the Creative Commons Attribution 3.0 License. Atmospheric Chemistry and Physics

Organic matter represents an important frac- tion of the fine particle aerosol, yet our knowledge of the roles of organics in the activation of aerosol particles into cloud droplets is poor. A cloud condensation nucleus (CCN) counter is used to examine the relative growth rates of cloud droplets for case studies from field measurements on the North Pacific Ocean and in a coniferous forest. A model of the condensational growth of water droplets, on particles dissolving according to their solubility in water, is used to simulate the initial scattering of the droplets as they grow in the CCN counter. Simulations of the growth rates of fine particles sampled in the marine boundary layer of the North Pacific Ocean shows no evidence of natural marine organic material contributing to the CCN water uptake but there is an indication of an influence from organics from diesel ship emissions on the size distribution of sulphate and the abil- ity of these particles to act as CCN. Simulations of the ob- servations of water uptake on biogenic organic aerosol par- ticles sampled in a coniferous forest indicate an impact of the organic on the water uptake rates, but one that is still smaller than that of pure sulphate. The existence of organics becomes important in determining the water uptake as the or- ganic mass increases relative to sulphate. The values of the organic component of the hygroscopicity parameter that describes the CCN activity were found to be negligible for the marine particles and 0.02-0.05 for the forest particles.

North Pacific Ocean and in a coniferous forest. A model of the condensational growth of water droplets, on particles dissolving according to their solubility in water, is used to simulate the initial scattering of the droplets as they grow in the CCN counter. Simulations of the growth rates of fine particles sampled in the marine boundary layer of the North Pacific Ocean indicate that the main influence of the marine organic mate-10 rial on the water uptake rate is from its effect on the size distribution of the sulphate. Simulations of the observations of water uptake on biogenic organic aerosol particles sampled in a coniferous forest indicate an impact of the organic on the water uptake rates, but one that is still smaller than that of pure sulphate. The solubility of the organic becomes an important factor in determining the water uptake as the organic mass in-

Introduction
One of the largest uncertainties in determining the human impact on the global climate 20 is the effect of aerosol particles on clouds or the indirect effect (IPCC, 2007). Atmospheric aerosol particles have varying abilities to nucleate cloud droplets depending on the size and composition of the particles. Historically, it was believed that primarily highly soluble inorganic species, such as sulphate and sea salt, contributed to cloud droplet growth. More recently, it has become apparent that organics are a significant Introduction Interactive Discussion initial growth of aerosols as they activate into cloud droplets is examined using measurements from two field projects, one sampling from a ship on the North Pacific Ocean as part of the Canadian SOLAS study (C-SOLAS) and one sampling in Golden Ears Provincial Park (GE) on the west coast of Canada. These sites provide strong contrasts, a marine environment with the aerosol mass dominated by sulphate (Phinney 5 et al., 2006) and a forest site with a small urban influence and dominated by organic aerosol (Shantz et al., 2004). A model that treats changes in droplet growth with respect to time and the Mie scattering off of these droplets is used to simulate the CCN measurements and to study the differences of the two contrasting aerosol regimes on the nucleation of cloud droplets. 10 As it can be difficult to characterize the water activity of organic aerosols due to their complexity and insufficient information about their chemical composition, a single hygroscopicity parameter that considers both hygroscopic and CCN activity (κ: Petters and Kreidenweis, 2007) was also determined. Using this parameter can simplify the physical and chemical information needed for describing activation in models.

Particle chemical constituents
During the C-SOLAS study, a Quadrupole Aerodyne Aerosol Mass Spectrometer (Q-AMS, see Jayne et al., 2000) provided the size distribution of sulphate, nitrate, ammonium, methanesulphonic acid (MSA) and total organics in the aerosol. In the Q-AMS, the particles are focused into a narrow beam and directed into an oven, which 5 vapourises the particles. The vapourised components of the particles are ionized by electron impact. The resulting ion fragments are detected with a quadrupole mass spectrometer. A comprehensive table of fragmentation patterns is used to combine the fragments over 300 amu into mass concentrations of the above major compounds. The MSA fragmentation patterns were first implemented in this study and were determined 10 in laboratory experiments (Phinney et al., 2006). In addition to the total sampled particle masses, the composition is also determined as a function of particle size. The particles are sized based on the time it takes for a particle to travel from the inlet to the detection point. This size is measured as a vacuum aerodynamic diameter (D va ). A comprehensive description of the design of the Q-AMS, its operation, quantification 15 methods and calibration procedures are given elsewhere (Alfarra et al., 2004;Allan et al., 2003a;Allan et al., 2003b;Jayne et al., 2000;Jimenez et al., 2003). Also during C-SOLAS, a Micro-Orifice, Uniform-Deposit Impactor (MOUDI) was used to collect particles for major ion analysis using ion chromatography (IC). This was done using Teflon filters on each of the 12 stages of the MOUDI. Despite its name, the 20 MOUDI was operated in a static mode as opposed to rotating mode that gives rise to the uniform-deposit. The static operation was sufficient as the filters were completely extracted in de-ionized water prior to analysis by IC. The MOUDI data in this work was used to determine the contribution from sea salt. The agreement between the Q-AMS measurements and the MOUDI results for chemical species measured using Introduction Particles for analysis of total mass concentrations of particulate organic carbon (OC), black carbon (BC) and water-soluble organic carbon (WSOC) were collected on single Quartz filters. The BC and OC concentrations were determined by a modified NIOSH method using a Sunset Labs Thermal Optical Transmission Instrument (Lee et al.,5 2003; Sharma et al., 2002). Water-soluble organic carbon (WSOC) was also measured in samples from the quartz filters as described by Shantz et al. (2004).

Water uptake: cloud condensation nucleus counter
A University of Wyoming Model MA100 static thermal gradient cloud condensation nucleus counter (CCNc) was used for the CCN measurements during both studies (Shantz et al., 2003;Snider et al., 2006). Light scattered off the growing droplets within the supersaturated environment of the chamber is detected as a voltage from a photodiode. This voltage is proportional to both the number concentrations and sizes of the growing droplets. For a given supersaturation, the size of the droplets is dependent on their composition, and thus it is difficult to distinguish the effects of number and size 15 from the voltage measurement unless one is relatively constant. For a given particle composition, the difference between the maximum voltage (which is determined when the droplets are at their largest size prior to falling out of the detection region) and the baseline voltage (the signal prior to the exposure of the particles to a water supersaturation in the chamber) is proportional to the number concentration of the droplets. Introduction  Shantz et al. (2003) and examines the variation in the light scattered by the particles and growing droplets. The rate of change in the scattered light, represented by the time-resolved voltage, is proportional to the growth rate into cloud droplets. CCNc supersaturations should be experimentally determined and not assumed to be 5 the theoretical values (Frank et al., 2007). The effective supersaturations (hereafter: S eff ) within the CCNc were estimated to be 0.19-0.50% using nearly monodisperse ammonium sulphate particles generated in the laboratory, as discussed in Shantz et al. (2003). All supersaturations discussed here are based on these calibrations. The evaluation of S eff results in the largest uncertainty in these measurements. The errors 10 in S eff are estimated as approximately ±15%. It is difficult to evaluate how this error affects the growth rates from the observations, and so this uncertainty is applied in a model and reflected as the maximum and minimum limits of S eff .

Model description
3.1 Simulations of cloud droplet growth rates in the CCNc

15
An adiabatic cloud parcel model (hereafter called the "ACP model") described by Leaitch et al. (1986) was modified to include partially soluble material (Lohmann et al., 2004;Shantz et al., 2003) and to simulate the growth rates in the CCNc (hereafter referred to as the "CCNc model") (Shantz et al., 2003). The difference between the ACP model and the CCNc model is how the supersaturation (S) is calculated. In the 20 ACP model, the rate of cooling of the air parcel is balanced by the rate of water uptake to determine S. In the CCNc model, S is specified by a transient rise (that occurs in the CCNc after the flush of particles into the CCNc chamber) followed by a constant S, in an attempt to mimic the variation of S in the chamber. The transient S is determined from comparisons of simulated growth rates with those measured for pure, Introduction The mass accommodation coefficient (α c ) can have a strong effect on the simulated condensation rate of water, and this is an outstanding uncertainty in quantifying the indirect effect of aerosols on climate forcing. There are a wide range of α c values used in the literature spanning two orders of magnitude, from 0.01 to 1 (e.g. Davidovits et al., 2004;Laaksonen et al., 2005;Marek and Straub, 2001;Mozurkewich, 1986) and 5 the references therein). For cases where an organic film may form on the surface of the cloud droplet, it has been suggested that the value of α c could be as low as 10 −5 (Chuang, 2003;Feingold and Chuang, 2002). Ruehl et al. (2008) determined α c for field measurements (normalized to α c for laboratory ammonium sulphate) and found that 60% of ambient CCN grew similarly to ammonium sulphate but a number of cases 10 had a low α c compared to ammonium sulphate. We did not have a measurement of the droplet size from this CCNc as was the case for Ruehl et al. (2008) so we were unable to determine α c using their method. Also, α c can not be determined from the growth rates here, as attempted by Stroud et al. (2007), because of the uncertainties in these CCNc's such as turbulence and edge effects. Consequently, α c is assumed to be unity 15 in all of the present simulations as suggested in many of the above references. What this means for this study is that although the effect of the organic on water uptake is represented primarily by the dissolution of the organic, any real impact of the organic could alternatively or in some combination be through effects on α c or the surface tension. Introduction

Simulations of organics
All the organics in the model are assumed to resemble adipic acid in terms of the physical and chemical properties, except for solubility. The latter is varied. This assumption is used because the CCN and water uptake properties of adipic acid (Table 1) have 5 been widely studied (e.g., Bilde and Svenningsson, 2004;Broekhuizen et al., 2004;Corrigan and Novakov, 1999;Cruz and Pandis, 1997;Ervens et al., 2004;Hori et al., 2003;Joutsensaari et al., 2001;Prenni et al., 2001;Raymond and Pandis, 2002;Shulman et al., 1996) and adipic acid has been observed in the atmosphere (Grosjean et al., 1978;Kawamura and Yasui, 2005;Wang et al., 2006). The determination of the hygroscopicity parameter κ (Petters and Kreidenweis, 2007) (Sect. 5) is also a means of evaluating this assumption. Sensitivity to solubility is examined because the previous laboratory work and simulations suggest that solubility is an important factor when dealing with organics (Shantz et al., 2003) and it is better constrained than other properties such as the surface ten-15 sion and the accommodation coefficient. Thus, the dissolution of the organic in the model depends on its solubility and is based on its equilibrium with the amount of water present on the particles and droplets at each time step, as described in Shantz et al. (2003). The dissolution kinetics discussed in Asa-Awuku and Nenes (2007), which is more important at low temperatures, was not accounted for in this work. The surface 20 tension is defined according to Ervens et al. (2004) and is shown in Table 1. This does not take into account the possibility of a lower solubility film-forming organic compound (Feingold and Chuang, 2002), the effects of which are uncertain (Kanakidou et al., 2005). Over the oceans, sulphate and methanesulphonic acid (MSA) are formed from the oxidation of dimethyl-sulphide (CH 3 SCH 3 ) (Seinfeld and Pandis, 1998). Natural organic material, other than MSA, is also present in fine particles in marine environments 15 (e.g. O'Dowd et al., 2004;Phinney et al., 2006), and ships contribute primary organic particles as well as SO 2 (e.g. Hobbs et al., 2000). In the absence of significant anthropogenic aerosols, either from long-range transport or from ship emissions, the marine aerosol consists primarily of sulphate, MSA and sea salt as well as low concentrations of organics. This was the scenario for most of the C-SOLAS measurements (Phinney Introduction The Q-AMS time series of mass concentrations of total particulate sulphur (sulphate + MSA) and total organics for C-SOLAS are shown in Fig. 1. Periods of fumigation from the ship's exhaust, during times of mooring, have been removed as described by Phinney et al. (2006). With the exception of the 18-19 July period, sulphur dominates 5 the fine particle aerosol. The ammonium mass concentrations (not shown) were low over the time series (on average, the ammonium to sulphate ratio was 0.2) and the sulphate was mostly acidic (Phinney et al., 2006). For 18-19 July, there was a substantial increase in the organics. The wind speeds were relatively low during this time and the air trajectories looped over the ocean. The 10 Q-AMS mass spectrum for this period resembles a mix of a diesel signature and the background organic (Phinney et al., 2006;Phinney et al., 2008 1 ). It is believed that in this case the aerosol was impacted by emissions of ship traffic in the region. Three periods selected for the present study of CCN droplet growth analysis are highlighted in Fig. 1. On 16 July, the fine particle aerosol is made up of 90% sulphate 15 and MSA and 10% organic material. On 27 July, the organic fraction was <10%. The third case, 18 July, is during the peak of the period of highest organic fraction, up to 30%.

Simulations of 27 July 2002, from C-SOLAS
The observed particle number distribution for 27 July is shown in Fig. 2a. It is approxi-20 mated by three lognormal distributions centred at 0.045 µm (mode 1), 0.16 µm (mode 2), and 0.65 µm (mode 3). The chemical mass size distributions from the Q-AMS are compared with the mass size distribution estimated from the SMPS and PCASP in Fig. 2b. The majority of the mass is in sizes >0.1 µm, and there is reasonable agreement between the physical and Q-AMS measurements for mode 2. 25 The Q-AMS measurements lack sufficient sensitivity to resolve the mass in mode 1, so it is assumed to be 100% sulphuric acid. Both the Q-AMS and the measurements ACPD 8,2008 Droplet growth on organic aerosols from field measurements N. C. Shantz et al. Interactive Discussion from the MOUDI indicate the mode 2 particles are composed mostly of acidic sulphate with a smaller amount of MSA. As the Q-AMS total organic is below detection limit in both the time-of-flight and mass measurement modes of the Q-AMS, the organic is assumed to be zero in this case. Assuming that the MSA behaves as sulphate, then mode 2 is modelled as 100% sulphuric acid. Mode 3 was dominated by sea salt Phinney et al., 2006), and it is assumed to consist entirely of NaCl. Table 3 shows the assumptions of the chemical breakdown for all modes for all of the field case studies. Figure 3 shows the scattering cross-sectional area from the simulations compared with the observed voltage normalized to the base voltage from the CCNc. The scale of the right hand axis has been adjusted to match the simulated growth curves with the 10 observed values. As this is a case of nearly pure polydisperse sulphate, this scaling is used for the remaining cases and provides a reference for the other cases to nearly pure sulphuric acid. The error bars reflect the model runs performed at the S eff ±15%, giving the maximum and minimum range in S eff (Sect. 2.3). The minimum activation diameters from the simulations are indicated with vertical lines for the different chamber 15 S eff (Fig. 2a). At the lowest S eff , 0.19%, particles in modes 2 and 3 are activated. For S eff =0.34%, a small fraction of particles are activated in mode 1. At S eff =0.50%, just under 50% of the mode 1 particles are activated. Sensitivity tests with the model indicate that including up to 15% soluble or insoluble organic or changing the sulphuric acid in mode 1 to ammonium bisulphate slightly decreases the growth rate at S eff =0.34% 20 and 0.50%. These same sensitivity tests for mode 2 do not alter the growth rate at all. Overall, because the organic mass was low and these composition sensitivity tests do not modify the results substantially, we believe that this is a good case to use as a reference for the comparisons between the CCNc observed growth rates and CCNc model results. The next case study (16 July) will be used to verify this reference case. Interactive Discussion number concentration in mode 1. Note that on 16 July, mode 2 is split into two modes, 2a at 0.1 µm and 2b at 0.22 µm. Both 2a and 2b are assumed to have the same chemical composition. The Q-AMS measurements for mode 1 are below detection limit. Mode 2 is dominated by acidic sulphate, and organics comprise 10% of the mode 2 mass, exclusive 5 of MSA. There were no MOUDI measurements on 16 July, so mode 3 is assumed to be NaCl, as consistent with the rest of the study.
The simulated and observed CCN growth rates for 16 July are shown in Fig. 5. With the same right-hand axis scaling as used in Fig. 3, the simulated and observed growth rates compare closely, lending confidence that this scaling properly reflects a sulphate 10 reference for the remaining cases. It does not matter what solubility is assumed for the organic; soluble and insoluble organics gave the same results.

Simulations of 18 July 2002, from C-SOLAS
The measured and modelled size distributions for 18 July are shown in Fig. 6a, the case with the highest organic mass concentrations. As discussed in Sect. 4.1.1, the mass 15 spectrum of the organic in this case indicates that there was a significant contribution to this organic from diffuse diesel emissions, and the source was possibly from ships crossing the region. The number distribution is represented in the simulations by four modes: mode 1a centred at 0.035 µm, mode 1b at 0.08 µm (hereafter mode 1a and mode 1b will be referred to as "mode 1"), mode 2 at 0.19 µm and mode 3 at 0.65 µm. 20 The observations of Phinney et al. (2008 1 ) show that the geometric diameter of mode 1b is larger than the mode diameters on 27 and 16 July due to the sulphate condensing on primary organics from the diesel emissions. Figure 6b shows the mass distributions. More information about the chemical composition of mode 1 is available from the Q-AMS for 18 July than 27 or 16 July because 25 mode 1 is larger in size and mass concentration. From the Q-AMS measurements, mode 1 is assumed to be composed of 55% sulphuric acid and 45% organic, and mode 2 is 70% sulphuric acid and 30% organic. These mixtures are assumed to be 8206 ACPD 8,2008 Droplet growth on organic aerosols from field measurements N. C. Shantz et al. MOUDI measurements show mode 3 to be a combination of sulphate, sodium nitrate and sodium chloride. Because all three are quite soluble, the choice is insignificant, and mode 3 is modelled here as pure NaCl. Figure 7a shows the simulations and the CCNc observations of the droplet growth.

5
The S eff error bars are not shown in this case (for simplicity), but all results fall within that uncertainty. All organics in the distribution are assumed to have the same solubility. The shaded areas encompass the growth rates for assumptions ranging between relatively soluble (solubility=200 g L −1 ) and insoluble organic (solubility=10 −6 g L −1 ). The soluble case dissolves readily and accumulates more water early on in the simulation; 10 hence the slightly higher growth curves for the soluble case. Figure 6a shows the minimum particle diameters activated at the different S eff values (the vertical lines) for the insoluble organic case. The soluble organic case activates at the same or very slightly smaller sizes than for the insoluble case.
At S eff =0.19%, all of mode 2 is activating, and there is little difference between the 15 soluble and insoluble organic growth curves because there is only 30% organic in this mode and a small number concentration (the growth is less sensitive to changes in the chemical composition in such a situation). There is slightly more sensitivity to the choice of soluble and insoluble scenarios at S eff =0.34% and 0.5%, because the activation is occurring in regions of the size distribution where the number concentrations are 20 changing more rapidly with size and the organic fraction is higher. Still the relative differences are very small and the results from this case agree with those of Abdul-Razzak and Ghan (2005) where any solubility assumption is a close enough approximation as long as the organic is not assumed to behave as an inorganic. To illustrate the effect of assuming the organic is equivalent to an inorganic, and because that assumption has 25 been used in some in previous closure studies (e.g., Conant et al., 2004), the results of Fig. 7a are reproduced for the case where the entire distribution is assumed to be sulphuric acid (Fig. 7b). In this case, the simulated CCN scattering cross sections are significantly higher than the corresponding observations, and most of the simulated 8207 ACPD 8,2008 Droplet growth on organic aerosols from field measurements N. C. Shantz et al. points lie at the edge of or outside the uncertainty defined by the chamber supersaturation. This result indicates the significant error that arises from assuming the organic portion behaves similarly to the inorganic portion, and underscores the need to treat the components of the aerosol separately. The results shown in Fig. 7 indicate that the organic reduces the effectiveness of the 5 individual particle as a CCN compared to pure inorganics. However, compared with the cases from 27 and 16 July, considerably more of the mode 1 particles activate on 18 July at the higher S eff values. This is why the measured growth rates are higher on 18 July than the other two cases. In this case the organic appears to have acted as a substrate for the sulphate to condense upon (e.g. Virkkula et al., 1999;Phinney et al., 10 2008 1 ), redefining the size distribution of the sulphate. Thus, on 18 July the organic appears to be ineffective at water uptake but it contributes to the formation of a broader, more effective CCN distribution by redistributing the sulphate.

Golden Ears Provincial Park (GE) field study
The GE provincial park study was part of the Pacific 2001 field campaign (e.g. Li, 2004).

15
The park (covered predominantly with tall coniferous trees) is on the north slope (203 m elevation) of the Lower Fraser Valley and about 45 km northeast of Vancouver, British Columbia. Aerosol and trace gas measurements were made in the southern part of the park from 6-11 August 2001. Table 4 summarizes the measurements made during this project that have been discussed previously by Shantz et al. (2004). Note all times 20 mentioned with respect to GE are in local time (i.e. Pacific Daylight Savings Time). The goal of the present study is to examine the cloud-forming properties from this forest aerosol that is also influenced by urban sources.  Fig. 8. The aerosol was dominated by OC through the sampling period, and increasingly towards the end of the period. The organic aerosol is believed to be a mix of primary and secondary organics from both anthropogenic and biogenic sources; sulphate was likely an ammonium salt (Shantz et al., 2004). Figure 8 also shows water-soluble organic carbon (WSOC) concentration. Because a range of solubilities 5 is tested, WSOC is simply provided as a reference. WSOC is 20-60% of the total OC, which is in the range of what was observed in other studies for polluted, rural and background sites (Carvalho et al., 2003;Decesari et al., 2000;Decesari et al., 2001;Saxena and Hildemann, 1996;Zappoli et al., 1999). The particulate matter volume for particles >0.1 µm determined from the size dis-10 tribution measurements is shown in Fig. 8. If we assume a mass density such as 1.0-1.5 g cm −3 , there is reasonable agreement between the particulate matter volume and the sum of the OC plus sulphate mass concentrations indicating that most of the OC and sulphate was in the fine particles. The sulphate concentrations at GE were similar or lower than measured over the Pacific Ocean.

15
The CCNc deltaV (proportional to CCN number concentration) at S eff =0.19% switches from being above the particle volume estimate for the first half of the study to being below the particle volume from 9 August onwards (Fig. 8). Although sulphate is low compared to the organic throughout the study, the sulphate to organic ratio is significantly higher during the period prior to 9 August (17%) than after 9 August (4%). 20 The change in the sulphate to organic ratio is a major reason for the change in the correspondence of the CCN with the particle volume, and it indicates a reduced effectiveness of the organic relative to the sulphate as CCN. To examine this further, two periods are selected representing before and after 9 August. These periods are highlighted in Fig. 8  Interactive Discussion to these observations using 3 modes as shown in Fig. 9b. There is one dominant number mode at about 0.034 µm diameter with a very high number concentration, a shoulder mode near 0.07 µm diameter, and a large particle mode at 0.36 µm.
For the purpose of modelling the CCN growth rates in this case, all of the inorganic ions (sulphate, ammonium, nitrate, sodium and chloride) are assumed to be 5 ammonium sulphate. All 3 modes are assumed to have the same composition as each other based on the filter measurements from 7 August: 76% organic, 4% BC (an insoluble organic with molecular weight=12.01 g mol −1 , index of refraction=(1.82, 0.74) and density=2 g cm −3 ) and 20% ammonium sulphate. Particles are assumed to be internally mixed based on hygroscopic growth measurements at this site (Aklilu and Mozurkewich, 2004;Shantz et al., 2004). Figure 10 shows the scattering cross-sectional area from the model output compared with the voltage from the CCNc observations. Absorption by the BC is included in the calculation, however its influence is negligible because it is present in such small quantities. As for the C-SOLAS 18 July case, the shaded areas encompass the growth 15 rates between relatively soluble and insoluble organic. The vertical lines on Fig. 9b show the minimum particle diameters activated for the organics with a solubility of 10 −6 g L −1 and at all 3 S eff 's, these fall within the shoulder mode 2. The 200 g L −1 organic activates at slightly smaller sizes (not shown), but still only activates minimum diameters within mode 2. Mode 1 does not activate regardless of the organic solubility. 20 However, the shoulder mode 2 contains a large number of particles (and mass) that may explain the larger growth rates here than during the C-SOLAS cases. The differences between the modelled soluble (upper curves of shaded sections) and insoluble (lower curves) organic cases are much larger than for the C-SOLAS cases, indicating the importance of organic solubility with increasing organic fraction in the 25 aerosol. For a highly soluble organic, the observations should match the upper curves. The insoluble organic simulation has only sulphate influencing the droplet growth and if the organic does not contribute at all, the observations will match the lower curves.
Since the observed growth rates are much higher than the modelled insoluble curves 8210 ACPD 8,2008 Droplet growth on organic aerosols from field measurements N. C. Shantz et al. Interactive Discussion and are also substantially higher than those observed for the C-SOLAS cases which had similar or higher sulphate concentrations, the organic aerosol makes a significant contribution to the water uptake. Simulated growth rates for an organic with an effective solubility of about 10 g L −1 best match the observed growth rates for 7 August. Note that if we change our initial 5 assumptions, this effective solubility may be altered, such as the sulphate and organic may not be distributed equally across all sizes or the organic may have affected the surface tension or mass accommodation coefficient in ways that are not properly referenced in the model.
In these simulations, with ammonium sulphate internally mixed with organics, am-10 monium sulphate dissolves initially and dominates the droplet growth until more water collects on the droplets allowing the slightly soluble organics to start to dissolve. At that point the organic may also contribute to the droplet growth. The WSOC/OC fraction is approximately 25% (Fig. 8). Simulations run with 20% ammonium sulphate, 20% organic with 200 g L −1 solubility and 60% insoluble organic 15 to match the WSOC measurements show reasonable agreement between these simulations and the observations. The use of a lower solubility (i.e. <200 g L −1 ) worsens the agreement. Figure 11a shows the number and volume distributions for 11 August. The modelled 20 size distribution, represented by 3 modes, is shown in Fig. 11b. Although the total number concentration is less than on 7 August, there are more particles in mode 2 on 11 August than on 7 August and the volume is substantially higher on 11 August. The increased mass is explained by the condensation of secondary organic aerosol as suggested by the gradual increase in particle mass from 9-11 August. Most of the ad-25 ditional material on 11 August is suspected to be due to the oxidation of monoterpenes as measured concentrations of a known product of terpene oxidation, cis-pinonic acid, also increased during this period (Cheng et al., 2004;Shantz et al., 2004 Interactive Discussion is assumed to have the same composition, an internal mixture of 85% organic, 2% BC and 13% ammonium sulphate (note that the sum of all inorganic ions is assumed to be ammonium sulphate). Figure 12 shows the scattering cross-sectional area from the simulations compared with the observations of CCNc droplet growth rates. The vertical lines in Fig. 11b show 5 the minimum activated diameters at the 3 S eff 's (none of the mode 1 particles activate, even when the organic is assumed to be soluble). The growth rates and the number of particles activated are much greater for 11 August than for any of the previously discussed cases.

Simulations of 11 August 2001, from the GE study
At S eff =0.50% the observed growth curves are close to the insoluble curve, suggest-10 ing a smaller contribution from the organics to the water uptake on 11 August than 7 August at this supersaturation. The observed growth rates are higher than the modelled insoluble cases at S eff =0.34%. The difference between the modelled results at these two S eff values is similar to those from 7 August. For S eff =0.5%, there may have been competition of water vapour between the particles in the CCNc at the highest super-15 saturation. Sensitivity tests suggest that an aerosol composed of soluble inorganics may deplete the supersaturation substantially but slightly soluble organics do not as they don't take up water to the same extent. If, however, the organics are soluble, competition for water vapour may drive the supersaturation to a lower value especially at S eff =0.5%, thus taking up less water, scattering less light and showing a lower voltage. 20 This may indicate a limitation to the growth rate measurement. Overall, the modelled growth rates best agree with the observations for organics with an effective solubility of 5 g L −1 , slightly less soluble than 7 August. Keep in mind this is an effective solubility and by changing the initial assumptions, we could find a different value for solubility.
Despite the fact that we found a lower overall effective solubility from the simulations 25 for 11 August compared to 7 August, the WSOC/OC fraction is higher on 11 August compared to 7 August (over 50% - Fig. 8 insoluble organic based on the WSOC measurements. Reasonable agreement with the observations is achieved for this situation. However, there are uncertainties in these measurements such as the long filter sampling time, variations in the chemical composition with respect to size and assumptions about the organic. These simulations are fairly sensitive to the amount of sulphate present and if we decrease the amount of 5 ammonium sulphate from 13% to 7%, the WSOC solubility would be higher. Overall, these results indicate that the organic is less effective as CCN compared to sulphate, but in these situations with relatively low sulphate, the organic particles do act as CCN. Even with lower sulphate masses on 11 August, there is an increase in the number of CCN as there is an increase in the particle number concentration, and similar to C-SOLAS 18 July, it is suspected that the influence of the organic on the sulphate contributes to the increased growth rates. This argument depends on whether the organic coats the sulphate or whether the sulphate is readily available to the water vapour as this model assumes. 5 Implementing a single parameter representing water uptake (κ) 15 Recently Petters and Kreidenweis (2007) discussed the hygroscopicity parameter κ that encompasses water uptake capabilities (both above and below supersaturation) and simplifies the physical and chemical properties needed for aerosols in cloud microphysical models. Because the majority of organic constituents are unknown and thus properties such as molecular weight and solubility are also unknown, a κ value is 20 useful in describing the water activity of aerosols from different sources. A version of the present CCNc model using κ was produced (hereafter "κ-CCN model") in order to compare estimated the total κ values (κ tot ) for the five case studies discussed above and to determine the organic κ values (κ org ) from the field measurements.
For the nearly pure sulphate case from C-SOLAS (27 July), the best agreement be-25 tween the CCNc model and the κ-CCN model is for κ tot =0.7 (see Table 5 for a summary of all κ values determined) and a slightly lower value of κ tot =0.65 for 16 July due to the  (2007) for H 2 SO 4 (κ=0.9), which may be due to the differences in the way the water activity is calculated in the CCNc model compared to Petters and Kreidenweis (2007). In Fig. 13a, the results for modelling C-SOLAS 18 July using κ are shown. This case, 5 with a marine aerosol mixed with organic, shows the best agreement with κ tot =0.35.
The κ tot can be constrained in this case to 0.2-0.6. Petters and Kreidenweis (2007) show that for aerosols that consist of a few chemical species, the individual chemical κ values can be weighted by the volume fraction to determine the κ tot (using κ tot = ε i κ i where ε i is the volume fraction and i represents 10 the individual species). Using the measured aerosol composition from the Q-AMS for 18 July (assuming sulphate has κ=0.7 based on 27 July results) and κ tot =0.35, then κ org is found to be zero. In other words, for 18 July over the North Pacific Ocean the organic does not contribute to the water uptake. GE forest 7 August show the best agreement between the observed and modelled 15 growth rates using κ tot =0.16 (Fig. 13b). Assuming the ammonium sulphate has a κ=0.61 (Petters and Kreidenweis, 2007), κ org is estimated as 0.05. For GE 11 August, the estimated κ tot is 0.08, which falls in the range of κ values provided by Petters and Kreidenweis (2007) for the oxidation products of monoterpenes. The 11 August κ org is estimated to be 0.02, lower than the κ org for 7 August. The measurements of water 20 uptake at relative humidities <100% that Aklilu and Mozurkewich (2004) made during the GE study gave values of κ tot (called "b" in their paper) in the range of 0.04 to 0.10, with most values between 0.05 and 0.07. These fall within the uncertainty range of the values determined here. The uncertainties in our measurements do not allow the absolute determination of κ, but based on the time series, it seems that κ is decreasing 25 over time during the GE study based on these 2 case studies (consistent with Aklilu and Mozurkewich, 2004). This could reflect a change in the properties of the organic aerosol over time, or it might be the result of a change in the nature of the mixing of the organic and inorganic components.

Summary and conclusions
A cloud condensation nucleus counter (CCNc) is used to investigate the role of organics as CCN by comparing the measured growth rates of cloud droplets to those simulated using a kinetic model (CCNc model). Five cases of ambient measurements of CCN, particle size distribution and particle chemistry are examined, three from the 5 North Pacific Ocean and two from a forest in British Columbia, Canada. One of the marine cases, for which the aerosol was dominated by sulphate (27 July 2002) is used as a reference for the CCNc model. This reference case is verified using a second marine case also dominated by sulphate (16 July 2002). In these cases with the fine particle mass ≥90% inorganic, the organic is of no consequence.

10
The organic fraction is higher in the third marine case (30% of the total mass -18 July 2002). Evidence suggests that this aerosol was a mix of somewhat aged ship emissions and marine aerosol. The comparison of the observations and simulations indicates that the sulphate in these aerosol particles was responsible for most of the water uptake. However, the organics influenced the CCN growth rates by enhancing the 15 number concentration of particles containing sulphate and redistributing the sulphate to larger and more CCN active sizes.
Two forest cases (from Golden Ears Provincial Park) had higher number concentrations, much higher organic mass fractions (80-90% of the total mass), and similar or smaller sulphate mass concentrations than the marine cases. The observed CCN 20 growth rates from these two cases are much higher than for the marine cases. This is due to an increase in the number concentrations and some influence of the organic material on the water uptake. The best fits of the simulations to the observations were obtained for organics with solubilities of 5-10 g L −1 . This result assumes an accommodation coefficient of unity and a specified surface tension. Differences in those two 25 quantities from the modelled values might also contribute to the water uptake.
Values of the parameter κ (Petters and Kreidenweis, 2007)  Interactive Discussion are 0.35-0.7 and the organic components are found to be negligible. For the forest aerosol, the total κ value is found to be 0.08-0.16 with an organic κ of 0.02-0.05. The observations suggest that the κ values decreased with time during the GE study as the aerosol accumulated a larger organic mass.
Acknowledgements. For the Golden Ears field study (Pacific 2001), the authors would like to ACPD 8,2008 Droplet growth on organic aerosols from field measurements N. C. Shantz et al. Cloud condensation nuclei activation of limited solubility organic aerosol, Atmos. Environ., 40, 605-617, 2006. Hegg, D. A., Gao, S., Hoppel, W., Frick, G., Caffrey, P., Leaitch, W. R., Shantz, N., Ambrusko, J., and Albrechcinski, T.: Laboratory studies of the efficiency of selected organic aerosols as CCN, Atmos. Res., 58, 155-166, 2001.  Heat Mass Transf., 44, 39-53, 2001. Marshall, J., Lohmann, U., Leaitch, W. R., Shantz, N., Phinney, L., Toom-Sauntry, D., and Sharma, S.: Optical properties of aerosol particles over the northeast Pacific, J. Appl. Meteorol., 44, 1206Meteorol., 44, -1220Meteorol., 44, , 2005. Medina, J., Nenes, A., Sotiropoulou, R. E. P., Cottrell, L. D., Ziemba, L. D., Beckman, P. J.,      The κ org are found by using the measured composition of the aerosol and κ tot and applying the mixing rule κ= ε i κ i . c The range of κ org is determined as those that fall within the error bars of S eff ±15%. The x-axis is the sample time in seconds as the particles activate and grow into cloud droplets. The observed growth rates at 3 supersaturations (shown as a voltage) from the CCNc are the thicker lines, corresponding to the y-axis on the right. The other curves with the many symbols are the modelled curves, on the left y-axis shown as a scattering cross section. These y-axes are scaled to match in this case. The error bars are simulations performed at the effective supersaturation ±15% to reflect the error in the calculated supersaturation.  (Fig. 3) is tested in this case to give confidence in these comparisons between the observations and simulations. See Fig. 3 8,2008 Droplet growth on organic aerosols from field measurements N. C. Shantz et al.  ACPD 8,2008 Droplet growth on organic aerosols from field measurements N. C. Shantz et al.