Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Measurements of black carbon properties during cloud, biomass burning, and free tropospheric conditions… Schroder, Jason C. 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2014_february_schroder_jason.pdf [ 2.88MB ]
Metadata
JSON: 24-1.0167086.json
JSON-LD: 24-1.0167086-ld.json
RDF/XML (Pretty): 24-1.0167086-rdf.xml
RDF/JSON: 24-1.0167086-rdf.json
Turtle: 24-1.0167086-turtle.txt
N-Triples: 24-1.0167086-rdf-ntriples.txt
Original Record: 24-1.0167086-source.json
Full Text
24-1.0167086-fulltext.txt
Citation
24-1.0167086.ris

Full Text

MEASUREMENTS OF BLACK CARBONPROPERTIES DURING CLOUD, BIOMASSBURNING, AND FREE TROPOSPHERICCONDITIONS AT A MARINE BOUNDARYLAYER SITE AND HIGH ELEVATIONMOUNTAIN SITEbyJason C. SchroderB. Sc., Northern Arizona University, 2001A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFDoctor of PhilosophyinTHE FACULTY OF GRADUATE AND POSTDOCTORALSTUDIES(Chemistry)The University of British Columbia(Vancouver)December 2014© Jason C. Schroder, 2014AbstractBlack carbon is a subset of the total atmospheric aerosol population that is formedin the incomplete combustion of fossil fuels, biofuels, and biomass. This researchfocused on the properties of black carbon particles measured in the boundary layerand free troposphere, as well as the activation of black carbon particles by clouddroplets. The primary motivation for this research is to increase our understandingof the properties of black carbon under these different atmospheric conditions.A single particle soot photometer was used to study properties of black car-bon particles incorporated into cloud droplets at two field locations: 1) a marineboundary layer site, and 2) a high elevation mountain site. At both sites, a sizedependence on the fraction of black carbon incorporated into cloud droplets wasobserved; and for small (<100 nm) diameters, black carbon was efficiently incor-porated into droplets. In addition, at the marine boundary layer site, thick coatingswere observed on the small diameter black carbon particles that were incorporatedinto the droplets, which was consistent with theory.The single particle soot photometer was also used at a high elevation mountainsite to investigate properties of black carbon from biomass burning and black car-iibon within the free troposphere. The average mass concentration of black carbonwas found to be significantly higher (≈ 9x) during periods of biomass burningthan within the free troposphere, yet had similar mass median diameters. Coatingthicknesses of black carbon containing particles during the two subsets of datawere also investigated. Average coating thicknesses for black carbon core diame-ters between 140 to 160 nm was 55 nm when sampling in the free troposphere, butapproximately 32 nm when sampling air masses influenced by biomass burning.The results presented in this dissertation increase our understanding of theproperties of black carbon particles and how they vary as a function of locationand type of air mass sampled. This information can be used to further constraincomputer models that are used to predict how black carbon can affect climate.iiiPrefaceChapter 4 is a co-authored journal article that has been accepted and publishedto an online discussion forum, and is awaiting peer review. Chapters 5 and 6are being prepared for submission to peer-reviewed journals as co-authored jour-nal articles. The details of my contributions to each research chapter, mentionedabove, are outlined below.Chapter 4 (first author of a published journal article) J. C. Schroder, S. J.Hanna, R. L. Modini, A. L. Corrigan, A. M. Macdonald, K. J. Noone, L. M.Russell, W. R. Leatich, and A. K. Bertram. Size-resolved observations of refrac-tory black carbon in cloud droplets at a marine boundary layer site, AtmosphericChemistry and Physics Discussion, 14: 11447-11491, 2014.• I formulated the research questions in collaboration with my supervisor.• I calibrated, optimized, operated, and maintained two SP2s throughout theduration of this field study.• I performed all of the data analysis in this publication with exception to theback trajectory model simulations, refractory black carbon coating thick-ivness, and aerosol mass fraction analysis.• I prepared all of the figures in this publication, with exception to Figure 4.3showing the back trajectory analysis.• Writing of the text for this publication was in collaboration with my super-visor.• Contributions from co-authors:– Dr. S. Hanna performed the HYSPLIT back trajectory model sim-ulations and provided the data used for the refractory black carboncoating thickness analysis.– Dr. R. Modini and A. Corrigan provided the HR-ToF-AMS data usedfor the aerosol mass fractions analysis as well as the SEMS bulk aerosolparticle count data.– Dr. K. Noone provided consultation on the theory and calculationsperformed on the counteflow virtual impactor.– Dr. R. Leaitch provided consultation on the cloud microphysics forthe clouds reported in this chapter.Chapter 5 (first author of a journal article in preparation for submission to apeer-reviewed journal) J. C. Schroder, S. J. Hanna, S. Sharma, S. Sjostedt, K.J. Noone, A. M. Macdonald, W. R. Leaitch, and A. K. Bertram. Size-resolvedobservations of refractory black carbon particles in cloud droplets at a Canadianhigh elevation site, 2014.v• I formulated the research questions in collaboration with my supervisor.• I calibrated, optimized, operated, and maintained one of the SP2s used dur-ing this field study.• I performed all of the data analysis in this publication with exception to theback trajectory model simulations.• I prepared all of the figures in this publication, with exception to Figure 5.2showing the back trajectories analysis.• Writing of the text for this publication was in collaboration with my super-visor.• Contributions from co-authors:– Dr. S. Sharma and Dr. R. Leaitch calibrated and operated the secondSP2 used during this field study.– Dr. S. Hanna performed the HYSPLIT back trajectory model simula-tions for this publication.– Dr. S. Sjostedt performed the data analysis for the acetonitrile mea-surements– Dr. K. Noone provided consultation on the theory and calculationsperformed on the counteflow virtual impactor.– Dr. R. Leaitch provided consultation on the cloud microphysics forthe clouds reported in this chapter.viChapter 6 (second author of a journal article in preparation for submissionto a peer-reviewed journal) S. J. Hanna, J. C. Schroder, A. M. Macdonald, W.R. Leaitch, and A. K. Bertram, Measurements of refractory black carbon at theWhistler high elevation site from 2009-2012, 2014• I calibrated, optimized, operated, and maintained the SP2s used in the 2010and 2012 datasets.• Dr. S. Hanna and I shared the data analysis for this manuscript. Specifically,I performed the initial data analysis used in the size distributions and massconcentration analysis reported in the manuscript.• I prepared all the figures and tables within this chapter.• Writing of the text for this chapter was in collaboration with Dr. S. Hannaand my supervisor.viiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xixList of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxiList of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxxvAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxxviii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 The Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Aerosols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Black Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4viii1.3.1 Black carbon emissions . . . . . . . . . . . . . . . . . . . 51.4 The Effects of Aerosols on Climate . . . . . . . . . . . . . . . . . 61.4.1 Radiative forcing . . . . . . . . . . . . . . . . . . . . . . 61.4.2 Direct effects on radiative forcing . . . . . . . . . . . . . 61.4.3 Indirect effects on radiative forcing . . . . . . . . . . . . . 71.5 Effects of Black Carbon on Climate . . . . . . . . . . . . . . . . 81.5.1 Direct effects of black carbon on climate . . . . . . . . . . 81.5.2 Indirect effects of black carbon on climate . . . . . . . . . 81.6 Dissertation Research Goals . . . . . . . . . . . . . . . . . . . . 91.7 Overview of Dissertation . . . . . . . . . . . . . . . . . . . . . . 101.8 Chapter 1 Figures and Tables . . . . . . . . . . . . . . . . . . . . 122 Formation of Cloud Droplets and kappa-Ko¨hler Theory . . . . . . . 152.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.2 Theory Describing Droplet Formation for Insoluble Particles Coatedwith Soluble Material . . . . . . . . . . . . . . . . . . . . . . . . 162.3 Chapter 2 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . 193 Single Particle Soot Photometer . . . . . . . . . . . . . . . . . . . . 203.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203.2 Theory of Operation . . . . . . . . . . . . . . . . . . . . . . . . . 203.3 Mass Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . 223.4 Volume Equivalent Diameter . . . . . . . . . . . . . . . . . . . . 233.5 Refractory Black Carbon Coating Analysis . . . . . . . . . . . . . 24ix3.5.1 Step 1: Determining the scattering amplitude of coatedrefractory black carbon particles . . . . . . . . . . . . . . 253.5.1.1 Leading edge only fitting procedure . . . . . . . 263.5.2 Step 2: Relating calculated Mie scattering amplitudes toan SP2 instrument response (i.e. calibration of the elasticscattering detectors) . . . . . . . . . . . . . . . . . . . . . 283.5.3 Step 3: Determining the coating thickness of a coated re-fractory black carbon particle using a core and shell Miescattering model . . . . . . . . . . . . . . . . . . . . . . . 283.5.4 SP2 optical detection limits and the implications for deter-mining average coating thickness . . . . . . . . . . . . . . 293.6 Chapter 3 Figures . . . . . . . . . . . . . . . . . . . . . . . . . . 314 Size-Resolved Observations of Refractory Black Carbon Particlesin Cloud Droplets at a Marine Boundary Layer Site . . . . . . . . . 354.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354.2 Sampling Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384.2.1 Site description . . . . . . . . . . . . . . . . . . . . . . . 384.2.2 Inlets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394.3 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404.3.1 Counterflow virtual impactor . . . . . . . . . . . . . . . . 404.3.1.1 Theory of operation . . . . . . . . . . . . . . . 404.3.1.2 Calculating a CVI cut-size . . . . . . . . . . . . 41x4.3.1.3 Calculating a CVI enhancement factor . . . . . . 434.3.1.4 CVI cut-size and enhancement factor measuredat La Jolla, CA . . . . . . . . . . . . . . . . . . 444.3.2 Refractory black carbon measurements . . . . . . . . . . . 444.3.2.1 Refractory black carbon mass measurements . . 444.3.2.2 Refractory black carbon coating thickness mea-surements . . . . . . . . . . . . . . . . . . . . . 454.3.3 Size distribution measurements of the bulk aerosol . . . . 464.3.4 Aerosol mass spectrometry . . . . . . . . . . . . . . . . . 464.3.5 Back trajectories . . . . . . . . . . . . . . . . . . . . . . 474.3.6 Cloud properties . . . . . . . . . . . . . . . . . . . . . . 474.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . 484.4.1 Back trajectories . . . . . . . . . . . . . . . . . . . . . . 484.4.2 Meteorological conditions and cloud properties . . . . . . 494.4.3 Size distributions . . . . . . . . . . . . . . . . . . . . . . 504.4.3.1 Size distributions measured from the total inlet(BulkAeroTot and rBCTot) . . . . . . . . . . . . 504.4.3.2 Size distributions measured from the residual in-let (BulkAeroRes and rBCRes) . . . . . . . . . . 524.4.4 Size-resolved activated fractions . . . . . . . . . . . . . . 534.4.5 Coating thickness of refractory black carbon residuals . . . 544.4.6 In-cloud aqueous phase chemistry . . . . . . . . . . . . . 554.4.7 Size distributions of rBCRes(Core+Coating) . . . . . . . . 56xi4.4.8 Comparison of rBCRes as a function of size with predic-tions based on kappa-Ko¨hler theory . . . . . . . . . . . . 574.4.8.1 Bulk aerosol composition . . . . . . . . . . . . 574.4.8.2 The critical diameter for activation of the bulkaerosol in the cloud droplets sampled . . . . . . 584.4.8.3 Critical supersaturation for the cloud dropletssampled . . . . . . . . . . . . . . . . . . . . . . 584.4.8.4 Predictions of the critical diameter for activationof refractory black carbon cores . . . . . . . . . 604.5 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . 624.6 Chapter 4 Figures and Tables . . . . . . . . . . . . . . . . . . . . 645 Size-Resolved Activation of Refractory Black Carbon in Cloud Dropletsat a Canadian High Elevation Site . . . . . . . . . . . . . . . . . . . 765.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 765.2 Sampling Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . 785.2.1 Site description . . . . . . . . . . . . . . . . . . . . . . . 785.2.2 Inlets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 785.3 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805.3.1 Refractory black carbon measurements . . . . . . . . . . . 805.3.2 Size distribution measurements of the bulk aerosol . . . . 805.3.3 Back trajectories . . . . . . . . . . . . . . . . . . . . . . 815.3.4 Cloud properties . . . . . . . . . . . . . . . . . . . . . . 81xii5.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . 825.4.1 Back trajectories . . . . . . . . . . . . . . . . . . . . . . 825.4.2 Meteorological conditions and cloud properties . . . . . . 835.4.3 Size distributions . . . . . . . . . . . . . . . . . . . . . . 845.4.3.1 Size distributions measured from the total inlet(BulkAeroTot and rBCTot) . . . . . . . . . . . . 845.4.3.2 Size distributions measured from the residual in-let (BulkAeroRes and rBCRes) . . . . . . . . . . 855.4.4 Size-resolved activated fractions . . . . . . . . . . . . . . 865.5 Comparison with Previous Measurements . . . . . . . . . . . . . 875.6 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . 895.7 Chapter 5 Figures and Tables . . . . . . . . . . . . . . . . . . . . 916 Measurements of Refractory Black Carbon at a High Elevation Moun-tain Site during 2009, 2010, and 2012 . . . . . . . . . . . . . . . . . . 1006.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1006.2 Site, Sampling and Analysis . . . . . . . . . . . . . . . . . . . . 1026.2.1 Site description . . . . . . . . . . . . . . . . . . . . . . . 1026.2.2 Refractory black carbon mass measurements . . . . . . . . 1026.2.3 Refractory black carbon coating thickness measurements . 1036.2.4 Total aerosol measurements . . . . . . . . . . . . . . . . . 1046.2.5 Back trajectories . . . . . . . . . . . . . . . . . . . . . . 1046.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . 105xiii6.3.1 Size distributions . . . . . . . . . . . . . . . . . . . . . . 1056.3.2 Refractory black carbon measurements at WHI . . . . . . 1066.3.3 Identifying periods of biomass burning sampling . . . . . 1076.3.3.1 Refractory black carbon mass concentrations dur-ing periods of biomass burning . . . . . . . . . 1076.3.3.2 Refractory black carbon coating thicknesses dur-ing periods of biomass burning . . . . . . . . . 1086.3.4 Identifying periods of free tropospheric sampling . . . . . 1096.3.4.1 Refractory black carbon mass concentrations inthe free troposphere . . . . . . . . . . . . . . . 1106.3.4.2 Refractory black carbon coating thicknesses inthe free troposphere . . . . . . . . . . . . . . . 1116.4 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . 1116.5 Chapter 6 Figures and Tables . . . . . . . . . . . . . . . . . . . . 1157 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1227.1 Activation of Refractory Black Carbon into Liquid Water CloudDroplets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1227.2 Refractory Black Carbon Properties During Biomass Burning andFree Troposphere Sampling . . . . . . . . . . . . . . . . . . . . . 1267.3 Considerations for Future Work . . . . . . . . . . . . . . . . . . . 128Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130xivAppendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148Appendix A HR-ToF-AMS Ion Pairing Scheme used in Chapter 4 . . . 148Appendix B Calculation of the Droplet Transmission Factor Throughthe CVI used in Chapter 4 . . . . . . . . . . . . . . . . . . . . . . . . 150Appendix C Calculation of the Droplet Transmission Factor Throughthe CVI used in Chapter 5 . . . . . . . . . . . . . . . . . . . . . . . . 153xvList of TablesTable 1.1 Best-estimates of BC emissions in Tg yr-1 with available esti-mate ranges in parentheses (adapted from Bond et al. (2013)Table 6). Energy related estimates are from energy-relatedcombustion sources while open burn refers to open vegetative,or biomass burning estimates. . . . . . . . . . . . . . . . . . . 12Table 4.1 Summary of cloud microphysical properties showing the aver-age CVI cut-size (CVI−D50) where the uncertainty stems fromthe calculated cut-size (see Section 4.3.1.2 for details); averageliquid water content (LWC) and one standard deviation; andthe cloud droplet number (CDNCTot) and volume (VolTot) con-centrations for droplets with diameters between 2-50 µm. Alsoshown is the number(CDNCSampCDNCTot)and volume(VolSampVolTot)frac-tions of droplets sampled, where CDNCSamp and VolSamp arethe number and volume concentrations, respectively, for thefraction of droplets sampled. . . . . . . . . . . . . . . . . . . . 74xviTable 4.2 Averaged number (N) and mass (M) concentrations, modal pa-rameters Dg and σg for aerosol and rBC particles during thetwo cloud events measured at Mt. Soledad. The subscriptsTot and Res represent measurements made from the total andresidual inlets respectively. . . . . . . . . . . . . . . . . . . . 75Table 5.1 Summary of cloud microphysical properties showing the datesand times sampled; the average CVI cut-size (CVI−D50), wherethe uncertainty comes from the calculated cut-size; averageliquid water content (LWC) and one standard deviation; totalcloud droplet number concentration (CDNCTot); and the num-ber fraction of droplets sampled (CDNCSamp/CDNCTot), whereCDNCSamp is the droplet number concentration greater than theCVI cut-size. . . . . . . . . . . . . . . . . . . . . . . . . . . . 98Table 5.2 Average number (N) and mass (M) concentrations, modal pa-rameters Dg and σg for bulk aerosol and rBC particles duringClouds 1 and 2. The subscripts Tot and Res represents mea-surements made from the total and residual inlets respectively.All values reported for residual particles have been correctedfor the CVI enhancement (see Section 4.3.1.3) and droplet losses(see Appendix C) . . . . . . . . . . . . . . . . . . . . . . . . . 98xviiTable 5.3 Summary of average black carbon mass activated fractions (AF)measured at other remote mountain locations (adapted fromCozic et al. (2007)), as well as from this study. . . . . . . . . . 99Table 6.1 Summary of basic statistics for rBC mass concentrations andmass distributions measured at WHI for the full record period,periods of biomass burning (BB), and periods of free tropo-sphere sampling (FT). . . . . . . . . . . . . . . . . . . . . . . 120Table 6.2 Comparison of rBC mass concentrations measured in the freetroposphere from this study and several other locations. . . . . 121xviiiList of FiguresFigure 1.1 Vertical structure of the atmosphere (top panel) for a typi-cal midlatitude temperature profile (adapted from Wallace andHobbs (2006)), and a schematic showing the components ofthe troposphere (bottom panel). . . . . . . . . . . . . . . . . . 13Figure 1.2 Schematic showing the different mechanisms by which aerosolscan affect climate (adapted from IPCC-AR4 (2007) Figures2.10 and 7.20). The thickness of the lines represent the mag-nitude of the radiation. . . . . . . . . . . . . . . . . . . . . . 14Figure 2.1 Panel A is an example Ko¨hler curve for a 50 nm dry (NH4)2SO4particle using a κ value of 0.61 (Petters and Kreidenweis,2007). Panel B is the critical supersaturation (SC) as a functionof ammonium sulfate dry diameter. . . . . . . . . . . . . . . . 19Figure 3.1 Schematic of the SP2 showing the three main components: 1)aerosol inlet; 2) intracavity laser; and 3) the two scattering andtwo incandescence detectors. . . . . . . . . . . . . . . . . . . 31xixFigure 3.2 Example mass calibration plot for the broadband incandes-cence channel of the SP2. This plot was created from pre(red circles) and post (blue squares) campaign Aquadag datacollected on one of the SP2s used in Chapter 4. The calibra-tion data was fit to a second order polynomial function (solidblack line) for the combined pre and post measurement datato create the equation used to determine the mass of an un-known rBC particle (shown in the box below the plot). Alsoshown are the 95% prediction bands (pink dashed lines) corre-sponding to the polynomial fit. The X axis is plotted in termsof Aquadag mass on the bottom axis and in terms of volumeequivalent diameter (VED) on the top axis. . . . . . . . . . . 32Figure 3.3 Raw signals produced by the SP2 from a theoretical scatter-ing only particle in panel A and a theoretical coated absorb-ing rBC particle in panel B. For both particle types, the sig-nals recorded from the broadband incandescence detector (redlines), scattering detector (blue lines), and split-detector (greenlines) are shown. The reconstructed leading edge only (LEO)scattering signal (blue dashed line) is also shown in panelB. The center of the laser beam is represented by the blackdashed line and the point at which the signal recorded fromthe split-detector crosses zero is indicated by the solid blackline. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33xxFigure 3.4 An example calibration plot for the elastic scattering detectorcreated from 200 and 300 nm PSL particles. Each data pointis the median of ≈ 25,000 particles and the error bars indicateone standard deviation. The black line is a linear fit with aforced intercept at zero.The units for both axes are arbitrary. . 34Figure 4.1 Schematic showing the configuration of the inlets and instru-mentation housed in the shipping container. . . . . . . . . . . 64Figure 4.2 Schematic of the CVI showing ambient air (green dashed lines)being drawn towards the CVI probe. The supply flow (redlines) are shown to permeate the porous frit (blue dotted re-gion) creating the counterflow (orange lines) and sample flow(purple dash-dot line). The theoretical stagnation plane (blackdashed line) and CVI geometry parameters needed to calcu-late the stopping distance of a droplet entering the CVI probe(see Section 4.3.1 for further details) are also shown and la-beled 1 to 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . 65xxiFigure 4.3 In-cloud HYSPLIT 96hr back trajectories ending at hourly in-tervals for Cloud 2 (12 June 21:00 to 13 June 12:00 PDT) inpanels A and C, and Cloud 3 (17 June 21:00 to 18 June 08:00PDT) in panels B and D. All back trajectories started at 10 ma.g.l.. Darker yellow regions on land in panels A and B indi-cate densely developed urban areas containing 50,000 or morepeople (United States Census Bureau). Panels C and D showthe vertical profiles over the same hourly intervals shown inpanels A and B. . . . . . . . . . . . . . . . . . . . . . . . . . 66Figure 4.4 Time series data for both Cloud 2 (left side) and Cloud 3 (rightside) showing; liquid water content (LWC, blue trace) and am-bient temperature (red trace) in panel A; wind speed and direc-tion in panel B; cloud droplet number size distributions withthe CVI−D50 (black trace) overlaid in panel C; the numbersize distribution for the total aerosol in panel D, and the resid-ual aerosol in panel E. All data shown are five-minute averagesand meet the criteria discussed in the text. . . . . . . . . . . . 67xxiiFigure 4.5 Average cloud droplet number size distributions for Cloud 2(panel A) and Cloud 3 (panel B) measured by the FM-100(black circles) and fit to a lognormal distribution function (blackdashed lines). The average cloud droplet volume distributions(blue squares) and lognormal fits (blue dashed lines) are alsoshown for each cloud event. The CVI−D50 is indicated oneach panel by a red dashed line. . . . . . . . . . . . . . . . . 68Figure 4.6 Summary of the averaged number size distributions for Cloud2 (panels A and C) and Cloud 3 (panels B and D) for the totalaerosol (red solid lines); residual aerosol (red dashed lines);total rBC as a function of core diameter (black solid lines);and residual rBC as a function of core diameter (black dashedlines). Both the aerosol and rBC for each cloud event areshown in two ways, a log scale (panels A and B) to high-light the relative differences between the aerosol and rBC aswell as normalized to the respective maximum value (panels Cand D) to highlight the shift in size distributions. All residualdistributions have been corrected for CVI enhancement (seeSection 4.3.1.3 and droplet losses (see Appendix B). . . . . . 69xxiiiFigure 4.7 Shown in panels A and B are the mean size dependent ac-tivated fraction (AF) for the aerosol (red circles), and rBC(black triangles) for Clouds 2 and 3, respectively, where theerror bars represent one standard deviation (1σstd) of the meanAF . The bottom axes represent particle diameter for the aerosoland core diameter for rBC. Since the fraction of the clouddroplets sampled by the CVI was less than 100%, the calcu-lated activated fractions should be considered as lower limitsto the total AF during the two cloud events. Shown in pan-els C and D are the averaged rBC coating thicknesses (bluecircle) in nm with 1σstd as the error bars. Since only 50%of the rBC containing particles detected with the SP2Res weresuccessfully fit with the LEO fitting procedure, the coatingthicknesses shown in panels C and D are only from a subset(50%) of the rBC particles measured with the SP2Res. . . . . . 70xxivFigure 4.8 Normalized number size distributions from Cloud 2 (panel A)and Cloud 3 (panel B) for residual rBC as a function of particlediameter [rBCRes(Core+Coating)]. The residual bulk aerosoldistribution (BulkAeroRes) and rBC distribution as a functionof core diameter [rBCRes(Core)] are also shown for compari-son. Since only 50% of the rBC containing particles detectedwith the SP2Res were successfully fit with the LEO fitting pro-cedure, the rBCRes(Core+Coating) shown in panels A and Bare only from a subset (50%) of the rBC particles measuredwith the SP2Res. . . . . . . . . . . . . . . . . . . . . . . . . . 71Figure 4.9 Sub-micrometer non-refractory average aerosol mass fractionsfor Clouds 2 and 3 based on an ion-pairing scheme (see Sec-tion 4.4.8.1 and Appendix A) and measured from a high reso-lution time-of-flight aerosol mass spectrometer. . . . . . . . . 72Figure 4.10 Panel A shows the critical supersaturation (SC, black lines)as a function of particle dry diameter based on measured HR-ToF-AMS bulk compositions and an ion-pairing scheme. PanelB shows SC as a function of rBC core diameters with coatingthicknesses ranging from 0-200 nm. In panel B, the coatingsare assumed to have the same composition as the bulk residualaerosol (Figure 4.9). The solid lines are for Cloud 2 and thedashed lines are for Cloud 3. . . . . . . . . . . . . . . . . . . 73xxvFigure 5.1 Schematic showing the configuration of the inlets and relevantinstrumentation used in this study. . . . . . . . . . . . . . . . 91Figure 5.2 HYSPLIT 24hr back trajectories. Panels A and C are trajec-tories ending at hourly intervals for Cloud 1 (2 July 10:00 to12:00 PST); panels B and D are trajectories ending at hourlyintervals for Cloud 2 (12 July 07:00 to 11:00 PST). All backtrajectories started at 10 m above ground level. Red verticaltriangles are fires reported by the Canadian Wildland Fire In-formation System (CWFIS) and red horizontal triangles arefires detected from MODIS for all fires within 24 hours priorto the start of the cloud. Panels C and D show the verticalprofiles over the same hourly intervals shown in panels A and B. 92Figure 5.3 Time series data for Cloud 1 (left side) and Cloud 2 (rightside) showing; liquid water content (LWC) in panel A; clouddroplet number size distributions with the CVI−D50 (blacktrace) overlaid in panel B; the number size distributions for thetotal bulk aerosol in panel C, the residual bulk aerosol in panelD, and acetonitrile mixing ratio in panel E. All data shown inpanels A to D are one-minute averages and meet the criteriadiscussed in Section 5.4.2. Data in panel E are 15-minute av-erages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93xxviFigure 5.4 Average cloud droplet number size distributions for Cloud 1in panel A and Cloud 2 in panel B (black circles) and fit to alognormal distribution function (black lines). The CVI−D50is indicated in each panel by a red line. . . . . . . . . . . . . . 94Figure 5.5 Summary of the averaged number size distributions for Cloud1 (panels A and B), and Cloud 2 (panels C and D) for thetotal bulk aerosol (red solid lines); residual bulk aerosol (reddashed lines); total rBC as a function of core diameter (blacksolid line); and residual rBC as a function of core diameter(black dashed lines). Both the aerosol and rBC for each cloudare shown in two ways, a log scale (panels A and C) as wellas normalized to the respective maximum values (panels Band D). All residual distributions have been corrected for theCVI enhancement (see Section 4.3.1.3) and droplet losses (seeAppendix C) . . . . . . . . . . . . . . . . . . . . . . . . . . 95xxviiFigure 5.6 Mean size dependent activated fractions (AF) for the bulkaerosol (red circles) and rBC (black triangles) for Clouds 1and 2 in panels A and B, respectively. The error bars representone standard deviation of the mean activated fraction withineach 10 nm bin. The bottom axis represents particle diame-ter for the bulk aerosol and core diameter for rBC. Since thefraction of cloud droplets sampled by the CVI was less than100%, the calculated activated fractions should be consideredas lower limits to the total activated fraction. . . . . . . . . . . 96Figure 5.7 Averaged rBC mass distributions as a function of rBC corediameters during Cloud 1 (panel A) and Cloud 2 (panel B)for the total rBC (rBCTot, red circles) and the residual rBC(rBCRes, black triangles). The solid lines are fits using a log-normal function. . . . . . . . . . . . . . . . . . . . . . . . . 97Figure 6.1 Normalized mass distributions for periods of free tropospheresampling and biomass burning. Lines are single lognormal fits. 115Figure 6.2 Measured rBC mass concentrations for 2009, 2010, and 2012with periods of biomass burning marked by red shaded areas. . 116Figure 6.3 Histograms of measured rBC mass concentration at WHI forthe full record (panel A), periods of biomass burning (panelB), and periods of free troposphere sampling (panel C). Allconcentrations are 10 minute averages. . . . . . . . . . . . . . 117xxviiiFigure 6.4 Identification of biomass burning periods in 2009 (left side)and 2010 (right side). Panel A Shows rBC mass concentra-tion as measured by the SP2 (binned into 10 minute intervals);panel B shows the ratio of organic material (OM) to OM plussulfate (SO4), where a ratio of > 0.7 is considered to be fromnon-anthropogenic sources; panel C shows the raw ion sig-nals of the levoglucosan fragments at m/z 57, 60, and 73 asmeasured by the ACSM in 2009 and the C-ToF-AMS in 2010.Ion signals that were greater than 5σstd were considered to beinfluenced by biomass burning. Panel D shows the numberof fires reported by day within 200 km of Whistler Mountain(CWFIS). The red boxes indicate the portion of the data in-cluded from biomass burning. . . . . . . . . . . . . . . . . . 118Figure 6.5 2D histograms of coating thicknesses and core diameters forall rBC containing particles measured during the biomass burn-ing period from July 26-28, 2010 (panel A), and the period offree troposphere sampling in April-May 2012 (panel C). Onlyparticles with detectable scattering signals are shown in pan-els A and C. Coating thickness frequency distributions for rBCcores with VED from 140-160 nm (represented on panels Aand C as black dashed lines) are shown in panel B for biomassburning periods, and panel D for free troposphere periods. . . 119xxixFigure B.1 Correlation plots between the cloud droplet number concen-tration (CDNC) greater than the CVI−D50 (CDNC > D50)and the enhancement factor (EF) corrected residual numberconcentration greater than 100 nm (NRes>100nm) in cm-3. . . . 152Figure C.1 Correlation plots for Clouds 1 and 2 in panels A and B, respec-tively, of the cloud droplet number concentration (CDNC) fordroplets that are greater than the CVI cut-size as a function ofthe enhancement factor corrected residual number concentra-tion. The slope of the linear fits to the correlation data rep-resent the scaling factor needed to account for droplet lossesand 1/slope represents the droplet transmission. . . . . . . . . 155xxxList of SymbolsA amplitude of SP2 scattering signalAF activated fractionaw water activityB baseline offset for Gaussian distributionC fraction of CVI orifice outer radius used in Equation 4.4CF size-resolved instrument sensitivity correction factorCVI−D50 CVI cut-sizeD cloud droplet diameterDg lognormal distribution mean parameterDp dry particle diameterDpc critical activation diameterDstop droplet stopping distance within CVIxxxiDT droplet transmission factor through CVIEF CVI enhancement factorε constant used in Equation 4.5 (0.158)ηair viscosity of airF1 CVI counterflowF2 CVI sample flowi ith componentκ hygroscopicity parameterκBulk hygroscopicity parameter for internally mixed particleλ wavelengthLcur length of gas stream lines curving around CVI probe tipLmin length from the CVI probe tip to the top of the porous fritLpor length from the top of the CVI porous frit to the end of the stag-nation planeLstag overall length to the end of the stagnation planeMRes residual particle mass concentrationMTot total (residual + interstitial) particle mass concentrationxxxiiMw molecular weight of waterµ Gaussian distribution mean parameterµsp2 center of the SP2 laser beamm/z mass to charge ration number of molesNRes residual particle number concentrationNTot total (residual + interstitial) particle number concentrationR universal gas constantrd cloud droplet radiusri inner radius of CVI orificero outer radius of CVI orificeRed droplet Reynolds numberρair density of airρd droplet densityρw density of waterρbc density of black carbonxxxiiiσ Gaussian distribution width parameterσsp2 width of SP2 laser beamσg lognormal distribution width parameterσs/a surface tension at the air/surface interfaceσstd standard deviationσw surface tension of waterS supersaturationSC critical supersaturationT temperaturet timeυ volume fractionV∞ air intake velocity of CVIVED volume equivalent diameterX length of CVI porous fritx SP2 peak heightxxxivList of AbbreviationsABL atmospheric boundary layerACSM aerosol chemical speciation monitorAPD avalanche photodiodea.g.l. above ground levela.m.s.l. above mean sea levelBC black carbonBulkAeroRes residual bulk aerosol (entire aerosol population, including BC)BulkAeroTot total (residual + interstitial) bulk aerosol (entire aerosol popula-tion, including BC)CCN cloud condensation nucleiCDNC cloud droplet number concentrationCPC condensation particle counterxxxvC-ToF-AMS C-mode time-of-flight aerosol mass spectrometerCVI counterflow virtual impactorCWFIS Canadian wildland fire information systemDMA differential mobility analyzerFT free troposphereFM fog monitorHIPPO HIAPER pole-to-pole observationsHR-ToF-AMS high-resolution time-of-flight aerosol mass spectrometerHYSPLIT hybrid single particle Lagrangian integrated trajectory modelIPCC Intergovernmental Panel on Climate ChangeLEO leading edge onlyLWC liquid water contentMODIS moderate resolution imaging sprectroradiometerNOAA National Oceanic and Atmospheric AdministrationOM organic aerosol, or matterOPC optical particle counterxxxviPDT pacific daylight savings timePMT photomultiplier tubePSL polystyrene latex beadPST pacific standard timerBC refractory black carbonrBCRes residual rBC particlesrBCTot total (residual + interstitial) rBC particlesSEMS scanning electrical mobility spectrometerSMPS scanning mobility particle sizerSP2 single particle soot photometerSP2Res SP2 measuring residual particlesSP2Tot SP2 measuring total (residual + interstitial) particlesUHSAS ultra-high sensitivity aerosol spectrometerWACS Whistler aerosol and cloud studyWHI Whistler high elevation siteWPS wide range particle spectrometerxxxviiAcknowledgmentsA dissertation that focuses on field measurements of atmospheric aerosols takesan immense amount of effort from a great number of individuals. None more sothan my advisor, Dr. Allan Bertram, who provided an open atmosphere for me tobegin learning the complex field of atmospheric chemistry and grow as a scientistin general. Allan, I appreciate the direction you provided, your patience in myignorance, and the knowledge you imparted as I journeyed through the muddledwaters of black carbon cloud interactions.To all the past and present Bertram group members: Emily, Pedro, Donna,Michael, Sarah, Yuan, Meng, MiJung, Stephan, Song, Aidan, Rich, Lyndsay,Vickie, Yuri, Ce´dric, Ryan, and James, thank you for helping to create a workenvironment that was fun, stimulating, and most of all memorable. I would liketo especially thank Emily, Sarah, Pedro, and Michael who received the brunt ofmy incessant questioning. I am forever indebted to all of you for the kindness,patience, and knowledge each of you shared with me.All of the work presented in this dissertation would not have been possiblewithout the help and expertise of all the field campaign contributors, of whichxxxviiithere are too many to list, but I thank each and everyone of you for your contribu-tion and for helping me accomplish this immense task. I would like to particularlythank Richard for everything you taught me in the field and your invaluable men-torship along the way, and Anne Marie, for your insight into CVIs and for doinganything and everything in your power to collect the best dataset possible. Forspending hours with me on the phone diagnosing SP2 detector issues, and in gen-eral, helping me understand the complex world of SP2 measurements, I wouldlike to thank Subu. Finally, to Rob and Ran who spent hours with me holed up ina shipping container at ungodly hours of the night, thanks for the entertainmentand friendships that have ensued.A most special appreciation goes to my mother, Patricia Schroder, who in-stilled in me, at a very early age, that I can accomplish anything I set my mind to.Mom, assuming all goes well here, this is just another testament to the constantsupport and unconditional love you’ve shown me over the years, regardless of myendeavor. Without this I would not be the person I am today and only exemplifieswhat a wonderful person and amazing mother you truly are. Also, to my late un-cle, Mr. Ron Gill, without whom none of this would have come to fruition. Ron,you plucked me from a path that was surely headed for despair and showed methat the world is a beautiful place. I wholeheartedly believe that it was throughthis that I have ended up where I am today, a doctoral candidate in atmosphericchemistry.To my children, Dillon and Zoe¨, you have brought so much joy into my lifethat words could never justify. Dillon, throughout this journey you have taughtxxxixme, and continue to teach me, the virtue of patience. Your humor often remindsme that life doesn’t always have to be serious and sometimes playing catch withyou is way more important than analyzing data. To my precious little Zoe¨ girl,your constant source of undying affection has picked me up more times than Ican count, and your thoughtful nature never ceases to amaze me from such ayoung girl. To both of you, I am eternally grateful, and if there is one thing Ihope you’ve learned by watching me go through this, is that you too can alsoaccomplish anything if you put your mind to it.Finally, to the love of my life, Mrs. Lisa Schroder. What you have sacrificedfor me to accomplish this task will take me multiple lifetimes to repay. Your never-ending support, pursuit for harmony, calm nature, warm hugs, and acceptance arejust a few of your traits that have been my rock. It is because of you that all ofthis was possible, and as a result, I hereby grant you the most distinguished award(created by your late father) of PhT (Put him Through). Luv, I am so thankful foreverything you’ve given to me. There is no one else on the planet I would haverather endured this with than you!xlTo my wife, Lisa, son, Dillon, and daughter, Zoe¨ who are constant remindersthat in the pursuit of becoming a better person there is no time like right now.xliChapter 1Introduction1.1 The AtmosphereThe Earth‘s atmosphere is divided into several layers as seen in Figure 1.1 toppanel, which shows a temperature profile as a function of altitude that typicallyoccurs near midlatitude regions (between approximately 23° and 66° for bothhemispheres) (Wallace and Hobbs, 2006). The troposphere is the bottom layerand is characterized by a decrease in temperature with an increase in altitude. Thislayer contains approximately 85% of the total mass within the atmosphere (Jacob,1999), and extends from the Earth’s surface to approximately 11 km (dependingon latitude). The troposphere can be further classified into two regions known asthe atmospheric boundary layer (ABL) and the free troposphere (FT) (Figure 1.1bottom panel). The ABL is the lowest portion of the troposphere, typically from0-1 km, where the frictional forces of the Earth’s surface acting on wind direction1and velocity are greatest, causing turbulent vertical mixing of atmospheric com-ponents (Stull, 2003). The altitude at which the Earth’s frictional forces on winddirection and velocity diminish is called the FT. Within the FT vertical mixing isslow (Stull, 2003).The atmosphere is made up of gaseous, liquid, and solid materials that areemitted from both natural and anthropogenic sources (e.g. volcanic eruptions,dust storms, forest fires, and vehicular exhaust). It is within the troposphere thatthese components can most greatly influence the quality of life for humans byaltering such parameters as visibility, or air quality, which in turn can impacthuman health. All of the work performed in this document takes place in thetroposphere.1.2 AerosolsThe liquid and solid components, mentioned above, are referred to as particulatematter, or aerosol when that particulate matter is suspended in a gaseous medium(Finlayson-Pitts and Pitts, 2000), such as air. Although the rigorous definition ofaerosol accounts for both the particles and the gas it is suspended in, it is com-mon in atmospheric literature to use the word aerosol and particle interchangeably.Therefore, for the remainder of this dissertation the word aerosol will refer to theparticles alone, without reference to the gas. Aerosols are ubiquitous in the atmo-sphere and exist in a wide range of sizes, number concentrations, and chemicalcompositions. The exact chemical and physical properties of aerosol particles aregoverned by the source of emission and any chemical, or physical changes that2occur while the particles are suspended in the atmosphere.Aerosols within an air mass can exist in different mixing states. The mix-ing state describes the extent to which multiple different aerosols from differentsources have been atmospherically processed to form internal mixtures of compo-sition within a single aerosol. The two extreme cases of mixing state are classifiedas externally mixed and internally mixed. Externally mixed aerosols exist whenaerosols from multiple sources are present in the same air mass as independentparticles with chemical compositions that are characteristic of the initial source.Alternatively, several aerosols from different sources can be combined into a sin-gle aerosol particle through processes such as coagulation or condensation. Theseare considered to be internally mixed as the chemical composition of a single par-ticle is now a mixture of species from each of the different sources. For example,a black carbon particle that has had ammonium sulfate condensed on its surfaceis considered to be an internally mixed particle.The presence of complex mixtures of aerosols that have different sizes, chem-ical compositions, and mixing states within a given air mass creates a dynamicsystem in which atmospheric gases and aerosols interact under different meteo-rological conditions (e.g. temperature, pressure, relative humidity). Variationsin the meteorological conditions as well as the size and chemical composition ofthe aerosols are important parameters as they greatly affect the amount of timean aerosol spends suspended within the atmosphere (Seinfeld and Pandis, 2006),referred to as the lifetime of an aerosol. The greater the aerosol‘s lifetime, thegreater the impact it can have on climate (see Section 1.4 for further discussion).31.3 Black CarbonBlack Carbon (BC) is a subset of the aerosol population that is emitted as a resultof incomplete combustion. Historically, the term BC has not had a standardizeddefinition and has been used, in the literature, to represent various componentsof the total carbonaceous material present in the atmosphere (Bond et al., 2013;Petzold et al., 2013). Here, I adopt the recommendations laid out by Bond et al.(2013), namely that BC is only produced in flames and is distinguished from otherforms of atmospheric carbonaceous material by the following four physical prop-erties:1. BC strongly absorbs visible light.2. BC retains its form at high temperatures (referred to as refractory).3. BC is insoluble in water, organic solvents, and other components of atmo-spheric aerosols.4. BC is an agglomerate of small carbon spherules.Because the measurements of BC contained in this document were carried outwith a single particle soot photometer (SP2, see Chapter Chapter 3) I further adoptthe convention suggested by Petzold et al. (2013) and refer to the BC measuredby the SP2 as refractory black carbon, or rBC.41.3.1 Black carbon emissionsBC emissions come from both naturally occurring and anthropogenic sources.The only natural source of BC in the atmosphere is biomass burning initiated bynatural causes, such as lightning. However, biomass burning BC is also emittedfrom anthropogenic sources such as wood based cook stoves, or stoves used forheating, as well as the burning of agricultural waste, or prescribe forest fires usedin forest management.Table 1.1 was adapted from Bond et al. (2013), and lists the best-estimates ofBC emissions from the year 2000 and the year 1750, which is often used as prein-dustrial conditions. BC emissions are segregated into those emissions that areassociated with anthropogenic energy related BC and open vegetative, or biomassburning emissions (Bond et al., 2013). Approximately 4.8 Tg of BC were emittedin the year 2000 as a result of energy related anthropogenic combustion sources,an 1100% increase from preindustrial conditions. A much smaller yet still sig-nificant increase of 180% was observed for open burning BC emission sources.The estimates for open burning listed here however, do not necessarily accountfor agricultural waste burning due to a lack of direct measurements of such pro-cesses (Bond et al., 2013). Combining BC emissions from both sources resultsin a total increase of BC emissions from preindustrial levels of 436%. However,based on the ranges of total BC emissions estimated from all sources in the year2000 (shown in parenthesis in Table 1.1) a considerable uncertainty is associatedwith these estimates. These large uncertainties have also been reported in otherstudies (e.g. Bond et al., 2004; Koch et al., 2009).51.4 The Effects of Aerosols on ClimateAs mentioned in Section 1.2, aerosols can affect climate, and they can do so di-rectly and indirectly. Figure 1.2 pictorially describes the different affects thataerosols can have on climate.1.4.1 Radiative forcingThe metric used by atmospheric scientists to quantify a particular aerosol‘s effecton climate is known as radiative forcing. The Intergovernmental Panel on ClimateChange (IPCC) defines, in general terms, radiative forcing as “the net change inthe energy balance of the Earth’s atmospheric system as a result of some imposedperturbation, such as an aerosol. It is usually expressed in watts per square meter,averaged over a particular period of time, and quantifies the energy imbalance thatoccurs when the imposed change takes place” (IPCC-AR5, 2013). It is throughinteractions with solar radiation and clouds that aerosols can impose either a cool-ing or warming effect on the Earth’s atmospheric system. An aerosol that has apositive radiative forcing is considered to be a warming forcer, while an aerosolthat has a negative radiative forcing is considered to be a cooling forcer.1.4.2 Direct effects on radiative forcingAerosols can directly affect climate by absorbing, or scattering incoming solarradiation (Figure 1.2). When incoming solar radiation is scattered by aerosolsthe net amount of radiation reaching the Earth’s surface is reduced, leading to anegative radiative forcing and cooling of the Earth’s atmospheric system (IPCC-6AR5, 2013). Opposite to scattering, when incoming solar radiation is absorbed byaerosols the energy is emitted in the form of heat, leading to a positive radiativeforcing and a warming of the Earth’s atmospheric system (IPCC-AR5, 2013).1.4.3 Indirect effects on radiative forcingAerosols can also affect climate indirectly through a variety of interactions withclouds. The indirect effects on climate are classified into two categories; the 1stindirect effect, or cloud-albedo effect, and the 2nd indirect effect, or cloud-lifetimeeffect. Both of these effects are represented pictorially in Figure 1.2. In order foran aerosol to impact climate indirectly it must have the ability to act as a cloudcondensation nuclei (CCN), which is a requirement to form cloud droplets (seeChapter 2 for further details on cloud droplet formation).The cloud-albedo effect is characterized by an increase in the cloud dropletnumber concentration (CDNC) as a result of an increase in the number of CCNactive aerosols. The cloud-albedo effect assumes the amount of liquid water avail-able remains the same as the clean cloud conditions (see Figure 1.2). The increasein number of CDNC increases the reflectivity, or albedo of the cloud. An increasein the cloud-albedo reduces the net radiation reaching the Earth’s surface produc-ing a negative radiative forcing and hence a net cooling effect on climate.The cloud-lifetime effect, on the other hand, is the result of the decrease indroplet size, as a result of an increase in the number of CCN active aerosols. Adecrease in droplet size results in a decrease in the amount of precipitation andtherefore, an increase in the lifetime of the cloud. An increased lifetime, in turn,7reduces the amount of solar radiation reaching the Earth’s surface also producinga negative radiative forcing.1.5 Effects of Black Carbon on Climate1.5.1 Direct effects of black carbon on climateBC is a strong absorber of all wavelengths of visible light (Bond et al., 2013). Infact, BC is the atmospheric species with the strongest known absorption of solarradiation at these wavelengths (Bond et al., 2013). As a result of such efficientabsorption, BC has the potential to significantly influence climate directly. Basedon the current best-estimates the BC direct radiative forcing, from all present daysources, is +0.88 W m-2, however this number has a 90% uncertainty. Accountingfor this uncertainty places BC direct radiative forcing estimates between +0.17to +1.48 W m-2 (Bond et al., 2013). The uncertainty in the BC direct radiativeforcing estimates stems from uncertainties in the emission rates of BC, spatial andtemporal distributions of BC, as well as removal rates of BC from the atmospherethrough such processes as wet or dry deposition (Bond et al., 2013).1.5.2 Indirect effects of black carbon on climateBC particles have been shown to acquire hydrophyilic coatings (Bond et al., 2013).Once coated with a hydrophyilic substance they can become CCN active, andtherefore contribute to the indirect effects on climate (Bond et al., 2013). The best-estimates of the indirect radiative forcing of BC, based on current knowledge, is8+0.23 W m-2 (Bond et al., 2013). However, similar to the direct radiative forcingestimate, this indirect radiative forcing estimate has a 90% uncertainty associatedwith it, thus giving a range of -0.47 to +1.0 W m-2 (Bond et al., 2013). Bond et al.(2013) remark that it is a lack of scientific understanding in the BC-cloud effectsthat is the largest source of uncertainty in calculating the contribution of BC toclimate change.1.6 Dissertation Research GoalsTo better understand the role of BC in climate, and to better quantify the directand indirect effects of BC on climate, the incorporation of rBC particles into clouddroplets as a function of rBC particle size was investigated at two locations: 1)La Jolla, CA, and 2) Whistler, BC. Measurements of rBC in cloud droplets as afunction of size is important to test our current understanding of the incorporationof rBC particles into cloud droplets and to test our understanding of the CCNproperties of rBC. A better understanding of the CCN properties of BC shouldeventually translate to better estimates of the direct and indirect effects of BC onclimate.In addition to investigating the incorporation of rBC in cloud droplets, thisthesis also describes measurements of rBC at the Whistler High Elevation Site(WHI) conducted during 2009, 2010, and 2012. This included measurements ofrBC from biomass burning and rBC in the free troposphere. From these studies,properties of rBC, such as coating thickness and size distributions, that are neededwhen calculating the direct effect of rBC on climate were determined. The three9year record also serves as a bench-mark that can be used to test current atmo-spheric models that predict concentrations of BC in the atmosphere.1.7 Overview of DissertationChapter 1 (this chapter) gives an introduction to atmospheric aerosols, includingBC, and their potential effect on climate. It also outlines the research goals of thisdissertation. Chapter 2 discusses the theory used to determine if an aerosol willactivate to form a cloud droplet, and Chapter 3 describes the operational theoryand data analysis procedures for the SP2, the primary instrument used in thisresearch. Chapters 4-6 are the research chapters. For example, Wang et al. (2014)compared modelled BC concentrations with ambient measurements of BC andconcluded that most models overestimate atmospheric BC concentrations, likelyindicating that activation of BC into cloud droplets is not simulated correctly,and therefore contribute to the uncertainties of the direct and indirect estimates.By directly measuring CCN properties of BC in cloud droplets at two distinctlocations, models can be further constrained to more accurately determine theindirect effects of BC on climate.Chapter 4 investigates the size-resolved activation of rBC measured at a ma-rine boundary layer site during two liquid water cloud events. The size-resolvedactivated fractions of rBC are compared to the activated fractions of the bulkaerosol, and the average coating thickness of rBC containing particles are alsoreported and discussed. In addition, the observed activation of rBC at this site iscompared with kappa-Ko¨hler theory in order to validate the theory with observa-10tions from a natural environment. Chapter 5 also investigates the size-resolvedactivation of rBC in liquid water clouds, but at a high elevation mountain site.Comparisons between the size-resolved activated fractions of rBC and the size-resolved activated fractions of the bulk aerosol are also discussed in Chapter 5.Chapter 6 presents results from a three year period of measurements taken atthe high elevation mountain site, and discusses properties (size distributions, massconcentrations and distributions, as well as coating thicknesses) of rBC measuredduring two types of sampling conditions: 1) during biomass burning episodes and2) during free tropospheric sampling conditions. Finally, Chapter 7 summarizesthe research with some general conclusions and considerations for future work.111.8 Chapter 1 Figures and TablesTable 1.1: Best-estimates of BC emissions in Tg yr-1 with available esti-mate ranges in parentheses (adapted from Bond et al. (2013) Table 6).Energy related estimates are from energy-related combustion sourceswhile open burn refers to open vegetative, or biomass burning estimates.Time frame BC EmissionsYear 2000Energy related 4.8 (1.2-15)Open burning 2.8 (0.8-14)Total 7.5 (2-29)Year 1750Energy related 0.4 (-)Open burning 1.0 (-)Total 1.4 (-)12Figure 1.1: Vertical structure of the atmosphere (top panel) for a typical mid-latitude temperature profile (adapted from Wallace and Hobbs (2006)),and a schematic showing the components of the troposphere (bottompanel).13Figure 1.2: Schematic showing the different mechanisms by which aerosolscan affect climate (adapted from IPCC-AR4 (2007) Figures 2.10 and7.20). The thickness of the lines represent the magnitude of the radia-tion.14Chapter 2Formation of Cloud Droplets andkappa-Ko¨hler Theory2.1 IntroductionGiven the numerous pathways by which aerosols can interact with clouds to affectclimate, it is important to determine a given aerosol’s ability to act as a CCN(see Section 1.4.3). The following section describes the details of how Ko¨hlertheory can be used to determine if an aerosol will act as a CCN, and provides thenecessary background theory for the rBC cloud studies described in Chapters 4and 5.152.2 Theory Describing Droplet Formation forInsoluble Particles Coated with SolubleMaterialThe supersaturation of water vapor (S) over a droplet containing a dissolved solutecan be described by the following equation (Seinfeld and Pandis, 2006):S =[awexp(4σs/aMwRTρwD)−1]x100, (2.1)where aw is the activity of water in a solution, σs/a is the surface tension of theair/surface interface, Mw is the molecular weight of water, R is the universal gasconstant, T is temperature, ρw is the density of water, and D is the droplet diameter.S is the supersaturation and is defined as the saturation minus one and is usuallyexpressed as a percent. For example, a supersaturation, S, of 1% correspondsto a relative humidity of 101%. A more convenient form of Equation 2.1 thatincorporates a single hygroscopicity parameter kappa (κ) for the aerosol (Pettersand Kreidenweis, 2007), has been widely adopted (e.g., Chang et al., 2010; Duseket al., 2011; Moore et al., 2012; Petters et al., 2009; Prenni et al., 2007; Rose et al.,2010), and can be expressed as:S =[D3−D3pD3−D3p(1−κ)exp(4σMwρwRTD)−1]x100, (2.2)where D is still the droplet diameter; Dp is the dry particle diameter; σ is thedroplet surface tension, and is assumed to be that of water (0.072 J m-2); and κis a compositionally specific parameter that describes an aerosol’s hygroscopic-16ity. For a complex internally mixed aerosol particle, such as an insoluble particlecoated with soluble material, the hygroscopicity of the entire particle (κBulk) canbe calculated by following a simple mixing rule as follows:κBulk =∑iυiκi, (2.3)where υi is the volume fraction and κi is the hygroscopicity parameter of theith component present in an internally mixed aerosol (Petters and Kreidenweis,2007). A large value for κBulk would indicate a strongly hygroscopic aerosol thatwould therefore be an efficient CCN.At a constant Dp, Figure 2.1 can be used to evaluate the relationship betweenS and droplet growth. Figure 2.1 panel A shows an example of this relationshipfor ammonium sulfate, a ubiquitous atmospheric aerosol, where supersaturation isplotted as a function of droplet diameter. Equation 2.2 panel A is often referred toas a Ko¨hler curve. The maximum of the Ko¨hler curve for a given dry diameter ofan aerosol with hygroscopicity κ represents the minimum, or critical supersatura-tion (SC) required for that specific aerosol to be activated into a cloud droplet. Atany supersaturation ≥ SC the droplet will continue to grow indefinitely, or at leastuntil it grows large enough to rain out. SC can be determined from Equation 2.2as a function of aerosol dry diameter, to produce a plot similar to that shown inFigure 2.1 panel B. In this figure, SC is plotted as a function of aerosol dry di-ameter for ammonium sulfate using a κ value of 0.61 (Petters and Kreidenweis,2007). Figure 2.1 panel B can then be used to predict the minimum, or critical dry17diameter (Dpc) of an aerosol with hygroscopicity κ that will act as a CCN, if S ofthe cloud is known. Alternatively, if the Dpc is known, in a given cloud for a givenaerosol species, a plot similar to that in Figure 2.1 panel B can be used to predictthe maximum SC reached in the cloud.In summary, kappa-Ko¨hler theory can be used to predict whether or not anaerosol with a given diameter and hygroscopicity will activate to form a clouddroplet at a given S. However, Figure 2.1 panel B also indicates that for a givenaerosol at a constant S not all sizes of that aerosol will activate to form droplets.As an example, if ammonium sulfate were exposed to a supersaturation of ap-proximately 0.2% all diameters ≥≈90 nm would activate, while all diameters <90 nm would not activate to form cloud droplets. Therefore, a cloud will have bothactivated particles, which have particles incorporated into cloud droplets, as wellas non-activated particles referred to as interstitial particles. This is representedpictorially in Figure 1.2, where the cloud droplets are represented by grey circleswith smaller black circles inside, and the interstitial particles are represented bysmall red circles.182.3 Chapter 2 FiguresFigure 2.1: Panel A is an example Ko¨hler curve for a 50 nm dry (NH4)2SO4particle using a κ value of 0.61 (Petters and Kreidenweis, 2007). PanelB is the critical supersaturation (SC) as a function of ammonium sulfatedry diameter.19Chapter 3Single Particle Soot Photometer3.1 IntroductionThe primary instrument used to measure the mass, number and coating propertiesof rBC, throughout all of the research chapters in this dissertation, was a singleparticle soot photometer (SP2). Since the SP2 was the principal instrument usedthroughout the course of this thesis work, the following sections describe in detailthe SP2 theory of operation and procedures for data analysis.3.2 Theory of OperationThe SP2 (Droplet Measurement Technologies, Boulder, CO) is a commerciallymanufactured instrument designed to measure the mass of individual rBC parti-cles via laser induced incandescence (Stephens et al., 2003). A schematic of theSP2 is shown in Figure 3.1. The instrument is comprised of an aerosol inlet, an in-20tracavity laser and four detectors. Ambient aerosol particles are drawn in throughthe aerosol inlet at a flow rate of ≈ 120 cm3 min-1 and introduced into a highintensity (≈ 1 MW cm-2), intracavity, continuous Nd:YAG laser beam. A diodelaser, operating at a wavelength (λ ) of 808 nm, is used to pump a Nd:YAG dopedcrystal to produce laser light with an output λ=1064 nm in a TEM00 mode. Oncean aerosol particle intersects the intracavity laser, two events can occur: 1) theparticle can elastically scatter light and/or 2) the particle can absorb laser light.By measuring the elastically scattered light and by measuring the incandescenceof the particle due to absorbed laser light, the SP2 can be used to obtain the rBCmass, rBC volume equivalent diameter (VED), and the thickness of any coatingsurrounding individual rBC particles. Each of these measurements is discussed infurther detail in Sections 3.3 to 3.5 below.Two types of SP2s were used during the field studies discussed in this disser-tation. One is a four channel instrument, which records either a low-gain, or ahigh-gain signal for each of the four detectors shown in Figure 3.1 at any giventime. The second type of SP2 is an eight channel instrument that simultaneouslyrecords both the high and low gain signals for each of the four detectors shownin Figure 3.1, thus extending the size and mass detection limits. The detectionlimits for each of the specific SP2s used during the experiments discussed in thisdissertation will be addressed in the corresponding chapters for which they wereused.213.3 Mass CalibrationIf a particle efficiently absorbs at λ=1064 nm, such as rBC, it will rapidly heat toit’s vaporization, or boiling point, temperature (≈ 4000K for rBC) as it traversesthe laser beam profile. Once a rBC particle reaches its boiling point temperature,incandescence occurs, and the visible light emitted is recorded by two Photomulti-plier Tubes (PMT), labeled ‘Ch1 PMT Broadband Incandescence’ and ‘Ch2 PMTNarrowband Incandescence’ in Figure 3.1. The incandescence signals measuredby the SP2 are only from the refractory portion of black carbon. Both of the in-candescent PMT detectors are positioned to collect light over a solid angle of ≈pi/2 sr. The broadband incandescent detector is optically filtered to collect lightwith wavelengths between approximately 350 to 800 nm, whereas the narrow-band incandescent detector is filtered to only collect light between wavelengths of630 to 800 nm. The ratio of the broadband to narrowband signal can be used toextract the boiling point temperature by utilizing the theory of two-color pyrom-etry (Schwarz et al., 2006; Stephens et al., 2003). The boiling point temperatureis characteristic of the absorbing species and can be used as a means of particleidentification. However, since BC is considered to be the only species in the atmo-sphere that absorbs in the near-IR and gives a significant incandescence signal atthe wavelengths measured, all incandescent signals used in the analysis includedin this dissertation were taken to be from rBC (McMeeking et al., 2010; Schwarzet al., 2006).The amplitude of the signal recorded by the incandescent detectors is propor-22tional to the mass of rBC in a particle. The signal and rBC mass are related througha calibration curve created using a known reference material (see Figure 3.2 as anexample). All SP2s used in this study were calibrated with a reference materialrecommended by the manufacturer of the SP2. The calibration material is a col-loidal graphite called Aquadag® (Moteki et al., 2009). The Aquadag® is first sizeselected using a differential mobility analyzer (DMA) prior to being sampled bythe SP2. Data is recorded for ≈10,000 to 100,000 Aquadag® particles for eachparticle size selected by the DMA. The baseline corrected maximum signal am-plitude, referred to as peak height, of the signal recorded by the broadband incan-descent detector is determined for each particle. For a given size selected particlea distribution of peak heights is obtained. In order to find the average peak heightfor a given size the distribution of peak heights is fit with a Gaussian function.Once the average peak height for a given size-selected particle is determined,the corresponding mobility diameter is converted to a mass using the size depen-dent effective densities reported by Gysel and Crosier (2007). The proportional-ity between broadband incandescent peak height and mass is then determined byfitting a second order polynomial function to the data (Figure 3.2). The fit param-eters are then used to convert measured peak heights to mass for rBC particles inthe atmosphere.3.4 Volume Equivalent DiameterOnce the mass is determined, the volume equivalent diameter (VED) for eachindividual rBC particle can be calculated by assuming a density (ρbc) of 1.8 g23cm-3 (Bond and Bergstrom, 2006) and solving the following equation for VED:Mass =43pi(VED2)3ρbc. (3.1)3.5 Refractory Black Carbon Coating AnalysisUpon intersecting the intracavity laser beam, a particle will elastically scatterlight, and this scattered light is collected by two avalanche photodiodes (APDs)positioned to capture light over a solid angle of ≈ pi/2 sr (Schwarz et al., 2006).These two scattering detectors are both optically filtered to only collect light be-tween wavelengths of approximately 850 to 1200 nm, and are labeled ‘Ch0 APDScattering’ and ‘Ch3 APD Split Scattering’ in Figure 3.1 . The elastically scat-tered light detected by these two APDs can be used to obtain a measure of anycoating thickness on an absorbing rBC particle.The procedure for determining the coating thickness on a coated rBC parti-cle requires three steps: 1) the amplitude of the scattering signal from a coatedrBC particle is determined; 2) a relationship between the calculated Mie scatter-ing amplitude and the SP2 instrument response is established; and 3) a core andshell Mie scattering model is used to determine the coating thickness required toproduce the same scattering amplitude as determined in step 1. Each of these stepsis discussed in detail below.243.5.1 Step 1: Determining the scattering amplitude of coatedrefractory black carbon particlesSince the SP2 laser beam is in a TEM00 mode and exhibits a Gaussian profilethe scattering signal from a non-absorbing (referred to as scattering only) particlethat intersects the laser will also exhibit Gaussian behavior, where the maximumamplitude will occur at the most intense region of the beam, which for a Gaussianprofile is at the center. As an example, the scattering signal from a theoreticalscattering only particle is shown in Figure 3.3 panel A (blue trace), where themaximum amplitude is shown to occur at the center of the beam.When a rBC particle that has a non-refractive coating, such as sulfate, inter-sects the laser beam the coating will begin to vaporize before incandescence ofthe rBC core occurs (Schwarz et al., 2006). Therefore, when the coated particlereaches the center of the laser beam the scattering amplitude no longer representsthe scattering produced from a particle with a diameter equal to its core plus itscoating. Figure 3.3 panel B shows the SP2 signals recorded from a theoreticalabsorbing rBC particle, where the scattering signal is shown as a solid blue line,and indicates that the maximum scattering amplitude occurs prior to reaching thecenter of the beam and therefore no longer represents the true size of the particlebeing sampled. However, for a short time before the coating begins to vaporizethe scattering signal measured does represent the scattering properties of the un-affected coated rBC particle (Gao et al., 2007). Therefore, since the laser beamis fixed in space and has a Gaussian profile the full scattering signal of the coatedrBC particle can be reconstructed by using only the leading edge of the measured25scattering signal in order to determine the true scattering amplitude of the particle.This technique is called the leading edge only (LEO) technique and was first im-plemented, in conjunction with SP2 data, by Gao et al. (2007). The procedure fordetermining a coated rBC particle’s scattering amplitude using the LEO techniqueis outline below.3.5.1.1 Leading edge only fitting procedureThe scattering signal from a scattering only particle exhibits Gaussian behaviorand is therefore, mathematically represented by the following equation:F(x) = A∗ exp−(x−µsp22σsp2)2+B, (3.2)where A is the maximum scattering amplitude, and with respect to time, x is theposition of a particle within the laser beam at time t, µsp2 is the position of a par-ticle at the center of the beam, σsp2 is the width of the beam, and B is the baselineoffset. Since the SP2 laser has a stable Gaussian profile and is fixed in space σsp2and µsp2 are also fixed and can be characterized by using non-absorbing, purelyscattering, polystyrene latex (PSL) beads. The beam width is found by collectingdata for 50,000 to 100,000 PSL particles and each signal is fit with a Gaussianfunction to extract the σsp2 parameter. The average σsp2 from all PSL particles isthen used as the fixed beam width.The incorporation of a split-detector (labeled ‘Ch3 APD Split Scattering’ inFigure 3.1) has a physical gap in between two detecting surfaces, which providesan absolute positional reference for a particle in the laser beam, relative to the26beam center (Gao et al., 2007). The signal produced by the split-detector froma theoretical particle is shown in Figure 3.3 panels A and B as a sold green line.The scattering signal drops to zero when the scattered light falls on the gap in thedetector. The signal shape seen in Figure 3.3 is produced when the initial portionof the scattering signal, prior to crossing the split in detecting surfaces, is digitallyinverted in order to determine the position of the particle at the point at whichthe signal crosses zero (zero-crossing) by means of linear interpolation. UsingPSL particles, an average offset of the time from the zero-crossing to the timethe particle reaches the center (referred to as the zero-crossing-to-peak time) ofthe beam can be determined. This offset is then combined with the zero-crossingtime of an unknown rBC particle to determine the position at which the maximumscattering amplitude would occur had the coating not volatilized, µsp2.The point at which the scattering signal is 5% of the maximum laser intensityoccurs defines the last point in the leading edge used to reconstruct the full scat-tering properties (Gao et al., 2007). Similar to finding the zero-crossing-to-peaktime offset the ending point of the leading edge is relative to the zero-crossing,and the constant offset needed can be determined from PSL data.Once σsp2, µsp2, and the leading edge data points are known, Equation 3.2reduces to a linear function, where the slope represents the amplitude and the in-tercept is the baseline offset. Thus, using this method the true scattering amplitudeof coated rBC can be determined on a single particle level.273.5.2 Step 2: Relating calculated Mie scattering amplitudes toan SP2 instrument response (i.e. calibration of theelastic scattering detectors)In Section 3.5.3 a core and shell Mie model is used to predict scattering amplitudesof coated rBC particles. However, prior to using the Mie model, it is necessaryto be able to relate the calculated Mie scattering amplitude to the SP2 signal re-sponse. This is achieved by collecting data for different sizes of PSL particlesand calculating the scattering amplitude as determined by the LEO technique aswell as calculating the scattering amplitude by using Mie scattering calculations(Leinonen Mie Code developed by (Ma¨tzler, 2002a,b)). A PSL refractive indexof 1.59-0.0i was used in the Mie scattering calculation. Plotting the amplitude asdetermined by the LEO technique as a function of the amplitude as determined byMie calculations and fitting the data to a linear function provides the parametersneeded to relate the core and shell Mie scattering amplitudes to the SP2 signalresponse. An example calibration plot is shown in Figure 3.4, where data wereonly available for two PSL sizes, 200 and 300 nm.3.5.3 Step 3: Determining the coating thickness of a coatedrefractory black carbon particle using a core and shellMie scattering modelFor any rBC containing particles sampled by the SP2, the full scattering signalderives from both the rBC core and any coating surrounding the particle. Sincethe mass of the rBC is known from the incandescence signal (and can be usedto calculate the core VED) and the maximum scattering amplitude is determined28from the LEO technique, a core and shell Mie scattering model can be employedto determine what coating thickness would give the scattering amplitude as deter-mined from the LEO technique (Metcalf et al., 2012; Schwarz et al., 2008b). Inthis work a core and shell Mie scattering model was used to construct a lookuptable for core diameters from 60 to 220 nm (in 1 nm increments) and shell thick-nesses from 0 to 360 nm (in 1 nm increments). The complex index of refractionused for the core was 1.95-0.79i (Bond and Bergstrom, 2006) and for the shell was1.5-0.0i, which is consistent with that of dry sulfate or sodium chloride (Metcalfet al., 2012; Schwarz et al., 2008a,b). When calculating coating thicknesses, theparticles were idealized as a pure BC core uniformly coated with a non-absorbingmaterial, although the actual particle morphology may be more complicated (Sed-lacek et al., 2012). The calculated amplitudes determined from the core and shellMie scattering model were all scaled by the calibration plot shown in Figure 3.4.3.5.4 SP2 optical detection limits and the implications fordetermining average coating thicknessThe optical detection limits of the SP2 impose some restrictions on determiningaverage coating thickness across the full range of rBC core sizes that can be mea-sured by incandescence. These optical detection limits impose restrictions on thecoating analysis for both small and large rBC-containing particles.For small particles, although the incandescence measurements can size rBCcores down to≈ 70 nm, the SP2 optical detection limit means that scattering frombare rBC cores below ≈ 100 nm cannot be measured. As particle size decreases29below 100 nm, thicker and thicker coatings are required to produce a measurablescattering signal. As a result, particles with smaller rBC cores and thin or nocoatings have no scattering signal that can be used in the LEO fitting procedureand the coating thickness on them cannot be determined. In this analysis anyparticle with no measurable scattering was assumed to have a coating thickness of0 (i.e. they were assumed to be bare rBC cores). As a result of this assumption,the average coating thickness reported for particles below 100 nm is a lower limit.For large rBC cores the optical detectors become saturated when even rela-tively modest coatings are present. For example, the scattering from a 220 nmrBC core with a coating thickness of 40 nm would saturate the SP2 optical de-tector. Since the contribution from particles with large scattering signals is notincluded in the calculation of average coating thickness, it follows that the coatingthickness for particles larger than ≈ 100 nm is considered an underestimation.303.6 Chapter 3 FiguresFigure 3.1: Schematic of the SP2 showing the three main components: 1)aerosol inlet; 2) intracavity laser; and 3) the two scattering and twoincandescence detectors.31Figure 3.2: Example mass calibration plot for the broadband incandescencechannel of the SP2. This plot was created from pre (red circles) andpost (blue squares) campaign Aquadag data collected on one of theSP2s used in Chapter 4. The calibration data was fit to a second orderpolynomial function (solid black line) for the combined pre and postmeasurement data to create the equation used to determine the massof an unknown rBC particle (shown in the box below the plot). Alsoshown are the 95% prediction bands (pink dashed lines) correspondingto the polynomial fit. The X axis is plotted in terms of Aquadag masson the bottom axis and in terms of volume equivalent diameter (VED)on the top axis.32Figure 3.3: Raw signals produced by the SP2 from a theoretical scatteringonly particle in panel A and a theoretical coated absorbing rBC parti-cle in panel B. For both particle types, the signals recorded from thebroadband incandescence detector (red lines), scattering detector (bluelines), and split-detector (green lines) are shown. The reconstructedleading edge only (LEO) scattering signal (blue dashed line) is alsoshown in panel B. The center of the laser beam is represented by theblack dashed line and the point at which the signal recorded from thesplit-detector crosses zero is indicated by the solid black line.33Figure 3.4: An example calibration plot for the elastic scattering detectorcreated from 200 and 300 nm PSL particles. Each data point is themedian of ≈ 25,000 particles and the error bars indicate one stan-dard deviation. The black line is a linear fit with a forced interceptat zero.The units for both axes are arbitrary.34Chapter 4Size-Resolved Observations ofRefractory Black Carbon Particlesin Cloud Droplets at a MarineBoundary Layer Site4.1 IntroductionBC particles, which typically have sizes less than 1 µm, are emitted into the atmo-sphere through incomplete combustion of fossil fuels or biomass burning (Bondet al., 2013) (see Section 1.3) . When first emitted into the atmosphere these par-ticles are thought to mainly be hydrophobic. During their atmospheric lifecyclehydrophillic substances, such as sulfate or water soluble organics, can form a coat-35ing surrounding the BC cores (Ching et al., 2012; Metcalf et al., 2012; Schwarzet al., 2008b), enabling the particles to act as CCN (see Chapter 2). Modelingstudies have shown that this process can occur in the boundary layer on the orderof hours during the daytime (Riemer et al., 2004, 2010).By absorbing solar radiation or by acting as CCN, BC particles can influenceclimate both directly and indirectly (see Section 1.5 and Bond et al. (2013); Wang(2013); Wang et al. (2013); Zhuang et al. (2010)). To predict both the direct andindirect effects of BC on climate, a good understanding of the CCN ability of BCis needed (Koch et al., 2011; Vignati et al., 2010; Wang, 2013). When describingthe CCN properties of BC in atmospheric models, different approaches have beenapplied. Often BC is initially assumed to be CCN inactive and is converted toa CCN active species after a prescribed time and at a constant efficiency (Kochet al., 2011; Vignati et al., 2010; Wang, 2013). Alternatively, the CCN propertiesof particles containing BC in models have been described by Ko¨hler theory (seeChapter 2 and Ching et al. (2012); Jacobson (2012); Koch et al. (2011); Riemeret al. (2010)) .The CCN ability of BC particles has been studied in both laboratory and fieldstudies (Cozic et al., 2007; Dusek et al., 2011; Hallberg et al., 1992, 1994; Hen-ning et al., 2010, 2012; Hitzenberger et al., 2000, 2001; Kasper-Giebl et al., 2000;Koehler et al., 2009; Kuwata et al., 2009; Noone et al., 1992; Petters et al., 2009;Petzold et al., 2005; Popovicheva et al., 2011; Sellegri et al., 2003; Verheggenet al., 2007). Laboratory studies have examined both coated and uncoated BCparticles (Dusek et al., 2011; Henning et al., 2010, 2012; Koehler et al., 2009;36Petters et al., 2009; Petzold et al., 2005; Popovicheva et al., 2011). These stud-ies have shown that for uncoated flame or spark generated BC the particles arenot activated into cloud droplets even with supersaturations of ≥ 1% (Henninget al., 2012). On the other hand, once BC is coated with hygroscopic material thesupersaturation required for activation is significantly decreased, a decrease thatcan be described under laboratory conditions by Ko¨hler theory (Chapter 2 and(Henning et al., 2010, 2012; Petzold et al., 2005; Popovicheva et al., 2011)). Fieldmeasurements of the activated fraction of BC in cloud droplets have also inves-tigated the CCN ability of BC particles. Most measurements have shown that asBC particles age in the atmosphere the fraction incorporated into cloud dropletsincreases (Cozic et al., 2007; Hitzenberger et al., 2000, 2001; Kasper-Giebl et al.,2000; Sellegri et al., 2003; Verheggen et al., 2007). An exception to this trendis the work done by Granat et al. (2010) which showed that after several daysof travel in the winter subtropical marine environment, soot retained much of itshydrophobic properties. However, the conclusions stated in this study assumedthat hygroscopicity controlled the removal of BC through wet scavenging, whena more important factor is particle size. Furthermore, the washout ratio, which isthe metric used to interpolate scavenging efficiencies, was computed for BC massand not number. Therefore, direct comparisons of their work and the work pre-sented here are difficult to make. Some field studies (Hallberg et al., 1992, 1994)have shown that 10-20% of atmospheric BC can activate at supersaturations ofbetween ≈ 0.2-0.5%. Several of the studies mentioned in the preceding discus-sion used different measurement techniques than the one presented in this chapter.37Slowik et al. (2007) presents a detailed inter-comparison of different methods formeasuring BC and possible corrections needed for making direct comparisons be-tween differing techniques. None of the data reported here have been correctedfor sampling technique.Although the activated fraction of BC in cloud droplets has been measured atsome locations, the true contribution of BC to CCN in the atmosphere is unknown,yet potentially significant (Chen et al., 2010). Measurements of the activated frac-tion of BC as a function of size are important to test our current understanding ofthe incorporation of BC particles into cloud droplets. In this study we measuredrBC as a function of size in cloud residuals at a marine boundary layer site (251m amsl) in La Jolla, CA during 2012 and compared the results with rBC as afunction of size measured from a total inlet that sampled both cloud residuals andinterstitial particles. Coating thicknesses on rBC cores in the cloud residuals werealso determined. The measurements of rBC as a function of size in cloud residualswere further compared to predictions with kappa-Ko¨hler theory (see Chapter 2).4.2 Sampling Site4.2.1 Site descriptionThe sampling site was located below the peak of Mt. Soledad (251 m above meansea level (a.m.s.l.)), which is 3 km from the coast of the Pacific Ocean in La Jolla,CA (32.8400°N, 117.2769°W). The city of La Jolla is predominately residentialwith a population of approximately 43,000 people and situated 24 km north of38San Diego (population 1.3 million), the closest urban center.Data were collected from 27 May to 18 June, 2012 using instruments housed ina modified shipping container. A total of three stratocumulus clouds were sampledduring this time frame. The first cloud event was rejected from analysis due to aninstrumental error. The second cloud event occurred from 12 June 2012 20:43 to13 June 2012 11:35 PDT, and hereinafter called Cloud 2. The third cloud eventtook place from 17 June 2012 20:36 to 18 June 2012 07:52 PDT, and called Cloud3 for the remainder of the document.4.2.2 InletsTwo inlets, referred to as the total inlet and residual inlet, were used during thisstudy (Figure 4.1). The total inlet measured both interstitial and cloud residualparticles during cloud events. This heated inlet was designed and built followingthe specifications reported by Bates et al. (2002) and therefore assumed to havethe same transmission efficiency, namely > 95% for particles <6.5 µm.The intake of the residual inlet was a CVI (see Section 4.3.1) that enabledthe sampling of cloud droplets without interstitial particles, thus only the residualparticles of the cloud droplets were sampled. This inlet was used only duringcloudy periods and was connected to a branch of the total inlet by a 3-way valve(Figure 4.1). During a cloud event, the valve was manually switched so that clouddroplets were sampled through the CVI and cloud residuals were measured byinstrumentation connected downstream of the valve. At times when no cloudswere present the valve was switched such that all instruments sampled ambient39particles.Since much of the analysis performed in this study is based on a measuredratio of particle number concentration it was necessary to ensure that there wereno significant losses of particles due to the inlet configuration. Therefore, parti-cle losses from diffusion, sedimentation, turbulent inertial deposition and inertialdeposition from both bends and contractions for the total and residual inlets (as-suming cloud free sampling) were calculated using the Particle Loss Calculator(Von der Weiden et al., 2009) and found to be < 2% for particles with diametersbetween 0.07 and 1 µm, covering the size range used for this analysis.4.3 Experimental4.3.1 Counterflow virtual impactor4.3.1.1 Theory of operationIn order to separate cloud droplets from interstitial particles a counterflow virtualimpactor (CVI) was used. A CVI separates droplets from interstitial particles byusing inertial separation. A schematic of the CVI probe is shown in Figure 4.2.During operation, the CVI probe is placed inside a wind tunnel, which is con-nected to a vacuum pump that accelerates droplet laden air (shown as green linesin Figure 4.2) to high velocities (≈100 m s-1). Since the air contains differentsizes of droplets and particles, the process of accelerating this air mass to a con-stant velocity imparts different inertial forces on the droplets and particles. Thevacuum intake draws the accelerated air towards the CVI probe, which is made up40of two concentric cylindrical tubes (see Figure 4.2). The tip of the outer cylinderprovides the inlet where cloudy air is sampled through an orifice with an outerradius of ro. The inner concentric cylinder has a porous frit at its tip; which al-lows a supply gas (shown as red lines in Figure 4.2) to permeate through. Theinner cylinder is also connected to a vacuum pump downstream, which createsthe sample flow (shown as a purple line in Figure 4.2), the counterflow (shown asyellow lines in Figure 4.2), and the stagnation plane, which is a theoretical planewithin the porous frit that droplets must cross in order to be sampled (shown as ablack dashed line in Figure 4.2). Based on a CVI’s specific geometry and the flowrates used a minimum droplet diameter exists that will be sampled by the CVI.Once a droplet enters the CVI probe tip the water component of the droplet, aswell as any dissolved volatile gasses, begin to evaporate upon impacting with thewarm dry counterflow of gas (often purified air or N2). Droplets that are sampledby the CVI are further evaporated by heated sections of the sample tubing, whichare typically held at a constant temperature of 40°C. Therefore, the instrumenta-tion downstream of the CVI samples only the residuals of the cloud droplets thatremain after evaporation.4.3.1.2 Calculating a CVI cut-sizeThe minimum droplet diameter sampled by a CVI is considered to be the CVIcut-size (CVI−D50) and is the diameter for which a droplet’s stopping distance(Dstop) is greater than the length to the stagnation plane (Lstag) (Anderson et al.,411993; Noone et al., 1988), or:CVI−D50 = Dstop > Lstag, (4.1)where Lstag is comprised of three components, and is equal to:Lstag = Lmin +Lpor +Lcur, (4.2)where Lmin (see #2 labeled in Figure 4.2) is the amount of dead-space that existsbetween the CVI orifice and the top of the porous frit, Lpor (see #3 labeled inFigure 4.2) is the length from the top of the porous frit to the stagnation plane,which is related to the counterflow (F1), sample flow (F2), and the length of theentire porous frit (X , see #4 labeled in Figure 4.2) by Equation 4.3:Lpor =[F1−F2F1]X , (4.3)Lcur is a parameterization (Anderson et al., 1993) that estimates the length aparticle must traverse to deviate from the high velocity streamlines curving aroundthe CVI probe tip in order to enter the CVI orifice, and is considered a linearfunction of the outer radius (ro) of the tip:Lcur =Cro, (4.4)where C can range from 0 to 1 (Anderson et al., 1993; Noone et al., 1988) and42represents the uncertainty of the estimate. In this document, CVI−D50 is alwaysreported at C=0.5 and the uncertainty of the cut-size is calculated by averaging thedifferences calculated when using the lower and upper bounds of C=0 and C=1,respectively.DStop can be calculated as follows (Serafini, 1954):DStop =rdρd3ε 32 ρair[Re1/3d ε1/2−pi2+ tan−1(Re−1/3d ε−1/2)], (4.5)where rd is the droplet radius, ρd is the droplet density, ε is a constant (0.158), ρairis the density of air, and Red is the droplet Reynolds number, which is representedby:Red =2ρdrdV∞ηair, (4.6)where V∞ is the CVI air intake velocity, and ηair is the viscosity of air.4.3.1.3 Calculating a CVI enhancement factorSampling through a CVI will also enhance the particle concentrations relative tothe initial ambient concentrations (Noone et al., 1988; Serafini, 1954). Therefore,measurements made downstream of a CVI need to be corrected by what’s referredto as the enhancement factor (EF), which can be calculated using the followingequation:EF =V∞pir2iF2, (4.7)where ri is the inner radius of the CVI probe tip.The CVI used during this study, as well as the study discussed in Chapter 543was based on the design by (Noone et al., 1988) with the following geometry: anouter radius (ro) of 0.9 cm; an inner radius (ri) of 0.3 cm; and a 9.5 cm long porousfrit (X).4.3.1.4 CVI cut-size and enhancement factor measured at La Jolla, CABased on Equation 4.5 above, a CVI−D50 of 11.5±0.7 µm and 11.6±0.7 µm forCloud 2 and Cloud 3, respectively, were calculated for the two clouds sampled inthis study. In addition, using Equation 4.7, calculated EF values were 7.1 and 7.4for Cloud 2 and Cloud 3, respectively.4.3.2 Refractory black carbon measurements4.3.2.1 Refractory black carbon mass measurementsRefractory black carbon (rBC) mass was measured from the total inlet and theresidual inlet using two separate SP2s (Droplet Measurement Technologies, Boul-der, CO). These instruments are referred to as the total SP2 (SP2Tot) and the resid-ual SP2 (SP2Res). The location of these instruments is shown in Figure 4.1. TheSP2 has been described in detail in Chapter 3 and elsewhere (Moteki and Kondo,2008; Schwarz et al., 2006; Stephens et al., 2003). The two SP2s used in this studywere calibrated pre and post-campaign following the procedure discussed in Sec-tion 3.3. The calibration parameters used to determine mass were taken from asecond order polynomial fit of the combined pre and post-campaign data. A VEDwas also determined from the measured mass assuming a BC density of σbc=1.8g cm-3 (Bond and Bergstrom, 2006). The SP2Res was a 4 channel instrument with44a detection range of 70-220 nm, whereas the SP2Tot was an 8 channel instrumentwith a detection range of 70-558 nm.4.3.2.2 Refractory black carbon coating thickness measurementsThe information from the two scattering APDs (see Figure 1.1) was used to de-termine the coating thicknesses on rBC cores as described in Section 3.5 and byGao et al. (2007). Due to a failure of the split detector on the SP2 connected to thetotal inlet, coating thicknesses on rBC cores were only determined from the datacollected with the SP2 connected to the residual inlet. To determine coating thick-nesses the scattering amplitudes for rBC containing particles were determinedusing the LEO fitting method (see Section 3.5 and Gao et al. (2007)). In short, theearly part of the elastic scattering signal (up to 5% of the maximum laser intensity(Gao et al., 2007)) was fit to a Gaussian function and the maximum scatteringamplitude was retrieved from the fit.As mentioned above, only data collected with the SP2 connected to the resid-ual inlet were analyzed for coating thickness. Approximately 99% of the rBCcontaining particles detected with the SP2Res had elastic scattering signals abovethe background signals on the APDs. Nevertheless, only 50% of the rBC contain-ing particles detected with the SP2Res were successfully fit with the LEO fittingprocedure. Hence, coating information reported here is only from a subset of therBC particles measured with the SP2Res. Failure to fit the elastic scattering signalswith the LEO procedure was mainly due to either; a) large scattering signals thatsaturated the APD detector, or b) time dependent scattering signals were above 5%45of the maximum laser intensity at time zero. 11% and 27% of all rBC containingparticles detected with the SP2Res had a failure due to a) and b), respectively.4.3.3 Size distribution measurements of the bulk aerosolTwo instruments were used to measure particle size distributions (Figure 4.1).Size distributions of particles sampled from the total inlet were determined witha scanning electrical mobility spectrometer (SEMS, model 2002, BMI, Hayward,CA), which counted particles into 61 discrete size bins from 0.01-1 µm with a 5-minute scan time interval. Size distributions of particles sampled from the residualinlet were determined with a scanning mobility particle sizer (SMPS, model 3034,TSI, St. Paul, MN), which recorded particle counts into 55 size bins from 10-487nm with a 3-minute scan time interval. Both the SEMS and SMPS operate basedon the coupling of a size selecting DMA and a condensational growth particlecounter (CPC). During cloud free sampling the SMPS and SEMS agreed to within4% for particles between 70 and 400 nm.4.3.4 Aerosol mass spectrometryAn online high-resolution time-of-flight aerosol mass spectrometer (HR-ToF-AMS,Aerodyne Reserch Inc., Billerica, MA) was operated downstream of the CVI onthe residual inlet to characterize the chemical composition of the cloud dropletresiduals. The HR-Tof-AMS measured non-refractory, sub-micrometer aerosolchemical composition at high time resolution (DeCarlo et al., 2006). Here weonly consider data measured by the HR-ToF-AMS in its mass-spectrum and V-46modes of operation. These data were recorded as 2 minute averages every 4-6minutes, depending on how many other modes of operation (W-mode, light scat-tering) the instrument was alternating between. Size-resolved composition datafor the residual particles measured by the HR-ToF-AMS in time-of-flight modeare not considered here since the signal was generally at or below the detectionlimit. Standard quantification procedures (Allan et al., 2004) were applied to themass spectra measured by the HR-ToF-AMS to determine the relative concen-trations of the non-refractory species (organic, nitrate, sulfate, ammonium andchloride) typically reported by aerosol mass spectrometry.4.3.5 Back trajectoriesAir mass back trajectories were calculated using the National Oceanic and At-mospheric Administration (NOAA) Hybrid Single Particle Lagrangian IntegratedTrajectory Model (HYSPLIT) (Draxler and Rolph, 2013; Rolph, 2013). All trajec-tory calculations used the National Centers for Environmental Predictions EDASmeteorological dataset. Trajectories were calculated starting at 10 m above groundlevel (a.g.l.), 96 hours backwards in time, and at hourly intervals throughout theentire period of cloud sampling.4.3.6 Cloud propertiesA fog monitor (FM-100, model 100, Droplet Measurement Technologies, Boul-der, CO), which is a forward scattering optical spectrometer, was stationed on topof the shipping container providing an in-situ measurement of CDNC in 20 dis-47crete size bins ranging from 2-50 µm, whilst simultaneously monitoring the liquidwater content (LWC) present with a 1 second time resolution (Eugster et al., 2006).4.4 Results and Discussion4.4.1 Back trajectoriesThe back trajectories for Cloud 2 (Figure 4.3 panels A and C) show that the airmass spent most of the previous 96 hours over the Pacific Ocean and arrived atthe sampling site from a northwesterly direction. During the first part of Cloud 2(12 June 21:00 to 13 June 08:00 PDT) the back trajectories became progressivelymore northerly and the air mass began traveling near large populated urban re-gions. Towards the end of the cloud event (at ≈ 09:00 PDT on 13 June) the windsshifted to southwesterly. Based on the back trajectories the air mass for Cloud2 traveled ≈ 40-50 km over land before reaching the sampling site. In addition,the air mass spent a significant amount of time close to the ocean surface prior tobeing lifted up to the sampling site (Figure 4.3 panel C).The back trajectories for Cloud 3 (Figure 4.3 panels B and D) also show thatthe air mass spent the majority of the previous 96 hours over the Pacific Oceanbefore arriving at the site. At the start of Cloud 3 (17 June 21:00 to 22:00 PDT) theair mass arrived from the northwest. Throughout the remainder of the cloud event(17 June 23:00 to 18 June 08:00 PDT) the air mass continued to shift farther northand by the end of the cloud event (18 June 08:00 PDT) the air mass was travelingsouth along the coastline before arriving at the sampling site. The back trajectories48indicate that the air mass traveled ≈10-20 km over land prior to arriving at thesampling site. Similar to Cloud 2, the air spent a significant amount of time closeto the ocean surface prior to being lifted up to the sampling location (Figure 4.3panel D). Since the trajectories during both clouds are close to the coastline fora period of time, it is likely these air masses contained both marine particles andanthropogenic emissions.4.4.2 Meteorological conditions and cloud propertiesFor the purposes of this study, the data were classified as in-cloud and includedfor analysis if they met the following criteria: 1) the five-minute-averaged CVIcounterflow was within ±5σstd of the mean counterflow to ensure only periods ofstable CVI flows were included; and 2) the five-minute-averaged LWC was greaterthan 0.05 g m-3 to remove periods of entrainment, or “patchy”regions of the cloudas much as possible (Cozic et al., 2007).The measured cloud properties as a function of time are shown in Figure 4.4(panels A to C), where Cloud 2 is shown on the left side and Cloud 3 is shown onthe right side of the plot. Cloud 2 was characterized by an average temperature of13.4±0.2 °C with light and variable winds (0.5±0.2 m s-1 and 190±90°respectively),and an average LWC of 0.13±0.07 g m-3. During the middle portion of Cloud 2(13 June 01:00 to 02:00 PDT) the droplet distributions clearly show an intervalwhere the number of droplets above the CVI−D50 (black trace overlaid on panelC) increases significantly, which coincides, in time, with a sharp increase in LWC.Cloud 3 was characterized by an average temperature of 15.2±0.1 °C, northerly49(330±10°) winds with an average speed of 1.5±0.4 m s-1, and an average LWCof 0.09±0.02 g m-3.The cloud droplet number and volume size distributions, averaged over theentire event, are shown in Figure 4.5, and further summarized in Table 4.1. Cloud3 had a CDNC of 146 cm-3, a factor of two higher than during Cloud 2 (68 cm-3).From the calculated CVI −D50 and the fits to the droplet size distribution(Figure 4.5) the number and volume fraction of droplets sampled by the CVI weredetermined. The results are summarized in Table 4.1. During Cloud 2 the numberfraction of droplets larger than the CVI−D50 was about 38% and for Cloud 3the fraction sampled was about 24%. Since only the larger droplets were sampledby the CVI during these two cloud events, the results presented herein are onlyrepresentative of the larger droplet population.4.4.3 Size distributionsAverage size distributions of the bulk aerosol particles and rBC particles measuredfrom the total and residual inlets for both Cloud 2 and Cloud 3 are shown inFigure 4.6. Data are plotted in two ways: on a log scale, in panels A and B, andnormalized to the respective maximum, in panels C and D. Table 4.2 summarizesthe results obtained from the size distribution analysis.4.4.3.1 Size distributions measured from the total inlet (BulkAeroTot andrBCTot)The average size distributions of the bulk aerosol measured from the total inlet (re-ferred to as BulkAeroTot for the remainder of the document) and the average size50distributions of rBC measured from the total inlet (referred to as rBCTot for the re-mainder of the document) are shown in Figure 4.6. The BulkAeroTot distributionsfor both clouds show evidence of at least two overlapping modes. A single modelognormal distribution function was fit to the data yielding mean geometric diam-eters (Dg) of 108 and 81 nm with geometric standard deviations (σg) of 1.58 and1.70 for Cloud 2 and Cloud 3 respectively. Integration of the number distributionduring Cloud 2 results in a total number concentration (NTot) for the bulk aerosolof 981 cm-3. Likewise, NTot during Cloud 3 was measured to be 994 cm-3. Previ-ous measurements in the marine boundary layer have classified the environmentas “clean” marine if the number of particles is ≤ 300-500 cm-3 and “polluted”marine if the number concentrations are ≥ 400-1500 cm-3 (Andreae, 2009; Bateset al., 2000; Glantz and Noone, 2000; Hawkins et al., 2010; O’Dowd et al., 2001;Pirjola and O’Dowd, 2000; Twohy et al., 2005). Thus, the particle concentrationsmeasured at Mt. Soledad, in addition to the back trajectories, suggest that for bothclouds the air masses can be classified as polluted marine aerosols. The size dis-tributions of the BulkAeroTot as a function of time are also included in Figure 4.4panel D for comparison.The rBCTot size distributions for each cloud are shown in Figure 4.6. The Dgfor rBCTot (assuming the number distributions are lognormal) during both eventslies somewhere in the nucleation mode at <70 nm, which is outside the detectionrange of the SP2. Integration of the rBCTot distributions, from 70-220 nm, yieldsan NTot of 75 cm−3 during Cloud 2 and 62 cm-3 in Cloud 3. Assuming a rBCdensity of σbc=1.8 g cm-3 the total mass concentrations (MTot) are 73 and 62 ng51m-3 for Cloud 2 and Cloud 3 respectively (Table 4.2). The rBCTot mass concentra-tions observed at Mt. Soledad were higher than concentrations measured in cleanmarine air (Cooke et al., 1997; Shank et al., 2012), but considerably lower thanconcentrations measured in most urban environments (see Table 1 in Metcalf et al.(2012)).4.4.3.2 Size distributions measured from the residual inlet (BulkAeroResand rBCRes)Average size distributions of the bulk aerosol measured from the residual inlet(referred to as BulkAeroRes for the remainder of the document) are also shown inFigure 4.6. The size distributions of the BulkAeroRes indicate that it was mostlythe larger particles of the BulkAeroTot distributions that were incorporated intothe sampled cloud droplets. The size distributions of BulkAeroRes as a function oftime for both cloud events are included in Figure 4.4 panel E for comparison.The size distributions for BulkAeroRes shown in Figure 4.6 have a local min-imum at 110 nm for Cloud 2 and 90 nm for Cloud 3. The particles observed atsizes less than the local minima may be due to droplet “splash”, or due to a leak inthe CVI. (Pekour and Cziczo, 2011; Schwarzenboeck et al., 2000; Vidaurre et al.,2011).Figure 4.6 also shows the size distributions of the rBC residuals (rBCRes) mea-sured with the CVI. Figure 4.6 shows that rBC cores smaller than 100 nm areincorporated into cloud droplets. In addition, Figure 4.6 shows that most of therBCRes have effective sizes of less than 100 nm and that the rBCRes are overalllarger than the rBCTot. Fitting the rBCRes size distributions (assuming these dis-52tributions are lognormal) results in mean geometric diameters of 87 and 81 nm forClouds 2 and 3 respectively.4.4.4 Size-resolved activated fractionsThe size-resolved activated fraction [AF(Dp)] for rBC and the bulk aerosol werecalculated by taking the ratio of the number size distributions measured from theresidual inlet and the number size distributions measured from the total inlet. Priorto calculating AF(Dp), a spline interpolation algorithm was applied to the rBC andbulk aerosol number size distributions. After a spline interpolation was applied tothe data, the following equation was used to calculate the size-resolved activatedfraction:AF(Dp) =NRes(Dp)∗CF(Dp)NTot(Dp)∗EF ∗DT, (4.8)where NRes(Dp) is the number of residual particles as a function of size, CF(Dp)is the size-resolved instrument sensitivity correction factor, NTot(Dp) is the num-ber of particles measured from the total inlet as a function of size, EF is the CVIenhancement factor (see Section 4.3.1.3) and DT is the droplet transmission fac-tor through the CVI. Calculations of the droplet transmission factor are given inAppendix B and plotted in Figure B.1. CF(Dp), which corrects for variances ininstrument detection efficiencies, were determined from a 12hr period of cloudfree air on 5 June 2012 for the bulk aerosol and from side by side ambient sam-pling of room air during the post-campaign calibration for rBC.The AF(Dp) for the bulk aerosol and rBC are presented in Figure 4.7, wherethe error bars represent one standard deviation (1σstd) of the AF for each 10 nm53size bin. The AF as a function of size for the bulk aerosol is similar to previousmeasurements in similar clouds (Hallberg et al., 1994). Figure 4.7 panels A andB show that the AF of rBC cores is significant, even for core diameters of ≤ 100nm. Figure 4.7 panels A and B also show that during both clouds the AF for rBCcores is larger than the AF for the bulk aerosol at diameters <≈150 nm. These re-sults can be explained by the presence of large coatings surrounding the core (seeSection 4.4.5). Since the fraction of droplets sampled by the CVI was <100%, thecalculated AF should be considered as a lower limit to the total fraction activatedduring these two cloud events.4.4.5 Coating thickness of refractory black carbon residualsCoating thicknesses were determined using a core and shell Mie model (see Sec-tion 3.5) and are shown in Figure 4.7 panels C and D, where the error bars rep-resent 1σstd and the symbols represent the averages. When calculating coatingthicknesses, the particles were idealized as a pure BC core uniformly coated witha non-absorbing material, although the actual particle morphology may be morecomplicated (Sedlacek et al., 2012). As mentioned in Section 4.3.2.2, only 50% ofthe rBC containing particles detected with the SP2Res were successfully fit withthe LEO fitting procedure. Hence, coating information reported below is onlyfrom a subset of the rBC residual particles measured.Both cloud events show a similar trend, namely that the coating thicknessesare larger at smaller rBC core diameters than larger rBC core diameters, which hasalso been observed elsewhere (Metcalf et al., 2012). At small core diameters (70-54100 nm) the coating thicknesses range from roughly 45-115 nm, while at largercore diameters (200-215 nm) the coatings range from roughly 0-60 nm. Coat-ing thicknesses measured in this study fall between values measured in previousstudies. For example, coating thicknesses ranging from 20±10 nm, in fresh urbanplumes (Schwarz et al., 2008a), up to 188±31 nm on more aged rBC particles(Metcalf et al., 2012) have been reported.4.4.6 In-cloud aqueous phase chemistryIn the discussion above, we assume the coatings on the rBC cores were presentbefore incorporation into the cloud droplets. However, some of the coating ma-terial may have formed after the rBC cores were incorporated into the cloud byaqueous phase chemistry. Although stratus cumulus decks can persists for hours,large eddy simulations of non-precipitating stratocumulus suggest that in-cloudresidence times of air parcels can be on the order of 12 minutes (Stevens et al.,1995). Based on this work, we assume that an upper limit of 1 hour for the cloudsmeasured in this study is reasonable. Whether significant aqueous phase chem-istry can occur on this time scale depends on the level of SO2 and oxidants. WhenSO2 is absorbed by a cloud droplet, it partitions in different forms as a function ofpH: at lower pH values, the primary aqueous phase oxidant of dissolved SO2, orS(IV) is H2O2; for higher pH values, ozone and catalyzed aerobic oxidation areimportant oxidation pathways. Depending on the pH and available oxidants, con-version of the dissolved S(IV) to S(VI) can be fast or slow. The absorption of largeamounts of HNO3 or N2O5 can reduce the pH significantly, which will require the55presence of H2O2 in order to significantly convert S(IV) to S(VI). Since the cloudsat Soledad were mostly overnight, the primary oxidant (H2O2) was likely lower(and probably near zero). Also, analysis of the cloud water indicated that nitratewas high, and thus the pH was probably low. Based on these factors, we suspectthat aqueous phase production of sulfate was not large. However, a quantitativeestimate of the sulfate produced is not possible since the measurements of SO2and H2O2 were not performed at this site. Based on the discussion above, we as-sumed that a large fraction of the coatings were present before rBC particles wereincorporated into the cloud droplets. However, the production of coating materialfrom in-cloud aqueous phase chemistry cannot be ruled out.4.4.7 Size distributions of rBCRes(Core+Coating)For rBC particles measured from the residual inlet, a core diameter (determinedfrom the incandescence signal) and a coating thickness (determined from the Miescattering calculations discussed in Section 3.5) were combined to calculate theoverall rBC particle diameter (Core + Coating). These data are plotted in Fig-ure 4.8, where the distributions are shown to be shifted to larger sizes when com-pared with the distributions plotted as a function of core diameter [rBCRes(Core)]alone. These results suggest that the coated rBC particle diameters were likelythe dominant factor in activation ability over composition. This is a result that hasalso been seen elsewhere (Fierce et al., 2013). For example, Fierce et al. (2013)investigated how different particle characteristics (e.g. size, supersaturation, and24 hour aging) at emission influence CCN activity and found that for SC ¿ 0.2%56CCN activity had a greater sensitivity to emission size than composition.4.4.8 Comparison of rBCRes as a function of size withpredictions based on kappa-Ko¨hler theorySections 4.4.5 and 4.4.7 provide a qualitative explanation for why rBC coressmaller than 100 nm are incorporated into cloud droplets, namely that rBC coreshave thick coatings which lead to overall particle diameters >100 nm. In the fol-lowing, we expand on this qualitative explanation by carrying out a quantitativeanalysis showing that the presence of rBC cores smaller than 100 nm in the cloudresiduals is consistent with kappa-Ko¨hler theory. This quatitative analysis con-sists of the following steps: 1) an estimation of the bulk aerosol composition; 2)an estimation of the critical diameter for activation of the cloud droplets sampled;3) an estimation of the critical supersaturation required to form the droplets sam-pled; and 4) a prediction of the critical diameter for activation of rBC cores. Steps1-3 are required to carry out the predictions in step 4.4.4.8.1 Bulk aerosol compositionAn HR-Tof-AMS was used to measure the bulk aerosol composition downstreamof the CVI. Five species (organic, nitrate, sulfate, ammonium and chloride) werequantitatively differentiated. Then, based on a simplified ion-pairing scheme sim-ilar to Gysel and Crosier (2007) (see Appendix A), the mass fractions of am-monium nitrate, ammonium sulfate, ammonium bisulfate, sulfuric acid, and am-monium chloride were calculated. The results of these calculations are shownin Figure 4.9. In order to determine the bulk aerosol hygroscopicity (see Equa-57tion 2.3), mass fractions, of these individual components, were first converted tovolume fractions using an organic density of 1.4 g cm-3 (Moore et al., 2012) anddensities reported in Lide (2001) for the inorganic salts.4.4.8.2 The critical diameter for activation of the bulk aerosol in the clouddroplets sampledThe critical diameter for activation (Dpc) of the bulk aerosol is often calculated byintegrating the residual number distribution from the largest to the smallest diam-eters until the number concentration equals the CDNC sampled (see for exampleHersey et al. (2013)). Using this method, Dpc was found to be 241 nm and 239nm for Cloud 2 and Cloud 3 respectively. Note these Dpc apply only to the clouddroplets sampled (i.e. cloud droplets >≈11 µm). Different Dpc values would beexpected if the entire droplet population were sampled.4.4.8.3 Critical supersaturation for the cloud droplets sampledTo estimate the critical supersaturation (SC) for the formation of cloud dropletssampled during the two cloud events (i.e. cloud droplets >≈11 µm) the singleparameter kappa-Ko¨hler model was used (see Chapter 2). The individual κ valuesused in Equation 2.3 for the species discussed in Section 4.4.8.1 were; 0.1 fororganic (Lance et al., 2013; Moore et al., 2012; Rose et al., 2010); 0.67 for am-monium nitrate (Petters and Kreidenweis, 2007); 0.61 for ammonium sulfate andammonium bisulfate (Petters and Kreidenweis, 2007; Wu et al., 2013); and 0.71for sulfuric acid, which is the average of the range reported in Shantz et al. (2008).The value for ammonium chloride κ was calculated according to Equation A28 in58Rose et al. (2008) using a Van’t Hoff factor of 2.Using the above κi values and the values for υi, discussed in Section 4.4.8.1,in Equation 2.3, κBulk values of 0.50 and 0.41 were calculated for Cloud 2 andCloud 3 respectively. The values determined during this study are consistent withthe values suggested by Andreae and Rosenfeld (2008) for Cloud 2 and lower thanthe values suggested for Cloud 3, for marine aerosols.Shown in Figure 4.10 panel A are plots of SC as a function of dry diameterfor Cloud 2 (solid line) and Cloud 3 (dashed line) calculated using Equation 2.2and Equation 2.3. Combining Dpc (see Section 4.4.8.2) with the results plotted inFigure 4.10, SC for the cloud droplets sampled can be determined. The points atwhich Dpc intersects with the calculated SC traces shown in Figure 4.10, result inestimations of SC values of approximately 0.05% for both clouds. Note, these SCapply only to the droplets sampled by the CVI. Different SC values would be ex-pected if the entire droplet population had been sampled. However, it also shouldbe noted that theory predicts that the largest droplets in the distribution shouldhave been the first to form, thus formed on particles activated at the lowest super-saturations. The SC calculated for the coated rBC particles shown in Figure 4.10are consistent with this theory.In this determination of SC several assumptions were made, which are ad-dressed separately below: 1) the predominant mechanism for incorporation ofparticles into droplets was nucleation scavenging, and influences by impactionwere negligible (Noone et al., 1992); 2) the contribution from sea salt aerosolscould be neglected. Based on sea salt mass concentrations measured by the HR-59Tof-AMS behind the CVI and calibrated against collocated ion chromatographymeasurements following a procedure similar to that introduced by Ovadnevaiteet al. (2012), we estimated an upper limit of approximately 15% for the sea saltmass fraction of cloud residuals. By including an estimated sea salt mass fractionof 15% into the bulk aerosol composition, bulk kappa values of 0.57 and 0.49were calculated and used to estimate SC. Using these kappa values reduced thereported estimated SC of 0.05% by <8%; 3) we assumed that the particles wereinternally mixed and the composition did not depend on size. Since, during thisstudy, the size dependent HR-ToF-AMS data were at or below the detection limitwe could not determine if the composition was dependent on size. Additionally,no measurements of the bulk aerosol mixing state were carried out; and 4) we as-sumed that the entire fraction of organics was water soluble and represented by aκ of 0.1. To determine if SC was sensitive to this value, κ for organics was variedfrom 0-0.2, which is roughly consistent with the range of κ values reported in theliterature for organics (Chang et al., 2010; Lathem et al., 2013; Mei et al., 2013).Over this range of κ values, SC varied by <4%.4.4.8.4 Predictions of the critical diameter for activation of refractoryblack carbon coresIn Figure 4.10 panel B the critical diameter for activation of rBC cores, for clouddroplets sampled, is calculated using kappa-Ko¨hler theory (see Chapter 2) andassuming coating thicknesses ranging from 0-200 nm, which covers the rangeof coating thicknesses measured. In these calculations the composition of thecoating was assumed to be the same as determined by the HR-ToF-AMS (see60Section 4.4.8.1), and the rBC cores were assumed to be insoluble with a κ = 0(Rose et al., 2010). As expected, in Figure 4.10 panel B, SC decreases as thecoating thickness increases at a constant rBC core diameter. Figure 4.10 panelB also shows that if the SC is >≈ 0.05%, for both clouds, the critical diameterfor activation of the rBC cores is <50 nm for all cores with coating ≥100 nm.The combinations of core + thicknesses shown in Figure 4.10 can be translatedinto the equivalent volume fractions and kappa values, assuming the coating hasa kappa of 0.5 and the core has a kappa=0. The smallest kappa in this figure isfor a 1 µm core with a 25 nm coating, κ=0.07; however the size is so large thecritical supersaturation is very low anyways despite the low kappa. The smallesttotal particle size (and highest critical supersaturation) represented in Figure 4.10is a particle with a 50 nm core and 25 nm coating, total 100 nm diameter; thecorresponding kappa is 0.44. For cores between 50 and 200 nm and thicknessesbetween ≈ 25 to 200 nm, kappas range from ≈ 0.35 to 0.5. As can be seen forthe two lines in Panel A this much variability in overall kappa does not create alarge change in critical supersaturation. This suggests that the uncertainties of theestimated SC should have minimal influence on the predicted critical activationdiameter of rBC using kappa-Ko¨hler theory. Furthermore, the results from kappa-Ko¨hler theory are consistent with the observations of rBC cores of 100 nm andless being activated into the sampled cloud droplets.614.5 Summary and ConclusionsCloud residuals were measured during two cloud events at the top of Mt. Soledadin La Jolla, CA. Back trajectories showed that air masses for both cloud eventsspent at least 96 hours over the Pacific Ocean and traveled near, or over popu-lated regions before arriving on site. Based on measured bulk aerosol numberconcentrations the two air masses sampled were classified as polluted marine air,a classification consistent with the back trajectories and measured concentrationsof black carbon mass.The size distributions of the bulk aerosol residuals were shifted to larger sizesfor both cloud events compared to the size distributions measured from the to-tal inlet. The size distributions of rBC cloud residuals were also shifted towardslarger diameters when compared to the size distributions of rBC measured fromthe total inlet. The measurements of cloud residuals clearly show that rBC coresless than 100 nm can be incorporated into cloud droplets, assuming that the coat-ings surrounding these cores are at least moderately hygroscopic and sufficientlylarge enough to increase the overall particle diameter so that they can contributeto the CCN at low supersaturations. The activated fraction of 70-80 nm rBC coreswas 0.01 and 0.045 for Cloud 2 and Cloud 3 respectively. Since the fraction ofcloud droplets sampled by the CVI was less than 100%, the measured activatedfractions of rBC are lower limits to the total fraction of rBC activated during thesetwo cloud events. The coating analysis of the SP2 data shows that the rBC coresthat were activated into cloud droplets had thick coatings, with average coating62thicknesses of≈75 nm at core diameters between 70-80 nm and≈29 nm coatingsfor core diameters between 200-210 nm. Furthermore, the presence of rBC coresless than 100 nm in cloud residuals is consistent with kappa-Ko¨hler theory andthe measured rBC coating thicknesses.634.6 Chapter 4 Figures and TablesFigure 4.1: Schematic showing the configuration of the inlets and instru-mentation housed in the shipping container.64Figure 4.2: Schematic of the CVI showing ambient air (green dashed lines)being drawn towards the CVI probe. The supply flow (red lines) areshown to permeate the porous frit (blue dotted region) creating thecounterflow (orange lines) and sample flow (purple dash-dot line). Thetheoretical stagnation plane (black dashed line) and CVI geometry pa-rameters needed to calculate the stopping distance of a droplet enteringthe CVI probe (see Section 4.3.1 for further details) are also shown andlabeled 1 to 4.65Figure 4.3: In-cloud HYSPLIT 96hr back trajectories ending at hourly inter-vals for Cloud 2 (12 June 21:00 to 13 June 12:00 PDT) in panels A andC, and Cloud 3 (17 June 21:00 to 18 June 08:00 PDT) in panels B andD. All back trajectories started at 10 m a.g.l.. Darker yellow regions onland in panels A and B indicate densely developed urban areas contain-ing 50,000 or more people (United States Census Bureau). Panels Cand D show the vertical profiles over the same hourly intervals shownin panels A and B.66Figure 4.4: Time series data for both Cloud 2 (left side) and Cloud 3 (rightside) showing; liquid water content (LWC, blue trace) and ambienttemperature (red trace) in panel A; wind speed and direction in panelB; cloud droplet number size distributions with the CVI−D50 (blacktrace) overlaid in panel C; the number size distribution for the totalaerosol in panel D, and the residual aerosol in panel E. All data shownare five-minute averages and meet the criteria discussed in the text.67Figure 4.5: Average cloud droplet number size distributions for Cloud 2(panel A) and Cloud 3 (panel B) measured by the FM-100 (black cir-cles) and fit to a lognormal distribution function (black dashed lines).The average cloud droplet volume distributions (blue squares) and log-normal fits (blue dashed lines) are also shown for each cloud event.The CVI−D50 is indicated on each panel by a red dashed line.68Figure 4.6: Summary of the averaged number size distributions for Cloud 2(panels A and C) and Cloud 3 (panels B and D) for the total aerosol(red solid lines); residual aerosol (red dashed lines); total rBC as afunction of core diameter (black solid lines); and residual rBC as afunction of core diameter (black dashed lines). Both the aerosol andrBC for each cloud event are shown in two ways, a log scale (panelsA and B) to highlight the relative differences between the aerosol andrBC as well as normalized to the respective maximum value (panels Cand D) to highlight the shift in size distributions. All residual distri-butions have been corrected for CVI enhancement (see Section 4.3.1.3and droplet losses (see Appendix B).69Figure 4.7: Shown in panels A and B are the mean size dependent activatedfraction (AF) for the aerosol (red circles), and rBC (black triangles)for Clouds 2 and 3, respectively, where the error bars represent onestandard deviation (1σstd) of the mean AF . The bottom axes representparticle diameter for the aerosol and core diameter for rBC. Since thefraction of the cloud droplets sampled by the CVI was less than 100%,the calculated activated fractions should be considered as lower limitsto the total AF during the two cloud events. Shown in panels C andD are the averaged rBC coating thicknesses (blue circle) in nm with1σstd as the error bars. Since only 50% of the rBC containing particlesdetected with the SP2Res were successfully fit with the LEO fittingprocedure, the coating thicknesses shown in panels C and D are onlyfrom a subset (50%) of the rBC particles measured with the SP2Res.70Figure 4.8: Normalized number size distributions from Cloud 2 (panel A)and Cloud 3 (panel B) for residual rBC as a function of particle di-ameter [rBCRes(Core+Coating)]. The residual bulk aerosol distribu-tion (BulkAeroRes) and rBC distribution as a function of core diameter[rBCRes(Core)] are also shown for comparison. Since only 50% of therBC containing particles detected with the SP2Res were successfullyfit with the LEO fitting procedure, the rBCRes(Core+Coating) shownin panels A and B are only from a subset (50%) of the rBC particlesmeasured with the SP2Res.71Figure 4.9: Sub-micrometer non-refractory average aerosol mass fractionsfor Clouds 2 and 3 based on an ion-pairing scheme (see Section 4.4.8.1and Appendix A) and measured from a high resolution time-of-flightaerosol mass spectrometer.72Figure 4.10: Panel A shows the critical supersaturation (SC, black lines) asa function of particle dry diameter based on measured HR-ToF-AMSbulk compositions and an ion-pairing scheme. Panel B shows SC asa function of rBC core diameters with coating thicknesses rangingfrom 0-200 nm. In panel B, the coatings are assumed to have thesame composition as the bulk residual aerosol (Figure 4.9). The solidlines are for Cloud 2 and the dashed lines are for Cloud 3.73Table 4.1: Summary of cloud microphysical properties showing the average CVI cut-size (CVI−D50) wherethe uncertainty stems from the calculated cut-size (see Section 4.3.1.2 for details); average liquid watercontent (LWC) and one standard deviation; and the cloud droplet number (CDNCTot) and volume (VolTot)concentrations for droplets with diameters between 2-50 µm. Also shown is the number(CDNCSampCDNCTot)and volume(VolSampVolTot)fractions of droplets sampled, where CDNCSamp and VolSamp are the number andvolume concentrations, respectively, for the fraction of droplets sampled.Cloud Date CVI−D50 LWC CDNCTot VolTot# Sampled (µm) (g m-3) (cm-3)(CDNCSampCDNCTot)(µm3 m-3)(VolSampVolTot)12-13 June 2013 11.5 0.1322043-1135 PDT ±0.72 ±0.0767.67 0.38 1.24E5 0.9117-18 June 2013 11.6 0.0932036-0752 PDT ±0.72 ±0.02145.8 0.24 8.88E40.6874Table 4.2: Averaged number (N) and mass (M) concentrations, modal pa-rameters Dg and σg for aerosol and rBC particles during the two cloudevents measured at Mt. Soledad. The subscripts Tot and Res representmeasurements made from the total and residual inlets respectively.Cloud 2 Cloud 3Aerosol rBC Aerosol rBCNTot (cm-3) 980.8 75.24 994.0 62.13MTot (ng m-3) - 73.41 - 61.83Dg,Tot (nm) 107.7 <70 80.54 <70σg,Tot 1.577 - 1.703 -NRes (cm-3) 43.46 2.000 83.15 3.86MRes (ng m-3) - 2.741 - 4.735Dg,Res (nm) 331.9 87.30 269.2 80.72σg,Res 1.187 1.259 1.281 1.26875Chapter 5Size-Resolved Activation ofRefractory Black Carbon in CloudDroplets at a Canadian HighElevation Site5.1 IntroductionAs mentioned in the previous chapter, several studies have looked at the activatedfraction of BC mass in cloud droplets (Chy´lek et al., 1996; Cozic et al., 2007;Hitzenberger et al., 2000, 2001; Kasper-Giebl et al., 2000; Sellegri et al., 2003).Reported values of the average BC mass activated fractions range from 0.06 to0.8 (Cozic et al., 2007) (and references within), where lower values were typi-76cally seen at urban sites closer to anthropogenic emission sources, and the higheractivated fractions were seen at more remote higher elevation sites where it isthought that the air mass is aged to some extent providing the BC an opportunityto form a hydrophilic coating. However, not all data are consistent with this pic-ture. For example, Granat et al. (2010) has shown that BC was three times lessefficiently activated compared to sulfate, even after a travel time of several days,suggesting that BC containing particles from India have retained much of their hy-drophobicity after a travel time of several days. However, the conclusions statedin this study assumed that hygroscopicity controlled the removal of BC throughwet scavenging, when a more important factor is particle size. Furthermore, thewashout ratio, which is the metric used to interpolate scavenging efficiencies, wascomputed for BC mass and not number. Therefore, direct comparisons of theirwork and the work presented here are difficult to make.Even though there have now been several studies on the activated fractionof BC mass in cloud droplets in the natural environment, only one size-resolvedfield study (Schroder et al., 2014), to the best of our knowledge, has been carriedout (see Chapter 4). Additional studies on the activated fractions of BC in clouddroplets as a function of size are useful to test the current understanding of BCactivation in cloud droplets and the CCN properties of BC in the atmosphere.In this study we present data collected at a remote mountain site during theWhistler Aerosol and Cloud Study (WACS) that took place in 2010 (Macdonaldet al., 2014) (manuscript in preparation). The size-resolved activated fractions ofrBC were measured and compared to the size-resolved activated fractions of the77bulk aerosol during two cloud events.5.2 Sampling Site5.2.1 Site descriptionSampling took place during the Whistler Aerosol and Cloud Study (WACS) dur-ing 2010, and occurred at the Whistler High Elevation Site (WHI). The WHI islocated at the peak of Whistler mountain (50.06°N, 122.96°W) at an elevation of2182 m a.m.s.l.. This site is approximately 120 km north of Vancouver, BritishColumbia, its nearest major urban center, and sits above the Whistler Village (650m a.m.s.l.). A full description of the site and instruments present during this cam-paign can be found in the campaign overview paper (Macdonald et al., 2014).Briefly, there were two facilities used to house the instrumentation; a lift oper-ator‘s hut, which includes a year round monitoring station maintained by Envi-ronment Canada (Ahlm et al., 2013; Leaitch et al., 2011; Macdonald et al., 2011;McKendry et al., 2010; Takahama et al., 2011); and a shipping container that wasused only during the campaign. The cloud portion of the study took place from 1July to 28 July 2010.5.2.2 InletsThree inlets were used in this study (Figure 5.1), the total aerosol inlet and residualinlet were connected to instrumentation housed in the shipping container by a 3-way valve. During cloud-free periods of sampling the valve was switched so78that ambient air was sampled through the total aerosol inlet (Figure 5.1). Whenclouds were present the valve was manually switched so that droplet residualswere sampled through the residual inlet (Figure 5.1), which included a CVI as itsintake (see Section 4.3.1). The total aerosol, which is the sum of both the residualand interstitial particles, was sampled through the total aerosol inlet (Figure 5.1).Droplets sampled through the total inlet were dried by evaporation as they entereda temperature controlled room, held at approximately room temperature (20-25°C). The third inlet, labeled total rBC inlet in Figure 5.1, was the inlet connectedto the lift operator‘s hut and was used to measure the total rBC. Similar to the totalinlet, droplets sampled through the total rBC inlet were dried by evaporation asthey entered the temperature controlled room, which on average was at 23.5 °C.A CVI (see Section 4.3.1) sampled cloud droplets but excluded interstitialparticles based on the principle of inertial separation (Noone et al., 1988). Afterseparation from interstitial particles the droplets were dried by a warm dry coun-terflow as well as heated sections of the sample tubing, leaving only the residualparticles of the cloud droplets to be sampled by downstream instrumentation.The CVI −D50 for the two clouds sampled during this study were calcu-lated based on Equation 4.5 and were found to be approximately 10 µm for bothclouds. Sampling through the CVI enhanced the particle concentrations sampledwhen compared to ambient concentrations. Enhancement factors (EF) for the twoclouds sampled were calculated to be approximately 11 and 9 using Equation 4.7.To account for droplet losses during CVI sampling a droplet transmission (DT )factor was determined (see Appendix C). All residual distributions have been cor-79rected for both the enhancement of particles and droplet losses.5.3 Experimental5.3.1 Refractory black carbon measurementsTwo SP2s (Droplet Measurement Technologies, Boulder, CO) were used to mea-sure rBC (Figure 5.1). One SP2 was connected to the total rBC inlet and is referredto as the total SP2 (SP2Tot), and the second SP2 sampled from the residual inletand is referred to as the residual SP2 (SP2Res) (see Chapter 3 for a more thor-ough description and details on the SP2 theory of operation). Both the SP2Totand the SP2Res used in this study had a rBC size detection range of 70 to 210 nmVED. As discussed in Section 3.5 the SP2 can be used to measure the thicknessof any coating surrounding the core of a rBC particle if the elastic scattered lightfrom the particles is measured with a split detector. However, the split detector onthe SP2 connected to the residual inlet was not functioning throughout this study.Therefore, no coating thickness analysis was possible.5.3.2 Size distribution measurements of the bulk aerosolBulk aerosol size distributions (here bulk aerosol refers to all aerosol particles re-gardless of composition) from the total aerosol inlet were measured with a widerange particle spectrometer (Liu et al., 2010a) (WPS, model 1000XP, MSP, Shore-view, MN), which combines a DMA coupled to a CPC and an optical particlecounter (OPC) to provide particle counts for particles with diameters between8010 nm to 10 µm. For this analysis, particles with diameters of up to 500 nmwere taken from the DMA CPC counts and all diameters greater than 500 nmwere taken from the counts recorded by the OPC portion of the instrument. Sizedistributions of the residual bulk aerosol, sampled from the residual inlet, weremeasured with an ultra-high sensitivity aerosol spectrometer (UHSAS, DropletMeasurements Technologies, Boulder, CO). The UHSAS sampled particles be-tween 0.06 and 1 µm every 10 seconds. These measurements have been discussedin further detail by Pierce et al. (2012) and Macdonald et al. (2014).5.3.3 Back trajectoriesAir mass back trajectories were obtained using the NOAA HYSPLIT model (Draxlerand Rolph, 2012; Rolph, 2012). For the trajectory calculations the National Cen-ters for Environmental Predictions EDAS meteorological dataset was used. Alltrajectories were calculated at 10 m a.g.l., 24 hours backwards in time and athourly intervals for the entire duration of cloud sampling.5.3.4 Cloud propertiesA fog monitor (FM-100, model 100, Droplet Measurement Technologies, Boul-der, CO), which is a forward scattering optical spectrometer, was stationed on topof the container providing an in-situ measurement of the cloud droplet numberconcentration (CDNC) and the liquid water content (LWC) present. Droplets withdiameters between 2 and 50 µm were counted into 20 discrete size bins with aone-second time resolution (Eugster et al., 2006).815.4 Results and DiscussionTwo cloud events are analyzed in this study: the first cloud event took place on 2July 2010 10:48 to 12:02 PST and is referred to as Cloud 1, and the second cloudevent was sampled from 12 July 2010 07:38 to 11:15 PST and is referred to asCloud 2.5.4.1 Back trajectoriesThe 24 hour back trajectories that arrived at the onset of each cloud event as wellas at every subsequent hour until the end of the event are shown in Figure 5.2.According to Figure 5.2 panels A and B, the trajectories during Cloud 1 arrivedon-site roughly from the west, whereas the trajectories for Cloud 2 arrived at thepeak from the northwest. The trajectories calculated for Cloud 1 and 2 did not passover any heavily populated regions, suggesting little influence from recent freshurban emissions. The corresponding vertical profiles for each of the air massessampled are shown in Figure 5.2 panels C and D. In general, both air masses showsome level of vertical lifting prior to being sampled.Forest fires reported within the previous 24 hours by the Canadian WildlandFire Information System (CWFIS, vertical red triangle) as well as fires detectedby the Moderate Resolution Imaging Spectroradiometer (MODIS, horizontal redtriangle) are also plotted on Figure 5.2 panels A and B to indicate if the air massessampled may have been influenced by biomass burning aerosols. No fires werereported within the 24 hours prior to the sampling of Cloud 1, indicating therewas likely no influence from biomass burning aerosols immediately preceding this82event. However, several forest fires were reported within the 24 hours prior to thesampling of Cloud 2 and within roughly 500 km of the sampling site, suggestingthe possibility of some biomass burning influence. Trace gas measurements ofacetonitrile (Macdonald et al., 2014), a marker for biomass burning (Holzingeret al., 2005), show a sharp rise in the mixing ratio at the beginning of Cloud 2,further suggesting that the air mass sampled during Cloud 2 most likely had someinfluence from biomass burning aerosols.5.4.2 Meteorological conditions and cloud propertiesData were classified as in-cloud and included for analysis if the one-minute-averaged LWC was greater than 0.05 g m-3 and if the one-minute-averaged CVIcounterflow rate varied by less than ±5σstd of the mean counterflow rate. Thisfiltering of the data removed periods of unstable flows and regions of “patchy”clouds as much as possible (Cozic et al., 2007).Measured cloud properties as a function of time are shown in Figure 5.3 panelsA and B, where Cloud 1 is shown on the left side and Cloud 2 is shown the rightside of the plot. Panel A shows the LWC as a function of time for each of theclouds. Averaging the data shown in panel A gave LWC values of 0.07±0.04and 0.11±0.06 g m-3 for Clouds 1 and 2, respectively. Figure 5.3 panel B showsthe droplet number distributions as a function of time for each of the clouds withthe calculated CVI−D50 overlaid as a solid black line. The time-dependent datashown in Figure 5.3 panel B were used to calculate the average droplet numberdistributions for each cloud event (Figure 5.4). Integration of the lognormal fits83between 2 and 50 µm in Figure 5.4 gave cloud droplet number concentrations(CDNC) of 194 cm-3 and 258 cm-3 for Clouds 1 and 2, respectively. Applyingthe CVI−D50 to the droplet distributions shown in Figure 5.4 indicates that 28%and 41% of the droplet number were sampled during Clouds 1 and 2 respectively.Sampling only a fraction of the cloud droplet number distribution indicates thatthe results presented in this study apply only to the larger diameter cloud droplets.Averaged relevant properties of Clouds 1 and 2 are summarized in Table 5.1 forreference.5.4.3 Size distributionsAverage size distributions of the bulk aerosol and rBC particles, measured fromthe total and residual inlets for both clouds are presented in Figure 5.5. Dataare plotted in two ways; on a log scale (panels A and C), and normalized to therespective maximums (panels B and D). Table 5.2 summarizes the results of thesize distribution analysis.5.4.3.1 Size distributions measured from the total inlet (BulkAeroTot andrBCTot)The average size distribution of the bulk aerosol measured with the total aerosolinlet (referred to here as BulkAeroTot) and the average size distributions of rBCmeasured with the total rBC inlet (referred to here as rBCTot) are shown in Fig-ure 5.5. The BulkAeroTot size distributions (red solid lines in Figure 5.5) for bothclouds show two distinct modes with Dg, as determined by fitting a lognormalfunction to the data, of 78 and 47 nm for the smaller diameter mode and 148 and84137 nm for the larger diameter mode (Table 5.2) for Cloud 1 and 2 respectively.Integration of these number distributions from 10 nm to 1 µm resulted in totalnumber concentrations (NTot) of 736 and 376 cm-3 for Clouds 1 and 2, respec-tively. The size distributions of the BulkAeroTot as a function of time are alsoshown in Figure 5.3 panel C for comparison.The average rBCTot size distributions for each cloud are also shown in Fig-ure 5.5 (black solid lines). Assuming that the rBC distributions follow a lognor-mal function the Dg values are less than 70 nm, which is below the detection limitof the SP2s. Integration of the rBC number distributions, for core VEDs between70 and 210 nm, results in an NTot of 8.1 cm-3 during Cloud 1 and 1.6 cm-3 duringCloud 2. The total mass (MTot) of rBC was found to be 8.9 and 2.1 ng m-3 forClouds 1 and 2, respectively, assuming a density of σbc=1.8 g cm-3.5.4.3.2 Size distributions measured from the residual inlet (BulkAeroResand rBCRes)Average size distributions of the bulk aerosol measured from the residual inlet(referred to here as BulkAeroRes), are shown in Figure 5.5 (red dashed lines). Ap-plying lognormal fits to the BulkAeroRes yields Dg values of 122 and 128 nm forClouds 1 and 2, respectively. The Dg values found for the BulkAeroRes distribu-tions indicate that the majority of the particles activated into cloud droplets werefrom the larger diameter mode of the BulkAeroTot size distributions. Integrationof the BulkAeroRes size distributions between 60 nm and 1 µm gave measuredcloud residual number concentrations (NRes) of 91 and 77 cm-3 during Clouds 1and 2, respectively.85Average rBCRes distributions are shown in Figure 5.5 (black dashed lines),and, assuming these distributions follow a lognormal function, the Dg valueswould be below the detection limit of the SP2 (70 nm), similar to the rBCTot.Integration of these distributions, for rBC particles with core VEDs between 70and 210 nm, result in NRes values of 2 cm-3 and 0.3 cm-3, and residual massconcentrations (MRes) were found to be 2.1 and 0.4 ng m-3 for Clouds 1 and 2,respectively.5.4.4 Size-resolved activated fractionsThe bulk aerosol and rBC size-resolved activated fraction, AF(DP), was calcu-lated similar to Section 4.4.4 using Equation 4.8. In order to account for differ-ences in bin widths counted by the WPS and UHSAS a spline interpolation algo-rithm was applied to all distributions prior to calculating AF(DP). The CF(DP)in this study represents the size-resolved correction factor that accounts for dif-ferences in instrument sensitivity, which were calculated by taking the ratio ofthe total bulk aerosol number distribution measured from the WPS to the resid-ual number distribution measured from the UHSAS during periods of cloud-freesampling, when the two instruments should have been measuring the same con-centrations. The determination of DT in this study is discussed in Appendix C.Figure 5.6 shows the calculated AF(DP) for the bulk aerosol (red circles) andrBC (black triangles) for both clouds sampled, where AF(DP) of the bulk aerosolwas plotted as a function of particle diameter and the rBC size-resolved activatedfractions were plotted as a function of rBC core VED. The error bars shown in86Figure 5.6 are taken as 1σstd of the mean AF determined within each size bin. TheAF values show a dependence on size for both the bulk aerosol and rBC duringboth clouds. The AF of rBC ranged from approximately 0.2 at the smallest sizebin (70 to 80 nm) in Cloud 1 to 0.6 at the largest size bin of 200 to 210 nm duringCloud 2. Since the fraction of droplets sampled by the CVI was <100%, thecalculated AF should be considered as a lower limit to the total fraction activatedduring these two cloud events.The size-resolved activated fractions of rBC cores shown in Figure 5.6 alsoindicate that even the small 70 nm rBC cores are efficiently activated. Although,since rBC core diameter is not the only factor that needs to be considered in CCNability and the coating thicknesses could not be determined for these rBC cores,the extent to which the overall particle size influences the activation of these par-ticles could not be investigated. Since 70 nm cores (uncoated) are only expectedto be activated at relatively high cloud supersaturations (≈1%, see Figure 4.10), itis likely that the rBC cores here were coated with soluble material. During Cloud1 the rBC activated more efficiently than the bulk aerosol for all sizes. This ob-servation also strongly suggests that the rBC particles during Cloud 1 were coatedwith soluble material.5.5 Comparison with Previous MeasurementsThe study described in Chapter 4 is the only other study in which the size-resolvedactivation of rBC was measured. In that study, there was also a significant quantityof smaller (sub-100 nm) diameter rBC particles that were activated into cloud87droplets, a result that is consistent with the current study. The previous studyshowed that rBC cores with diameters less than 100 nm that were activated incloud droplets had thick coatings ranging from approximately 45 to 115 nm.Although there has only been one previous study to measure size resolved acti-vation of rBC in cloud droplets, there have been several previous studies that haveinvestigated the average BC mass activated fraction (Cozic et al., 2007; Hitzen-berger et al., 2000, 2001; Kasper-Giebl et al., 2000; Sellegri et al., 2003). How-ever, in these studies size-resolved information was not obtained, and only thetotal activated fraction of BC mass over the entire size range measured was re-ported. For comparison purposes, we calculated the average BC mass activatedfractions for both clouds measured at Whistler by taking the ratio of the rBC resid-ual mass (MRes) to the rBC total mass (MTot). Using this method, and values forMRes and MTot shown in Table 5.2, resulted in average BC mass activated frac-tions of 0.24 and 0.19 for Clouds 1 and 2 respectively. Since less than 100% ofthe droplet number distribution was sampled for both Clouds 1 and 2, these acti-vated fractions of rBC mass are likely lower limits to the true activated fractions.Another uncertainty, when determining the mass activated fractions, is the frac-tion of the total mass sampled by the SP2 instruments. Shown in Figure 5.7 arethe rBC mass distributions sampled from both the residual and total inlets. Byassuming that the mass distributions are lognormal, Figure 5.7 suggests that theSP2s captured most, but not all, of the rBC mass during both of the cloud events.In Table 5.3 we compare the activated fractions of rBC mass measured herewith the average activated fraction of BC mass determined at other remote moun-88tain sites. The average activated fractions of rBC mass for clouds measured atWhistler are smaller than the values reported at other remote mountain sites. Thedifferences may be due to the fact that less than 100% of the droplet number dis-tributions were sampled in the current study and therefore, the activated fractionsof rBC mass likely represent lower limits.As discussed above, the air mass for Cloud 2 was likely influenced by biomassburning particles. A few studies have shown that freshly emitted biomass burn-ing aerosols can have large diameters as well as reasonably high κ values makingthem fairly hygroscopic, and therefore efficient CCN (see Chapter 2) at or near thepoint of emission (Petters et al., 2009; Pratt et al., 2011). Andreae and Rosenfeld(2008) summarized previous studies of CCN properties of biomass burning parti-cles and concluded that, in general, the CCN efficiency of biomass burning parti-cles increases with aging. Combining these previously reported results, biomassburning aerosols, whether freshly emitted or aged, can act as efficient CCN. Thesize-resolved activated fractions measured here, for both bulk aerosol and rBC,appear to be consistent with the conclusions from these previous studies.5.6 Summary and ConclusionsThe size-resolved AF of rBC was measured for two clouds at the peak of WhistlerMountain during the WACS-2010 campaign. Back trajectories calculated for bothclouds showed that the air masses sampled did not travel over or near highly popu-lated areas within the previous 24 hours. The first cloud sampled had no influencefrom biomass burning aerosols within the previous 24 hours. However, 24 hours89prior to sampling Cloud 2, several forest fires were reported or detected withinthe vicinity of the sampling site. Elevated mixing ratios of acetonitrile were alsodetected at the beginning of the cloud, indicating that the air mass sampled duringCloud 2 likely had at least some influence from biomass burning aerosols.The fractions of bulk aerosol and rBC that activated both show a dependenceon size from the two clouds measured. The AF of rBC ranged from approximately0.2 at the smallest size bin (70 to 80 nm) to 0.6 at the largest size bin (200 to 210nm). Since the fraction of droplets sampled by the CVI was less than 100%, thecalculated AF should be considered as lower limits to the total fractions activatedduring these two cloud events. The size-resolved AF of rBC cores shown in Fig-ure 5.6 indicate that even the small 70 nm rBC cores are efficiently activated.Since 70 nm uncoated cores are only expected to be activated at relatively highcloud supersaturations, it is likely that the rBC cores were coated with soluble ma-terial. In addition, during Cloud 1 the fraction of rBC cores was more efficientlyactivated than the bulk aerosol for all diameters measured. This result stronglysuggests that large hygroscopic coatings surrounded these cores since, as seen inChapter 4, the overall rBC particle diameter likely is the dominant factor in de-termining the CCN ability. These results should be useful for testing models andparameterizations used to predict activation of BC in cloud droplets. For example,the size resolved activated fractions presented in this chapter could be used to testsimulated activated fractions at similar supersaturations and similar elevations.905.7 Chapter 5 Figures and TablesFigure 5.1: Schematic showing the configuration of the inlets and relevantinstrumentation used in this study.91Figure 5.2: HYSPLIT 24hr back trajectories. Panels A and C are trajecto-ries ending at hourly intervals for Cloud 1 (2 July 10:00 to 12:00 PST);panels B and D are trajectories ending at hourly intervals for Cloud 2(12 July 07:00 to 11:00 PST). All back trajectories started at 10 mabove ground level. Red vertical triangles are fires reported by theCanadian Wildland Fire Information System (CWFIS) and red hori-zontal triangles are fires detected from MODIS for all fires within 24hours prior to the start of the cloud. Panels C and D show the verticalprofiles over the same hourly intervals shown in panels A and B.92Figure 5.3: Time series data for Cloud 1 (left side) and Cloud 2 (right side)showing; liquid water content (LWC) in panel A; cloud droplet numbersize distributions with the CVI−D50 (black trace) overlaid in panel B;the number size distributions for the total bulk aerosol in panel C, theresidual bulk aerosol in panel D, and acetonitrile mixing ratio in panelE. All data shown in panels A to D are one-minute averages and meetthe criteria discussed in Section 5.4.2. Data in panel E are 15-minuteaverages.93Figure 5.4: Average cloud droplet number size distributions for Cloud 1 inpanel A and Cloud 2 in panel B (black circles) and fit to a lognormaldistribution function (black lines). The CVI−D50 is indicated in eachpanel by a red line.94Figure 5.5: Summary of the averaged number size distributions for Cloud1 (panels A and B), and Cloud 2 (panels C and D) for the total bulkaerosol (red solid lines); residual bulk aerosol (red dashed lines); totalrBC as a function of core diameter (black solid line); and residual rBCas a function of core diameter (black dashed lines). Both the aerosoland rBC for each cloud are shown in two ways, a log scale (panels Aand C) as well as normalized to the respective maximum values (panelsB and D). All residual distributions have been corrected for the CVIenhancement (see Section 4.3.1.3) and droplet losses (see AppendixC)95Figure 5.6: Mean size dependent activated fractions (AF) for the bulkaerosol (red circles) and rBC (black triangles) for Clouds 1 and 2 inpanels A and B, respectively. The error bars represent one standarddeviation of the mean activated fraction within each 10 nm bin. Thebottom axis represents particle diameter for the bulk aerosol and corediameter for rBC. Since the fraction of cloud droplets sampled by theCVI was less than 100%, the calculated activated fractions should beconsidered as lower limits to the total activated fraction.96Figure 5.7: Averaged rBC mass distributions as a function of rBC core diam-eters during Cloud 1 (panel A) and Cloud 2 (panel B) for the total rBC(rBCTot, red circles) and the residual rBC (rBCRes, black triangles).The solid lines are fits using a lognormal function.97Table 5.1: Summary of cloud microphysical properties showing the datesand times sampled; the average CVI cut-size (CVI−D50), where theuncertainty comes from the calculated cut-size; average liquid watercontent (LWC) and one standard deviation; total cloud droplet numberconcentration (CDNCTot); and the number fraction of droplets sampled(CDNCSamp/CDNCTot), where CDNCSamp is the droplet number con-centration greater than the CVI cut-size.Cloud Date CVI−D50 LWC CDNCTot# Sampled (µm) (g m-3) (cm-3)(CDNCSampCDNCTot)2 July 2010 9.67 0.0711048-1202 PST ±0.08 ±0.04194.14 0.2812 July 2010 9.63 0.1120738-1115 PST ±0.48 ±0.06258.36 0.41Table 5.2: Average number (N) and mass (M) concentrations, modal param-eters Dg and σg for bulk aerosol and rBC particles during Clouds 1and 2. The subscripts Tot and Res represents measurements made fromthe total and residual inlets respectively. All values reported for resid-ual particles have been corrected for the CVI enhancement (see Sec-tion 4.3.1.3) and droplet losses (see Appendix C)Cloud 1 Cloud 2Aerosol rBC Aerosol rBCNTot (cm-3) 736.3a 8.1c 375.6a 1.6cMTot (ng m-3) - 8.9c - 2.1cMode 1: 77.8 (1.4) Mode 1: 46.8 (1.3)Dg(σg)Tot (nm) Mode 2: 148.3 (1.4) <70 Mode 2: 137.1 (1.3) <70NRes (cm-3) 90.5b 2.0c 77.2b 0.3cMRes (ng m-3) - 2.1c - 0.4cDg(σg)Res (nm) 122.2 (1.2) <70 127.9 (1.2) <70a10 to 1000 nmb60 to 1000 nmc70 to 220 nm98Table 5.3: Summary of average black carbon mass activated fractions (AF) measured at other remote moun-tain locations (adapted from Cozic et al. (2007)), as well as from this study.Location AF Site Elevation Reference DatesPuy de Doˆme (France) 0.33 1465 m Sellegri et al. (2003) Feb. - Apr. 2001Rax (Austria) 0.54 1644 m Hitzenberger et al. (2001) Apr. 1999 & Mar. 20000.45 Kasper-Giebl et al. (2000) Sept. 1995Mt. Sonnblick (Switzerland)0.753106 mHitzenberger et al. (2000) Sept 1996 & Apr. - May 1997Junfraujoch (Switzerland) 0.61 3850 m Cozic et al. (2007) Jul. - Aug. 20040.24 2 July 2010Mt. Whistler0.192182 m This work12 July 201099Chapter 6Measurements of Refractory BlackCarbon at a High ElevationMountain Site during 2009, 2010,and 20126.1 IntroductionIn order to assess the role of BC particles in climate and health, knowledge ofthe properties of BC (e.g. size distribution, mass loading, and mixing state) inthe atmosphere is needed, and this information needs to be represented accuratelyin atmospheric models. In this study we present measurements of rBC at theWhistler High Elevation Site (WHI) in British Columbia, Canada. The data were100collected during the summertime of 2009 and 2010, as well as the spring of 2012.WHI is a remote station which is often in the free troposphere (FT). It sees littleinfluence from any nearby large urban centers (Macdonald et al., 2011), but inthe summer it can be impacted by local and regional biomass burning (Takahamaet al., 2011).A number of studies have measured BC mass concentrations in remote loca-tions. In the INTEX-B campaign a SP2 was used to measure rBC during flightsover the US Pacific Northwest (Dunlea et al., 2009). In the INTEX-B campaign,the sampled air was classified into several groups, one of which was the back-ground (free troposphere) aerosol. Mass concentrations in the FT were low rel-ative to air masses influenced by regional sources and averaged 90 ng m-3. Inthe HIAPER Pole-to-Pole Observations (HIPPO) project, five sets of flights wereconducted between 2009 and 2011 over the remote Pacific with an SP2 on board(Kipling et al., 2013; Schwarz et al., 2010b, 2013; Shen et al., 2014). In these,rBC mass concentrations ranged from 1-40 ng m-3 for latitudes of 20-60 °N andaltitudes up to 5km. Measurements of rBC mass concentration have also beenperformed at the remote Jungfraujoch station in Switzerland (3580 m a.m.s.l.).There rBC mass concentrations in late winter ranged from median values of 8.11to 24.08 ng m-3 (scaled by a factor of≈1.6 for STP) for FT background or bound-ary layer aerosol respectively (Liu et al., 2010b)In this study a seasonal and time-of-day approach is combined with back tra-jectory analysis to separate periods of FT sampling from the general dataset. Peri-ods of biomass burning influence are also identified using chemical speciation of101the total aerosol.6.2 Site, Sampling and Analysis6.2.1 Site descriptionSampling for these experiments also took place at WHI (see Section 5.2.1). SinceWhistler Mountain is the location of a large ski resort, the periodic arrival ofsnow-grooming machines, snowmobiles, and trucks near the sampling site resultsin very sharp spikes in the rBC mass and number concentrations. The data havebeen filtered to remove these spikes and, although it is believed that all of thesehave been removed from the dataset, the possibility that some lesser influences re-main means that the data presented represents an upper limit for mass and numberconcentration.6.2.2 Refractory black carbon mass measurementsFor this study two SP2s (see Chapter 3 for details on the SP2) were used during thethree collection periods . SP2-1 was used during 2009 and 2010 and SP2-2 wasused during 2012. Both SP2s were calibrated with Aquadag® using the procedurediscussed in Section 3.3. Using a BC density of σbc=1.8 g cm-3 VEDs were alsocalculated from the measured masses of ambient particles using Equation 3.1.One issue that arises with the SP2 is reduced detection efficiency at smallerrBC diameters (Laborde et al., 2012; Schwarz et al., 2010a). In this study SP2-1had 100% detection efficiency down to an rBC mass of 1.26 fg (110 nmVED) and102reached 50% detection efficiency at 0.41 fg (76 nm VED). SP2-2 had 100% de-tection efficiency down to an rBC mass of 0.68 fg (90 nm VED) and reached 50%detection efficiency at 0.26 fg (66 nm VED). Because of this we have used 0.41fg (76 nm VED) as a lower detection limit. At this mass the detection efficiencyof SP2-2 was approximately 80%. We have not attempted to correct for this issuesince the fraction of the total rBC mass arising from these small particles is notknown at all times, and because the shape of the detection efficiency curve canvary with particle morphology and mixing state (Laborde et al., 2012).6.2.3 Refractory black carbon coating thicknessmeasurementsrBC particles were analyzed to determine the thickness of any coating surroundingthe core. Section 3.5 describes, in detail, the procedure used to calculate thecoating thickness.In this work a coating analysis was done for two subsets of the full data record.These were particles measured during periods of biomass burning influence andparticles measured during periods of FT sampling.For the biomass burning periods, approximately 30% of the leading-edge fitsfailed due to the time dependent scattering signal being above 5% of the maximumlaser intensity at a time of zero. Another 5% of the particles were rejected fromthe coating analysis because the leading edge fit gave a result that was abovethe scattering detector saturation limit and therefore outside the calibration range.This means that the coating thicknesses were determined for only 65% of the rBC103particles detected.For the FT sampling periods, only 6% of the leading-edge fits failed due tothe time dependent scattering signal exceeding 5% of maximum laser intensity ata time of zero. Another 3% failed when the leading edge fit gave a result abovethe scattering detector saturation limit. In this case the mixing state analysis wasdone for 91% of the particles detected.6.2.4 Total aerosol measurementsIn order to identify periods of biomass burning, an Aerodyne aerosol chemicalspeciation Monitor (ACSM) (Ng et al., 2011) was used for chemical speciation ofthe total aerosol during the measurements in 2009 and 2012. The ACSM is part ofthe regular suite of instruments operating at the WHI site (Takahama et al., 2011)and is similar to the Aerodyne Aerosol Mass Spectrometer (Jayne et al., 2000),with the main differences being that it does not measure particle size and it hasreduced sensitivity.In 2010 the ACSM was offline and total aerosol speciation was performedwith a C-mode time-of-flight aerosol mass spectrometer (C-ToF-AMS, AerodyneResearch). The C-ToF-AMS measures aerodynamic diameter and provides highermass resolution and increased sensitivity relative to the ACSM.6.2.5 Back trajectoriesAir mass back-trajectories were calculated using the NOAA HYSPLIT model(Draxler and Rolph, 2012; Rolph, 2012) with the GDAS 1-degree meteorologi-104cal dataset. This has a 3-hour time resolution, a 1-degree horizontal resolution,and 23 vertical layers. The back trajectories were calculated at the elevation ofWhistler peak (2182 m a.m.s.l.) as well as at 200 m below the peak to account foruncertainties in the initial conditions.6.3 Results and Discussion6.3.1 Size distributionsThe range of rBC diameters measured in this study was 76-220 nmVED. It is pos-sible, however, to estimate what portion of the rBC mass is not being measured.If we assume that the mass distribution of rBC containing particles is lognormalwith a single mode we can fit the distributions measured by the SP2 and estimatewhat portion of the actual mass distribution is outside of our measurement range.This was done for two subsets of the full rBC record at WHI: all particles frombiomass burning periods, and all particles from FT periods (Figure 6.1). The frac-tion of the distribution being sampled was 55% and 60% for biomass burning andFT periods respectively. All mass concentrations reported here have been scaledby a factor of 1.8 to account for the fraction of mass not sampled.In Figure 6.1 the red trace shows the mass distribution for all rBC containingparticles collected during periods of biomass burning. The full distribution (fromthe fit) has a mass median diameter of 218 nm and σg of 1.6 (Table 6.1). This is inreasonable agreement with other SP2 measurements of biomass burning particles.Kondo et al. (2011) observed a mass mean diameter of 207±31 nm for biomass105burning plumes of Asian origin measured over Alaska and a mass mean diameterof 187±10 nm for biomass burning plumes a few hours old measured over north-ern Canada. Sahu et al. (2012) measured a mass mean diameter of 193±16 nm forfresh biomass burning plumes in California and Schwarz et al. (2008a) measureda mass mean diameter of 210 nm for fresh biomass burning plumes over Texas.The mass distribution for periods of FT sampling is also shown in Figure 6.1(blue trace). The full distribution (from the fit) has a mass median diameter of185 nm and a σg of 1.7 (Table 6.1). Liu et al. (2010b) measured a geometricmean diameter of 220-240 nm for FT rBC particles sampled at the Jungfraujochstation in late winter. They noted that the diameters measured during periods of FTsampling did not differ significantly between periods of FT and non-FT sampling.In addition, Metcalf et al. (2012) found a mass median diameter of 161±41 nmfrom rBC measurements made in the FT over Los Angeles, CA.6.3.2 Refractory black carbon measurements at WHIrBC mass concentrations measured at WHI, binned into 10 minute intervals, isshown in Figure 6.2. Data are available for June-August of 2009; June-July of2010; and April-May of 2012. Figure 6.3, panel A, shows a histogram of rBCmass concentrations for the full measurement record. Over the full measurementperiod, the median rBC concentration was aproximately 15 ng m-3 with 10th and90th percentile values of about 2 and 42 ng m-3 respectively (Table 6.1).1066.3.3 Identifying periods of biomass burning samplingPeriods where the sampled particulate matter was influenced by biomass burningare indicated by the shaded regions in Figure 6.2 and highlighted by the boxed re-gion in Figure 6.4. These were isolated from the general dataset using a procedurebased on that of Takahama et al. (2011). First, non-anthropogenic sources wereseparated from anthropogenic sources using the ratio of organic material (OM)to organic material plus sulfate aerosol (OM+SO4), as measured by the ACSMor C-ToF-AMS. Periods dominated by non-anthropogenic aerosols are character-ized by an OM/(OM+SO4) ratio of greater than 70%. Second, periods of biomassburning were then separated from biogenic sources by the increased concentration(> 5σstd of back ground concentrations) of levoclucosan fragments in the ACSMor C-ToF-AMS signal at m/z 57, 60, and 73 (see Figure 6.4 and Figure 6.5).Levoglucosan is a commonly used tracer for biomass burning aerosols (Simoneitet al., 1999; Takahama et al., 2011). Third, to ensure there were forest fires withinthe area, the number of forest fires reported by the CWFIS within 200 km of WHIwas also taken into account.6.3.3.1 Refractory black carbon mass concentrations during periods ofbiomass burningPeriods where the sampled particulate matter was produced by biomass burningare indicated by the shaded region in Figure 6.2. Higher rBC mass concentrationswere observed during biomass burning events than during any other period. Fig-ure 6.3 panel B shows a histogram of rBC mass concentrations for all periods of107biomass burning influence. These periods had a mean concentration of approxi-mately 124 ng m-3 and a median concentration of about 94 ng m-3 with 10th and90th percentile values of about 41 and 234 ng m-3 respectively (Table 6.1).6.3.3.2 Refractory black carbon coating thicknesses during periods ofbiomass burningFigure 6.5 panel A shows a 2D histogram of coating thickness as a function ofrBC core diameter for all rBC containing particles measured during the biomassburning period that occurred from July 26-28, 2010. A scattering calibration wasnot done for the 2009 data, so coating thicknesses could not be determined for the2009 biomass burning period.The SP2 scattering detectors limit the range for which coating thicknesses canbe measured. Small bare, or thinly coated, particles are below the threshold fordetection of scattered light, while large rBC particles with thick coatings are abovethe saturation limit of the instrument. These limits are evident in Figure 6.5 panelA, where the data are cut off in the lower left and upper right areas of the plot.We‘ve set the lower limits in Figure 6.5 panel A at 2σstd above the minimumdetectable scattering signal of the instrument to avoid the random uncertainty inthe scattering from any given particle resulting in a calculated scattering amplitudebelow our detection limits.A frequency distribution of coating thicknesses for rBC core particles withdiameters between 140-160 nm VED (see Section 6.2.3) measured during theperiod of biomass burning is shown in Figure 6.5 panel B. The average coatingthickness during this period was 55 nm with 10th and 90th percentile values of 3108and 100 nm respectively. This average coating thickness was within the range ofcoating thickness observed by others. For example, Laborde et al. (2013) foundan average coating thickness of biomass burning influenced rBC in Paris of 15nm for 200 nm rBC core VED, and Schwarz et al. (2008a) observed an averagecoating thickness of 65 nm for rBC core VED between 190 and 210 nm.6.3.4 Identifying periods of free tropospheric samplingMacdonald and colleagues considered a number of strategies for determiningwhen WHI is in the FT (Macdonald et al., 2011). They examined diurnal cyclesin ozone, CO, and water vapor as well as measurements of atmospheric stabilityin the air mass extending from the bottom of Whistler valley up to the peak. In allcases the data showed that in the nighttime (20:00 to 08:00 PST) Whistler Peak isalmost entirely free of boundary layer influences, a conclusion that is supportedby more detailed meteorological analysis of the area (Gallagher et al., 2011), andby approaches employed for other mountain-top sites (Andrews et al., 2011).Following their work, we have restricted our analysis for the FT to the 2012April-May nighttime data (20:00-08:00 PST) as a conservative approach to elimi-nating boundary layer influences. As a further measure to eliminate possible localor regional influences we have combined this seasonal and time-of-day approachwith analysis of air mass back trajectories generated with HYSPLIT (see Sec-tion 6.2.5). Ten day back trajectories were calculated for every hour from 20:00to 08:00 PST and analyzed so that air masses could be classified as non-FT or FT.In order to qualify as a FT air mass, free of valley influence, the back trajectory109had to show the air mass having spent less than 12 of its final 48 hours at pres-sures above 900hPa (about 1km in altitude) (Macdonald et al., 2011). Once thiscondition was met, the SP2 data collected from 30 minutes prior until 30 minutesafter that air masses arrival time was classified as FT.6.3.4.1 Refractory black carbon mass concentrations in the freetroposphereFigure 6.3, panel C, shows a histogram of measured rBC mass concentrations forperiods when Whistler was determined to be in the FT. These periods had a meanrBC mass concentration of about 14 ng m-3, and a median rBC mass concentrationof about 11 ng m-3 with 10th and 90th percentile values of around 1 and 46 ng m-3respectively (Table 6.1).Table 6.2 summarizes the rBC mass concentrations measured for the FT in thisstudy, and also gives values for some comparable measurements made at otherlocations. The Jungfraujoch high elevation site in Switzerland is located at 3580m a.m.s.l and like WHI, it is well removed from significant pollution sources(Cozic et al., 2007). Free troposphere rBC concentrations measured in late winterat Jungfraujoch are similar to those measured in April and May of 2012 at WHI.They are slightly lower than WHI for general FT periods (median values of about11 ng m-3 for WHI and about 8 ng m-3 for Jungfraujoch) and drop lower forperiods when the Jungfraujoch station is influenced by precipitation (median valueof around 3 ng m-3 for Jungfraujoch) (Liu et al., 2010b). rBC mass concentrationsmeasured in the FT over the US northwest coast during INTEX-B (average of 90ng m-3) are considerably higher than those measured in this study. In the HIPPO110campaigns, rBC mass concentrations in the central Pacific from 20°N-60°N (0-10km altitude) ranged from 0.1 to 40 ng m-3 depending on season and altitude.6.3.4.2 Refractory black carbon coating thicknesses in the free troposphereFigure 6.5 panel C shows a 2D histogram of coating thickness as a function of rBCcore diameter for all rBC containing particles measured during the FT samplingperiods in April-May of 2012. As in Figure 6.5 panel A, we‘ve set the lower limitsin Figure 6.5 panel C at +2σstd of the minimum detectable scattering signal of theinstrument.From the coating thickness frequency distribution of rBC particles with coreVEDs in the 140-160 nm range (see Figure 6.5 panel C), the average coatingthickness was 32 nm with 10th and 90th percentile values of 4 and 60 nm respec-tively. A study that reports on the rBC coating thicknesses measured in the FTwas a study conducted by Metcalf et al. (2012). This study measured the coatingthicknesses of rBC particles in the FT near the Los Angeles basin, and found thatthe mean coating thickness over LA was 188±31 nm for all rBC core diametersbetween 80-227 nm VED.6.4 Summary and ConclusionsA 3 year period of rBC measurements from a high elevation mountain site atWhistler, BC were investigated with a SP2. From the full 3 year rBC record, twosubsets of data were identified; 1) periods that were influenced by biomass burningand 2) periods of free tropospheric sampling. Air masses that were influenced111by biomass burning were identified by using measurements of aerosol chemicalspeciation, while free tropospheric periods were identified by using a seasonaland time-of-day approach. Each subset of data was analyzed to investigate thesize distributions, mass concentrations, and coating thicknesses of rBC containingparticles.Fitting lognormal functions to the rBC core mass size distributions resultedin median mass diameters of 208 nm for periods of biomass burning influence,a median mass diameter that was consistent with other measurements made ofbiomass burning influenced rBC. The median mass diameter found for periods offree tropospheric sampling was 192 nm. The mass size distributions during bothperiods of sampling show a similar spread in the distribution, where the width pa-rameter σg was found to be 1.6 and 1.7 for biomass burning and free troposphericsampling periods respectively.The median mass concentration found for periods of free tropospheric sam-pling was appoximately 11 ng m-3, which was significantly lower than free tro-pospheric air measured in the northwestern part of the US. However, the medianmass concentrations measured at the remote high elevation site at Whistler agreedto within 75% of mass concentrations measured at another remote mountain highelevation site (Jungfraujoch).Coating thicknesses of rBC containing particles during the two subsets of datawere investigated, and frequency distributions of coating thickness were investi-gated for rBC particles with core diameters between 140 and 160 nm. The coat-ing frequency distribution for the biomass burning period had a median coating112thickness of about 55 nm and was quite broad in shape where the 10th and 90thpercentiles were 5 and 106 nm respectively. The coating thickness frequency dis-tribution during free tropospheric sampling had a median thickness of nearly halfthat found for the biomass burning period, at approximately 32 nm, and was muchnarrower in shape with 10th and 90th percentiles of 6 and 63 nm respectively.The results found from this study contribute to the body of evidence for mea-surements of refractory black carbon at remote mountain sites and under differentsampling conditions. Adding to this body of evidence aids in increasing our un-derstand of how black carbon properties (e.g. size distributions, mass concentra-tions, and coating thicknesses) vary as a result of location as well as the type ofair mass. Increasing our understanding of these rBC properties can then be usedto further constrain computer models that are used to predict how rBC can affectclimate. For example, Wang et al. (2014) used atmospheric BC data collected overthe ocean and in the free troposphere to constrain the application of a commonlyused global climate model (GEOS-Chem) coupled to a radiative transfer modelto infer a direct radiative forcing of BC to be 0.19 W m-2, with an uncertaintyrange of 0.17-0.31 W m-2. They found that this large range of uncertainty wasprimarily a result of the uncertainty in BC distribution in the atmosphere. Fur-thermore, the constrained estimate of 0.19 W m-2 they computed was lower (by asmuch as 21%) than results of other modeled estimates reported in the literature,which was primarily driven by BC concentrations over the oceans and in the freetroposphere. Data from the study presented in this chapter could also be appliedin a similar manner as was done in Wang et al. (2014) to constrain direct radiative113forcing estimates at similar locations as Mt. Whistler.1146.5 Chapter 6 Figures and TablesFigure 6.1: Normalized mass distributions for periods of free tropospheresampling and biomass burning. Lines are single lognormal fits.115Figure 6.2: Measured rBC mass concentrations for 2009, 2010, and 2012with periods of biomass burning marked by red shaded areas.116Figure 6.3: Histograms of measured rBC mass concentration at WHI for thefull record (panel A), periods of biomass burning (panel B), and peri-ods of free troposphere sampling (panel C). All concentrations are 10minute averages.117Figure 6.4: Identification of biomass burning periods in 2009 (left side) and2010 (right side). Panel A Shows rBC mass concentration as measuredby the SP2 (binned into 10 minute intervals); panel B shows the ratio oforganic material (OM) to OM plus sulfate (SO4), where a ratio of > 0.7is considered to be from non-anthropogenic sources; panel C shows theraw ion signals of the levoglucosan fragments at m/z 57, 60, and 73as measured by the ACSM in 2009 and the C-ToF-AMS in 2010. Ionsignals that were greater than 5σstd were considered to be influencedby biomass burning. Panel D shows the number of fires reported byday within 200 km of Whistler Mountain (CWFIS). The red boxesindicate the portion of the data included from biomass burning.118Figure 6.5: 2D histograms of coating thicknesses and core diameters for allrBC containing particles measured during the biomass burning periodfrom July 26-28, 2010 (panel A), and the period of free tropospheresampling in April-May 2012 (panel C). Only particles with detectablescattering signals are shown in panels A and C. Coating thickness fre-quency distributions for rBC cores with VED from 140-160 nm (rep-resented on panels A and C as black dashed lines) are shown in panel Bfor biomass burning periods, and panel D for free troposphere periods.119Table 6.1: Summary of basic statistics for rBC mass concentrations and mass distributions measured at WHIfor the full record period, periods of biomass burning (BB), and periods of free troposphere sampling(FT).10th Percentile 50th Percentile 90th Percentile Mean MMDa[Mass] ng m-3 [Mass] ng m-3 [Mass] ng m-3 [Mass] ng m-3 (σg) nmFull record 2.31 14.99 43.05 23.20 -BB periods 40.92 94.03 234.09 123.91 218 (1.6)FT periods 1.05 10.57 32.63 14.16 185 (1.7)aMass median diameter120Table 6.2: Comparison of rBC mass concentrations measured in the free troposphere from this study andseveral other locations.Dataset 10th Percentile 50th Percentile 90th Percentile Mean Reference[Mass] ng m-3 [Mass] ng m-3 [Mass] ng m-3 [Mass] ng m-3WHI FT 1.05 10.57 32.63 14.16 This workJungfraujoch FT 0.62 8.11 27.44 13.24 Liu et al. (2010b)Jungfraujoch FTa 0.45 2.68 14.31 6.23 Liu et al. (2010b)Northwest Coast (USA) FT - - - 90 Dunlea et al. (2009)Central Pacificb - 0.1-40 - - Wang et al. (2014)aInfluenced by precipitationbResults are from 20°N to 60°N, 0-10 km, and depend on the season.121Chapter 7Conclusions7.1 Activation of Refractory Black Carbon intoLiquid Water Cloud DropletsRefractory black carbon in liquid water cloud residuals was measured duringcloud events at a marine boundary layer site in La Jolla, CA, and at the top ofWhistler Mountain in Whistler, BC.In Chapter 4 two liquid water clouds were sampled at La Jolla, CA, and basedon the calculated back trajectories, bulk aerosol number concentrations, and rBCmass concentrations, the air masses were classified as polluted marine air. Num-ber size distributions were determined for both clouds, and it was found that thesize distributions of both the bulk aerosol and rBC measured in cloud residualsdisplayed a shift towards larger diameters when compared to the respective mea-surements made from the total inlet. The number size distributions of rBC resid-122uals also showed that rBC particles with small (< 100 nm) core diameters can beincorporated into cloud droplets and therefore contribute to the CCN population.The coating analysis of the SP2 data showed that the smaller rBC cores that wereactivated into cloud droplets had thick coatings, with average coating thicknessesof ≈75 nm at core diameters between 70-80 nm compared to ≈29 nm coatingsfor core diameters between 200-210 nm. These results suggest that for larger di-ameter rBC particles only a modest coating (≈ 25 nm) of hygroscopic material isneeded to become CCN active at relatively low critical supersaturations ( 0.05%).Furthermore, incorporating the coating into the rBC size distributions resulted ingreater similarities with the size distributions of the residual bulk aerosol mea-sured at the marine boundary layer site.Size-resolved activated fractions of rBC were determined and compared tosize-resolved activated fractions of the bulk aerosol, and it was found that duringboth clouds sampled, the activated fractions for rBC cores were larger than theactivated fractions for the bulk aerosol at diameters <≈150 nm. These resultswere explained by the presence of thick coatings surrounding the smaller rBCcores. In addition, the activated fractions of rBC particles with core diameters <100 nm were significant, where the activated fractions of the smallest diametersmeasured (70-80 nm), were found to be 0.01 and 0.05 for Cloud 2 and Cloud 3respectively. Since the fraction of droplets sampled by the CVI was <100%, thecalculated activated fractions found during this study should be considered lowerlimits to the total fractions activated.The presence of rBC core particles with diameters < 100 nm seen in the obser-123vations made at La Jolla were quantitatively validated using kappa-Ko¨hler theory.Based on the bulk aerosol chemical compositions and estimations of the criticalactivation diameters of the bulk aerosol, the cloud critical supersaturations werefound to be approximately 0.05% during the two clouds sampled in this study.Using the estimated critical supersaturation of the clouds sampled and applyingkappa-Ko¨hler theory to rBC particles with different coating thickness showed thatthe critical diameter for activation of rBC cores was <50 nm for all cores withcoating thicknesses ≥100 nm. This quantitative prediction using kappa-Ko¨hlertheory was consistent with the observations of rBC cores with diameters < 100nm being activated into the sampled cloud droplets, if the measured coating thick-nesses are factored in.Measurements of rBC in liquid water cloud droplets were also measured dur-ing two clouds at the peak of Whistler Mountain and presented in Chapter 5. Backtrajectory analysis showed that neither of the two clouds sampled had travelednear, or over any highly populated regions within 24 hours prior to being sam-pled. Combining calculated back trajectories with forest fire locations, reportedor detected by satellite, indicated that the first cloud sampled had no influencefrom biomass burning aerosols. However, it was concluded that the second cloudsampled most likely had some influence by biomass burning aerosols since therewere several forest fires located in the vicinity of the back trajectories. High levelsof the biomass burning marker, acetonitrile, further confirmed this conclusion.The number size distributions of the residual bulk aerosol measured at WhistlerMountain showed that the majority of the activated aerosol came from the larger124diameter mode of the total bulk aerosol. However, unlike the residual rBC mea-sured in La Jolla, the maximum of the residual rBC measured at Whistler Moun-tain was below the detection limit (≈70 nm) of the SP2. By assuming the rBCresidual distribution is lognormal, the mean diameter would therefore be < 70nm. This result is consistent with the observations of smaller rBC core diameterparticles being activated into liquid water droplets at Whistler Mountain.The average rBC mass fractions activated during the two clouds sampled atWhistler were found to be 0.24 and 0.19 for Clouds 1 and 2 respectively. These av-erage rBC mass fractions activated were lower than those values reported at otherremote mountain sites. The differences may have been due to the fact that lessthan 100% of the droplet number distributions were sampled during the Whistlerstudy and therefore, the rBC mass fractions activated were likely lower limits.Finally, the size-resolved activated fractions of both the bulk aerosol and rBCobserved at Whistler also showed a size dependence for the two clouds sampledat this remote mountain site. The activated fractions of rBC ranged from approx-imately 0.2 at the smallest size bin (70 to 80 nm) to 0.6 at the largest size bin(200 to 210 nm). Since the fraction of droplets sampled by the CVI during thisstudy was also less than 100%, the calculated activated fractions measured here,should also be considered lower limits to the total fractions activated. The resultsfrom the size-resolved activated fractions calculated from Whistler also showedthat smaller diameter (≈ 70 nm) rBC cores were significantly incorporated intothe droplets measured at this site. Since 70 nm uncoated cores are only expectedto be activated at relatively high cloud supersaturations (≈0.5%, see Figure 4.10),125it is likely that the rBC cores were coated with soluble material. In addition, dur-ing the first cloud sampled the rBC activated more efficiently than the bulk aerosolover all sizes measured, which strongly suggests the presence of hydrophilic coat-ings surrounded the rBC cores making the overall particle diameter larger andtherefore more easily activated.7.2 Refractory Black Carbon Properties DuringBiomass Burning and Free TroposphereSamplingIn Chapter 6, rBC measurements from the summer of 2009 and 2010 as well asthe spring of 2012 at a high elevation mountain site at Whistler were presented.From the three sampling periods, two subsets of data were identified: 1) periodsthat were influenced by biomass burning and 2) periods of free tropospheric sam-pling. Each subset of data was analyzed to investigate the size distributions, massconcentrations, and coating thicknesses of rBC containing particles.Fitting of the mass size distributions resulted in median mass diameters of 208nm for periods of biomass burning influence, and 192 nm for periods of free tro-posphere sampling. Both of these mass median diameters measured at the WHIsite were consistent with previously reported values. Median mass concentrationswere also investigated and found to be approximately 94 ng m-3 during periodsof biomass burning influence, and 11 ng m-3 during free troposphere sampling(scaled by a factor of 1.72 for STP). The free tropospheric rBC mass concentra-tions measured at WHI were significantly lower than similar measurements in the126northwestern part of the US, but agreed to within 65% of mass concentrationsmeasured at the remote mountain high elevation site on top of the Jungfraujoch.Based on the histograms of rBC coating thicknesses for rBC particles withcore diameters between 140-160 nm the average coating thickness during periodsof biomass burning (55 nm) was nearly two times thicker than the average coatingthickness measured during periods of free troposphere sampling (32 nm). A com-parison of the average coating thickness for all rBC core particles with diametersfrom 80-200 nm measured at WHI with another study that measured coated rBCparticles in the free troposphere over the Los Angeles basin showed that the rBCmeasured in the free troposphere at WHI had significantly thinner coatings.Figure 5 shown in Bond et al. (2013) shows the modelled SC of BC from bothdiesel and gasoline exhaust as a function of both particle diameter and BC massfraction. In general this figure shows that the overall particle diameter has a greatereffect on activation than the mass fraction of BC present in a mixed particle, at aconstant SC. More specifically, Figure 5 shows, at a constant SC of ≈0.05% theincrease in mass fraction of BC is negligible for particle diameters greater thanapproximately 200 nm. Although the source of rBC measured during the studiespresented in both Chapters 4 and 5 are most likely not pure engine exhaust, theresults presented in these chapters are consistent with the results of the modelshown in Figure 5 of Bond et al. (2013). The results presented in Chapter 4 areconsistent with these results as even for the smallest core diameters measuredat the marine boundary layer site the overall particle diameters were larger than200 nm, on average, when the coating thickness were taken into account. Figure 5127further shows that for small diameter BC (≈70-100 nm) at 70% BC mass fractionswould require very large (>1%) SC are required. Therefore, the conclusion inChapter 5 that the rBC cores activated at the high elevation mountain site had largecoatings surrounding the cores is supported by Figure 5 in Bond et al. (2013) sincethe supersaturations experienced at Whistler were likely much lower than 1%.7.3 Considerations for Future WorkBased on the high degree of uncertainty (≈ 90%) associated with current best-estimates on the radiative forcing of BC due to indirect effects, it is clear that morestudies similar to the ones presented in Chapters 4 and 5 are needed to evaluateand constrain model simulations. Although the research presented in these chap-ters does add to the body of knowledge on the properties of BC that drive CCNactivation, there were some limitations to the conclusions that could be drawn.Since only larger (>≈ 10-11 µm) droplets were investigated during these studies,the question remains of whether BC particles are activated into smaller dropletsin a similar manner to larger droplets. Future field studies with a CVI that isoptimized to sample smaller droplets would help to answer this question.Chapters 4 and 5 also showed that smaller (<≈ 100 nm) rBC particles can actas CCN when large hygroscopic coatings are present. A number of studies (e.g.,Metcalf et al., 2012; Rose et al., 2006; Schwarz et al., 2008a) have shown that asignificant fraction of the atmospheric BC number distributions are from smaller(< ≈ 100 nm) diameter particles and, if coated, these particle may substantiallyinfluence the CCN population. In the quantitative coating analysis presented in128Chapter 4, an assumption was made that the composition of the coating was iden-tical to that of the bulk aerosol for all rBC core diameters. It would be worthwhileto investigate the chemical compositions of coatings on activated BC particlesto test if this assumption is valid over all BC diameters. Such an investigationcould also increase our knowledge of what properties drive the CCN ability of at-mospheric BC. Furthermore, because such large uncertainties are associated withmodel estimates of the contribution of BC to the indirect effects on climate exist,it would be informative to design a field study that measures the exact componentsneeded by a specific model to simulate a cloud event. This type of study would al-low for direct comparison of ambient observations with model simulation, whichcould provide useful information on necessary model constraints and directionfor future ambient studies. It would also be informative to conduct similar studiesdiscussed in Chapters 4 and 5 at other sites and for different cloud types to betterunderstand the global role that clouds play in processing rBC.129BibliographyL. Ahlm, K. M. Shakya, L. M. Russell, J. C. Schroder, J. P. S. Wong, S. J.Sjostedt, K. L. Hayden, J. Liggio, J. J. B. Wentzell, H. A. Wiebe, C. Mihele,W. R. Leaitch, and A. M. Macdonald. Temperature-dependent accumulationmode particle and cloud nuclei concentrations from biogenic sources duringWACS 2010. Atmos. Chem. Phys., 13(6):3393–3407, Mar. 2013.doi:10.5194/acp-13-3393-2013. URLhttp://www.atmos-chem-phys.net/13/3393/2013/http://www.atmos-chem-phys.net/13/3393/2013/acp-13-3393-2013.pdf. → pages78J. D. Allan, A. E. Delia, H. Coe, K. N. Bower, M. Alfarra, J. L. Jimenez, A. M.Middlebrook, F. Drewnick, T. B. Onasch, M. R. Canagaratna, J. T. Jayne, andD. R. Worsnop. A generalised method for the extraction of chemicallyresolved mass spectra from Aerodyne aerosol mass spectrometer data. Journalof Aerosol Science, 35(7):909–922, July 2004.doi:10.1016/j.jaerosci.2004.02.007. → pages 47T. L. Anderson, R. J. Charlson, and D. S. Covert. Calibration of a CounterflowVirtual Impactor at Aerodynamic Diameters from 1 to 15 m. Aerosol Scienceand Technology, 19(3):317–329, 1993. → pages 41, 42M. Andreae. Correlation between cloud condensation nuclei concentration andaerosol optical thickness in remote and polluted regions. AtmosphericChemistry and Physics, 9(2):543–556, 2009. → pages 51M. Andreae and D. Rosenfeld. Aerosol cloud precipitation interactions. Part 1.The nature and sources of cloud-active aerosols. Earth-Science Reviews, 89(1-2):13–41, July 2008. doi:10.1016/j.earscirev.2008.03.001. → pages 59, 89E. Andrews, J. A. Ogren, P. Bonasoni, A. Marinoni, E. Cuevas, S. Rodrı´guez,130J. Y. Sun, D. A. Jaffe, E. V. Fischer, U. Baltensperger, E. Weingartner, M. C.Coen, S. Sharma, A. M. Macdonald, W. R. Leaitch, N.-H. Lin, P. Laj,T. Arsov, I. Kalapov, A. Jefferson, and P. Sheridan. Climatology of aerosolradiative properties in the free troposphere. Atmospheric Research, 102(4):365–393, 2011. doi:10.1016/j.atmosres.2011.08.017. → pages 109T. S. Bates, P. K. Quinn, D. S. Covert, D. J. Coffman, J. E. Johnson, andA. Wiedensohler. Aerosol physical properties and processes in the lowermarine boundary layer: a comparison of shipboard sub-micron data fromACE-1 and ACE-2. Tellus B, 52B:258–272, 2000. → pages 51T. S. Bates, Coffman D. J., D. S. Covert, and P. K. Quinn. Regional marineboundary layer aerosol size distributions in the Indian, Atlantic, and PacificOceans: A comparison of INDOEX measurements with ACE-1, ACE-2, andAerosols99. Journal of Geophysical Research, 107(D19):8026, 2002.doi:10.1029/2001JD001174. → pages 39T. C. Bond and R. W. Bergstrom. Light Absorption by Carbonaceous Particles:An Investigative Review. Aerosol Science and Technology, 40(1):27–67, Jan.2006. doi:10.1080/02786820500421521. → pages 24, 29, 44T. C. Bond, D. G. Streets, K. F. Yarber, S. M. Nelson, J. H. Woo, and Z. Klimont.A technology-based global inventory of black and organic carbon emissionsfrom combustion. Journal of Geophysical Research, 109(D14203):1–43,2004. doi:10.1029/2003JD003697. → pages 5T. C. Bond, S. J. Doherty, D. W. Fahey, P. M. Forster, T. Berntsen, B. J.DeAngelo, M. G. Flanner, S. Ghan, B. Ka¨rcher, D. Koch, S. Kinne, Y. Kondo,P. K. Quinn, M. C. Sarofim, M. G. Schultz, M. Schulz, C. Venkataraman,H. Zhang, S. Zhang, N. Bellouin, S. K. Guttikunda, P. K. Hopke, M. Z.Jacobson, J. W. Kaiser, Z. Klimont, U. Lohmann, J. P. Schwarz, D. Shindell,T. Storelvmo, S. G. Warren, and C. S. Zender. Bounding the role of blackcarbon in the climate system: A scientific assessment. Journal of GeophysicalResearch: Atmospheres, 118(11):5380–5552, June 2013.doi:10.1002/jgrd.50171. → pages xvi, 4, 5, 8, 9, 12, 35, 36, 127, 128R. Y.-W. Chang, J. G. Slowik, N. C. Shantz, A. Vlasenko, J. Liggio, S. J.Sjostedt, W. R. Leaitch, and J. P. D. Abbatt. The hygroscopicity parameter (κ)of ambient organic aerosol at a field site subject to biogenic and anthropogenicinfluences: relationship to degree of aerosol oxidation. Atmospheric Chemistry131and Physics, 10(11):5047–5064, June 2010. doi:10.5194/acp-10-5047-2010.→ pages 16, 60W. Chen, Y. H. Lee, P. J. Adams, A. Nenes, and J. H. Seinfeld. Will black carbonmitigation dampen aerosol indirect forcing? Geophysical Research Letters, 37(L09801):1–5, May 2010. doi:10.1029/2010GL042886. → pages 38J. Ching, N. Riemer, and M. West. Impacts of black carbon mixing state on blackcarbon nucleation scavenging: Insights from a particle-resolved model.Journal of Geophysical Research, 117(D23209):1–21, Dec. 2012.doi:10.1029/2012JD018269. → pages 36P. Chy´lek, C. M. Banic, B. Johnson, P. A. Damiano, G. A. Isaac, W. R. Leaitch,P. S. K. Liu, F. S. Boudala, B. Winter, and D. Ngo. Black carbon: Atmosphericconcentrations and cloud water content measurements over southern NovaScotia. Journal of Geophysical Research, 101(D22):29105–29110, 1996.doi:10.1029/95JD03433. → pages 76W. Cooke, S. Jennings, and T. Spain. Black carbon measurements at Mace Head,19891996. Journal of Geophysical Research, 102:1989–1996, 1997.doi:10.1029/97JD01430. → pages 52J. Cozic, B. Verheggen, S. Mertes, P. Connolly, K. Bower, A. Petzold,U. Baltensperger, and E. Weingartner. Scavenging of black carbon in mixedphase clouds at the high alpine site Jungfraujoch. Atmospheric Chemistry andPhysics, 7(7):1797–1807, 2007. → pages xviii, 36, 37, 49, 76, 83, 88, 99, 110P. F. DeCarlo, J. R. Kimmel, A. Trimborn, N. M. J., J. T. Jayne, A. C. Aiken,M. Gonin, K. Fuhrer, T. Horvath, K. S. Docherty, D. R. Worsnop, and J. L.Jimenez. Field-deployable, high-resolution, time-of-flight aerosol massspectrometer. Analytical Chemistry, 78(24):8281–8289, 2006.doi:10.1021/ac061249n. → pages 46R. Draxler and G. Rolph. HYSPLIT (HYbrid Single-Particle LangrangianIntegrated Trajectory) Model access via NOAA ARL READY Website .NOAA Air Resources Laboratory, Silver Spring, MD., 2012. URLhttp://ready.arl.noaa.gov/HYSPLIT.php. → pages 81, 104R. R. Draxler and G. D. Rolph. HYSPLIT (HYbrid Single-Particle LangrangianIntegrated Trajectory) Model access via NOAA ARL READY Website .132NOAA Air Resources Laboratory, Silver Spring, MD., 2013. URLhttp://ready.arl.noaa.gov/HYSPLIT.php. → pages 47E. J. Dunlea, P. F. DeCarlo, A. C. Aiken, J. R. Kimmel, R. E. Peltier, R. J. Weber,J. Tomlinson, D. R. Collins, Y. Shinozuka, C. S. McNaughton, S. G. Howell,A. D. Clarke, L. K. Emmons, E. C. Apel, G. G. Pfister, A. van Donkelaar,R. V. Martin, D. B. Millet, C. L. Heald, and J. L. Jimenez. Evolution of Asianaerosols during transpacific transport in INTEX-B. Atmospheric Chemistryand Physics, 9(19):7257–7287, Oct. 2009. doi:10.5194/acp-9-7257-2009. →pages 101, 121U. Dusek, G. P. Frank, A. Massling, K. Zeromskiene, Y. Iinuma, O. Schmid,G. Helas, T. Hennig, A. Wiedensohler, and M. O. Andreae. Water uptake bybiomass burning aerosol at sub- and supersaturated conditions: closure studiesand implications for the role of organics. Atmospheric Chemistry and Physics,11(18):9519–9532, Sept. 2011. doi:10.5194/acp-11-9519-2011. → pages 16,36W. Eugster, R. Burkard, F. Holwerda, F. N. Scatena, and L. Bruijnzeel.Characteristics of fog and fogwater fluxes in a Puerto Rican elfin cloud forest.Agricultural and Forest Meteorology, 139(3):288–306, Oct. 2006. → pages48, 81L. Fierce, N. Riemer, and T. C. Bond. When is cloud condensation nuclei activitysensitive to particle characteristics at emission? Journal of GeophysicalResearch: Atmospheres, 118(24):13,476–13,488, Dec. 2013. ISSN 2169897X.doi:10.1002/2013JD020608. URLhttp://doi.wiley.com/10.1002/2013JD020608. → pages 56B. J. Finlayson-Pitts and J. N. J. Pitts. Chemistry of the Upper and LowerAtmosphere. Academic Press, San Diego, 2000. ISBN 0-12-257060-x. →pages 2J. P. Gallagher, I. G. McKendry, A. M. Macdonald, and W. R. Leaitch. Seasonaland Diurnal Variations in Aerosol Concentration on Whistler Mountain:Boundary Layer Influence and Synoptic-Scale Controls. Journal of AppliedMeteorology and Climatology, 50(11):2210–2222, Nov. 2011.doi:10.1175/JAMC-D-11-028.1. → pages 109133R. S. Gao, J. P. Schwarz, K. K. Kelly, D. W. Fahey, L. a. Watts, T. L. Thompson,J. R. Spackman, J. G. Slowik, E. S. Cross, J. H. Han, P. Davidovits, T. B.Onasch, and D. R. Worsnop. A Novel Method for Estimating Light-ScatteringProperties of Soot Aerosols Using a Modified Single-Particle SootPhotometer. Aerosol Science and Technology, 41(2):125–135, 2007.doi:10.1080/02786820601118398. → pages 25, 26, 27, 45P. Glantz and K. J. Noone. A physically-based algorithm for estimating therelationship between aerosol mass and cloud droplet number. Tellus B, 52B:1216–1231, Nov. 2000. → pages 51L. Granat, J. E. Engstro¨m, S. Praveen, and H. Rodhe. Light absorbing material(soot) in rainwater and in aerosol particles in the Maldives. Journal ofGeophysical Research, 115(D16307):1–12, Aug. 2010.doi:10.1029/2009JD13768. → pages 37, 77M. Gysel and J. Crosier. Closure study between chemical composition andhygroscopic growth of aerosol particles during TORCH2. AtmosphericChemistry and Physics, 7(24):6131–6144, 2007. → pages 23, 57, 149A. Hallberg, J. A. Ogren, K. J. Noone, J. Heintzenberg, A. Berner, I. Solly,C. Kruisz, G. Reischl, S. Fuzzi, M. C. Facchini, H. C. Hansson,A. Wiedensohler, and I. B. Svenningsson. Phase partitioning for differentaerosol species in fog. Tellus B, 44B(5):545–555, 1992. → pages 36, 37A. Hallberg, J. A. Ogren, K. J. Noone, K. Okada, J. Heintzenberg, and I. B.Svenningsson. The influence of aerosol particle composition on cloud dropletformation. Journal of Atmospheric Chemistry, 19:153–171, 1994. → pages36, 37, 54L. N. Hawkins, L. M. Russell, D. S. Covert, P. K. Quinn, and T. S. Bates.Carboxylic acids, sulfates, and organosulfates in processed continental organicaerosol over the southeast Pacific Ocean during VOCALS-REx 2008. Journalof Geophysical Research, 115(D13201), July 2010.doi:10.1029/2009JD013276. → pages 51S. Henning, H. Wex, T. Hennig, A. Kiselev, J. R. Snider, D. Rose, U. Dusek,G. P. Frank, U. Po¨schl, A. Kristensson, M. Bilde, R. Tillmann,A. Kiendler-Scharr, T. F. Mentel, S. Walter, J. Schneider, C. Wennrich, andF. Stratmann. Soluble mass, hygroscopic growth, and droplet activation of134coated soot particles during LACIS Experiment in November (LExNo).Journal of Geophysical Research, 115(D11206):1–10, June 2010.doi:10.1029/2009JD012626. → pages 36, 37S. Henning, M. Ziese, A. Kiselev, H. Saathoff, O. Mo¨hler, T. F. Mentel,A. Buchholz, C. Spindler, V. Michaud, M. Monier, K. Sellegri, andF. Stratmann. Hygroscopic growth and droplet activation of soot particles:uncoated, succinic or sulfuric acid coated. Atmospheric Chemistry andPhysics, 12(10):4525–4537, May 2012. doi:10.5194/acp-12-4525-2012. →pages 36, 37S. P. Hersey, J. S. Craven, A. R. Metcalf, J. Lin, T. Lathem, K. J. Suski, J. F.Cahill, H. T. Duong, A. Sorooshian, H. H. Jonsson, M. Shiraiwa, A. Zuend,A. Nenes, K. A. Prather, R. C. Flagan, and J. H. Seinfeld. Composition andhygroscopicity of the Los Angeles Aerosol: CalNex. Journal of GeophysicalResearch: Atmospheres, 118(7):3016–3036, Apr. 2013.doi:10.1002/jgrd.50307. → pages 58R. Hitzenberger, A. Berner, R. Kromp, A. Kasper-Giebl, A. Limbeck,W. Tscherwenka, and H. Puxbaum. Black carbon and other species at ahigh-elevation European site (Mount Sonnblick, 3106 m, Austria):Concentrations and scavenging efficiencies. Journal of Geophysical Research,105(D20):24637–24645, 2000. → pages 36, 37, 76, 88, 99R. Hitzenberger, A. Berner, H. Glebl, K. Drobesch, A. Kasper-Giebl,M. Loeflund, H. Urban, and H. Puxbaum. Black carbon (BC) in alpineaerosols and cloud water - concentrations and scavenging efficiencies.Atmospheric Environment, 35(30):5135–5141, 2001. → pages 36, 37, 76, 88,99R. Holzinger, J. Williams, G. Salisbury, T. Klu¨pfel, M. de Reus, M. Traub, P. J.Crutzen, and J. Lelieveld. Oxygenated compounds in aged biomass burningplumes over the Eastern Mediterranean: evidence for strong secondaryproduction of methanol and acetone. Atmospheric Chemistry and Physics, 5(1):39–46, Jan. 2005. doi:10.5194/acp-5-39-2005. → pages 83IPCC-AR4. Climate Change 2007: The Physical Science Basis. Contributions ofWorking Group I to the Fourth Assessment Report of the IntergovernmentalPanel on Climate Change. Cambridge University Press, Cambridge, UnitedKingdom and New York, NY, 2007. → pages xix, 14135IPCC-AR5. CLIMATE CHANGE 2013: The Physical Science Basis. WorkingGroup I to the Contributions of the Intergovernmental Panel on ClimateChange. Cambridge University Press, Cambridge, United Kingdom and NewYork, NY, 2013. → pages 6, 7D. J. Jacob. Introduction to Atmospheric Chemistry. Princeton University Press,Princeton, NJ, 1999. → pages 1M. Z. Jacobson. Investigating cloud absorption effects: Global absorptionproperties of black carbon, tar balls, and soil dust in clouds and aerosols.Journal of Geophysical Research, 117(D06205):1–25, Mar. 2012.doi:10.1029/2011JD017218. → pages 36J. T. Jayne, D. C. Leard, X. Zhang, P. Davidovits, K. A. Smith, C. E. Kolb, andD. R. Worsnop. Development of an Aerosol Mass Spectrometer for Size andComposition Analysis of Submicron Particles. Aerosol Science andTechnology, 33(1):49–70, July 2000. doi:10.1080/027868200410840. →pages 104A. Kasper-Giebl, A. Koch, R. Hitzenberger, and H. Puxbaum. Scavengingefficiency of ’aerosol carbon’ and sulfate in soopercooled clouds at MountSonnblick (3106 m a.s.l., Austria). Journal of Atmospheric Chemistry, 35(1993):33–46, 2000. → pages 36, 37, 76, 88, 99Z. Kipling, P. Stier, J. P. Schwarz, A. E. Perring, J. R. Spackman, G. W. Mann,C. E. Johnson, and P. J. Telford. Constraints on aerosol processes in climatemodels from vertically-resolved aircraft observations of black carbon.Atmospheric Chemistry and Physics, 13(12):5969–5986, June 2013.doi:10.5194/acp-13-5969-2013. → pages 101D. Koch, M. Schulz, S. Kinne, C. McNaughton, J. R. Spackman, Y. Balkanski,S. Bauer, T. Berntsen, T. C. Bond, and Boucher. Evaluation of black carbonestimations in global aerosol models. Atmospheric Chemistry and Physics, 9:9001–9026, 2009. → pages 5D. Koch, Y. Balkanski, S. E. Bauer, R. C. Easter, S. Ferrachat, S. J. Ghan,C. Hoose, T. Iversen, A. Kirkeva˚ g, J. E. Kristjansson, X. Liu, U. Lohmann,S. Menon, J. Quaas, M. Schulz, O. Seland, T. Takemura, and N. Yan. Sootmicrophysical effects on liquid clouds, a multi-model investigation.136Atmospheric Chemistry and Physics, 11(3):1051–1064, Feb. 2011.doi:10.5194/acp-11-1051-2011. → pages 36K. A. Koehler, P. J. DeMott, S. M. Kreidenweis, O. B. Popovicheva, M. D.Petters, C. M. Carrico, E. D. Kireeva, T. D. Khokhlova, and N. K. Shonija.Cloud condensation nuclei and ice nucleation activity of hydrophobic andhydrophilic soot particles. Physical chemistry chemical physics, 11(36):7906–7920, Sept. 2009. doi:10.1039/b905334b. → pages 36Y. Kondo, N. Oshima, M. Kajino, R. Mikami, N. Moteki, N. Takegawa, R. L.Verma, Y. Kajii, S. Kato, and A. Takami. Emissions of black carbon in EastAsia estimated from observations at a remote site in the East China Sea.Journal of Geophysical Research, 116(D16201):1–14, Aug. 2011.doi:10.1029/2011JD015637. → pages 105M. Kuwata, Y. Kondo, and N. Takegawa. Critical condensed mass for activationof black carbon as cloud condensation nuclei in Tokyo. Journal ofGeophysical Research, 114(D20202):1–9, Oct. 2009.doi:10.1029/2009JD12086. → pages 36M. Laborde, M. Schnaiter, C. Linke, H. Saathoff, K.-H. Naumann, O. Mo¨hler,S. Berlenz, U. Wagner, J. W. Taylor, D. Liu, M. Flynn, J. D. Allan, H. Coe,K. Heimerl, F. Dahlko¨tter, B. Weinzierl, a. G. Wollny, M. Zanatta, J. Cozic,P. Laj, R. Hitzenberger, J. P. Schwarz, and M. Gysel. Single Particle SootPhotometer intercomparison at the AIDA chamber. Atmospheric MeasurementTechniques, 5(12):3077–3097, Dec. 2012. doi:10.5194/amt-5-3077-2012. →pages 102, 103M. Laborde, M. Crippa, T. Tritscher, Z. Jura´nyi, P. F. Decarlo,B. Temime-Roussel, N. Marchand, S. Eckhardt, A. Stohl, U. Baltensperger,A. S. H. Pre´voˆt, E. Weingartner, and M. Gysel. Black carbon physicalproperties and mixing state in the European megacity Paris. AtmosphericChemistry and Physics, 13(11):5831–5856, June 2013.doi:10.5194/acp-13-5831-2013. → pages 109S. Lance, T. Raatikainen, T. B. Onasch, D. R. Worsnop, X.-Y. Yu, M. L.Alexander, M. R. Stolzenburg, P. H. McMurry, J. N. Smith, and A. Nenes.Aerosol mixing state, hygroscopic growth and cloud activation efficiencyduring MIRAGE 2006. Atmospheric Chemistry and Physics, 13(9):5049–5062, May 2013. doi:10.5194/acp-13-5049-2013. → pages 58137T. L. Lathem, A. J. Beyersdorf, K. L. Thornhill, E. L. Winstead, M. J. Cubison,A. Hecobian, J. L. Jimenez, R. J. Weber, B. E. Anderson, and A. Nenes.Analysis of CCN activity of Arctic aerosol and Canadian biomass burningduring summer 2008. Atmospheric Chemistry and Physics, 13(5):2735–2756,Mar. 2013. doi:10.5194/acp-13-2735-2013. → pages 60W. R. Leaitch, A. M. Macdonald, P. C. Brickell, J. Liggio, S. J. Sjostedt,A. Vlasenko, J. W. Bottenheim, L. Huang, S.-M. Li, P. S. Liu,D. Toom-Sauntry, K. A. Hayden, S. Sharma, N. C. Shantz, H. A. Wiebe,W. Zhang, J. P. Abbatt, J. G. Slowik, R. Y.-W. Chang, L. M. Russell, R. E.Schwartz, S. Takahama, J. T. Jayne, and N. L. Ng. Temperature response ofthe submicron organic aerosol from temperate forests. AtmosphericEnvironment, 45(37):6696–6704, Dec. 2011.doi:10.1016/j.atmosenv.2011.08.047. → pages 78D. R. Lide, editor. CRC Handbook of Chemistry and Physics. CRC Press, 82edition, 2001. → pages 58B. Y. Liu, F. J. Romay, W. D. Dick, K.-S. Woo, and M. Chiruta. A Wide-RangeParticle Spectrometer for Aerosol Measurement from 0.010 m to 10 m.Aerosol and Air Quality Research, 10:125–139, 2010a.doi:10.4209/aaqr.2009.10.0062. → pages 80D. Liu, M. Flynn, M. Gysel, A. Targino, I. Crawford, K. Bower, T. Choularton,Z. Jura´nyi, M. Steinbacher, C. Hu¨glin, J. Curtius, M. Kampus, A. Petzold,E. Weingartner, U. Baltensperger, and H. Coe. Single particle characterizationof black carbon aerosols at a tropospheric alpine site in Switzerland.Atmospheric Chemistry and Physics, 10(15):7389–7407, 2010b.doi:10.5194/acp-10-7389-2010. → pages 101, 106, 110, 121A. M. Macdonald, K. G. Anlauf, W. R. Leaitch, E. Chan, and D. W. Tarasick.Interannual variability of ozone and carbon monoxide at the Whistler highelevation site: 20022006. Atmospheric Chemistry and Physics, 11(22):11431–11446, Nov. 2011. doi:10.5194/acp-11-11431-2011. → pages 78, 101,109, 110A. M. Macdonald, W. R. Leaitch, J. P. D. Abbatt, A. K. Betram, D. J. Cziczo,P. Herckes, A. K. Y. Lee, J. Liggio, K. J. Noone, J. R. Pierce, L. M. Russell,S. J. Sjostedt, K. Strawbridge, H. A. Wiebe, L. Ahlm, J. H. Buller,P. Campuzano-Jost, E. Chan, J. Corbin, A. Elford, B. J. Firanski, K. L.138Hayden, S. M. Li, P. S. K. Liu, C. Mihele, D. Toom-Sauntry, S. Sharma,A. Sheppard, J. C. Schroder, J. G. Slowik, G. Stupple, A. Vlasenko, C. D.Wainwright, D. Wang, Y. Wang, J. Wentzell, and J. P. S. Wong. An Overviewof WACS 2010: Biogenic Aerosol Formation, Mountain Flows and CCN.Manuscript, page In Preparation, 2014. → pages 77, 78, 81, 83C. Ma¨tzler. MATLAB Functions for Mie Scattering and Absorption Version 1 -Research Report No. 2002-08. 2002a. URLhttp://www.iap.unibe.ch/publications/download/201/en/. → pages 28C. Ma¨tzler. MATLAB Functions for Mie Scattering and Absorption Version 2 -Research Report No. 2002-11. 2002b. URLwww.iap.unibe.ch/publications/download/199/en/. → pages 28I. G. McKendry, J. Gallagher, P. Campuzano Jost, A. Bertram, K. Strawbridge,R. Leaitch, and A. M. Macdonald. Ground-based remote sensing of anelevated forest fire aerosol layer at Whistler, BC: implications forinterpretation of mountaintop chemistry. Atmospheric Chemistry and Physics,10(23):11921–11930, Dec. 2010. doi:10.5194/acp-10-11921-2010. → pages78G. R. McMeeking, T. Hamburger, D. Liu, M. Flynn, W. T. Morgan,M. Northway, E. J. Highwood, R. Krejci, J. D. Allan, A. Minikin, and H. Coe.Black carbon measurements in the boundary layer over western and northernEurope. Atmospheric Chemistry and Physics, 10(19):9393–9414, Oct. 2010.doi:10.5194/acp-10-9393-2010. → pages 22F. Mei, P. L. Hayes, A. Ortega, J. W. Taylor, J. D. Allan, J. Gilman, W. Kuster,J. de Gouw, J. L. Jimenez, and J. Wang. Droplet activation properties oforganic aerosols observed at an urban site during CalNex-LA. Journal ofGeophysical Research: Atmospheres, 118(7):2903–2917, Apr. 2013.doi:10.1002/jgrd.50285. → pages 60A. R. Metcalf, J. S. Craven, J. J. Ensberg, J. Brioude, W. Angevine,A. Sorooshian, H. T. Duong, H. H. Jonsson, R. C. Flagan, and J. H. Seinfeld.Black carbon aerosol over the Los Angeles Basin during CalNex. Journal ofGeophysical Research: Atmospheres, 117(D00V13):1–24, Nov. 2012.doi:10.1029/2011JD17255. → pages 29, 36, 52, 54, 55, 106, 111, 128139R. H. Moore, K. Cerully, R. Bahreini, C. A. Brock, A. M. Middlebrook, andA. Nenes. Hygroscopicity and composition of California CCN during summer2010. Journal of Geophysical Research: Atmospheres, 117(D00V12):1–14,Nov. 2012. doi:10.1029/2011JD017352. → pages 16, 58N. Moteki and Y. Kondo. Method to measure time-dependent scattering crosssections of particles evaporating in a laser beam. Journal of Aerosol Science,39:348 – 364, Dec. 2008. doi:10.1016/j.jaerosci.2007.12.002. → pages 44N. Moteki, Y. Kondo, N. Takegawa, and S.-i. Nakamura. Directional dependenceof thermal emission from nonspherical carbon particles. Journal of AerosolScience, 40(9):790–801, Sept. 2009. doi:10.1016/j.jaerosci.2009.05.003. →pages 23N. L. Ng, S. C. Herndon, A. Trimborn, M. R. Canagaratna, P. L. Croteau, T. B.Onasch, D. Sueper, D. R. Worsnop, Q. Zhang, Y. L. Sun, and J. T. Jayne. AnAerosol Chemical Speciation Monitor (ACSM) for Routine Monitoring of theComposition and Mass Concentrations of Ambient Aerosol. Aerosol Scienceand Technology, 45(7):780–794, July 2011.doi:10.1080/02786826.2011.560211. → pages 104K. J. Noone, J. a. Ogren, J. Heintzenberg, R. J. Charlson, and D. S. Covert.Design and Calibration of a Counterflow Virtual Impactor for Sampling ofAtmospheric Fog and Cloud Droplets. Aerosol Science and Technology, 8(3):235–244, Jan. 1988. → pages 42, 43, 44, 79K. J. Noone, J. A. Ogren, A. Hallberg, J. Heintzenberg, J. Stro¨m, H.-C. Hansson,B. Svenningsson, A. Wiedensohler, M. C. Facchinni, B. G. Arends, andA. Berner. Changes in aerosol-size- and phase distributions due to physicaland chemical processes in fog. Tellus, 44B:489–504, 1992. → pages 36, 59C. O’Dowd, E. Becker, and M. Kulmala. Mid-latitude North-Atlantic aerosolcharacteristics in clean and polluted air. Atmospheric Research, 58(3):167–185, 2001. → pages 51J. Ovadnevaite, D. Ceburnis, M. Canagaratna, H. Berresheim, J. Bialek,G. Martucci, D. R. Worsnop, and C. O’Dowd. On the effect of wind speed onsubmicron sea salt mass concentrations and source fluxes. Journal ofGeophysical Research, 117(D16201):1–11, Aug. 2012.doi:10.1029/2011JD017379. → pages 60140M. S. Pekour and D. J. Cziczo. Wake capture, particle breakup, and otherartifacts associated with counterflow virtual impaction. Aerosol Science andTechnology, 45:758–764, 2011. doi:10.1080/02786826.2011.558942. →pages 52M. D. Petters and S. M. Kreidenweis. A single parameter representation ofhygroscopic growth and cloud condensation nucleus activity. AtmosphericChemistry and Physics, 7(8):1961–1971, Apr. 2007. → pages xix, 16, 17, 19,58M. D. Petters, C. M. Carrico, S. M. Kreidenweis, A. J. Prenni, P. J. DeMott, J. L.Collett, and H. Moosmu¨ller. Cloud condensation nucleation activity ofbiomass burning aerosol. Journal of Geophysical Research, 114(D22205):1–16, Nov. 2009. doi:10.1029/2009JD012353. → pages 16, 36, 37, 89A. Petzold, M. Gysel, X. Vancassel, R. Hitzenberger, H. Puxbaum,S. Vrochticky, E. Weingartner, U. Baltensperger, and P. Mirabel. On theeffects of organic matter and sulphur-containing compounds on the CCNactivation of combustion particles. Atmospheric Chemistry and Physics, 5(12):3187–3203, Dec. 2005. → pages 36, 37A. Petzold, J. A. Ogren, M. Fiebig, P. Laj, S.-M. Li, U. Baltensperger,T. Holzer-Popp, S. Kinne, G. Pappalardo, N. Sugimoto, C. Wehrli,A. Wiedensohler, and X.-Y. Zhang. Recommendations for reporting ”blackcarbon” measurements. Atmospheric Chemistry and Physics, 13(16):8365–8379, Aug. 2013. doi:10.5194/acp-13-8365-2013. → pages 4J. R. Pierce, W. R. Leaitch, J. Liggio, D. M. Westervelt, C. D. Wainwright,J. P. D. Abbatt, L. Ahlm, W. Al-Basheer, D. J. Cziczo, K. L. Hayden, A. K. Y.Lee, S.-M. Li, L. M. Russell, S. J. Sjostedt, K. B. Strawbridge, M. Travis,A. Vlasenko, J. J. B. Wentzell, H. a. Wiebe, J. P. S. Wong, and a. M.Macdonald. Nucleation and condensational growth to CCN sizes during asustained pristine biogenic SOA event in a forested mountain valley.Atmospheric Chemistry and Physics, 12(7):3147–3163, Apr. 2012.doi:10.5194/acp-12-3147-2012. → pages 81L. Pirjola and C. O’Dowd. Can new particle formation occur in the clean marineboundary layer? Journal of Geophysical Research, 105(D21):26531–26546,2000. → pages 51141O. B. Popovicheva, N. M. Persiantseva, E. D. Kireeva, T. D. Khokhlova, andN. K. Shonija. Quantification of the hygroscopic effect of soot aging in theatmosphere: laboratory simulations. The Journal of Physical Chemistry. A,115(3):298–306, Jan. 2011. doi:10.1021/jp109238x. → pages 36, 37K. A. Pratt, S. M. Murphy, R. Subramanian, P. J. DeMott, G. L. Kok, T. Campos,D. C. Rogers, A. J. Prenni, a. J. Heymsfield, J. H. Seinfeld, and K. A. Prather.Flight-based chemical characterization of biomass burning aerosols within twoprescribed burn smoke plumes. Atmospheric Chemistry and Physics, 11(24):12549–12565, Dec. 2011. doi:10.5194/acp-11-12549-2011. → pages 89A. J. Prenni, M. D. Petters, S. M. Kreidenweis, P. J. DeMott, and P. J. Ziemann.Cloud droplet activation of secondary organic aerosol. Journal of GeophysicalResearch, 112(D10223):1–12, May 2007. doi:10.1029/2006JD007963. →pages 16N. Riemer, H. Vogel, and B. Vogel. Soot aging time scales in polluted regionsduring day and night. Atmospheric Chemistry and Physics, 4(7):1885–1893,2004. → pages 36N. Riemer, M. West, R. Zaveri, and R. Easter. Estimating black carbon agingtime-scales with a particle-resolved aerosol model. Journal of Aerosol Science,41(1):143–158, Jan. 2010. doi:10.1016/j.jaerosci.2009.08.009. → pages 36G. Rolph. Real-time Environmental Applications and Display sYstem (READY)Website, NOAA Air Resources Laboratory, Silver Spring, MD, 2012. URLhttp//ready.arl.noaa.gov. → pages 81, 104G. D. Rolph. Real-time Environmental Application and Display sYstem(READY) Website. NOAA Air Resources Laboratory, Silver Spring, MD.2013. URL http//ready.arl.noaa.gov. → pages 47D. Rose, B. Wehner, M. Ketzel, C. Engler, J. Voigtla¨nder, T. Tuch, andA. Wiedensohler. Atmospheric number size distributions of soot particles andestimation of emission factors. Atmospheric Chemistry and Physics, 6(4):1021–1031, 2006. → pages 128D. Rose, S. S. Gunthe, E. Mikhailov, G. P. Frank, U. Dusek, M. O. Andreae, andU. Po¨schl. Calibration and measurement uncertainties of a continuous-flowcloud condensation nuclei counter (DMT-CCNC): CCN activation of142ammonium sulfate and sodium chloride aerosol particles in theory andexperiment. Atmospheric Chemistry and Physics, 8(5):1153–1179, Feb. 2008.→ pages 59D. Rose, A. Nowak, P. Achtert, A. Wiedensohler, M. Hu, M. Shao, Y. Zhang,M. O. Andreae, and U. Po¨schl. Cloud condensation nuclei in polluted air andbiomass burning smoke near the mega-city Guangzhou, China Part 1:Size-resolved measurements and implications for the modeling of aerosolparticle hygroscopicity and CCN activity. Atmospheric Chemistry andPhysics, 10(7):3365–3383, Apr. 2010. → pages 16, 58, 61L. K. Sahu, Y. Kondo, N. Moteki, N. Takegawa, Y. Zhao, M. J. Cubison, J. L.Jimenez, S. Vay, G. S. Diskin, A. Wisthaler, T. Mikoviny, L. G. Huey, A. J.Weinheimer, and D. J. Knapp. Emission characteristics of black carbon inanthropogenic and biomass burning plumes over California duringARCTAS-CARB 2008. Journal of Geophysical Research, 117(D16302):1–20,Aug. 2012. doi:10.1029/2011JD017401. → pages 106J. C. Schroder, S. J. Hanna, R. L. Modini, A. L. Corrigan, A. M. Macdonald,K. J. Noone, L. M. Russell, W. R. Leaitch, and A. K. Bertram. Size-resolvedobservations of refractory black carbon particles in cloud droplets at a marineboundary layer site. Atmospheric Chemistry and Physics Discussions, 14:11447–11491, 2014. doi:10.5194/acp-14-11447-2014. → pages 77J. P. Schwarz, R. S. Gao, D. W. Fahey, D. S. Thomson, L. A. Watts, J. C. Wilson,J. M. Reeves, M. Darbeheshti, D. G. Baumgardner, G. L. Kok, S. H. Chung,M. Schulz, J. Hendricks, A. Lauer, B. Ka¨rcher, J. G. Slowik, K. H. Rosenlof,T. L. Thompson, a. O. Langford, M. Loewenstein, and K. C. Aikin.Single-particle measurements of midlatitude black carbon and light-scatteringaerosols from the boundary layer to the lower stratosphere. Journal ofGeophysical Research, 111(D16207):1–15, 2006.doi:10.1029/2006JD007076. → pages 22, 24, 25, 44J. P. Schwarz, R. S. Gao, J. R. Spackman, L. A. Watts, D. S. Thomson, D. W.Fahey, T. B. Ryerson, J. Peischl, J. S. Holloway, M. Trainer, G. J. Frost,T. Baynard, D. a. Lack, J. a. de Gouw, C. Warneke, and L. a. Del Negro.Measurement of the mixing state, mass, and optical size of individual blackcarbon particles in urban and biomass burning emissions. GeophysicalResearch Letters, 35(13):L13810, July 2008a. doi:10.1029/2008GL033968.→ pages 29, 55, 106, 109, 128143J. P. Schwarz, J. R. Spackman, D. W. Fahey, R. S. Gao, U. Lohmann, P. Stier,L. a. Watts, D. S. Thomson, D. A. Lack, L. Pfister, M. J. Mahoney,D. Baumgardner, J. C. Wilson, and J. M. Reeves. Coatings and theirenhancement of black carbon light absorption in the tropical atmosphere.Journal of Geophysical Research, 113(D03203):1–10, Feb. 2008b.doi:10.1029/2007JD009042. → pages 29, 36J. P. Schwarz, J. R. Spackman, R. S. Gao, a. E. Perring, E. Cross, T. B. Onasch,A. Ahern, W. Wrobel, P. Davidovits, J. Olfert, M. K. Dubey, C. Mazzoleni,and D. W. Fahey. The Detection Efficiency of the Single Particle SootPhotometer. Aerosol Science and Technology, 44(8):612–628, June 2010a.doi:10.1080/02786826.2010.481298. → pages 102J. P. Schwarz, J. R. Spackman, R. S. Gao, L. A. Watts, P. Stier, M. Schulz, S. M.Davis, S. C. Wofsy, and D. W. Fahey. Global-scale black carbon profilesobserved in the remote atmosphere and compared to models. GeophysicalResearch Letters, 37(18):L18812, Sept. 2010b. doi:10.1029/2010GL044372.→ pages 101J. P. Schwarz, B. H. Samset, A. E. Perring, J. R. Spackman, R. S. Gao, P. Stier,M. Schulz, F. L. Moore, E. A. Ray, and D. W. Fahey. Global-scale seasonallyresolved black carbon vertical profiles over the Pacific. Geophysical ResearchLetters, 40(20):5542–5547, Oct. 2013. doi:10.1002/2013GL057775. → pages101A. Schwarzenboeck, J. Heintzenberg, and S. Mertes. Incorporation of aerosolparticles between 25 and 850 nm into cloud elements: measurements with anew complementary sampling system. Atmospheric research, 52:241–260,2000. → pages 52A. J. Sedlacek, E. R. Lewis, L. Kleinman, J. Xu, and Q. Zhang. Determination ofand evidence for non-core-shell structure of particles containing black carbonusing the Single-Particle Soot Photometer (SP2). Geophysical ResearchLetters, 39(6):L06802, Mar. 2012. doi:10.1029/2012GL050905. → pages 29,54J. H. Seinfeld and S. N. Pandis. Atmospheric Chemistry and Physics: From AirPollution to Climate Change. John Wiley & Sons, Hoboken, NJ, 2nd edition,2006. ISBN 978-0-471-72017-1. → pages 3, 16144K. Sellegri, P. Laj, R. Dupuy, M. Legrand, S. Preunkert, and J. P. Putaud.Size-dependent scavenging efficiencies of multicomponent atmosphericaerosols in clouds. Journal of Geophysical Research, 108(D11):4334, 2003.doi:10.1029/2002JD002749. → pages 36, 37, 76, 88, 99J. S. Serafini. Impingement of water droplets on wedges and double-wedgeairfoils at supersonic speeds. NACA, TR 1159, 1954. → pages 43L. M. Shank, S. Howell, A. D. Clarke, S. Freitag, V. Brekhovskikh, V. Kapustin,C. McNaughton, T. Campos, and R. Wood. Organic matter and non-refractoryaerosol over the remote Southeast Pacific: oceanic and combustion sources.Atmospheric Chemistry and Physics, 12(1):557–576, Jan. 2012.doi:10.5194/ACP-12-557-2012. → pages 52N. C. Shantz, W. R. Leaitch, L. Phinney, M. Mozurkewich, andD. Toom-Sauntry. The effect of organic compounds on the growth rate ofcloud droplets in marine and forest settings. Atmospheric Chemistry andPhysics, 8(19):5869–5887, Apr. 2008. → pages 58Z. Shen, J. Liu, L. W. Horowitz, D. K. Henze, S. Fan, H. Levy II, D. L.Mauzerall, J.-T. Lin, and S. Tao. Analysis of transpacific transport of blackcarbon during HIPPO-3: implications for black carbon aging. AtmosphericChemistry and Physics Discussions, 14(1):505–540, 2014.doi:10.5194/acpd-14-505-2014. → pages 101B. Simoneit, J. Schauer, C. Nolte, D. Oros, V. Elias, M. Fraser, W. Rogge, andG. Cass. Levoglucosan, a tracer for cellulose in biomass burning andatmospheric particles. Atmospheric Environment, 33(2):173–182, Jan. 1999.doi:10.1016/S1352-2310(98)00145-9. → pages 107J. G. Slowik, E. S. Cross, J.-H. Han, P. Davidovits, T. B. Onasch, J. T. Jayne,L. R. Williams, M. R. Canagaratna, D. R. Worsnop, R. K. Chakrabarty,H. Moosmu¨ller, W. P. Arnott, J. P. Schwarz, R.-S. Gao, D. W. Fahey, G. L.Kok, and A. Petzold. An Inter-Comparison of Instruments Measuring BlackCarbon Content of Soot Particles. Aerosol Science and Technology, 41(3):295–314, Apr. 2007. ISSN 0278-6826. doi:10.1080/02786820701197078.URL http://www.tandfonline.com/doi/abs/10.1080/02786820701197078. →pages 38145M. Stephens, N. Turner, and J. Sandberg. Particle identification by laser-inducedincandescence in a solid-state laser cavity. Applied optics, 42(19):3726–3736,July 2003. → pages 20, 22, 44B. Stevens, G. Feingold, W. R. Cotton, and R. L. Walko. Elements of theMicrophysical Structure of Numerically Simulated NonprecipitatingStratocumulus. Journal of the Atmospheric Sciences, 53(7):980–1006, 1995.→ pages 55R. B. Stull. Introduction to Boundary Layer Meteorology. Kluwer AcademicPress, Dordrecht, Netherlands, 2003. doi:10.1007/978-94-009-3027-8. →pages 2S. Takahama, R. E. Schwartz, L. M. Russell, A. M. Macdonald, S. Sharma, andW. R. Leaitch. Organic functional groups in aerosol particles from burning andnon-burning forest emissions at a high-elevation mountain site. AtmosphericChemistry and Physics, 11(13):6367–6386, July 2011.doi:10.5194/acp-11-6367-2011. → pages 78, 101, 104, 107C. H. Twohy, M. D. Petters, J. R. Snider, B. Stevens, W. Tahnk, M. Wetzel, L. M.Russell, and F. Burnet. Evaluation of the aerosol indirect effect in marinestratocumulus clouds: Droplet number, size, liquid water path, and radiativeimpact. Journal of Geophysical Research, 110:1–16, 2005.doi:10.1029/2004JD005116. → pages 51B. Verheggen, J. Cozic, E. Weingartner, K. Bower, S. Mertes, P. Connolly,M. Gallagher, M. Flynn, T. Choularton, and U. Baltensperger. Aerosolpartitioning between the interstitial and the condensed phase in mixed-phaseclouds. Journal of Geophysical Research, 112(D23):1–13, Dec. 2007.doi:10.1029/2007JD008714. → pages 36, 37G. Vidaurre, J. Hallett, and D. C. Rogers. Airborne Measurement of Liquid andTotal Water Content. Journal of Atmospheric and Oceanic Technology, 28(9):1088–1103, Sept. 2011. → pages 52E. Vignati, M. Karl, M. Krol, J. Wilson, P. Stier, and F. Cavalli. Sources ofuncertainties in modelling black carbon at the global scale. AtmosphericChemistry and Physics, 10(6):2595–2611, Mar. 2010. → pages 36S.-L. Von der Weiden, F. Drewnick, and S. Borrmann. Particle Loss Calculator anew software tool for the assessment of the performance of aerosol inlet146systems. Atmospheric Measurement Techniques, 2(2):479–494, Sept. 2009. →pages 40J. M. Wallace and P. V. Hobbs. Atmospheric Science: An Introductory Survey.Academic Press, London, United Kingdom, 2nd edition, 2006. → pages xix,1, 13C. Wang. Impact of anthropogenic absorbing aerosols on clouds andprecipitation: A review of recent progresses. Atmospheric Research, 122:237–249, Mar. 2013. → pages 36Q. Wang, D. J. Jacob, J. R. Spackman, A. E. Perring, J. P. Schwarz, N. Moteki,E. a. Marais, C. Ge, J. Wang, and S. R. H. Barrett. Global budget and radiativeforcing of black carbon aerosol: Constraints from pole-to-pole (HIPPO)observations across the Pacific. Journal of Geophysical Research:Atmospheres, 119(1):195–206, Jan. 2014. doi:10.1002/2013JD020824. →pages 10, 113, 121Z. Wang, H. Zhang, J. Li, X. Jing, and P. Lu. Radiative forcing and climateresponse due to the presence of black carbon in cloud droplets. Journal ofGeophysical Research: Atmospheres, 118:1–14, May 2013.doi:10.1002/jgrd.50312. → pages 36Z. J. Wu, L. Poulain, S. Henning, K. Dieckmann, W. Birmili, M. Merkel, D. vanPinxteren, G. Spindler, K. Mu¨ller, F. Stratmann, H. Herrmann, andA. Wiedensohler. Relating particle hygroscopicity and CCN activity tochemical composition during the HCCT-2010 field campaign. AtmosphericChemistry and Physics, 13(16):7983–7996, Aug. 2013.doi:10.5194/acp-13-7983-2013. → pages 58B. L. Zhuang, L. Liu, F. H. Shen, T. J. Wang, and Y. Han. Semidirect radiativeforcing of internal mixed black carbon cloud droplet and its regional climaticeffect over China. Journal of Geophysical Research, 115(May):D00K19, Sept.2010. doi:10.1029/2009JD013165. → pages 36147Appendix AHR-ToF-AMS Ion Pairing Schemeused in Chapter 4The inorganic ions (NH+4 , NO–3, SO2–4 , Cl–) mass fractions measured with the HR-ToF-AMS were converted to mass fractions of ammonium nitrate, ammoniumsulfate, ammonium bisulfate, sulfuric acid, or ammonium chloride using the sim-148plified ion-pairing scheme below:nNH4NO3 = nNO−3nNH4Cl = nCl−n(NH4)2SO4 = max(0,nNH+4 − nSO2−4−nNO−3 −nCl−)nNH4HSO4 = min(2nSO2−4− nNH+4 +nNO−3 +nCl−,nNH+4 −nNO−3 −nCl−)nH2SO4 = max(0,nSO2−4−nNH+4 +nNO−3 +nCl−)Where n is the number of moles of that species. This ion scheme is the same as theone used by Gysel and Crosier (2007) except it has been modified to incorporateammonium chloride.149Appendix BCalculation of the DropletTransmission Factor Through theCVI used in Chapter 4The droplet transmission factor (DT ) through the CVI was determined by plot-ting the number of droplets measured by the FM-100 that were greater than theCVI−D50 as a function of the number of residual particles (see Figure B.1). Thenumber of residual particles plotted in Figure B.1 were corrected for the CVI en-hancement factor (EF) and only residual particles >100 nm were included sinceparticles smaller than this size were likely not due to nucleation scavenging (seeSection 4.4.3.2 for further discussion). From 1/slope in Figure B.1 panels A and B,the DT values were determined to be 26% during Cloud 2 and 45% during Cloud3. DT values less than 100% may be attributed to ; 1) particle or droplet losses in150the CVI; 2) incomplete drying of the droplets; 3) misalignment of droplets in thewind tunnel prior to entering the CVI; and 4) insufficient acceleration of some ofthe larger droplets.151Figure B.1: Correlation plots between the cloud droplet number concentra-tion (CDNC) greater than the CVI −D50 (CDNC > D50) and theenhancement factor (EF) corrected residual number concentrationgreater than 100 nm (NRes>100nm) in cm-3.152Appendix CCalculation of the DropletTransmission Factor Through theCVI used in Chapter 5To account for droplet losses during CVI sampling a droplet transmission (DT )factor was determined by plotting the CDNC that was greater than the CVI cut-size as a function of the number of residual particles counted by a CPC (TSI-3775), which counted all residual particles behind the CVI with diameters be-tween 10 nm and 1 µm (Figure C.1). The number of residual particles counted bythe CPC was corrected for the enhancement of particles by the CVI. The droplettransmissions were calculated from 1/slope of linear fits to the data shown in Fig-ure C.1 panels A and B, and were found to be 33 and 35% for Clouds 1 and 2, re-spectively. A DT value of less than 100% may be due to; particle or droplet losses153in the CVI or tubing downstream of the CVI; droplets improperly aligned withinthe wind tunnel prior to entering the CVI; incomplete drying of the droplets; andinsufficiently accelerating some of the larger droplets, which would result in re-jection of the droplet by the counterflow. Droplet transmission values reportedhere are similar to values reported in the study described in Chapter 4.154Figure C.1: Correlation plots for Clouds 1 and 2 in panels A and B, respec-tively, of the cloud droplet number concentration (CDNC) for dropletsthat are greater than the CVI cut-size as a function of the enhancementfactor corrected residual number concentration. The slope of the lin-ear fits to the correlation data represent the scaling factor needed toaccount for droplet losses and 1/slope represents the droplet transmis-sion.155

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0167086/manifest

Comment

Related Items