) aerosol retrievals. J Geophys Res 2005;110:D10S17. doi:10.1029/2004JD004548. Seinfeld JH, Pandis SN. Atmospheric chemistry and physics — from air pollution to climate change. Wiley-Interscience edition; 1998. Shettle EP, Fenn RW. Models of atmospheric aerosols and their optical properties, in: Optical Properties in the Atmosphere. 1976; AGARD-Cp-183, NTIS, ADA 028615. Shettle EP, Fenn RW. Models for the aerosol of the lower atmosphere and the effect of humidity variations on their optical properties. 1979; AFGL-TR-79-0214, Environmental Research Paper No 675, NTIS, ADA 085951. Stier P, Seinfeld JH, Kinne S, Feichter J, Boucher O. Impact of nonabsorbing anthropogenic aerosols on clearsky atmospheric absorption. J Geophys Res 2006;111:D18201. doi:10.1029/2006JD007147. Stier P, Seinfeld JH, Kinne S, Boucher O. Aerosol absorption and radiative forcing. Atmos Chem Phys 2007;7:5237–61. Stratmann F, Siebert H, Spindler G, Wehner B, Althausen D, Heintzenberg J, et al. Newparticle formation events in a continental boundary layer: first results from the SATURN experiment. Atmos Chem Phys 2003;3:1445–59. Subramanian R, Neil M, Donahue NM, Bernardo-Bricker A, Rogge WF, Robinson AL. Insights into the primary-secondary and regional-local contributions to organic aerosol and PM2.5 mass in Pittsburgh, Pennsylvania. Atmos Environ 2007;41: 7414–33. Sun Y, Zhang Q, Macdonald AM, Hayden K, Li SM, Liggio J, et al. Size-resolved aerosol chemistry on Whistler Mountain, Canada with a high-resolution aerosol mass spectrometer during INTEX-B. Atmos Chem Phys 2009;9:3095–111. Taubman BF, Hains JC, Thompson AM, Marufu LT, Doddrige BG, Stehr JW, et al. Aircraft vertical profiles of trace gas and aerosol pollution over the mid-Atlantic United States: statistics and meteorological cluster analysis. J Geophys Res 2006;111: D10S07. doi:10.1029/2005JD006196. Turpin BJ, Lim HJ. Species contributions to PM2.5 mass concentrations: revisiting common assumptions for estimating organic mass. Aerosol Sci Technol 2001;35: 602–10. Van de Hulst HC. Light scattering by small particles, (1957). . Reprinted byNew York, NY: Dover Publication; 1981. Van Dingenen R, Raes F, Putaud JP, Baltensperger U, Charron A, Facchini MC, et al. A European aerosol phenomenology-1: physical characteristics of particulate matter at kerbside, urban, rural and background sites in Europe. Atmos Environ 2004;38: 2561–77. Vedal S. Ambient particles and health: lines that divide. J Air Waste Manage Assoc 1997;47:551–81. Velazco-Roa MA, Thennadil SN. Estimation of optical constants from multiple-scattered light using approximations for single particle scattering characteristics. Applied Optics 2007;46(35):8453–60. Weingartner E, Saathoff H, Schnaiter M, Streit N, Bitnar B, Baltensperger U. Absorption of light by soot particles: determination of the absorption coefficient by means of aethalometers. J Aerosol Sci 2003;34:1445–63. World Climate Programme WCP-112. A preliminary cloudless standard atmosphere for radiation computation. Geneva: World Meteorological Organization, WMO/TD-No 24; 1986.
