• Albrecht, B. A., 1989: Aerosols, cloud microphysics, and fractional cloudiness. Science,245, 1227–1230.

  • Anderson, T. L., 1992: Optimization of a counterflow virtual impactor (CVI) for studying aerosol effects on cloud droplet number. Ph.D. dissertation, University of Washington, 232 pp.

  • ——, D. S. Covert, and R. J. Charlson, 1994: Cloud droplet number studies with a counterflow virtual impactor. J. Geophys. Res.,99, 8249–8256.

  • Anthoni, P. A., 1996: Carbon dioxide eddy flux measurements in complex terrain from a coniferous forest under the influence of marine air. M.S. thesis, Dept. of Atmospheric Science, Oregon State University, 79 pp. [Available from Valley Library, Oregon State University, Corvallis, OR 97331.].

  • Arends, B. G., G. P. A. Kos, W. Wobrock, D. Schell, K. J. Noone, S. Fuzzi, and S. Pahl, 1992: Comparison of techniques for measurement of fog liquid water content. Tellus,44B, 604–611.

  • Baker, M. B., A. Blyth, D. Carruthers, S. Caughey, and T. W. Choularton, 1982: Field studies of the effect of entrainment upon the structure of clouds at Great Dun Fell. Quart. J. Roy. Meteor. Soc.,108, 899–916.

  • Baumgardner, D., 1983: An analysis and comparison of five water droplet measuring instruments. J. Climate Appl. Meteor.,22, 891–910.

  • ——, W. Strapp, and J. E. Dye, 1985: Evaluation of the forward scattering spectrometer probe. Part II: Corrections for coincidence and dead-time losses. J. Atmos. Oceanic Technol.,2, 626–632.

  • Betts, A. K., 1990: Diurnal variation of California coastal stratocumulus from two days of boundary layer soundings. Tellus,42A, 302–304.

  • Blaskovic, M., R. Davies, and J. B. Snider, 1991: Diurnal variation of marine stratocumulus over San Nicolas Island during July 1987. Mon. Wea. Rev.,119, 1469–1478.

  • Blyth, A. M., T. W. Choularton, G. Fullerton, J. Latham, C. Mill, M. Smith, and I. Stromberg, 1980: The influence of entrainment on the evolution of cloud droplet spectra. Quart. J. Roy. Meteor. Soc.,106, 821–840.

  • Borys, R. D., D. Lowenthal, and D. Mitchell, 1995: CCN regulation of snowfall rates. Eos, Trans. Amer. Geophys. Union,76 (46), A112.

  • Bower, K. N., and T. W. Choularton, 1992: A parameterization of the effective radius of ice free clouds for use in global models. Atmos. Res.,27, 305–339.

  • ——, ——, J. Latham, J. Nelson, M. B. Baker, and J. Jensen, 1994:A parameterization of warm clouds for use in atmospheric circulation models. J. Atmos. Sci.,51, 2722–2732.

  • Brenguier, J. L., 1989: Coincidence and dead-time corrections for particle counters. Part II. High concentration measurements with an FSSP. J. Atmos. Oceanic Technol.,6, 585–598.

  • Cess, R. D., and Coauthors, 1989: Interpretation of cloud-climate feedback as produced by 14 atmospheric general circulation models. Science,245, 513–516.

  • Charlson, R. J., J. E. Lovelock, M. O. Andreae, and S. G. Warren, 1987: Oceanic phytoplankton, atmospheric sulphur, cloud albedo, and climate. Nature,326, 655–661.

  • ——, S. E. Schwartz, J. M. Hales, R. D. Cess, J. A. Coakley Jr., J. E. Hansen, and D. J. Hofmann, 1992: Climate forcing by anthropogenic aerosols. Science,255, 423–430.

  • Covert, D. S., 1988: Pacific marine background aerosol: Average ammonium to sulfate molar ratio equals 1. J. Geophys. Res.,93, 8455–8458.

  • Dabberdt, W. F., and T. W. Schlatter, 1996: Research opportunities from emerging atmospheric observing and modeling capabilities. Bull. Amer. Meteor. Soc.,77, 305–323.

  • Fitzgerald, J. W., 1975: Approximation formulas for the equilibrium size of an aerosol particle as a function of its dry size and composition and the ambient relative humidity. J. Appl. Meteor.,14, 1044–1049.

  • Frick, G. M., and W. A. Hoppel, 1993: Airship measurements of aerosol size distributions, cloud droplet spectra, and trace gas concentrations in the marine boundary layer. Bull. Amer. Meteor. Soc.,74, 2195–2202.

  • Gerber, H., 1991: Direct measurement of suspended particulate volume concentration and far infrared extinction coefficient with a laser difffraction instrument. Appl. Opt.,30, 4824–4831.

  • Hegg, D. A., 1994: Cloud condensation nucleus-sulfate mass relationship and cloud albedo. J. Geophys. Res.,99, 25 903–25 907.

  • ——, L. F. Radke, and P. V. Hobbs, 1991: Measurements of Aitken nuclei and cloud condensation nuclei in the marine atmosphere and their relation to the DMS-Cloud-Climate hypothesis. J. Geophys. Res.,96, 18 727–18 733.

  • Hobbs, P. V., 1993: Aerosol cloud interactions. Aerosol-Cloud-Climate Interactions, P. V. Hobbs, Ed., Academic Press, 33–73.

  • Hoppel, W. A., J. W. Fitzgerald, G. M. Frick, R. E. Larson, and E. J. Mack, 1990: Aerosol size distributions and optical properties found in the marine boundary layer over the Atlantic Ocean. J. Geophys. Res.,95, 3659–3686.

  • ——, G. M. Frick, J. W. Fitzgerald, and R. E. Larson, 1994a: Marine boundary layer measurements of new particle formation and the effects of nonprecipitating clouds have on aerosol size distribution. J. Geophys. Res.,99, 14 443–14 459.

  • ——, ——, ——, and B. J. Wattle, 1994b: A cloud chamber study of the effect that nonprecipitating water clouds have on the aerosol size distribution. Aerosol Sci. Technol.,20, 1–30.

  • ——, ——, and ——, 1996: Deducing droplet concentration and supersaturation in marine boundary layer clouds from surface aerosol measurements. J. Geophys. Res.,101, 26 553–26 565.

  • Hudson, J. G., 1983: Effects of CCN concentrations on stratus clouds. J. Atmos. Sci.,40, 480–486.

  • ——, and G. Svensson, 1995: Cloud microphysical relationships in California marine stratus. J. Appl. Meteor.,34, 2655–2666.

  • Kiehl, J. T., and B. P. Briegleb, 1993: The relative roles of sulfate aerosols and greenhouse gases in climate forcing. Science,260, 311–314.

  • ——, and H. Rodhe, 1995: Modeling geographical and seasonal forcing due to aerosols. Aerosol Forcing of Climate, Dahlem Workshop Report 17, R. J. Charlson and J. Heintzenberg, Eds., Wiley and Sons, 107–121.

  • King, M. D., 1993: Radiative properties of clouds. Aerosol-Cloud-Climate Interactions, P. V. Hobbs, Ed., Academic Press, 123–149.

  • Knudson, E. O., and K. T. Whitby, 1975: Aerosol classification by electric mobility: Apparatus, theory, and applications. J. Aerosol Sci.,6, 443–451.

  • Kowalski, A. S., 1996: Occult cloudwater deposition to a forest in complex terrain: Measurement and interpretation. Ph.D. dissertation, Oregon State University, 209 pp. [Available from Valley Library, Oregon State University, Corvallis, OR 97331.].

  • ——, P. M. Anthoni, R. J. Vong, A. C. Delany, and G. D. McLean, 1997: Deployment and evaluation of a system for ground-based measurement of cloud liquid water and turbulent fluxes. J. Atmos. Oceanic Technol.,14, 468–479.

  • Kulmala, M., A. Laaksonen, R. J. Charlson, and P. Korhonen, 1997:Clouds without supersaturation. Nature,388, 236–237.

  • Laucks, M., 1996: Quantifying the uncertainty in measurements of aerosol optical properties relevant to the direct shortwave forcing of climate. Ph.D. dissertation, University of Washington, 274 pp.

  • Leaitch, W. R., G. A. Isaac, J. W. Strapp, C. M. Banic, and H. A. Weibe, 1992: The relationship between cloud droplet number concentrations and anthropogenic pollution: Observations and climatic implications. J. Geophys. Res.,97, 2463–2474.

  • Liu, B. Y. H., and D. Y. H. Pui, 1974: A sub-micron aerosol standard and the primary, absolute calibration of the condensation nucleus counter. J. Coll. Interface Sci.,47, 155–171.

  • Martin, G. M., D. W. Johnson, and A. Spice, 1994: The measurement and parameterization of effect radius of droplets in warm stratocumulus. J. Atmos. Sci.,51, 1823–1842.

  • McInnes, L. M., D. S. Covert, P. K. Quinn, and M. S. Germani, 1994:Measurements of chloride depletion and sulfur enrichment in individual sea-salt particles collected from the marine boundary layer. J. Geophys. Res.,99, 8257–8268.

  • ——, P. K. Quinn, D. S. Covert, and T. L. Anderson, 1996: Gravimetric analysis, ionic composition, and associated water mass of the marine aerosol. Atmos. Environ.,30, 869–884.

  • ——, D. S. Covert, and B. M. Baker, 1997: The number of seasalt, sulfate, and carbonaceous particles in the marine atmosphere: Measurements consistent with the ambient size distribution. Tellus,49B, 300–313.

  • Nakajima, T., M. D. King, J. D. Spinherne, and L. F. Radke, 1991: Determination of the optical thickness and effective particle radius of clouds from reflected solar radiation measurements Part II: Marine stratocumulus observations. J. Atmos. Sci.,48, 728–750.

  • Noone, K. J., H.-C. Hansson, and R. K. A. M. Mallant, 1992: Droplet sampling from crosswinds: An inlet efficiency calibration. J. Aerosol Sci.,23, 153–164.

  • Noonkester, V. R., 1984: Droplet spectra observed in marine stratus cloud layers. J. Atmos. Sci.,41, 829–845.

  • Norment, H. G., 1987: Numerical studies of sampling efficiencies of the ASCME and PMS aspirator hydrometer measurement instruments. J. Atmos. Oceanic Technol.,4, 253–263.

  • Novakov, T., C. Rivera-Carpio, J. E. Penner, and C. F. Rogers, 1994:The effect of anthropogenic sulfate aerosols on marine cloud droplet concentrations. Tellus,46B, 132–141.

  • Penner, J. E., and Coauthors, 1994: Quantifying and minimizing the uncertainty of climate forcing by anthropogenic aerosols. Bull. Amer. Meteor. Soc.,75, 375–400.

  • Pruppacher, H. R., and J. D. Klett, 1997: Microphysics of Clouds and Precipitation. Reidel, 954 pp.

  • Schulman, M. L., M. C. Jacobson, R. J. Charlson, R. E. Synovec, and T. E. Young, 1996: Dissolution behavior and surface tension effects of organic compounds in nucleating cloud droplets. Geophys. Res. Lett.,23, 277–280.

  • Slingo, A., 1990: Sensitivity of the earth’s radiation budget to changes in low clouds. Nature,343, 49–51.

  • Smith, R. N. B., 1990: A scheme for predicting layer clouds and their water content in a general circulation model. Quart. J. Roy. Meteor. Soc.,116, 435–460.

  • Tang, I. N., and H. R. Munkelwitz, 1994: Water activities, densities, and refractive indices of aqueous sulfates and sodium nitrate droplets of atmospheric importance. J. Geophys. Res.,99, 18 801–18 808.

  • Twomey, S., 1977: The influence of pollution on the shortwave albedo of clouds. J. Atmos. Sci.,34, 1149–1152.

  • Vincent, J. H., 1989: Aerosol Sampling, Science and Practice. Wiley and Sons, 290 pp.

  • Vong, R. J., and A. S. Kowalski, 1995: Eddy correlation measurements of size-dependent cloud droplet turbulent fluxes to complex terrain. Tellus,47B, 331–352.

  • ——, H.-C. Hansson, H. B. Ross, D. S. Covert, and R. J. Charlson, 1988: Northeastern Pacific sub-micrometer aerosol and rainwater composition: A multivariate analysis. J. Geophys. Res.,93, 1625–1637.

  • ——, B. M. Baker, F. J. Brechtel, R. T. Collier, J. M. Harris, A. S. Kowalski, N. C. McDonald, and L. M. McInnes, 1997: Ionic and trace element composition of cloudwater collected on the Olympic Peninsula of Washington State. Atmos. Environ.,21, 1991–2001.

  • Wielicki, B. A., R. D. Cess, M. D. King, D. A. Randall, and E. F. Harrison, 1995: Misson to Planet Earth: Role of clouds and radiation in climate. Bull. Amer. Meteor. Soc.,76, 2125–2153.

  • View in gallery

    Map of the Olympic peninsula of western Washington State and location of the Cheeka Peak Atmospheric Observatory (CPO).

  • View in gallery

    Droplet number size distributions from the FSSP-100 for four cloud events. Droplet mean diameter (Dvm), number concentration (Nd), and LWC are given for each 30-min period.

  • View in gallery

    Aerosol number size distributions typical of clear air during onshore, marine flow. Consecutive 10-min spectra are shown with instrumentally defined CCN indicated as particles with Dp > 80 nm (to the right of the vertical line scribed at Dp = 0.08 μm).

  • View in gallery

    Simultaneous cloud droplet and aerosol (as Nccn) number concentrations for high LWC [(a): LWC > 0.25 g m−3] and all [(b): LWC > 0.05 g m−3] cloud periods. No corrections for droplet aspiration efficiency have been performed here. Also presented are regression lines that, for (a) and (b), respectively, have slopes of 0.90 and 0.60, Y intercepts of 167 cm−3 (both), and r2 values of 0.71 and 0.60.

  • View in gallery

    Time evolution of cloud droplet (Nd) and aerosol (as both corrected and uncorrected Ncnn) number concentrations for a 2-day period of alternately cloudy and clear air at CPO. Droplets are lost at the inlet of the aerosol stack beginning at Julian 251.0, a period with high wind speeds (Vamb = 8 m s−1) and large droplets (Dvm = 18 μm). The depicted “aspiration losses” (periods where uncorrected Nccn < Nd) generally are only important for events where Nd is low (see text).

  • View in gallery

    Simultaneous cloud droplet and aerosol (as Nccn) number concentrations for high LWC [(a): LWC > 0.25 g m−3] and low LWC [(b): 0.25 g m−3 > LWC > 0.05 g m−3] cloud periods. The values for Nccn have been corrected for droplet aspiration efficiency according to (2) and (3) in both panels. Also presented are regression lines that, for (a) and (b), respectively, have slopes of 0.91 and 0.45, Y intercepts of 111 and 134 cm−3, and r2 values of 0.80 and 0.36.

  • View in gallery

    Observed relationship between cloud droplet effective diameter (Deff) and aerosol number concentration, as Nccn, for high LWC cloud (LWC > 0.25 g m−3).

  • View in gallery

    Comparison of measured cloud droplet effective diameter (Deff) to values of Deff predicted from (1) using LWC and Nccn for high LWC cloud (LWC > 0.25 g m−3). A 1:1 line is plotted for comparison purposes.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 156 156 5
PDF Downloads 17 17 2

Simultaneous Observations of Aerosol and Cloud Droplet Size Spectra in Marine Stratocumulus

View More View Less
  • 1 College of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, Oregon
  • | 2 Department of Atmospheric Sciences, University of Washington, Seattle, Washington
© Get Permissions
Full access

Abstract

Simultaneous field measurements of aerosol and cloud droplet concentrations and droplet diameter were performed at a maritime site on the coast of Washington State. The aerosol and droplet spectra were compared for estimating cloud condensation nucleus concentration (Nccn) as the number of particles with diameters greater than 80 nm, that is, NccnN(Dp > 80 nm). Several analytical approaches were developed and applied to the data, including a stratification of the observations into periods of high and low liquid water content (LWC) based on a threshold value of 0.25 g m−3. The aerosol data were corrected for inertial losses of cloud droplets at the inlet using wind speed and droplet size; this correction improved the measured relationships between Nccn and droplet number concentration (Nd). These measurements, when coupled with the range of possible aerosol chemical compositions, imply a cloud supersaturation of 0.24%–0.31% at the Cheeka Peak sampling site during periods of high LWC.

The observations of droplet and aerosol spectra supported Twomey’s cloud brightening hypothesis in that Nccn was highly correlated (r2 = 0.8) with Nd in apparent 1:1 proportions. For the investigated range (50 cm−3 < Nd < 600 cm−3) droplet effective diameter (Deff) was very sensitive to variation in Nccn for 50 cm−3 < Nccn < 200 cm−3, somewhat sensitive for 200 cm−3 < Nccn < 400 cm−3, but not very sensitive to variation in aerosol number for Nccn > 400 cm−3. A model was applied to the aerosol and droplet data to predict droplet size, as Deff, from N−0.33ccn and LWC. Predicted values for Deff agreed (r2 = 0.8) with Deff determined directly from the cloud droplet spectra, suggesting that this approach should be useful in climate modeling for predicting cloud droplet size from knowledge of Nccn and LWC.

Corresponding author address: Dr. Richard Vong, COAS-AtS, Oregon State University, Corvallis, OR 97331-5503.

Email: vong@oce.orst.edu

Abstract

Simultaneous field measurements of aerosol and cloud droplet concentrations and droplet diameter were performed at a maritime site on the coast of Washington State. The aerosol and droplet spectra were compared for estimating cloud condensation nucleus concentration (Nccn) as the number of particles with diameters greater than 80 nm, that is, NccnN(Dp > 80 nm). Several analytical approaches were developed and applied to the data, including a stratification of the observations into periods of high and low liquid water content (LWC) based on a threshold value of 0.25 g m−3. The aerosol data were corrected for inertial losses of cloud droplets at the inlet using wind speed and droplet size; this correction improved the measured relationships between Nccn and droplet number concentration (Nd). These measurements, when coupled with the range of possible aerosol chemical compositions, imply a cloud supersaturation of 0.24%–0.31% at the Cheeka Peak sampling site during periods of high LWC.

The observations of droplet and aerosol spectra supported Twomey’s cloud brightening hypothesis in that Nccn was highly correlated (r2 = 0.8) with Nd in apparent 1:1 proportions. For the investigated range (50 cm−3 < Nd < 600 cm−3) droplet effective diameter (Deff) was very sensitive to variation in Nccn for 50 cm−3 < Nccn < 200 cm−3, somewhat sensitive for 200 cm−3 < Nccn < 400 cm−3, but not very sensitive to variation in aerosol number for Nccn > 400 cm−3. A model was applied to the aerosol and droplet data to predict droplet size, as Deff, from N−0.33ccn and LWC. Predicted values for Deff agreed (r2 = 0.8) with Deff determined directly from the cloud droplet spectra, suggesting that this approach should be useful in climate modeling for predicting cloud droplet size from knowledge of Nccn and LWC.

Corresponding author address: Dr. Richard Vong, COAS-AtS, Oregon State University, Corvallis, OR 97331-5503.

Email: vong@oce.orst.edu

1. Introduction

The presence of atmospheric aerosol is thought to affect global climate through both direct and indirect interactions with solar radiation. This “aerosol-mediated” climate forcing includes the direct scattering and absorption of sunlight by particles in clear air (Charlson et al. 1992) and the influence of a subset of particles, termed cloud condensation nuclei (CCN), on cloud albedo through their control of cloud droplet size, cloud extent, and lifetime (Charlson et al. 1987; Albrecht 1989). There is aerosol microphysical and climatological evidence to support the hypothesis that the direct aerosol effect results in a climate cooling that partly offsets the increased warming expected from absorption of outgoing infrared radiation by greenhouse gases such as CO2, CH4, N2O, and chlorofluorocarbons (Kiehl and Briegleb 1993; Kiehl and Rodhe 1995).

The indirect effect of aerosol on climate derives from the fact that, for warm clouds, cloud optical thickness, and hence cloud albedo, depends on cloud total liquid water path and droplet size (Twomey 1977; Hobbs 1993). For a fixed cloud liquid water content (LWC), droplet number concentration (Nd) and droplet diameter (Dd) are inversely related such that Nd can be thought of as a control on cloud albedo. Furthermore, since cloud droplets form through condensation of water vapor onto aerosol particles greater than some critical diameter and solubility, cloud albedo ought to vary with CCN number concentration (Nccn) for a given cloud LWC. Not enough is known of the aerosol and cloud physics parameters involved to calculate the magnitude of this effect (Penner et al. 1994).

a. Observations of Nd and Nccn

Field data from aircraft studies (Hudson 1983; Hudson and Svensson 1995; Hegg et al. 1991; Hegg 1994;Martin et al. 1994; Leaitch et al. 1992; Bower et al. 1994) have shown that Nccn and Nd are similar in nonprecipitating clouds over the world’s oceans. Hudson (1983) reported a positive correlation between Nccn and Nd for aircraft sampling of clouds west of San Diego. Hegg et al. (1991) reported that Nccn (at 1% supersaturation) and Nd were “highly correlated” in marine boundary layer (MBL) stratus sampled by aircraft off the Washington coast. Martin et al. (1994) reported a one-to-one relationship between aerosol (0.05 μm < Dp < 3 μm) and droplet concentrations from several aircraft studies over the Pacific and Atlantic Oceans (during experiments FIRE and ASTEX).

Unfortunately, aircraft investigations of this type are expensive and it can be difficult to obtain large datasets for characterizing CCN and cloud climatologies. While satellite climatologies of cloud properties do provide temporal and spatial coverage, these investigations generally need in situ observations for comparison and validation purposes (e.g., Nakajima et al. 1991).

Recent ground-based cloud studies have not supported the idea that Nccn determines Nd. Anderson et al. (1994) did not observe a good correlation (r = 0.35) between Nccn and Nd when sampling droplets at a surface site (Cheeka Peak) on the Washington coast. They concluded that meteorological rather than aerosol parameters were the major controlling factors on Nd. Furthermore, Novakov et al. (1994) observed little or no sensitivity in Nd to Nccn at a surface site in Puerto Rico.

b. Modeling the relationship between Nd and Nccn

Current efforts to model climate change use general circulation models (GCMs) that include radiative transfer calculations based on forcings associated with trace gases such as CO2, clouds, and aerosol. The treatment of clouds in these GCM models has proven to be particularly important (Cess et al. 1989; Wielicki et al. 1995). Some GCMs simulate the shortwave cooling due to the indirect effect of aerosol on clouds by essentially specifying cloud albedo and coverage (Slingo 1990), but there is increasing effort to predict cloud presence and LWC from thermodynamic variables (Smith 1995) and calculate cloud albedo based on assumed values for precloud CCN concentrations (Bower and Choularton 1992; Bower et al. 1994).

Various measures of mean droplet diameter can be used to describe droplet populations. Effective diameter (Deff) is defined by the ratio of droplet volume to surface area [Deff∫ D3dn(Dd) dDd/ ∫ D2dn(Dd) dDd, where n(Dd) is the droplet number size distribution]. Volume mean diameter (Dvm) is defined as the size of the droplet with the average volume (Dvm = [6L/(πNd)]0.33, where L is the cloud liquid water content in g m−3). Effective diameter Deff is considered to be a key integral parameter for calculating the albedo of clouds (Slingo 1990; King 1993) but is somewhat difficult to determine because in general it must be calculated from droplet size distributions. In contrast, Dvm has no direct use in determining cloud radiative properties but can be predicted theoretically from the thermodynamics of adiabatic ascent and an estimate of Nd.

Martin et al. (1994) demonstrate that these two measures of droplet size, Deff and Dvm, can be related (i.e., D3eff = kD3vm) such that the slope for the regression (k) is a characteristic of the droplet spectral shape. Martin et al. determined that, for FIRE and ASTEX clouds, k = 0.80 for marine air, but k = 0.67 for continental air.

Given Dvm calculated from LWC and Nd, an empirical relationship between Dvm and Deff, and a relationship between CCN and droplet number concentrations (e.g., NccnNd), one can determine Deff directly from LWC and Nccn (Martin et al. 1994) as
DeffLπkNccn0.33

c. Relationships among Nd, supersaturation, and aerosol composition

Besides the sensitivity of Deff and Nd to Nccn, one would like to know the water vapor supersaturation (S) for marine stratus for use in models of cloud processes. In general, values for S in a cloud must be inferred because direct measurements currently are not possible (Dabberdt and Schlatter 1996). Sampling of stratiform clouds over the North Pacific by Hudson and Svensson (1995) and Hoppel et al. (1996) suggest that 0.1% < S < 0.4%.

CCN are defined as a subset of the aerosol particles that are sufficiently large and hygroscopic to nucleate water droplets at a specified value for S. Because of the dynamic nature of clouds, CCN or nucleated droplet concentration in a cloud parcel of some average S will not necessarily be the same as instrumentally measured CCN concentration at the same S. Supersaturation in a cloud parcel is reached by a combination of adiabatic or radiative cooling and turbulent mixing from some subsaturation relative humidity (RH) over time periods of tens of seconds to minutes or more. In a thermal gradient diffusion cloud chamber the time allowed for particle and droplet growth is generally much shorter. Thus, dissolution and hygroscopic growth of particles and droplets to water vapor equilibrium may not be achieved if there are surface active films or slightly soluble compounds present. This would result in an instrumental CCN that is an underestimate of cloud CCN. Turbulent mixing in clouds may produce fast, local changes in S above and below the average S in a cloud parcel that are not represented by CCN measurements.

Estimates of CCN number, diameter, and S have been made by measuring the droplets that nucleate at various S within an isothermal haze chamber or a continuous thermal gradient diffusion cloud chamber and comparing these values of Nccn to independently observed Nd (Hudson 1983; Hudson and Svensson 1994). Anderson (1992) and Anderson et al. (1994)) define Nccn as the number of residual aerosol particles counted after sampling by a counterflow virtual impactor that rejects droplets smaller than some minimum Dd.

Recent work by Hoppel et al. (1996) suggests that Nd and the effective maximum cloud supersaturation (Smax) can be inferred from ground-based measurements of the aerosol size distribution. This approach utilizes the frequently observed separation of the aerosol submicrometric number size distribution into interstitial and cloud droplet residue modes as the basis for estimating the minimum, or critical, diameter of aerosol (Dc) that previously has activated to form a cloud droplet in upwind nonprecipitating clouds (Hoppel et al. 1994a; Hoppel et al. 1994b; Hoppel et al. 1990; Frick and Hoppel 1993).

It is important to understand the aerosol chemical composition when comparing CCN and droplets because the ability of a particle to nucleate at a given value of S depends on composition. Observations at marine sites such as the Cheeka Peak Atmospheric Observatory (CPO) (Covert 1988; McInnes et al. 1994; McInnes et al. 1996; McInnes et al. 1997) suggest that in unpolluted marine air most of the aerosol particles contain sulfur compounds such as NH4HSO4 and (NH4)2SO4 with very little hygrophobic component. Often, however, 10%–70% of submicrometric aerosol mass is not characterized as the typical sulfate, nitrate, sea salt chemical composition. Soot is not present in appreciable quantities; rather, a variety of organics are thought to compose the unanalyzed mass for marine aerosol sampled at CPO. The stoichiometry of cloudwater collected at CPO is consisent with a largely hygroscopic aerosol composition in that sulfate dominates the sum of anions while three organics contribute <25% of total anions (Vong et al. 1997). While the nucleation process takes longer for organics than sulfates (Schulman et al. 1996), only aerosol insoluble volume fractions larger than 40% would increase S by more than 0.1% for a given dry aerosol size (Dc) of interest. In contrast, the incorporation of soluble gases such as HNO3 into cloud droplets reduces the value of Smax necessary for droplet activation but this effect is important mainly for urban areas (Kulmula et al. 1997). Thus, for marine aerosol at CPO where the hygrophobic fraction is typically less than 40%, Smax is primarily a function of aerosol size (Pruppacher and Klett 1997) and, at most, weakly dependent on variations in aerosol composition.

d. Field sampling and data analysis considerations

Observed differences in the correlation between Nccn and Nd for aircraft (Hudson 1983; Martin et al. 1994; Hegg et al. 1991) and ground-based (Anderson et al. 1994; Novakov et al. 1994) sampling studies could reflect differences in the types of clouds, or portions of clouds, that are sampled by aircraft compared to ground-based cloud studies. Aircraft sampling of clouds ordinarily follow decisions about which clouds to sample and what height within a cloud to sample; these datasets generally do not include samples from cloud base, edges, or cloud top. In contrast, ground-based cloud studies typically involve the collection of microphysical data on any available cloudy time periods. In that regard a data stratification according to what portion of a cloud is actually sampled is of interest.

A related issue arises because cloud-top entrainment of subsaturated air represents a process that may, at times, particularly confound any correlation between Nd and Nccn in that it results in a dilution of cloud LWC and evaporation of some or all droplets (decreases Nd for a given Nccn). Previous analyses demonstrate that the relative variability in droplet number (Blyth et al. 1980;Baker et al. 1982) and LWC (Vong and Kowalski 1995) increase near cloud boundaries and with entrainment.

Here a simple stratification criterion is utilized for reducing the effects of cloud edge and entrainment on droplet spectra based on the fact that both LWC and the relative variability in LWC produce similar groupings of cloud events (cf. Vong and Kowlaski 1995). Cloud LWC (with various threshold values) is used to stratify cloudy periods into high (LWC > 0.25 g m −3) and low LWC clouds as a means of eliminating time periods when LWC departs from adiabatic values (i.e., periods when cloud-top entrainment, cloud edge, or cloud base are influencing the droplet spectra). Cloud events with high LWC at Cheeka Peak display smaller values for the relative variability in both LWC and Nd than do the low LWC events. A similar LWC stratification for comparing Nccn and Nd was utilized by Borys et al. (1995). The effectiveness of this simple approach to data stratification is borne out by the observations themselves, as demonstrated later.

The effects of precipitation on Nd (Albrecht 1989) are minimized in that only data for nonprecipitating (as well as “high LWC”) cloud are utilized for comparing Nccn and Nd. After these two data stratification steps, we use measurements taken near the coast of the Olympic peninsula of Washington State to attempt to provide realistic estimates of Nccn, Nd, and Deff over the world’s oceans. We then utilize these measurements to evaluate a model of the dependence of cloud droplet diameter on CCN number concentration.

2. Experimental methods

The Cloud and Aerosol Chemistry Experiment (CACHE) was conducted at the Cheeka Peak Atmospheric Observatory (CPO) on the Olympic peninsula of Washington State. The site is located at 460-m elevation and 4 km inland from the northeastern Pacific Ocean at the summit of a ridge. Three intensive measurement campaigns were conducted at CPO: 10 April–25 May 1993 (CACHE-1), 14 April–27 May 1994 (CACHE-2), and 8 August–23 September 1994 (CACHE-3); see Fig. 1 for a map of this area.

A number of physical and chemical measurements were performed to investigate cloud droplet chemical composition as a function of droplet size (Vong et al. 1997), aerosol chemical composition (McInnes et al. 1996; McInnes et al. 1997), the eddy flux of cloud droplets to a fir forest (Vong and Kowalski 1995; Kowalski et al. 1997; Kowalski 1996), CO2 eddy fluxes (Anthoni 1996), and the relationships among Nccn, Nd, and Deff.

Sampling and analysis of rainwater, aerosol, and cloudwater composition (Vong et al. 1988; Covert 1988;McInnes et al. 1996; Vong et al. 1997) have demonstrated that Cheeka Peak receives clean, marine air during onshore (westerly quadrant) flow conditions. The 4-km upwind fetch over the Washington coast traverses sparsely inhabited, undulating terrain covered by a dense silver fir forest.

Cloud droplet size spectra and droplet concentration were measured at 10 m above ground level (AGL) from a tower using an aspirated forward-scattering spectrometer probe (FSSP) (PMS model FSSP-100-ER, see Baumgardner 1983; Baumgardner et al. 1985) with an additional measure of cloud liquid water content obtained from an open path, laser-diffraction instrument (GSI model PVM-100; Gerber 1991). The FSSP was mounted on a rotatable boom extending 2 m upwind of the tower and the axis of the inlet was kept within ±5° of the measured wind direction in order to minimize inlet losses related to large droplet inertia (Norment 1987; Vincent 1989). A fan aspirates droplets into the inlet aperture at 8 m s−1; a tapered inlet increases the air velocity to 24 m s−1 in the FSSP’s sensing volume. The FSSP sensing area was increased (by 30%) to 0.8 mm2 by the manufacturer for fast acquisition of droplet spectra at low values of Nd. This laser cross-sectional area was measured prior to CACHE-1 and again between CACHE-2 and -3 to ensure accuracy in the calculated air volume that is used to determine Nd. The FSSP was sampled at 5–10 Hz using a commerical pulse height analyzer (PMS model 1058B), a 30386/387 PC, and software written in BASIC.

The PVM-100 also was mounted on the rotatable boom and oriented with its sensing axis normal to the prevailing wind direction. The PVM calibration was checked regularly using an optical filter. The calibration signal from the optical disk was previously related to LWC in a wet wind tunnel by the manufacturer (Gerber 1991). The background in the PVM LWC data was corrected by subtracting the most recent signal collected during cloud-free conditions as indicated both by visual observation and the FSSP output.

The FSSP data were converted from number to volume distributions (2 μm < Dd < 47 μm for 3-μm diameter intervals, or 2 μm < Dd < 32 μm for 2-μm diameter intervals) using the probe’s dynamic sampling volume and corrections for the probe activity (Baumgardner 1983; Brenguier 1989; Kowalski et al. 1997). The FSSP size determination was verified in the field using glass beads (accounting for the index of refraction) and confirmed in the laboratory by the University of Washington Cloud and Aerosol Physics Group in Seattle and by PMS in Boulder.

Uncertainties in the FSSP operation were quantified by repeatedly comparing its integrated LWC to that for the collocated PVM; there was excellent agreement (slope = 1.0, r2 = 0.98, the FSSP LWC was 0.02 g m−3 higher than the PVM on average; Kowalski et al. 1997) between values reported by the two instruments under field conditions for the range in Nd reported here (cf. Arends et al. 1992). These comparisons did suggest that increased LWC uncertainty exists at Nd > 600 cm−3, presumably due to coincidence errors in the FSSP (activity > 35%); at high Nd the difference between the FSSP and PVM LWC was more than 20%. The FSSP errors in Nd due to counting statistics are neglible for Nd > 50 cm−3 (Hudson and Svensson 1995). The agreement in LWC, taken in conjunction with good results from the FSSP size calibrations, is taken as evidence that the FSSP produced values of Nd that were accurate within ±10% for Nd < 600 cm−3.

Aerosol number concentration (Np) and diameter (Dp) were measured at CPO using three techniques to cover the 20-nm–10-μm diameter range of interest as potential CCN. A differential mobility analyzer (DMA; TSI model 3071) selected particles (used here for 0.02 μm < Dp < 0.5 μm) that were passed to a condensation particle counter (CPC; TSI model 3760) for determining particle number concentration (Knudson and Whitby 1975; Liu and Piu 1974). An optical particle counter (OPC; PMS LASX) sensed aerosol in the upper end of the accumulation mode (used here for 0.5 μm < Dp < 1.2 μm). An aerodynamic particle sizer (APS; TSI model 3300) sensed coarse particles (1.2 μm < Dp < 7 μm) during CACHE-2 and -3. Most of the particles that are treated as instrumental CCN in this work were counted using the DMA rather than the OPC or APS.

Aerosol number-size distributions, dN/d logD (cm−3), were measured every 10 min for Dp > 20 nm. The DMA-generated portion of the spectra was considered to be more accurate than the OPC data with respect to Dp and Np because the diameter determination is not dependent on refractive index and because the integral of the measured DMA size distribution was within ±10% of the directly measured particle concentration with Dp > 20 nm. The OPC particle sizes, initially based on the refractive index of latex calibration spheres, were corrected by comparison to a second calibration with ammonium sulfate. This refractive index correction produced a match between the OPC and DMA number concentrations for overlappping, mutually sensitive diameter intervals (0.13 μm < Dp < 0.5 μm) and corresponds well to recent refractive index measurements by Tang and Munkelwitz (1994). The APS data were corrected from aerodynamic to geometric diameter by dividing the calibration particle sizes by the square root of particle density (1.7 g cm−3), thus achieving spectral overlap with the OPC data up to 1.2-μm diameter.

Ambient air was heated 10°C or more while passing through a vertical sampling stack such that aerosol were sensed by the DMA, OPC, and APS at sizes corresponding to an RH of 20%. Aspiration of this aerosol involved sweeping 1 m3 min−1 of ambient air through a 90° turn into the stack’s inlet (20-cm diameter). This“whole air inlet” was intended to aspire and evaporate any cloud droplets (Dd < 20 μm) such that, ideally, the reported aerosol spectra ought to represent both CCN and interstitial aerosol particles.

The FSSP, PVM, and aerosol stack were located near a 3D sonic anemometer (Applied Technology model SWS-211-3K) for measurement of ambient wind speed and direction. Vertical profiles from twice-daily rawindsondes obtained for the National Weather Service (NWS) Quillayute station located 20 km south of the Cheeka Peak site were utilized for estimating cloud-top heights (Kowalski 1996). An upwind meterological station was established at 200 m above mean sea level (MSL) upwind of CPO to provide temperature, RH, winds, and net radiation; these data were used in estimating cloud-base height (Kowalski 1996) in preference to the Quillayute NWS data, which were deemed to be less representative of cloud base in the complex terrain near CPO. Precipitation and drizzle were sensed using a rain gage and a wetness detector (Vaisala DRD-11A) and supplemented by subjective methods of drizzle detection (Vong et al. 1997).

3. Data analysis methodology

a. Estimating CCN concentration from aerosol size spectra

For the relatively simple hygroscopic aerosol composition that is typical of CPO during westerly flow, the aerosol particles that serve as atmospheric CCN can be identified from aerosol spectra and an estimate of the maximum supersaturation (Smax) of the clouds. Köhler curves relate Smax to the minimum particle “dry” diameter (Dc) (Pruppacher and Klett 1997; Fitzgerald 1975; Hobbs 1993; Hoppel et al. 1996; Kulmula et al. 1997) that will nucleate to form a cloud droplet and experience condensation growth.

Aerosol size spectra, obtained as dN/d logD, were integrated to obtain Nccn for the expected values of Smax. Thus, when CCN were defined instrumentally as particles greater than 40-, 60-, or 80-nm diameter, the integration produced three estimates of CCN number concentration corresponding to values of Smax of 0.67%, 0.37%, and 0.24%, respectively, for aerosol composed of pure ammonium bisulfate (Hoppel et al. 1996). The estimates of Smax corresponding to a given Dc increase by 29% when the aerosol is composed of 40% insoluble material such as soot; under this assumed composition (an unusual case for marine air at CPO) the 80-nm Dc corresponds to a value for Smax of 0.31%.

Each estimate for N(Dp > Dc) was then compared to observed cloud droplet number concentration (Nd) and the correlations of Nd with NccnN(Dp > Dc) were determined. This method assumes that determining N(Dp > Dc) based on the best estimate of the effective ambient S (i.e., Dc) will maximize the correlation between N(Dp > Dc) and Nd.

b. Characterization of droplet inlet losses

While submicrometric aerosol particles typically can be sampled from ambient air without major artifacts due to inlet losses, cloud droplets possess sufficient inertia to make aspiration losses a concern.

Since the FSSP inlet was oriented into the wind, it was very efficient (94%–98%) in aspiring cloud droplets. The values of Nd measured by the FSSP were not very sensitive to variations in ambient wind speed (9 m s−1Vamb ≥ 3 m s−1) that are typical of CPO.

In contrast to the FSSP inlet, flow modeling of the vertical stack that served as the aerosol inlet shows that droplets (with Dd > 10 μm) are lost to a significant degree (Laucks 1996). During cloud episodes, this inlet is intended to aspire both aerosol and cloud droplets so that the aerosol instrumentation can characterize the sum of CCN (from evaporated droplets) and interstitial particles in this air.

The aspiration efficiency of the aerosol inlet for droplets for sampling at a 90° angle to the ambient wind is droplet size and wind speed dependent (Vincent 1989) and can be expressed empirically as
DdR0.5
where R = Vamb/Vstack, Vstack = vertical stack air velocity (m s−1), Stk = Stokes number = (ρD2dVamb)/(18ηLc), Lc = inlet characteristic length (diameter), η = dynamic viscosity of air, and ρ = density of water.

Measured vertical wind speed components at CPO averaged ∼1 m s−1. Thus, the droplet sampling attack angle was probably 100°–110° for the aerosol stack rather than 90°, and (2) would underestimate droplet losses under these conditions. Wind tunnel measurements by Noone et al. (1992) support the validity of (2) for 90° sampling of droplets but it has not been experimentally verified for droplets approaching at angles >90°. For these reasons, droplet losses have been calculated for 90° to approximate their magnitude.

To account for droplet losses at the aerosol stack inlet in estimates of Nccn, the quantity [1 − ε(Dd)], that is, droplets that are lost, was determined from (2) and used to correct the measured particle concentrations (Nccn) based on observed wind speeds and ambient FSSP droplet size distributions. Subsequently we compare both the directly measured and the corrected values of Nccn to uncorrected values of Nd.

4. Results and discussion

a. Cloud climatology

CPO was immersed in cloud about 40% of all hours during CACHE. Clouds were present twice as frequently at night as during afternoon hours, similar to the results of other investigations (Betts 1990; Blaskovic et al. 1991). Cloud-base altitude upwind of CPO was typically between 200 and 300 m MSL (for 50% of the events) or lower (for 45% of the events) during the high LWC episodes selected for analysis of CCN–droplet relationships. Cloud-top heights were 500–600 m MSL during summer campaign periods, consistent with the low-level marine boundary layer inversion associated with the Pacific high. During the spring and autumn cloud top varied widely from 500 m MSL with stable conditions to as high as 4000 m during periods associated with frontal passages. The typical LWC at 10 m AGL was 0.1–0.6 g m−3 with 100 cm−3 < Nd < 450 cm−3 and 11 μm < Dvm < 16 μm for most cloud events. Adiabatic LWC was observed during many high LWC events but departures (∼10% of LWC in magnitude) also occurred; periods with cloud base below 200 m prevented determination of values for adiabatic LWC for about half of all hours (Kowalski 1996).

Figure 2 presents cloud droplet size distributions for high LWC periods at CPO. The type of narrow, monomodal, droplet number distributions that are depicted for 23, 28, and 29 August were frequently observed (Vong and Kowalski 1995); such spectral shapes are considered typical of nonentraining marine clouds (Pruppacher and Klett 1997; Noonkester 1984; Martin et al. 1994). Occasionally cloud events were sampled that displayed a broader distribution with a secondary peak at smaller diameter (e.g., 31 August).

Here Deff was calculated from the droplet spectra and compared to Dvm, demonstrating the excellent correspondance (r2 = 0.98, k = 0.78 ± 0.005, zero intercept) between these two measures of drop size. The similarity of the values for k from Martin et al. (1994) (k = 0.80) and here reflects similar droplet size spectral shapes for marine air over the Atlantic and Pacific Oceans.

b. Instrumental definition of CCN

Figure 3, typical of clear air aerosol size spectra, demonstrates that the aerosol size distributions at CPO exhibit the same separation into interstitial and cloud residue modes that has been reported by Hoppel et al. (1990), Hoppel et al. (1994a), Hoppel et al. (1994b), and Hoppel et al. (1996) for marine air; at CPO the observed minimum in aerosol dN/d logD for clear air occurs at 90 nm ⩽ Dp ⩽ 100 nm. The diameter of this aerosol spectral minimum reflects the smallest particles that have been incorporated into clouds upwind of CPO according to Hoppel et al. (1996); the CACHE minima in dN/d logD occurs at a slightly smaller diameter than the 110-nm value reported by Frick and Hoppel (1993) for airship flights off the Oregon coast.

Treating the data for cloudy periods as a whole, the use of an aerosol Dc of 80 nm for defining Nccn at CPO maximized the N(Dp > Dc) correlation with Nd (high LWC cloud, Fig. 4a). No optimization of Dc for different cloud events was attempted. This value of Dc is consistent with Hoppel’s interpretation of clear air aerosol spectra such as Fig. 3. The minimum particle diameter for CCN should be smaller at CPO than that deduced from the minimum in dN/d logD because S is enhanced at the CPO site compared to offshore, upwind stratocumulus cloud fields due to orographic lifting. Thus, an“orographic effect” exists for sampling ground-based clouds in complex terrain in that the definition of particles that can act as CCN extends to smaller particles for the surface site than for the stratocumulus cloud upwind over the Pacific Ocean. Higher values of Smax are expected due to lifting and enhanced vertical velocities at sites located in complex terrain compared to upwind stratiform cloud over the oceans. This orographic effect for ground-base cloud sampling produced an estimated reduction of 10–20 nm in the minimum particle size that activates in the cloud (from 90–100 to 80 nm). Thus, aerosol particles activate at CPO to become cloud droplets at a slightly smaller dry diameter than is characteristic of the upwind stratocumulus clouds.

The selected 80-nm minimum diameter for estimating Nccn, when taken in conjunction with the primarily hygroscopic aerosol composition, implies that CCN are defined here as aerosols that activate to become cloud droplets at 0.24% supersaturation (Hoppel et al. 1996). This value represents the maximum value of S for typical cloud events at CPO. Alternatively, the clear air minimum in dN/d logD at Dp ⩽ 100 nm suggests an upwind value of Smax ≥ 0.17% over the Pacifc Ocean (based on a 100% hygroscopic aerosol composition).

The appropriateness of Smax derived from this instrumental definition of CCN depends on the precision with which Nd and Nccn are determined and the absence of substantial hygrophobic fractions within the marine aerosol. We believe that this 80-nm average value for Dc is well determined because of the relatively simple and consistent aerosol chemical composition at CPO during marine onshore flow and because comparisons of the aerosol and cloud microphysical instrumentation with other methods produced results consistent with these assumptions.

c. Effect of data stratifications

Cloud events at CPO occur with both marine and continental airmass trajectories (Vong et al. 1997) but marine clouds are more likely to have optical properties that are sensitive to changes in aerosol concentration (Charlson et al. 1987; Bower et al. 1994) than continental clouds. Furthermore, continental clouds are more likely to contain hygrophobic materials that confound our instrumental definiton of Nccn. Given that the FSSP-100 was optimized for sampling for marine clouds (better precision for Nd ⩽ 600 cm−3), the results presented herein focus on marine air where Nccn ⩽ 600 cm−3. A few observations of cloud events with high Nccn are discussed for comparison purposes with the caveat that these data contain additional uncertainties.

Prior to analysis, the aerosol and droplet data were stratified by LWC and precipitation occurrence with the goal of selecting high LWC, nonprecipitating cloud for examing Nccn relationships with Nd and Deff. A number of stratified analyses were performed with the data using various LWC thresholds (in the range 0.15 ⩽ LWC ⩽ 0.4 g m−3) for defining high and low LWC cloud. These analyses demonstrated that high LWC cloud had a more consistent relationship between Nccn and Nd than did“all” or low LWC cloud events (Figs. 4a,b). The correlation between Nccn and Nd was typically twice as high as for high LWC as for low LWC cloud; low LWC cloud tended to have lower values of Nd for a given value of Nccn. In contrast to the result for aerosol number, the LWC correlation with Nd was up to four times as high for low LWC cloud periods as for high LWC cloud (r2 = 0.15) periods using the selected 0.25 g m−3 LWC threshold.

A strong correlation of Nd with LWC (as with low LWC cloud during CACHE) can be considered symptomatic of a lack of correlation between Nd and Nccn because other processes besides nucleation become important to Nd near cloud boundaries or during periods of entrainment or drizzle. At higher altitudes in a cloud, the maximum value of Nd would have been previously reached for that cloud and Nd would display a weaker correlation with LWC than lower in the cloud. Thus, aircraft studies are more likely to see correlation between Nccn and Nd because of the sampling strategy that involves flying through the middle of a cloud, thus obtaining samples of relatively high LWC cloud (LWC may approach adiabatic) while at the surface typically all types of cloud are sampled.

The selection of the value of 0.25 g m−3 for the LWC threshold for stratifying cloud periods is arbitrary (thresholds as low as 0.15 g m−3 or as high as 0.40 g m−3 did not dramatically change these results) but represents a reasonable separation of the observations into data sets of comparable sample size but contrasting meteorological processes and statistical properties. Quantatitive assessment of relationships among Nccn, Nd, and Deff herein are based on analyses of cloud with LWC ≥ 0.25 g m−3.

After the data stratification by LWC and precipitation occurrence, a total of 397 simultaneous 10-min observations Nd and Nccn existed for high LWC cloud with 352 additional data pairs available for low LWC cloud;these 749 10-min simultaneous observations of cloud and aerosol spectra and concentrations composed most of the data analyzed herein. Certain consecutive observations were used to evaluate the aerosol stack inlet (i.e., precloud Nccn [≡N(Dp > 80 nm)] was compared to incloud Nd).

d. Nd relationships to Nccn

Figure 5 presents a time series of 10-min values of Nd and Nccn through a 2-day time period of particularly uniform meteorological conditions when CPO was alternately cloudy and clear. The relative consistency of Nccn during the clear periods within these 2 days and similarity to Nd (all ∼150 cm−3) during the periods of high LWC cloud serve as an indirect measure of the comparability of data from the aerosol and droplet instrumentation. However, the observed low Nccn values during cloudy periods are not possible and serve as a reminder that cloud droplets, and thus Nccn, sometimes are lost at the inlet before being sensed by the aerosol instrumentation. Although cloud droplets also are lost by deposition to the upwind forest (Vong and Kowalski 1995; Kowalski et al. 1997), no correction for this effect has been attempted here; such “incloud” losses ought to affect Nd and Nccn in an identical manner in this study.

When droplet losses at the aerosol stack inlet are estimated using (2) with ambient droplet size spectra and Vamb, typically 20% of Nd is lost at 4 m s−1, 35% at 6 m s−1, and about 50% for 8 m s−1 wind speeds. This droplet loss is strongly droplet size dependent with typically 10% losses for Dd = 10 μm, 25% losses for Dd = 15 μm, but 40% losses for Dd = 20 μm.

Given evidence such as Fig. 5 for the loss of droplets upstream of the particle instrumentation, (2) was applied to the ambient droplet populations for determining the stack aspiration efficiency, ε(Dd), and an improved estimate for Nccn was obtained as
NccnNDpmeasuredNdDd

Figure 5 also presents the corrected values of Nccn using (3), these corrections were intended to account for droplet losses at the aerosol whole air inlet. Although the corrected values appear to be more physically reasonable, they too are smaller than the theoretical value (Nd). In particular there is a considerable loss of cloud droplets at the aerosol stack inlet when Dd > 15 μm and Vamb > 6 m s−1 (most of the period; Julian dates 251 to 251.7 in Fig. 5).

Figure 6a presents the relationship between corrected Nccn and Nd for high LWC cloud. These “corrected” values for Nccn are considered here to be superior in that the Y intercept has now decreased substantially and the slope has increased to near the expected value of 1.0. The plot is virtually unchanged at high values of Nccn because, as will be demonstrated below, high Nccn corresponds to relatively small cloud droplets; these small droplets are aspired efficiently. Figure 6b demonstrates the Nccn relationship with Nd for low LWC cloud after correction for aspiration efficency; this slope is about half that displayed in Fig. 6a. The relationship between Nccn and Nd data presented in Fig. 6 is not wind speed dependent, presumably because of the confounding influences of a wind-speed-dependent (and imperfect) inlet efficiency correction and an expected dependence of Smax on vertical velocity.

Thus, the inlet correction compensates for large cloud droplets that cannot make the ∼90° turn into the aerosol stack while hardly affecting Nccn when the cloud droplets are small. The use of corrected values for Nccn results in a relationship with Nd that approaches, but does not completely reach, the expected result of zero CCN for no droplets. Clearly, heterogeneous nucleation in liquid water clouds requires that NccnNd and, thus, we interprete the corrected values of Nccn to more accurately reflect the CCN concentration at CPO than the uncorrected values.

While it is likely that (2) and (3) underestimate droplet losses during aspiration, in part because the wind approaches the stack at angles greater than the modeled 90°, such calculations result in increasing fractions of Nd being incorporated into the calculation of corrected values of Nccn. Thus, any “perfect correction” for inlet losses would produce values for Nccn that would be increasingly dependent on Nd itself, making comparisons of Nccn and Nd rather unrewarding. However, from the corrections that were performed it appears that the “true” relationship between Nccn and Nd must approach a situation where 1.00 ≥ slope ≥ 0.91 and the Y intercept approaches zero (if aspiration losses did not exist or could be accounted for perfectly).

e. Observed sensitivity of Deff to variation in Nccn

Figure 7 presents the relationship between measured cloud droplet diameter (Deff) and inferred CCN concentrations, that is, corrected values of Nccn. The nonlinear relationship is close to the expected form from (1) (DeffN−0.33ccn) and results in a large sensitivity of Deff to Nccn at lower particle concentrations (Nccn < 200 cm−3) but less sensitivity at the highest concentrations (Nccn > 400 cm−3). These results are consistent with the idea that Nd, and therefore cloud albedo, would be most sensitive to increasing Nccn in the cleanest marine areas but less sensitive, if at all, in continental air masses impacted by emission sources.

As mentioned earlier, a few cloud episodes at CPO occurred for values of Nccn > 600 cm−3. These data were not used in these analyses because they are not representative of marine air and because the magnitude of the errors in measurements of Nd is increased for those cases due to the FSSP’s limited dynamic range. However, those data do extend to Nccn > 1400 cm−3 and indicate that further increases in Nd do not occur as Nccn increases in this concentration range. While coincidence errors in the FSSP (Kowalski et al. 1997; Baumgardner et al. 1985) might produce such a flattening in plots of Nccn versus Nd, consideration of the atmospheric processes involved suggest additional explanations that cannot be confirmed or refuted with these data. For example, the constancy of Nd for Nccn > 600 cm−3 could be due to increasing hygrophobic fractions of the typically continental aerosol (McInnes et al. 1996) associated with high Nccn at CPO (Vong et al. 1997). Furthermore, increasing competition for water vapor occurs with high Nccn such that the peak supersaturations that are achieved could be lower than for the lower Nccn cases associated with marine air. It is not likely that measurements with any single FSSP could accurately span the entire range of Nd of potential interest; thus the response of Nd and Deff to increasing Nccn in more polluted airmasses remains an open question at this time.

f. Predicting Deff from Nccn

Equation (1) forms a basis for the prediction of effective droplet diameter from Nccn. The predicted values of Deff, shown in Fig. 8, were generally in agreement (r2 = 0.80) with Deff calculated directly from the FSSP droplet spectra. Prediction errors were less than 3 μm when the droplets were small (Deff ⩽ 15 μm) but overpredictions of up to 8 μm occur for the largest droplet sizes (Deff = 23 μm). These overpredictions reflect our inability to completely correct Nccn for aspiration losses, rather than any inherent limitation in (1), because the prediction assumes that Nccn = Nd, whereas the data show Nccn > Nd for low values of Nd (large Deff).

While it would have been possible to substitute a derived relationship (Fig. 6a) between Nccn and Nd rather than equating them (to maximize the prediction skill for CACHE data alone) for (1), such an empirical relationship would reflect sampling at CPO rather than the application of these methods on a more widespread basis. For the “low LWC” cases at CPO, Fig. 6b suggests that (Nd ≈ 0.5Nccn), contrary to the assumption made in developing (1). In these regions (e.g., at cloud edges), modelers using (1) would overestimate Deff unless they first accounted for the effects of entrainment on Nd. Given the CACHE results, it seems likely that (1) can be appropriate for predicting droplet size (as Deff) in nonprecipitating clouds for cloud events corresponding to our high LWC cases and that other cases can be addressed with some modifications to describe Nd near cloud boundaries.

5. Summary and conclusions

Field measurements in marine stratocumulus performed from a coastal, mountaintop site during CACHE were used to investigate the microphysical properties of cloud droplets and the aerosol that serve as CCN with the goal of identifying a quantitative relationship between droplet size and CCN concentrations and developing predictive relationships between these quantities.

A data stratification procedure was developed and successfully applied to reduce the effects of meteorological variability on cloud properties. The stratification has a major impact on the expected proportionality between Nd and Nccn. Specifically, nonprecipitating, high LWC (>0.25 g m−3) cloud represented the ideal condition for examining CCN and droplet number relationships because the influences of cloud base and cloud-top processes on Nd were reduced. The success of this stratification helped reconcile published differences between data from aircraft studies, where selection of appropriate cloud sampling heights is typical, and surface sites, where sampling often includes periods where the instruments are near cloud base, edges, or even cloud top. Additionally, the use of inlet efficiency corrections for Nccn improved the measured Nd relationship with Nccn.

With the demonstrated success in using LWC to select appropriate cloud conditions, ground-based sites can be utilized for examining CCN–droplet relationships given that extensive datasets can be obtained and later stratified to avoid cloud base (S < Smax) and cloud top (entrainment may reduce Nd) where meteorological variability, rather than Nccn, dominates the variability in Nd. However, it was argued that the use of data from a surface site for characterizing marine boundary layer clouds must be done with a recognition that potential orographic influences exist that may modify some of the results. After recognizing this relatively small orographic effect, the CACHE aerosol and cloud spectral data (where Dc = 80 nm) are consistent with the approach of Hoppel et al. (1994a) and Hoppel et al. (1994b) where the minimum in the aerosol number size distribution at about 90–100 nm is considered to reflect the lower size limit of particles processed by upwind marine clouds over the Pacific Ocean.

Our dataset, collected over two seasons in two different years, strongly supports the theory that droplet diameter is controlled by CCN number concentration since measured values for NccnNd for a range of droplet concentrations (50 cm−3Nd ⩽ 600 cm−3). This result is supported by the correlation between these two measurements and by the “sense” of an inlet correction that makes NccnNd for low values of Nd.

Droplet effective diameter (Deff) was very sensitive to changes in CCN concentration for Nccn < 200 cm−3, somewhat less sensitive for 200 cm−3 < Nccn < 400 cm−3, and Deff was not very sensitive to changes in CCN concentration for Nccn > 400 cm−3. A simple model of droplet size, where DeffN−0.33ccn, was investigated using measurements and found to perform quite well as a means for estimating Deff. In the process it was confirmed that two measures of mean droplet size, Deff and Dvm (defined directly from LWC and Nd), are easily interconverted for marine spectra.

Based on a method that seeks the maximum correlation between aerosol and droplet concentrations, “accumulation mode” aerosol (2 μm > Dp > 0.1 μm) were shown to compose most, but not all, of the cloud-active nuclei at CPD. Some “nuclei mode” particles (100 nm > Dp > 80 nm) also were cloud active according to the data analyses. On that basis, the marine stratocumulus sampled at CPO were shown to have exhibited microphysical properties that are consistent with a peak supersaturation of 0.24% for aerosol composed of ammoniumbisulfate but 0.31% (an upper bound for Smax at CPO) for aerosol with 40% hygrophobic material internally mixed into the same particles.

Thus, this result provides further evidence that Twomey’s (1977) cloud brightening hypothesis is viable as a first approximation of the effect of aerosol on cloud droplet concentration and diameter for nonprecipitating marine clouds. It also suggests an analysis method for minimizing potentially confounding sources of variability. Cloud-active aerosol concentrations (CCN) control the number and size of cloud droplets in high LWC marine clouds observed at CPO.

Acknowledgments

This work was supported by National Science Foundation Grants ATM9118316 and ATM93112132 (Atmospheric Chemistry desk). We thank Dave Reinart and Andy Kowalski for Figs. 1 and 2.

REFERENCES

  • Albrecht, B. A., 1989: Aerosols, cloud microphysics, and fractional cloudiness. Science,245, 1227–1230.

  • Anderson, T. L., 1992: Optimization of a counterflow virtual impactor (CVI) for studying aerosol effects on cloud droplet number. Ph.D. dissertation, University of Washington, 232 pp.

  • ——, D. S. Covert, and R. J. Charlson, 1994: Cloud droplet number studies with a counterflow virtual impactor. J. Geophys. Res.,99, 8249–8256.

  • Anthoni, P. A., 1996: Carbon dioxide eddy flux measurements in complex terrain from a coniferous forest under the influence of marine air. M.S. thesis, Dept. of Atmospheric Science, Oregon State University, 79 pp. [Available from Valley Library, Oregon State University, Corvallis, OR 97331.].

  • Arends, B. G., G. P. A. Kos, W. Wobrock, D. Schell, K. J. Noone, S. Fuzzi, and S. Pahl, 1992: Comparison of techniques for measurement of fog liquid water content. Tellus,44B, 604–611.

  • Baker, M. B., A. Blyth, D. Carruthers, S. Caughey, and T. W. Choularton, 1982: Field studies of the effect of entrainment upon the structure of clouds at Great Dun Fell. Quart. J. Roy. Meteor. Soc.,108, 899–916.

  • Baumgardner, D., 1983: An analysis and comparison of five water droplet measuring instruments. J. Climate Appl. Meteor.,22, 891–910.

  • ——, W. Strapp, and J. E. Dye, 1985: Evaluation of the forward scattering spectrometer probe. Part II: Corrections for coincidence and dead-time losses. J. Atmos. Oceanic Technol.,2, 626–632.

  • Betts, A. K., 1990: Diurnal variation of California coastal stratocumulus from two days of boundary layer soundings. Tellus,42A, 302–304.

  • Blaskovic, M., R. Davies, and J. B. Snider, 1991: Diurnal variation of marine stratocumulus over San Nicolas Island during July 1987. Mon. Wea. Rev.,119, 1469–1478.

  • Blyth, A. M., T. W. Choularton, G. Fullerton, J. Latham, C. Mill, M. Smith, and I. Stromberg, 1980: The influence of entrainment on the evolution of cloud droplet spectra. Quart. J. Roy. Meteor. Soc.,106, 821–840.

  • Borys, R. D., D. Lowenthal, and D. Mitchell, 1995: CCN regulation of snowfall rates. Eos, Trans. Amer. Geophys. Union,76 (46), A112.

  • Bower, K. N., and T. W. Choularton, 1992: A parameterization of the effective radius of ice free clouds for use in global models. Atmos. Res.,27, 305–339.

  • ——, ——, J. Latham, J. Nelson, M. B. Baker, and J. Jensen, 1994:A parameterization of warm clouds for use in atmospheric circulation models. J. Atmos. Sci.,51, 2722–2732.

  • Brenguier, J. L., 1989: Coincidence and dead-time corrections for particle counters. Part II. High concentration measurements with an FSSP. J. Atmos. Oceanic Technol.,6, 585–598.

  • Cess, R. D., and Coauthors, 1989: Interpretation of cloud-climate feedback as produced by 14 atmospheric general circulation models. Science,245, 513–516.

  • Charlson, R. J., J. E. Lovelock, M. O. Andreae, and S. G. Warren, 1987: Oceanic phytoplankton, atmospheric sulphur, cloud albedo, and climate. Nature,326, 655–661.

  • ——, S. E. Schwartz, J. M. Hales, R. D. Cess, J. A. Coakley Jr., J. E. Hansen, and D. J. Hofmann, 1992: Climate forcing by anthropogenic aerosols. Science,255, 423–430.

  • Covert, D. S., 1988: Pacific marine background aerosol: Average ammonium to sulfate molar ratio equals 1. J. Geophys. Res.,93, 8455–8458.

  • Dabberdt, W. F., and T. W. Schlatter, 1996: Research opportunities from emerging atmospheric observing and modeling capabilities. Bull. Amer. Meteor. Soc.,77, 305–323.

  • Fitzgerald, J. W., 1975: Approximation formulas for the equilibrium size of an aerosol particle as a function of its dry size and composition and the ambient relative humidity. J. Appl. Meteor.,14, 1044–1049.

  • Frick, G. M., and W. A. Hoppel, 1993: Airship measurements of aerosol size distributions, cloud droplet spectra, and trace gas concentrations in the marine boundary layer. Bull. Amer. Meteor. Soc.,74, 2195–2202.

  • Gerber, H., 1991: Direct measurement of suspended particulate volume concentration and far infrared extinction coefficient with a laser difffraction instrument. Appl. Opt.,30, 4824–4831.

  • Hegg, D. A., 1994: Cloud condensation nucleus-sulfate mass relationship and cloud albedo. J. Geophys. Res.,99, 25 903–25 907.

  • ——, L. F. Radke, and P. V. Hobbs, 1991: Measurements of Aitken nuclei and cloud condensation nuclei in the marine atmosphere and their relation to the DMS-Cloud-Climate hypothesis. J. Geophys. Res.,96, 18 727–18 733.

  • Hobbs, P. V., 1993: Aerosol cloud interactions. Aerosol-Cloud-Climate Interactions, P. V. Hobbs, Ed., Academic Press, 33–73.

  • Hoppel, W. A., J. W. Fitzgerald, G. M. Frick, R. E. Larson, and E. J. Mack, 1990: Aerosol size distributions and optical properties found in the marine boundary layer over the Atlantic Ocean. J. Geophys. Res.,95, 3659–3686.

  • ——, G. M. Frick, J. W. Fitzgerald, and R. E. Larson, 1994a: Marine boundary layer measurements of new particle formation and the effects of nonprecipitating clouds have on aerosol size distribution. J. Geophys. Res.,99, 14 443–14 459.

  • ——, ——, ——, and B. J. Wattle, 1994b: A cloud chamber study of the effect that nonprecipitating water clouds have on the aerosol size distribution. Aerosol Sci. Technol.,20, 1–30.

  • ——, ——, and ——, 1996: Deducing droplet concentration and supersaturation in marine boundary layer clouds from surface aerosol measurements. J. Geophys. Res.,101, 26 553–26 565.

  • Hudson, J. G., 1983: Effects of CCN concentrations on stratus clouds. J. Atmos. Sci.,40, 480–486.

  • ——, and G. Svensson, 1995: Cloud microphysical relationships in California marine stratus. J. Appl. Meteor.,34, 2655–2666.

  • Kiehl, J. T., and B. P. Briegleb, 1993: The relative roles of sulfate aerosols and greenhouse gases in climate forcing. Science,260, 311–314.

  • ——, and H. Rodhe, 1995: Modeling geographical and seasonal forcing due to aerosols. Aerosol Forcing of Climate, Dahlem Workshop Report 17, R. J. Charlson and J. Heintzenberg, Eds., Wiley and Sons, 107–121.

  • King, M. D., 1993: Radiative properties of clouds. Aerosol-Cloud-Climate Interactions, P. V. Hobbs, Ed., Academic Press, 123–149.

  • Knudson, E. O., and K. T. Whitby, 1975: Aerosol classification by electric mobility: Apparatus, theory, and applications. J. Aerosol Sci.,6, 443–451.

  • Kowalski, A. S., 1996: Occult cloudwater deposition to a forest in complex terrain: Measurement and interpretation. Ph.D. dissertation, Oregon State University, 209 pp. [Available from Valley Library, Oregon State University, Corvallis, OR 97331.].

  • ——, P. M. Anthoni, R. J. Vong, A. C. Delany, and G. D. McLean, 1997: Deployment and evaluation of a system for ground-based measurement of cloud liquid water and turbulent fluxes. J. Atmos. Oceanic Technol.,14, 468–479.

  • Kulmala, M., A. Laaksonen, R. J. Charlson, and P. Korhonen, 1997:Clouds without supersaturation. Nature,388, 236–237.

  • Laucks, M., 1996: Quantifying the uncertainty in measurements of aerosol optical properties relevant to the direct shortwave forcing of climate. Ph.D. dissertation, University of Washington, 274 pp.

  • Leaitch, W. R., G. A. Isaac, J. W. Strapp, C. M. Banic, and H. A. Weibe, 1992: The relationship between cloud droplet number concentrations and anthropogenic pollution: Observations and climatic implications. J. Geophys. Res.,97, 2463–2474.

  • Liu, B. Y. H., and D. Y. H. Pui, 1974: A sub-micron aerosol standard and the primary, absolute calibration of the condensation nucleus counter. J. Coll. Interface Sci.,47, 155–171.

  • Martin, G. M., D. W. Johnson, and A. Spice, 1994: The measurement and parameterization of effect radius of droplets in warm stratocumulus. J. Atmos. Sci.,51, 1823–1842.

  • McInnes, L. M., D. S. Covert, P. K. Quinn, and M. S. Germani, 1994:Measurements of chloride depletion and sulfur enrichment in individual sea-salt particles collected from the marine boundary layer. J. Geophys. Res.,99, 8257–8268.

  • ——, P. K. Quinn, D. S. Covert, and T. L. Anderson, 1996: Gravimetric analysis, ionic composition, and associated water mass of the marine aerosol. Atmos. Environ.,30, 869–884.

  • ——, D. S. Covert, and B. M. Baker, 1997: The number of seasalt, sulfate, and carbonaceous particles in the marine atmosphere: Measurements consistent with the ambient size distribution. Tellus,49B, 300–313.

  • Nakajima, T., M. D. King, J. D. Spinherne, and L. F. Radke, 1991: Determination of the optical thickness and effective particle radius of clouds from reflected solar radiation measurements Part II: Marine stratocumulus observations. J. Atmos. Sci.,48, 728–750.

  • Noone, K. J., H.-C. Hansson, and R. K. A. M. Mallant, 1992: Droplet sampling from crosswinds: An inlet efficiency calibration. J. Aerosol Sci.,23, 153–164.

  • Noonkester, V. R., 1984: Droplet spectra observed in marine stratus cloud layers. J. Atmos. Sci.,41, 829–845.

  • Norment, H. G., 1987: Numerical studies of sampling efficiencies of the ASCME and PMS aspirator hydrometer measurement instruments. J. Atmos. Oceanic Technol.,4, 253–263.

  • Novakov, T., C. Rivera-Carpio, J. E. Penner, and C. F. Rogers, 1994:The effect of anthropogenic sulfate aerosols on marine cloud droplet concentrations. Tellus,46B, 132–141.

  • Penner, J. E., and Coauthors, 1994: Quantifying and minimizing the uncertainty of climate forcing by anthropogenic aerosols. Bull. Amer. Meteor. Soc.,75, 375–400.

  • Pruppacher, H. R., and J. D. Klett, 1997: Microphysics of Clouds and Precipitation. Reidel, 954 pp.

  • Schulman, M. L., M. C. Jacobson, R. J. Charlson, R. E. Synovec, and T. E. Young, 1996: Dissolution behavior and surface tension effects of organic compounds in nucleating cloud droplets. Geophys. Res. Lett.,23, 277–280.

  • Slingo, A., 1990: Sensitivity of the earth’s radiation budget to changes in low clouds. Nature,343, 49–51.

  • Smith, R. N. B., 1990: A scheme for predicting layer clouds and their water content in a general circulation model. Quart. J. Roy. Meteor. Soc.,116, 435–460.

  • Tang, I. N., and H. R. Munkelwitz, 1994: Water activities, densities, and refractive indices of aqueous sulfates and sodium nitrate droplets of atmospheric importance. J. Geophys. Res.,99, 18 801–18 808.

  • Twomey, S., 1977: The influence of pollution on the shortwave albedo of clouds. J. Atmos. Sci.,34, 1149–1152.

  • Vincent, J. H., 1989: Aerosol Sampling, Science and Practice. Wiley and Sons, 290 pp.

  • Vong, R. J., and A. S. Kowalski, 1995: Eddy correlation measurements of size-dependent cloud droplet turbulent fluxes to complex terrain. Tellus,47B, 331–352.

  • ——, H.-C. Hansson, H. B. Ross, D. S. Covert, and R. J. Charlson, 1988: Northeastern Pacific sub-micrometer aerosol and rainwater composition: A multivariate analysis. J. Geophys. Res.,93, 1625–1637.

  • ——, B. M. Baker, F. J. Brechtel, R. T. Collier, J. M. Harris, A. S. Kowalski, N. C. McDonald, and L. M. McInnes, 1997: Ionic and trace element composition of cloudwater collected on the Olympic Peninsula of Washington State. Atmos. Environ.,21, 1991–2001.

  • Wielicki, B. A., R. D. Cess, M. D. King, D. A. Randall, and E. F. Harrison, 1995: Misson to Planet Earth: Role of clouds and radiation in climate. Bull. Amer. Meteor. Soc.,76, 2125–2153.

Fig. 1.
Fig. 1.

Map of the Olympic peninsula of western Washington State and location of the Cheeka Peak Atmospheric Observatory (CPO).

Citation: Journal of the Atmospheric Sciences 55, 12; 10.1175/1520-0469(1998)055<2180:SOOAAC>2.0.CO;2

Fig. 2.
Fig. 2.

Droplet number size distributions from the FSSP-100 for four cloud events. Droplet mean diameter (Dvm), number concentration (Nd), and LWC are given for each 30-min period.

Citation: Journal of the Atmospheric Sciences 55, 12; 10.1175/1520-0469(1998)055<2180:SOOAAC>2.0.CO;2

Fig. 3.
Fig. 3.

Aerosol number size distributions typical of clear air during onshore, marine flow. Consecutive 10-min spectra are shown with instrumentally defined CCN indicated as particles with Dp > 80 nm (to the right of the vertical line scribed at Dp = 0.08 μm).

Citation: Journal of the Atmospheric Sciences 55, 12; 10.1175/1520-0469(1998)055<2180:SOOAAC>2.0.CO;2

Fig. 4.
Fig. 4.

Simultaneous cloud droplet and aerosol (as Nccn) number concentrations for high LWC [(a): LWC > 0.25 g m−3] and all [(b): LWC > 0.05 g m−3] cloud periods. No corrections for droplet aspiration efficiency have been performed here. Also presented are regression lines that, for (a) and (b), respectively, have slopes of 0.90 and 0.60, Y intercepts of 167 cm−3 (both), and r2 values of 0.71 and 0.60.

Citation: Journal of the Atmospheric Sciences 55, 12; 10.1175/1520-0469(1998)055<2180:SOOAAC>2.0.CO;2

Fig. 5.
Fig. 5.

Time evolution of cloud droplet (Nd) and aerosol (as both corrected and uncorrected Ncnn) number concentrations for a 2-day period of alternately cloudy and clear air at CPO. Droplets are lost at the inlet of the aerosol stack beginning at Julian 251.0, a period with high wind speeds (Vamb = 8 m s−1) and large droplets (Dvm = 18 μm). The depicted “aspiration losses” (periods where uncorrected Nccn < Nd) generally are only important for events where Nd is low (see text).

Citation: Journal of the Atmospheric Sciences 55, 12; 10.1175/1520-0469(1998)055<2180:SOOAAC>2.0.CO;2

Fig. 6.
Fig. 6.

Simultaneous cloud droplet and aerosol (as Nccn) number concentrations for high LWC [(a): LWC > 0.25 g m−3] and low LWC [(b): 0.25 g m−3 > LWC > 0.05 g m−3] cloud periods. The values for Nccn have been corrected for droplet aspiration efficiency according to (2) and (3) in both panels. Also presented are regression lines that, for (a) and (b), respectively, have slopes of 0.91 and 0.45, Y intercepts of 111 and 134 cm−3, and r2 values of 0.80 and 0.36.

Citation: Journal of the Atmospheric Sciences 55, 12; 10.1175/1520-0469(1998)055<2180:SOOAAC>2.0.CO;2

Fig. 7.
Fig. 7.

Observed relationship between cloud droplet effective diameter (Deff) and aerosol number concentration, as Nccn, for high LWC cloud (LWC > 0.25 g m−3).

Citation: Journal of the Atmospheric Sciences 55, 12; 10.1175/1520-0469(1998)055<2180:SOOAAC>2.0.CO;2

Fig. 8.
Fig. 8.

Comparison of measured cloud droplet effective diameter (Deff) to values of Deff predicted from (1) using LWC and Nccn for high LWC cloud (LWC > 0.25 g m−3). A 1:1 line is plotted for comparison purposes.

Citation: Journal of the Atmospheric Sciences 55, 12; 10.1175/1520-0469(1998)055<2180:SOOAAC>2.0.CO;2

Save