A brief overview of aerosols and cloud microphysics

Aerosol Classification Size Aitken particles r < 0. 1μm large particles 0. 1 ≤ r ≤ 1μm giant particles r > 1. 0μm where r is the radius of the aerosol particle.

Primary Aerosol Formation Mechanisms • Gas-to-particle conversion • Mechanical disintegration of the solid and liquid earth surface

Aerosol Formation over the oceans • The dominant source of aerosol particles over the ocean is due to the bursting of air bubbles that are present at the ocean’s surface.

Kohler curves Variations of the relative humidity and supersaturation adjacent to droplets of(1) pure water (blue), and adjacent to solution droplets containing the following fixed masses of salt: (2) 10**-19 kg of Na. Cl (red), (3) 10**-18 kg of Na. Cl (orange), (4) 10**-17 kg of Na. Cl (brown), (5) 10** -19 kg of (NH 4)2 SO 4 (green), and (6) 10**-18 kg of (NH 4)2 SO 4 (violet). Note the discontinuity in the ordinate at 100% relative humidity (adapted with modification from Rasool (1973))

Kappa as an aerosol parameter • The hygroscopicity parameter ( ) (Petters and Kreidenweis, 2007) is a proxy for chemical composition.

Cloud condensation nuclei (CCN) • Those aerosols which serve as nuclei upon which water vapor molecules condense are known as CCN. • The most commonly used device for measuring the concentration of CCN active at a given supersaturation is thermal diffusion chamber.

A simple relationship of CCN concentration • NCCN = c. Sk where NCCN = number of active CCN/cm 3.

Summary of CCN measurements • Continental air masses are richer in CCN than maritime air masses • Concentrations of CCN increase with supersaturation as expected. • Remote ocean air contains the fewest CCN • Typically, NCCN ≈ 100 cm− 3 or fewer at 1% supersaturation for maritime air masses • CCN concentrations also vary considerably with height. In general, one would expect CCN concentrations to diminish with height away from the ground in the troposphere as the air mass gets cleaner.

Primary CCN composition • (NH 4)2 SO 4 or NH 4 Cl are dominant over land. • Even over oceans at (s− 1) <~ 1%, chloride containing particles constitute less than 1% to the observed CCN concentration. • Sodium-containing particles may comprise the majority of CCN in marine fogs and stratus – i. e. low supersaturation clouds – (s− 1 < 0. 1%). • DMS GTP conversion to sulfates is the most important source of CCN over the oceans.

Size of CCN • Nucleation theory suggests the best CCN are large (greater than 0. 1 μm) and soluble thus we expect that CCN are giant or large aerosol. • However, experiments by Twomey in which air was filtered showed the CCN concentration did not fall off until filtered to about 0. 02 μm or well into Aitken range.

Growth of Population of Droplets by Condensation • The time rate of change of the supersaturation can be written as: • d. S/dt = P − C, where S is the supersaturation, P a production term and C a condensation term • d. S/dt = Q 1 dz/dt − Q 2 dχ/dt • DX/dt = d. M/dt]condensational growth rate X CCN]activated/ρa

Summary of simple model calculations • Maximum supersaturation is reached approximately 40 meters above cloud base. At this point the rate at which moisture is being made available by lifting (cooling causing condensation) is equal to the rate at which vapor is condensing onto CCN and droplets. Hence, the cloud droplet concentration is reached at this point and is equal to the number of CCN that were activated (passed over the peaks in their respective Kohler curves). Above this point this supersaturation decreases as the moisture is removed by condensation faster than it can be supplied due to lifting. Consequently, no more CCN are activated above this level. In cumulus clouds the maximum supersaturation is normally found within a few tens of meters above cloud base.

• The smaller droplets at activation grow more rapidly than the larger(initially) droplets. This, of course, is due to da/dt~ 1/a. Note that droplet growth is largest near the peak supersaturation as expected. AT 100 s the droplet size distribution becomes rather uniform as expected from condensational growth theory.

• The smallest aerosols become haze droplets initially since the supersaturation did not achieve large enough values for their activation. They subsequently evaporate in the presence of the large activated droplets. • Higher W’s lead to higher peak S’s and higher droplet concentrations for the same population of aerosols

Collision-Coalescence Growth • We define the collision efficiency E 1 as the ratio of the actual collision cross-section y 2 to the geometric cross-section (a 1 + a 2)2: E 1 =y 2 / (a 1 + a 2)2 • The collection kernel, K(a 1, a 2), is defined as K(a 1, a 2) = (a 1 + a 2)2(v 1 − v 2)E 1 E 2.

Consider a spectrum of droplets having a concentration (n(v)) of droplets of volume v. n(v, Δt) = n(v, 0) +∂/∂t(n(v, t))Δt , where Gain: Loss: where vc = v − v =implies v = vc + v. Note that ∂n(v)/∂t ~ n(vc)n(v) or n(v). This would imply the higher the droplet concentration the higher the rate of collection! But. . . for a given LWC higher concentrations →smaller droplets, smaller E’s.

Note if droplets are very small E is small Thus if CCN concentrations are large, for a given LWC, droplets are small and so is E

Giant cloud condensation nuclei (GCCN) • The aerosol distribution contains a number of giant or ultragiant aerosol particles which can act as the embryos for further coalescence growth (Houghton, 1938). • Feingold et al. (1999) performed large-eddy simulations of marine stratocumulus clouds using a detailed twomoment bin-resolving microphysics model with and without GCCN. • They found that observed GCCN concentrations of 10 -4 cm− 3 to 10− 2 cm− 3 are sufficient to transform nonprecipitating stratocumulus into a precipitating one, when CCN concentrations were in the range of 50 to 250 cm− 3. At lower CCN concentrations, drizzle is active anyway and the addition of GCCN had little impact.

• Thus GCCN at in opposition to enhanced CCN • CCN concentration variability dominates and GCCN moderates influence of CCN

Summary • In simplest terms, clouds most likely to produce warm rain are clean maritime and warm-based. The maritime, warm-based cloud has a lower concentration of activated CCN and, therefore, for a given LWC is more likely to have a broader droplet spectrum and produce a few big droplets. • A warm-based cloud has a large value of cloud-base saturation mixing ratio and, therefore, its potential for condensing a significant amount of LWC is greater. Such clouds are referred to as colloidally unstable clouds. • By contrast a continental, cold-based cloud activates a larger concentration of CCN and has a smaller saturation mixing ratio at cloud base. Thus, continental, cold-based clouds have a lower potential for liquid water production and must distribute its more limited LWC over more droplets further lowering its potential for creating precipitation. Such clouds are referred to as colloidally stable clouds.

How does air pollution impact a cloud’s ability to produce warm rain? • In general polluted air is characterized by higher values of CCN. In some polluted air CCN activated at 0. 75% supersaturation may be 2000 to 2500 cm− 3. The implication is that polluted air is colloidally stable and it should depress a cloud’s ability to produce warm rain. • However, are the clouds warm- or cold-based? Are the aerosols produced by pollution small and hygroscopic or is there a large number of GCCN (coalescence embroyos)? Does the anthropogenic activity also affect cloud dynamics? –produce bigger longer lasting clouds?

Aerosols and the ice-phase • Heterogeneous ice nucleation can occur by the following modes: • Vapor deposition nucleation • Condensation-freezing or sorption • Immersion-freezing • Contact nucleation

Concentrations of Ice Nuclei • Historically, ice nuclei concentrations have been assumed as predictors of ice particle concentrations. Laboratory studies in the 1940’s and 1950’s using systems for low temperature processing of ice nuclei collected on filters indicated that the variability in ice nuclei concentration was strongly a function of temperature.

Fletcher [1962] empirical formula:

Meyers et al

Cotton et al. [2003] generalization • Ni = NIN exp [12. 96 (Si - 1)], where NIN is a model forecast variable

Contact Ice Nuclei • Nic = exp(a + b(273. 15 − Tc))

De. Mott et al IN formula

History of ice nucleation parameterizations with basis in CFDC data • Meyers et al. (1992): nin= exp(12. 96(Si-1)- 0. 639) Ice supersaturation dependence only (no links to aerosol properties) • Phillips et al. (2008): (aerosol surface area, composition, Si, T) Lab based Scaling to “baseline” a. X = fdust, f. BC, fbio corrections IN conc. and sfc. area (PDA 08) • De. Mott et al. (2009): ( T, naer>0. 5 mm diameter) October 29, 2009 AAAR Annual Meeting Minneapolis, MN

• where a = 0. 00001968, b = -0. 0167(Tk 273. 15) + 0. 2877, Tk is cloud temperature in degrees • Kelvin, n-sub-aer, 0. 5 is the number concentration (scm-3) of aerosol particles with diameters larger than • 0. 5 μm, and n. IN, Tk is ice nuclei number concentration (std L-1)

Sources of Ice Nuclei • • In many of these studies, the dominant particle type that the ice crystal nucleated upon was a silicate particle(ie. desert dust). Clay particles Volcanic eruptions; not all volcanos are sources of IN. Studies indicate that Pacific volcanoes do not contain suitable particles in their eruptions to nucleate the ice phase. These volcanos emit molten lava as opposed to silicate particles as Mt. St. Helens did. Some combustion products. Industrial processes – steel mills, aluminum works, sulfide works, power plants. Many of these are sources of metal oxides which are good nucleating agents of ice. Certain strains of bacteria serve as excellent freezing nuclei. Schnell (1976) showed that ground up plant litter introduced into bulk water samples that were processed through drop freezing devices served as excellent freezing nuclei. The source of activation was a single strain of aerobic bacteria. Certain types of pollen serve as IN

Homogeneous freezing • Homogeneous freezing can occur at temperatures below roughly -33 C and by -40 C virtually all pure water drops freeze • The concentrations of pure water drops freezing is thus dependent upon the concentration of activated CCN • Moreover, homogeneous freezing can take place on deliquesced haze particles and therefore is related to concentrations of haze particles

Ice Multiplication • It has become increasingly evident that concentrations of ice crystals in “real” clouds are not always represented by the concentrations of IN measured or expected to be activated in such environments. In particular, it has been found that at temperatures warmer than -10◦C, the concentration of ice crystals can exceed the concentration of IN activated at cloud top temperature by as much as three or four orders of magnitude.

Ice Crystal Growth Mechanisms • Growth by Vapor Deposition • Growth by Riming • Aggregation

Vapor deposition • Vapor deposition is dependent upon the concentrations of activated IN or homogeneously frozen droplets/haze particles along with supersaturation with respect to ice, ice habits(temperature and SSi). • Riming growth is dependent upon the size and fall velocity of ice particles and liquid water content of droplets

Riming • The riming growth process is a collisioncoalescence process, analogous to the collisioncoalescence growth of liquid cloud droplets. The rate of growth of a single ice crystal of mass xi can be thus described as • E(xi /x) is the collection efficiency between the ice crystal xi and the cloud droplet x, f(x) is the number spectral density of cloud droplets of mass x ± δx.

Riming Efficiencies • Shown in Figure 10. 22 are theoretical efficiencies between plates and droplets. The plate radii for this example range from 147 μm to 404 μm (~ 0. 5 mm). • Note that since the droplet spectrum is dependent on the CCN population, E is dependent on the concentrations of CCN. • Other things being the same, the higher the CCN concentrations the smaller the cloud droplets, and the lower will be the collision efficiencies during riming. • Thus riming is more efficient in clean CCN environments and wet storms.

• Aggregation is a result of collision and coalescence of ice crystals • It is roughly a function of the square of the concentrations of ice crystals, their sizes and their stickiness •