PUBLICATIONS
Journal of Geophysical Research: Atmospheres
RESEARCH ARTICLE
10.1002/2016JD025082
Key Points:
• Numerical simulations of squall lines
that occurred during MC3E are used to
assess aerosol impacts on anvils
• Increases in aerosol concentration
exert nonmonotonic trends on mean
condensate and anvil mixing ratio
profiles
• Aerosol loading promotes monotonic
increases in anvil albedo and decreases
in cloud top longwave cooling
Correspondence to:
S. M. Saleeby,
[email protected]
Citation:
Saleeby, S. M., S. C. van den Heever,
P. J. Marinescu, S. M. Kreidenweis, and
P. J. DeMott (2016), Aerosol effects on
the anvil characteristics of mesoscale
convective systems, J. Geophys. Res.
Atmos., 121, 10,880–10,901, doi:10.1002/
2016JD025082.
Received 11 MAR 2016
Accepted 12 AUG 2016
Accepted article online 15 AUG 2016
Published online 26 SEP 2016
Aerosol effects on the anvil characteristics
of mesoscale convective systems
S. M. Saleeby1, S. C. van den Heever1, P. J. Marinescu1, S. M. Kreidenweis1, and P. J. DeMott1
1
Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado, USA
Abstract
Simulations of two mesoscale convective systems (MCSs) that occurred during the Midlatitude
Continental Convective Clouds Experiment were performed to examine the impact of aerosol number
concentration on the vertical distributions of liquid and ice condensate and the macrophysical, microphysical,
and radiative properties of the cirrus-anvil cloud shield. Analyses indicate that for an increase in aerosol
concentration from a clean continental to a highly polluted state, there was an increase in the rime collection
rate of cloud water, which led to less lofted cloud water. Aerosol-induced trends in the cloud mixing ratio profiles
were, however, nonmonotonic in the mixed phase region, such that a moderate increase in aerosol
concentration produced the greatest reduction in cloud water. Generally, less lofted cloud water led to less anvil
ice mixing ratio but more numerous, small ice crystals within the anvil. In spite of reduced anvil ice mixing ratio,
the anvil clouds exhibited greater areal coverage, increased albedo, reduced cloud top cooling, and reduced net
radiative flux, which led to an aerosol-induced warming (reduced cooling) effect in these squall lines.
1. Introduction
Mesoscale convective systems (MCSs) are ubiquitous, organized precipitation systems that occur in the tropics and midlatitudes. Over North America they may account for up to 70% of the warm season precipitation
[Fritsch et al., 1986]. They often form to the lee of the Rocky Mountains in association with preexisting convection or over the central Great Plains along surface boundaries such as the dryline [Jirak and Cotton, 2007].
MCSs play a critical role in the global energy budget by vertically redistributing heat and moisture [Cotton
et al., 1995]. Such vertical transport of moisture and condensate can lead to the generation of enormous cirrus anvil cloud shields [Maddox, 1980].
The most prolific type of MCS is the persistent elongated convective system (PECS) [Jirak et al., 2003], of which
the leading line, trailing stratiform (LLTS) squall line is the most common [Parker and Johnson, 2000]. They are
contrived of complexes of individual convective elements that form in an organized manner and exhibit a
contiguous stratiform-anvil region. Such squall lines are frequently long-lived as they persist in a selfreinforcing balance between vertical wind shear and cold pool dynamics [Rotunno et al., 1988].
Given the frequency of occurrence, size, and duration of MCSs, their associated upper level anvils can impose
a great influence on the radiation budget [Ramanathan and Collins, 1991; Fowler and Randall, 1994; Stephens,
2005; Stephens et al., 2008]. The expansive cirrus anvil clouds can be highly reflective, thus increasing albedo
and reducing downwelling shortwave radiation. Further, anvils tend to trap longwave radiation below their
cloud base and emit radiation to space from their high-level cloud tops. Both the macrophysical and microphysical properties of anvils, including areal coverage, ice water content, cloud thickness, and ice particle size
distributions, determine the associated radiation budget [Platt and Harshvardhan, 1988; Platt, 1989; Carrió
et al., 2007]. These cirrus radiative responses generally promote cloud top cooling and cloud base warming
[Webster and Stephens, 1980].
©2016. American Geophysical Union.
All Rights Reserved.
SALEEBY ET AL.
Squall line MCSs are composed of multiple elements including convective, stratiform, and anvil regions that
transition from deep anvil to thin cirrus [Feng et al., 2011]. Cloud radiative feedbacks can vary dramatically by
cloud type, and within such MCSs that contain a spectrum of cloud types, there may be a range of radiative
responses [Futyan et al., 2005; Chen et al., 2000]. Further, the radiative forcing of anvil clouds may vary depending
on the cloud microphysical and macrophysical characteristics [Chou et al., 2002; Lin et al., 2002; Futyan et al.,
2005; Morrison and Grabowski, 2011]. MCSs are not well represented in general circulation models (GCMs),
which may pose a serious limitation in establishing radiative balance and representing precipitation. As such,
one component of this study is to numerically simulate and investigate the overall mean vertical distribution
of ice and liquid condensate and their impact on the radiative forcing of midlatitude squall lines.
AEROSOL EFFECTS ON MCS ANVILS
10,880
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
Given the influence of condensate size spectra on the radiation budget of anvil clouds, we also seek to
examine the impacts of cloud droplet nucleating aerosols on squall lines. Over the U.S. Great Plains, surface
aerosol concentrations can be quite high and are influenced by anthropogenic surface emissions [Hand,
2011]. Long-range transport of biomass burning aerosols from Central America into the central plains has also
been shown to be a contributor to the aerosol population [Gebhart et al., 2001; Kreidenweis et al., 2001; Ferrare
et al., 2006]. Aerosols may act as cloud condensation nuclei (CCN) and ice nucleating particles (INP) and modify
the cloud droplet spectrum, which influences precipitation, the efficiency of vertical transport of condensate,
and upper level ice spectra. Given the level of uncertainly in aerosol indirect effects [Solomon et al., 2007],
understanding their range of influence on cloud microphysical processes, the vertical distribution of
water/ice, and the radiative impacts of MCS anvils is critical. Such knowledge is necessary for the development
of MCS parameterizations that need to consider condensate vertical distributions [Del Genio et al., 2012].
Satellite studies of convective clouds influenced by biomass burning aerosols indicate an increase in cloud
cover and colder cloud tops and allude to potential radiative effects [Lensky and Rosenfeld, 2006; Lin et al.,
2006]. Koren et al. [2010] examine aerosol effects on tropical anvils from satellite data and note that more polluted anvils are found at higher altitudes, which create a warming effect at the top of atmosphere (TOA).
However, such studies are limited in determining the details of the microphysical processes that link aerosols
to large-scale cloud properties. As such, numerical models are necessary tools for investigating causal relationships between aerosols and microphysics.
In modeling studies of deep convection, a change in aerosol concentration has been shown to impact hydrometeor size distributions which, in turn, perturb the rates of microphysical processes that influence the vertical
distribution of water/ice and precipitation rates [e.g., van den Heever and Cotton, 2007; van den Heever et al.,
2006, 2011; Carrió et al., 2007; Fan et al., 2007, 2010a; Tao et al., 2007, 2012; Morrison and Grabowski, 2011;
Morrison, 2012; Seigel et al., 2013; Storer et al., 2010; Storer and van den Heever, 2013; Lebo, 2014]. A change
in the anvil ice particle size may directly impact the albedo. A shift toward more numerous smaller ice particles,
through lofting and homogeneous freezing of more numerous smaller cloud droplets, tends to increase the
anvil albedo in a manner similar to the Twomey [1977] effect for liquid cloud droplets. Further, hygroscopic
aerosols impact precipitation formation processes [Albrecht, 1989], which determine whether precipitating
condensate is more efficiently deposited at the surface or lofted within updrafts and detrained at higher levels.
Previous modeling studies that discuss aerosol indirect effects on the upper level ice and anvil clouds produced by convection largely focus on isolated or idealized tropical deep convection [e.g., van den Heever
et al., 2006; Carrió et al., 2007; Tao et al., 2007; Fan et al., 2010a; Morrison, 2012] or idealized squall line environments [e.g., Khain et al., 2005; Tao et al., 2007; Seigel et al., 2013; Lebo, 2014]. There are fewer 3-D modeling
studies of aerosol indirect effects on observed squall lines [e.g., Lynn et al., 2005; Li et al., 2009; Tao et al., 2016]
due to their complex nature, and these tend to focus on precipitation responses rather than upper level cloud
microphysics [e.g., Li et al., 2009; Lynn et al., 2005].
Of the cited modeling studies that examine aerosol effects on upper level ice, some indicate that an
increase in aerosol concentration over a given range leads to monotonic increases in ice water content
and number concentration in the anvils [van den Heever et al., 2006; Carrió et al., 2007], while others denote
nonmonotonic responses to increasing aerosol concentration [Fan et al., 2007; Li et al., 2008; Carrió et al.,
2010; Loftus and Cotton, 2014]. Additionally, Khain et al. [2005] and Carrió et al. [2007] found an increase
in anvil lifetime via the persistence of small ice particles with low settling velocities. The aerosol effect on
anvil size, however, is less consistent among studies. Fan et al. [2010a] and Khain et al. [2005] report varying
increases in anvil size, while van den Heever et al. [2006] indicate smaller convective anvils with an increase
in aerosol concentration. With respect to radiative effects, Carrió et al. [2007] and Lee et al. [2009] indicate
changes in radiation associated with changes in ice characteristics of deep convective anvils. While much
knowledge has been gained from previous modeling studies of aerosol effects on deep convection, there
has been less emphasis on understanding the aerosol effects on hydrometeor characteristics, microphysical
processes, and cloud radiative properties in observed long-lived MCSs that can impact regional climate
through changes in net radiation flux.
This study seeks to gain understanding into the aerosol influence on MCS cirrus anvil radiative properties by
investigating the impacts of aerosol on hydrometeor characteristics and microphysical processes that generate
changes in cloud albedo, cloud top longwave radiation, and net radiative flux. This is accomplished through
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,881
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
numerical simulations of two squall line MCS
case studies using the Regional Atmospheric
Modeling System (RAMS) with a bin-emulating,
two-moment bulk microphysics scheme. The
simulated MCSs occurred during the
Midlatitude Continental Convective Clouds
Experiment (MC3E) [Jensen et al., 2016] in
May–June 2011. From these simulations we
seek to address the following questions with
respect to the overall bulk characterization of
squall line MCSs: (1) how does an increase in
the number concentration of cloud droplet
nucleating aerosols impact the mean vertical
distributions and spectra of liquid and ice
hydrometeors, (2) what are the dominant
microphysical processes involved in potential
Figure 1. Model grid configuration, with Grids 1, 2, and 3 at grid
changes to the hydrometeor profiles and how
spacings of 30 km, 6 km, and 1.2 km, respectively. Analyses are
performed for Grid 3.
do these change with aerosol loading, (3) how
do potential changes in upper level condensate induced by aerosol effects impact the characterization of the cirrus anvil shield and the associated radiative
effects, and (4) are the aerosol effects monotonic or nonmonotonic?
2. Model and Case Study Description
2.1. Model Description
The open source Colorado State University RAMS model [Cotton et al., 2003] version 6 was run over the central U.S. and southern Great Plains in a three-nested grid framework with successive grid spacing of 30 km,
6 km, and 1.2 km (Figure 1). A summary of the model configuration and a general description of the physics
packages used for simulations in this study are given in Table 1.
The RAMS microphysics model prognoses mass mixing ratio and number concentration of cloud droplets,
drizzle, rain, cloud ice, snow, aggregates, graupel, and hail [Walko et al., 1995; Meyers et al., 1997]. Aerosol
activation/cloud droplet nucleation is simulated according to Saleeby and Cotton [2004] and Saleeby and
van den Heever [2013]. Aerosol activation depends upon the environmental temperature, vertical velocity,
and aerosol concentration, size, and solubility. Aerosols are allowed to undergo nucleation scavenging and
wet and dry deposition, and they may be restored to the environment upon hydrometeor evaporation or
sublimation. The aerosol activation scheme is not limited to cloud base but, rather, permits secondary
Table 1. Summary of RAMS Model Grid Setup and Simulation Configuration
Model Aspect
Grid
Setting
Arakawa C grid [Mesinger and Arakawa, 1976]
Three nested grids:
Grid 1: 130 × 115 points, Δx = Δy = 30 km, Δt = 30 s
Grid 2: 302 × 227 points, Δx = Δy = 6 km, Δt = 7.5 s
Grid 3: 997 × 647 points, Δx = Δy = 1.2 km, Δt = 3.8 s
Initialization
Boundary conditions
Land surface model
Cumulus parameterization
Radiation scheme
Turbulence scheme
Microphysics scheme
SALEEBY ET AL.
60 Vertical Levels Δz = 50 m lowest level stretched
to 500 m aloft Model top at ~22 km AGL
GDAS-FNL reanalysis data for 20 May event RUC model analysis data for 23 May event
Lateral boundary nudging from gridded reanalysis [Davies, 1983].
LEAF-3 [Walko et al., 2000b]
Kain-Fritsch parameterization [Kain and Fritsch, 1993] on Grid 1
Two streams; hydrometeor sensitive [Harrington, 1997]
Horizontal diffusion via Smagorinsky [1963]; vertical diffusion via Mellor and Yamada [1974]
Two-moment bin-emulating bulk microphysics for eight hydrometeor species [Meyers et al., 1997; Saleeby and Cotton, 2004; Saleeby
and van den Heever, 2013]
AEROSOL EFFECTS ON MCS ANVILS
10,882
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
activation (activation of aerosols in cloud) as well, which can be important in deep convection. The flexibility in
this scheme stems from the ability of the model to track aerosol number and mass, which includes removing
aerosols upon activation and allowing for their regeneration upon hydrometeor evaporation/sublimation.
Unactivated aerosols, including regenerated aerosols, may be present in updrafts through transport upward
from cloud base or through entrainment. These aerosols would readily activate in the high-supersaturation
environment in convective updrafts.
The RAMS microphysics simulates cloud droplet autoconversion and accretion [Tzivion et al., 1987]; heterogeneous ice nucleation [DeMott et al., 2010]; homogeneous freezing of droplets [DeMott et al., 1994]; contactfreezing ice nucleation [Meyers et al., 1992]; secondary ice multiplication [Hallett and Mossop, 1974]; riming
of cloud droplets [Saleeby and Cotton, 2008]; heat and vapor diffusion [Walko et al., 2000a]; hydrometeor sedimentation [Feingold et al., 1998]; and shedding, melting, and ice aggregation [Walko et al., 1995; Meyers et al.,
1997]. Many of these microphysical processes, including riming and sedimentation, are represented by a binemulating approach in which the bulk hydrometeor gamma distributions are decomposed into size bins,
treated in a size specific manner for a given microphysical process, and then reassembled into gamma distributions. For example, the riming process uses size-specific collection efficiencies so that droplet collection in
the large bins is treated separately from droplet collection in the small bins [Saleeby and Cotton, 2008]. This
method in a bulk microphysics scheme offers a more realistic representation of microphysical processes without the cost of a full bin microphysics scheme. The RAMS model is freely available by personal request to the
authors or by download via the van den Heever Research Group webpage which can be accessed through
the CSU Atmospheric Science webpage.
For the suite of sensitivity experiments that are investigated in this study, the initialization of aerosol number
concentration was varied from clean continental conditions to highly polluted continental conditions. The
initial surface aerosol concentration was guided by surface-based CCN concentrations observed at the
Department of Energy's Atmospheric Radiation Measurement Program's Southern Great Plains site (ARM-SGP).
Figure 2a displays a time series of CCN number concentration measured at various supersaturations from 22 to
24 May 2011 during MC3E. The two shaded boxes indicate times of relatively clean and polluted conditions
observed during this time period while excluding the most highly polluted observations. Representative values
of 600 and 2000 cm 3 at the highest measured supersaturations (0.85% and 1.08%, respectively) were thus
denoted as clean and polluted conditions in this modeling study. To further explore the aerosol parameter
space, additional simulations, based on the highest CCN concentrations from the time series in Figure 2a, were
initialized with concentrations up to 4000 cm 3. Since our aerosol initialization was guided by surface point
observations, the aerosols were initialized horizontally homogeneously with an exponential decrease with
height (Figure 2b). The surface aerosol concentration represents the initial maximum value in the column.
The full suite of aerosol sensitivity simulations for each MCS event was initialized with maximum number concentrations of 600, 2000, 3000, and 4000 cm 3 (Figure 2b) and are denoted as CLE, POL2K, POL3K, and POL4K,
respectively, where CLE indicates a more clean profile and POL represents a more polluted environment. In this
study the aerosols mentioned above behave only as CCN; aerosols acting as INP are initialized in a separate
category to isolate the effects of hygroscopic aerosols, acting as CCN, on anvils.
The INP initialization vertical profile was guided by aircraft data from University of North Dakota (UND) Citation
flights during MC3E (Figure 2c). While observed INP were included in this study and heterogeneous ice nucleation (immersion-freezing) was active [DeMott et al., 2010], the initial INP profile was held constant in order to
isolate the effects of aerosols acting as CCN. The INP profile was initialized horizontally homogeneously with
the concentration of particles > 500 nm diameter, as measured by the aircraft observations (Figure 2c). INP
were treated diagnostically, such that new ice particle formation was only generated if the INP concentration
exceeded the grid cell concentration of cloud ice particles. There were no INP sources or sinks.
2.2. Case Study Description
The two MCS squall line events simulated and analyzed in this study occurred from 03:00 UTC to 15:00 UTC 20
May and 21:00 UTC 23 May to 09:00 UTC 24 May of 2011 during MC3E. These cases will, hereafter, be referred
to as the 20 May and 23 May events. These two particular MCSs have been discussed in a number of other
studies that examine particular elements of squall line convection [Tao et al., 2013; Lang et al., 2014; Fan
et al., 2015; Marinescu et al., 2016]. The 12 h analysis time period in each MCS event encompasses the main
development, mature, and dissipation stages of these MCSs [Marinescu et al., 2016]. Simulations were
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,883
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
initialized several hours prior to these
analysis periods in order to allow the
model to spin-up prior to the development of the initial squall line convection. In both events, convection
developed along a dryline in western
Oklahoma, grew upscale into LLTS
MCSs, and propagated in a mainly
eastward direction across the eastern
Oklahoma border.
Time snapshots of simulated and
observed cloud coverage, in Figure 3,
demonstrates that the model was
able to produce the time-evolving
cloud shields with representative
areal coverage and location. The
simulations analyzed herein are the
same set of RAMS simulations that
have been compared to observations
in more detail in the companion
paper by Marinescu et al. [2016],
wherein, they provide a thorough
comparison between these same
RAMS model simulations and the
squall line observations. They demonstrate that while some inconsistencies
exist, the simulations are able to accurately reproduce many of the
observed features of the squall lines,
including evolution, propagation,
precipitation rate, radar reflectivity,
and vertical velocity.
Figure 2. (a) Time series of CCN active at supersaturations of approximately
0.25, 0.3, 0.45, 0.65, 0.7, 0.85, and 1.08% from the ARM-SGP site, respectively,
from N_CCN_1 to N_CCN_7. (b) RAMS cloud droplet nucleating aerosol
initialization profiles. (c) Vertical profiles of large-aerosol number concentration, which may serve as INP, from UND-Citation flights during MC3E and the
RAMS model profile for INP initialization.
Since the simulations were able to
reproduce squall lines in both events
that exhibit many similarities in
space, time, and magnitude, we will
proceed in presenting key microphysical and dynamical properties and
processes that are modulated by
aerosol loading; further, we will
examine how these modulations
impact the anvil microphysical and
radiative characteristics.
3. Results of Aerosol Experiments
The following sections discuss a number of the dynamic and microphysical quantities representative of the
squall lines for both cases (20 May and 23 May). The majority of these quantities are plotted as spatial
and/or temporal averages across the domain during the 12 h period that encompasses the majority of the
lifetime of the MCSs. A presentation of MCS characteristics in this manner and an examination of the changes
in the bulk MCS characteristics as a complete entity, due to changes in aerosol concentration, helps to isolate
the strongest signals of aerosol influence over the entirety of the MCS and its lifecycle. While the focus is on
the bulk MCS characteristics, there are some analyzed fields that are more clearly understood when the
means are computed separately for the convective and stratiform regions of the squall lines as described
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,884
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
Figure 3. Simulated cloud cover (from simulation CLE) and GOES IR imagery at several times during the lifecycle of the two MCS events. Columns are labeled with the
associated dates and times, and rows are labeled to indicate panels displaying observed or simulated fields.
below. All vertical profiles and time series averages involving cloud, ice, and total condensate mixing ratios
include grid cells in which the associated mixing ratios exceed 0.01 g kg 1 so as to eliminate sampling
cloud-free regions.
3.1. Squall Line Separation Technique
For each simulation, the columns across the horizontal domain were partitioned into categories of convective
and stratiform, as is done in the model-based separation technique in Marinescu et al. [2016]. This methodology follows that from Churchill and Houze [1984], Lang et al. [2003], and Feng et al. [2011] with some minor
threshold modifications. Grid columns are considered convective if at least one of the following conditions
is satisfied: (1) precipitation rate > 25 mm h 1 or (2) |W| below the melting level > 3.0 m s 1, or (3) |W| above
the melting level > 5.0 m s 1 or (4) cloud mixing ratio below the melting level > 0.5 g kg 1 and W > 0.5 m s 1
or (5) a grid point in the column has a precipitation rate > 2 times that of the horizontal mean within a radius
of 20 km plus the precipitation rate > 1 mm h 1. A grid column is considered stratiform if it is not a convective
column, and the precipitation rate is > 0.05 mm h 1. For simplicity, the melting level here is taken as the
T = 273 K level. The results from this separation technique are used as needed in the analyses that follow. It
is clearly stated when the cloud column separation is applied; all other analyses are representative of the bulk
MCS characteristics.
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,885
Journal of Geophysical Research: Atmospheres
1
10.1002/2016JD025082
1
Figure 4. (a,d) Cloud droplet mixing ratio (g kg ), (b, e) number concentration (mg ), and (c, f) mean diameter (μm) for the 20 May (Figures 4a–4c) and 23 May
(Figures 4d–4f) cases. Shaded areas indicate the elevation zone inclusive of the cloud water trend reversal relative to CLE. The freezing levels are denoted by the
horizontal dashed lines.
3.2. Vertical Distribution of Cloud and Rain Properties
Differences in the number concentration of cloud droplet nucleating aerosols most directly impact the microphysics of any cloud system through modification of the cloud droplet distribution. All other changes to
microphysical fields and processes stem from this fundamental change in the droplet spectra, excepting
any direct changes due to the number of ice nucleating particles, which are not treated here. Figure 4 displays the vertical profiles of cloud droplet mixing ratio, number concentration, and mean diameter for both
MCS events. In both cases an increase in aerosol loading leads to a monotonic increase in cloud droplet concentration and a decrease in droplet size at all levels. The relationship between cloud droplet and aerosol
number concentration can be nonlinear depending on the activation fraction. This effect should be considered in the interpretation of responses in hydrometeor characteristics and microphysical processes to aerosol
loading. However, from examining the cloud number profiles (Figures 4b and 4e), the change in the maximum droplet number among sensitivity tests scales similarly to the differences in the initial maximum aerosol
concentrations. Thus, nonlinear activation processes may be playing only a minor role compared to other
nonlinear microphysical interactions.
The response to increasing aerosol in the cloud mixing ratio profiles is more complex. Below 5–6 km in both
events, there is a monotonic increase in cloud water mixing ratio with aerosol number concentration, which
results from an increase in the nucleation and vapor growth of new cloud droplets in the lower, warm-phase
portion of the clouds. Deep convection can produce large supersaturations, and an increase in aerosol number concentration allows more droplets to nucleate; the greater integrated surface area of the droplet population that is composed of more numerous droplets in high-aerosol environments supports greater total
vapor deposition growth through consumption of supersaturation.
In the upper portions of the liquid cloud (above ~5–6 km), however, this trend is conditionally reversed and
nonmonotonic, meaning that cloud water content decreases with a change in aerosol number from CLE to
POL2K; however, it then increases from POL2K to POL4K, while still remaining less than CLE. Since there is
a trend toward increasing cloud water mixing ratio at lower levels in response to an increase in aerosol concentration, the term “trend reversal” will refer to a decrease in associated condensate in response to aerosol
loading relative to the CLE case. The vertical location of the lowest noted altitude in the trend reversal in
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,886
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
cloud water varies among the aerosol sensitivity tests. The vertical zone over which the trend reversal in cloud
water begins (when moving upward from the surface) for all of the aerosol tests, is gray shaded in Figure 4.
The precise level at which the trend reversal occurs is a function of cloud water scavenging processes that are
discussed below. It should be noted that above ~9.5 km nearly all cloud droplets have homogeneously
frozen, thus leaving few grid points with liquid cloud in the sample average. As such, there is some noise
or biased peaks in the cloud water profiles at the highest few levels of the liquid cloud.
Since the cloud droplet number and mean diameters from these experiments follow consistent trends with
height while the mixing ratio trend changes with height, there must be variability in the competing source
and sink processes for the generation, transport, and/or scavenging of cloud water that is linked to differences in aerosol concentration. This will be explored in the next sections.
Changes in cloud droplet distributions and cloud water content due to aerosol number differences subsequently affect rain formation. Changes in the cloud droplet spectra directly impact the cloud droplet selfcollection and accretion processes for rain formation and growth [Tzivion et al., 1987], which in turn, impact
the mean profiles of cloud characteristics. However, rain may also be generated through melting of ice and
shedding from hail. These processes may vary between the convective and stratiform regions of the squall
line, and thus, it is pertinent to examine the mean rain characteristics, shown in Figure 5, separately for
these regions.
The rain mixing ratio trends are less apparent throughout the column compared to cloud water trends
between cases and between convective and stratiform regions. This lack of trend may indicate complex nonlinear interactions of competing processes (droplet self-collection, droplet accretion, shedding, melting, and
evaporation) that can impact rain. The only clear trend in mixing ratio exists below 2 km in both regions of the
20 May case. In this layer, there is greater rain water with increasing aerosol concentration. This is likely
related to the greater variation in rain number and size below 2 km in the 20 May case compared to the
23 May case.
In general, the production of rain water above the freezing level (dashed line on Figures 4a and 4d) is attributed to droplet self-collection and accretion of cloud water. The general rapid increase in rain mixing ratio
from ~4 km down to ~2 km is related to the sum of droplet self-collection, droplet accretion processes, shedding of rain from hail, and melting of ice particles below the freezing level. The reduction in rain mixing ratio
from ~2 km down to the surface is related to evaporation below cloud base.
The vertical profiles of rain droplet number concentration and size clearly demonstrate an aerosol effect
through largely monotonic trends for each case and region sampled. This is most noticeable in the rain number concentration profiles of the convective region in both MCS events. It should be noted that at the highest
levels of the liquid cloud (9–10 km), some variability arises in rain drop size and number due to sampling of
fewer rainy grid points, thus creating the sharp transition to zero values at the very top of the liquid cloud.
Generally, an increase in aerosol concentration leads to production of fewer but larger rain drops at nearly
all levels where liquid persists. While droplet self-collection is slowed due to smaller cloud droplet size in
the more polluted simulations, the fewer rain drops that form through cloud droplet self-collection have a
greater population of droplets to collect during rain sedimentation. This relationship between aerosol
concentration and rain drop number and size has been observed and simulated in other studies as well
[e.g., Altaratz et al., 2008; Berg et al., 2008; May et al., 2011; Saleeby et al., 2010, 2015; Sheffield et al.,
2015]. Above the freezing level, generation of rain drops is largely attributed to droplet self-collection and
accretion processes, which would tend to be greater at higher altitudes where cloud droplets are larger
and more readily collected. Below the freezing level and above cloud base, shedding and melting complicate
the determination of the exact source of rain drop formation. Below cloud base, shedding, melting, and
evaporation all impact rain drop number concentration in a net nonlinear manner.
The general increase in mean rain drop diameter from the top downward is likely attributed to accretion
growth of rain drops above cloud base. Below cloud base, evaporation impacts mean sizes. The CLE simulations have the smallest rain drops. These smaller drops are more likely to fully evaporate before reaching the
surface compared to the larger drops in the more polluted simulations. This evaporation of the smaller droplets in CLE leads to the continued increase in mean rain size. The more polluted simulations display a neutral
or slight decrease in rain size below cloud base. The overall evaporation across the size spectra must, therefore, be imposing a greater impact on the mean size than the full evaporation of the smaller droplets.
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,887
Journal of Geophysical Research: Atmospheres
1
10.1002/2016JD025082
1
Figure 5. Rain drop mixing ratio (g kg ) (a, d, g, and j), number concentration (L ) (b, e, h, and k), and mean diameter (mm) (c, f, i, and l) for the 20 May (Figures 5a–
5c and 5g–5i) and 23 May (Figures 5d–5f and 5j–5l) cases partitioned into averages over convective (Figures 5a–5f) and stratiform (Figures 5g–5l) regions. Estimated
mean cloud base is denoted by dashed lines on mixing ratio profiles.
While the changes in the cloud water spectra due to aerosol loading certainly impact the rain number and
size spectra, it is not clear that they impact the rain mixing ratio, whose effects could be masked by microphysical nonlinear interactions. Further, the rain mixing ratio profiles do not display a clear trend reversal as in the
profile of cloud mixing ratio; hence, it is helpful to (1) examine the total condensate and its vertical transport
in order to help shed some light on the trend reversal in cloud water and (2) isolate the mechanism leading to
reduced cloud water aloft in the more polluted scenarios.
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,888
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
An aerosol impact is present in the vertical distribution of total condensate within the MCSs. Figures 6a and
6b depicts the mean vertical profiles of total condensate mixing ratio. Both profiles display an increase in total
condensate at low levels for an increase in aerosol number. Above this lower level, the trend reversal is
weakly present from 5 to 12 km in the 20 May case and is more prominent in the 23 May case from 6 to
12 km. In the 23 May case, there exists a layer of neutral trend from 12 to 14 km. In both events there is a clear
trend reversal at the uppermost levels. The trend reversals vary between the two events, but they are present,
nonetheless. It is important to note that the trend reversal is greatest for the increase in aerosols from CLE to
POL2K. Additional increases in aerosol concentration beyond POL2K appear to reduce the magnitude of the
trend reversal and promote an increase in mixing ratio relative to POL2K.
3.3. Vertical Transport of Hydrometeors
To understand why there is less cloud water and total condensate mixing ratio above the level of the trend
reversal in the more polluted regimes, we first examine the difference in upward transport of total hydrometeor condensate relative to the CLE simulations (Figures 6c and 6d). The profiles from both MCSs reveal an
increase in upward condensate transport with aerosol concentration from near the surface up to ~6 km. In
the lower portion of the trend reversal layer (gray bar in Figures 6c and 6d), the change in condensate vertical
transport in the polluted scenarios is still positive in spite of a reduction in total condensate/cloud water. This
occurs since the removal of cloud condensate is not yet great enough to offset the positive effects of upward
condensate transport. Within the upper portion of the trend reversal layer, riming removes cloud water at a
greater rate than what is added via upward transport. These competing processes generate the observed
trend reversal layer in cloud water/total condensate. Above ~6 km, there is either neutral or reduced upward
condensate transport for an increase in aerosol concentration. These trends suggest that the aerosol-related
deficit in condensate aloft (Figures 6a and 6b) is linked to a reduction in vertical transport of condensate generated at lower levels.
With regards to vertical transport, the aerosol-induced change in sedimentation fall speed of hydrometeors
should also be considered. However, given that much of the vertical transport within MCSs occurs in convective cores, we suspect that changes in the net fall speed due to changes in hydrometeor size will be rather
small. This concept is also argued in Lebo [2014]. Further, fall speed changes would already be realized in
the instantaneous condensate mixing ratios values used to compute vertical transport. Hence, the representation of vertical transport of condensate (Figures 6c and 6d) is a good approximation of the ability of the
MCS to loft condensate.
Such changes in the trends in upward transport of condensate are associated with either changes in the
dynamics or microphysics of the system. Figures 6e and 6f displays vertical profiles of the percent difference
in updraft speed between CLE and the three polluted scenarios averaged over grid cells with upward motion
greater than 1 m s 1. The POL3K and POL4K profiles reveal modest increases in updraft strength in both
events on the order of a few percent, up to at least 12 km, as aerosol concentration increases. Though not
the focus of this study, aerosol influences on vertical motion have been attributed to warm and cold phase
invigoration processes [Khain et al., 2005; Koren et al., 2005; Khain et al., 2008; Koren et al., 2010; van den
Heever et al., 2011; Tao et al., 2012; Saleeby et al., 2015; Sheffield et al., 2015] as well as cold pool dynamics
and condensate loading [Khain et al., 2005; Storer et al., 2010; Lebo, 2014; Grant and van den Heever, 2015;
Storer and van den Heever, 2013].
At elevations above 12 km the difference in updraft speed becomes either negative (20 May) or less positive
(23 May), which suggests less invigorative support in the upper levels. The profiles of total condensate mixing
ratio (Figures 6a and 6b) and upward transport of condensate (Figures 6c and 6d) display evidence for
reduced lofting of condensate to the upper anvil levels in the more polluted cases. This reduces the potential
for ice vapor growth, which may reduce latent heating, buoyancy, and updraft strength at these uppermost
levels under more polluted conditions.
When comparing the two MCS events, one should note that the 20 May event is not as deep as the 23 May
event as evidenced by less condensate at the uppermost levels (Figures 6a and 6b). Further, the trend reversal in cloud water mixing ratio is more prominent and over a greater depth in the 20 May event (Figures 4a
and 4d). With considerably less condensate transport to the highest elevations, the 20 May event appears
more susceptible to changes in upper level updraft strength that may be linked to latent heating and
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,889
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
1
Figure 6. (a, b) Total condensate mixing ratio (g kg ), relative differences (%) in (c, d) vertical transport of total condensate
and (e, f) mean upward vertical motion for the 20 May (Figures 6a, 6c, and 6e) and 23 May (Figures 6b, 6d, and 6f) cases.
1
Shaded areas are same as Figure 4. Figures 6e and 6f were sampled over updrafts stronger than 1.0 m s .
buoyancy. As such, the mean change in updraft strength above 12 km in the 20 May event is negative for all
polluted scenarios (Figures 6e and 6f).
Regardless of the mechanism for the generation of stronger vertical motions from the surface to 12 km, the
upward condensate transport over most of the column in both events tends to be positively supported by
stronger updrafts. Since (1) the upward condensate transport trend with aerosol concentration (Figures 6c
and 6d) is positive up to 6 km and is neutral or negative above this, (2) the trend in cloud water
(Figures 4a and 4d) is positive below ~6 km and reverses above this level, and (3) the trend in updraft strength
(Figures 6e and 6f) is positive up to at least ~12 km, then the reversal in the trend in total condensate at ~6 km
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,890
Journal of Geophysical Research: Atmospheres
Figure 7. Riming rate of cloud water (g kg
Figure 4.
1
min
10.1002/2016JD025082
1
) for the (a) 20 May and (b) 23 May cases. Shaded areas are same as
(Figures 6a and 6b) is most likely linked to increased sinks or reduced sources in microphysical processes
rather than to changes in the dynamics.
3.4. Microphysical Processes
The budgets of the key microphysical processes were examined to help determine why cloud water mixing ratio
and total condensate demonstrate a positive trend with aerosol concentration below 6 km and a reversal in this
trend aloft. Of the microphysical processes examined (though not shown), the collection of cloud droplets by ice,
or riming, is the most influential process in cloud water removal. Previous studies also indicate that riming is one
of the most influential microphysical processes in deep convection [Morrison, 2012; Seigel et al., 2013; Storer and
van den Heever, 2013; Lebo, 2014; Marinescu et al., 2016]. Figure 7 displays a clear monotonic increase in the
amount of rimed cloud water as aerosol concentration is increased. Lebo [2014] also noted a similar aerosol
impact on the amount of rimed cloud water. The greatest difference in maximum riming rate occurs between
CLE and POL2K, with reduced increases in riming rate for additional aerosol loading. The elevation of greatest
riming rate and aerosol impact on riming varies slightly between MCS events, but it is generally maximized near
the top of the trend reversal layer. In spite of greater maximum riming rates (Figure 7) in the 23 May case, the
cloud water trend reversal (Figures 4a and 4d) is of smaller magnitude than in the 20 May case This indicates that
compared to the 20 May event, the riming process on 23 May does not scavenge cloud water at a fast enough
rate to offset the cloud water generation. Hence, the trend reversal is less dramatic in the 23 May event, which
implies that a relatively larger amount of cloud water is lofted to the upper levels on 23 May.
The riming rate response to aerosol loading indicates greater cloud water removal, which leads to less lofting
of water to the upper levels. It is important to note that an increase in the riming rate is both a function of the
total cloud water and the efficiency of riming. Both factors contribute to the riming profiles (Figure 7) and the
trend reversal profiles in cloud water (Figures 4a and 4d). Since the trend reversal in the cloud water profile
above 6 km is nonmonotonic for an increase in aerosol concentration, while the riming rate trend is generally
monotonic, additional factors beyond the amount of cloud water available to rime must be influencing the
cloud water profiles.
The total riming rate is a function of the size-dependent collection efficiency between ice and cloud droplets
and the total cloud water content. If liquid water content is constant among simulations, greater aerosol concentration would lead to more numerous, smaller droplets with smaller collection efficiency and reduced riming rates. Within our simulated events, however, in the more polluted scenarios, both the liquid water content
and the number concentration of droplets increase below the trend reversal level. This allows for only a moderate decrease in the mean cloud droplet size (Figures 4c and 4f), which translates to a moderate decrease in rime
collection efficiency. When the number concentration of cloud droplets increases by a greater amount than
does the decrease in collection efficiency with smaller droplets, then the riming rate will increase rather than
decrease. In our cases, this tends to occur in the middle to upper levels of the liquid cloud where mean cloud
droplet diameter is greater than ~20 μm (Figures 4c and 4f). Collection efficiencies between ice and cloud
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,891
Journal of Geophysical Research: Atmospheres
Figure 8. (a, d) Cloud ice mixing ratio (g kg
(Figures 8d–8f) cases.
1
), (b, e) number concentration (mg
10.1002/2016JD025082
1
), and (c, f) mean diameter (μm) for the 20 May (Figures 8a–8c) and 23 May
droplets change less dramatically between diameters of 20 and 40 μm than for smaller droplets [Wang and Ji,
2000; Saleeby and Cotton, 2008]. Saleeby and Cotton [2008] demonstrated steep slopes in the collection efficiency curves for droplets smaller than 20 μm and shallower slopes for droplets greater than 20 μm diameter.
The increase in the number concentration of droplets dominates over the decrease in collection efficiency when
the droplet sizes are relatively large, and total riming increases. Beyond the POL2K simulation, however, the droplets become smaller than 20 μm at all but the highest levels of the liquid cloud, and the decrease in collection
efficiency is more dramatic and dominates over the additional increase in droplet number; hence, there is a
weaker trend reversal in cloud water mixing ratio in POL3K and POL4K relative to POL2K. In spite of reduced
trend reversals in cloud water from POL2K to POL4K, the total riming rate (Figure 7) increases. Recall that the riming rate is both a function of riming efficiency and available cloud water to be rimed. The riming efficiency will
begin to decrease as droplets become smaller (i.e., POL2K to POL4K), but there is more generation of cloud water
(Figures 4a and 4d) with greater aerosol concentration in these cases and more numerous droplets to be rimed.
The number concentration of ice species may also impact riming rates. Though not shown, the number concentration of aggregates scales with the number concentration of cloud ice particles (Figures 8b and 8e) since
collisions between cloud ice particles generates aggregates. As such, in the more polluted regimes, there are
more numerous but smaller aggregates available to collect cloud droplets. Hail is also a prolific rimer, and the
number concentration of hail, not shown, tends to scale with the number of rain drops (Figure 5) since rain
drops act as hail embryos. As such, there are slightly fewer hail stones in the more polluted scenarios for collecting rime. While variability in the rimer species impacts riming rates, such variability is less substantial than
the aerosol-induced variability in cloud water and cloud droplet number and size.
Ultimately, the sources (primarily condensation) and sinks (primarily riming) of cloud water and the changes
to the droplet spectra impact the mixing ratio and spectra of cloud ice in the upper levels through
vertical transport.
3.5. Anvil Cloud Ice Characteristics
Under the assumption of unaltered INP concentrations, changes in the upper level cloud ice are highly influenced by changes in the cloud droplet spectra below and the transport of water to upper levels, where the
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,892
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
homogeneous freezing of cloud droplets is the most prolific source of cloud ice particles [Heymsfield et al.,
2005; Tao et al., 2012; Fan et al., 2010b]. Figure 8 displays the vertical profiles of cloud ice mixing ratio, number
concentration, and mean diameter. In both MCS events, there is an increase in cloud ice number concentration and decrease in size in association with the increase in aerosol number concentration. This change in the
cloud ice size distribution tends to reduce the generation of snow and aggregate mixing ratio in the anvil
clouds in the more polluted scenarios (not shown). However, aggregates that form from cloud ice selfcollection in more polluted conditions tend to be more numerous but smaller and with slower fall speeds.
This may influence the size and lifetime of anvil clouds.
The vertical profiles of cloud ice mixing ratio are determined by a balance between upward transport of condensate and subsequent precipitation of cloud ice. Cloud ice mixing ratio is reduced at all levels for POL2K
(Figures 8a and 8d). This result is most prominent in the 20 May event, which also exhibits the greater reduction in cloud water in the upper portions of the liquid cloud (Figures 4a and 4d). Recall that the reduction in
cloud water mixing ratio with increasing aerosol concentration within the upper levels of the liquid cloud was
most dramatic for an aerosol concentration increase from CLE to POL2K. Further increases in aerosol concentration greater than POL2K do not lead to either greater or monotonic reductions in cloud ice mixing ratio in
comparison to the CLE simulations. In the 20 May event, there is little variability in cloud ice mixing ratio
among the polluted simulations, while in the 23 May event the anvil cloud ice content increases, with respect
to CLE, from 12 to 15 km for aerosol simulations POL3K and POL4K. The pronounced reduction in upper anvil
level ice content in the 20 May event is likely linked to the more pronounced cloud water trend reversal
(Figure 4a) compared to the 23 May event (Figure 4b). When the cloud water mixing ratio profiles are not sufficiently modified by the riming removal process, then more cloud water is able to be transported to the
anvil levels.
As previously noted, the riming process in the model is represented in a bin-emulating manner [Saleeby and
Cotton., 2008]. This allows the larger droplets in the spectra (containing the majority of the mass) to be more
readily rimed than the smaller droplets (containing the majority of the number concentration), which have
lower collection efficiencies. This realistic representation of the riming process across the droplet spectra permits the numerous, small droplets to be vertically transported to the anvil and freeze. Meanwhile, scavenging
of the more massive droplets leads to less cloud mixing ratio transport to the anvil and less cloud ice mixing
ratio. Hence, there is a simultaneous reduction in cloud ice mixing ratio in the anvil alongside an increase in
cloud ice number concentration (of smaller ice crystals).
While there are many similar trends in cloud water and cloud ice mixing ratio profiles in response to increasing aerosol concentration, a blanket conclusion regarding the impact of these changes on anvil ice content is
difficult to make. The final outcome largely depends on the initial modification to the cloud droplet distribution and the strength of the microphysical processes that may prevent cloud water from being transported to
upper levels. The balance among changes in upward transport of condensate, precipitation rates, and microphysical sources and sinks determines the amount of anvil cloud ice mixing ratio. As mentioned earlier, there
are clear trends in anvil cloud ice number and size, which influence precipitation rates out of the anvil. Smaller
particles in the more polluted simulations have slower fall speeds and thus longer residence times. This
directly impacts the areal coverage of anvil level clouds.
3.6. Cloud Area
Figure 9 displays vertical profiles of the cloudy fraction of the domain and the percent change in the cloud
areal coverage relative to the CLE simulation. The profiles of cloud cover percentage (Figures 9a and 9b) place
the vertical locations of the maximum areal cloud coverage at just under 12 km for the 20 May case and just
over 12 km for the 23 May case. As such, we refer to the upper anvil as > 12 km and lower anvil as < 12 km.
The change in total squall line cloud coverage (Figures 9a and 9b) due to aerosol loading can be inferred from
the differences at the level of maximum cloud coverage (~12 km). Differences at this level and computations of
total cloud cover (not shown) indicate a net increase of up to only 1–2%, and the trend relative to CLE is not
monotonic. However, changes in cloud cover are more substantial in the upper anvil region in both events
(Figures 9c and 9d). Trends in the cloud cover percentage differences (Figures 9c and 9d) in the upper anvil
are largely monotonic with increasing aerosol. In both MCS events, at least a portion of the vertical profile of
areal coverage of upper anvil ice displays a substantial percentage increase in the more polluted scenarios.
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,893
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
Figure 9. (a, b) Vertical profiles of domain cloud fraction and (c, d) percentage difference in cloud coverage relative to CLE
for the 20 May (Figures 9a and 9c) and 23 May (Figures 9b and 9d) cases. Red dashed lines denote level of maximum cloud
cover as determined from Figures 9a and 9b, and blue dashed lines denote level of maximum cloud cover change as
determined from Figures 9c and 9d.
The 23 May event exhibits up to a 150% increase in anvil cloud area at the highest levels. The 20 May event peaks
with a 50% increase near 13 km, but it exhibits a decrease at the highest levels. The peak differences occur at
vertical levels where there are generally fewer clouds present (<10% domain coverage). However, there are still
significant cloud area increases closer to altitudes of 12 km, where the largest anvil extent is present.
The decrease in cloud cover at cloud top in the 20 May event is likely indicative of less anvil ice reaching the
upper levels. As mentioned previously, the 20 May event displays a greater cloud water loss in the upper
levels of the liquid cloud due to more effective cloud water removal by riming in the most polluted scenarios.
The updraft strength is also weaker above 11 km in the more polluted scenarios, thus providing less support
for upper level divergence and anvil ice detrainment at the highest altitudes. The increase in anvil area
between 12 and 15 km for more polluted cases occurs in spite of this effect and is thus, largely attributed
to longer residence times of smaller ice crystals in the polluted cases. The 23 May event compared to the
20 May event shows a greater increase in upper anvil cloud area, with respect to change in aerosol concentration, since this event contains more condensate throughout the column (Figures 6a and 6b) and more than
2 times the amount of cloud ice mixing ratio and number (Figure 8). As such, the divergent anvil in the 23 May
event has a greater chance of persisting with time and distance from the convective core before succumbing
to sublimation and dissipation. Further, in the 23 May case, the updrafts are stronger aloft in the more
polluted cases (Figure 6f). Hence, there is a likely contribution from enhanced upper level divergence
supplementing the influence of longer residence time of smaller ice crystals on the upper anvil area in the
more polluted simulations.
Fan et al. [2010a] also examine the modeled changes in anvil sizes induced by aerosol loading, but within
deep tropical convective storms, and Koren et al. [2010] examine aerosol effects on anvil sizes from a satellite
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,894
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
perspective. While our simulated conditions are different from these studies, they all reveal an increase in
anvil size with increased aerosol concentration. However, as noted above, the total cloud cover increased
only 1–2% while Fan et al. [2010a] reveal anvil increases of 15–100%, and Koren et al. [2010] show a substantial increase in the ratio of anvil to convective area. However, we do demonstrate a 25–150% increase in the
area of upper level anvil. It should be noted that, comparatively, our threshold ice water path for anvil cloud
identification was more constrained and less inclusive of very thin anvil, and our polluted simulations led to a
reduction in anvil ice mixing ratio rather than an increase, which is noted in Fan et al. [2010a]. In spite of the
differences in simulations and analyses methods between these studies, both demonstrate an increase in
anvil area with aerosol loading, though the magnitude of this increase is dependent on the threshold of
anvil cloud.
3.7. Impacts on Radiative Properties
Modifications of the anvil areal coverage and the ice crystal spectra can impact the radiative properties of the
upper level clouds, including cloud top albedo, outgoing longwave radiation, and radiative heating rates
[Platt, 1989; Jensen et al., 1994]. It has been shown thus far that in the more polluted scenarios the anvil level
cirrus tends to be composed of less total cloud ice mixing ratio but more numerous, smaller ice crystals with
larger extent.
The change in the ice crystal size spectra can impact the cloud albedo in a similar manner to Twomey [1977].
Figures 10a and 10f display plots of the mean squall line cloud top albedo. Trends are similar for both MCSs
and indicate a monotonic increase in cloud top albedo of up to 4–6% for greater aerosol concentrations due
to the presence of more numerous, smaller ice particles. This greater reflection of solar radiation imposes a
cooling effect. The amount of upwelling infrared radiation that is emitted to space at cloud top is also
impacted by the amount of upper level ice and the ice particle size spectra. Figures 10b and 10g display
the TOA outgoing longwave radiation (OLR) averaged over cloudy columns. There is a clear monotonic trend
in both events of reduced TOA-OLR with up to 6–8% reduction for the greatest aerosol concentration. This
response imposes a warming effect by trapping more upwelling longwave radiation. The decrease in TOAOLR may be tied to reduced emission of radiation from a higher elevation (colder temperature) due to
persistence of smaller ice crystals in the upper anvil. It may also be reduced by an increase in cloud opacity
resulting from more numerous, smaller crystals. Morrison and Grabowski [2011] also demonstrate an aerosolinduced increase in cloud top albedo and decrease in TOA-OLR that are of similar magnitude to the results
shown here. One key difference is that our results demonstrate these radiative responses in spite of reduced
anvil ice mixing ratio in more polluted cases, thus suggesting that changes in the anvil ice size spectra may be
more important than anvil ice mixing ratio.
The differences in upwelling and downwelling shortwave and longwave radiation combine to generate the net
vertical profile of radiative heating. Figures 10c and 10h display the vertical profiles of the radiative heating
rates averaged over the cloudy columns. The profiles in both events are dominated by cooling at upper levels.
Thick anvil and deep convection tends to exert a net cooling effect [Futyan et al., 2005; Chen et al., 2000]. The
aerosol effect on the heating rate profiles reveals an increase in elevation of the heating/cooling profile above
~10 km AGL. There is also an increase in the maximum cooling rate in the upper anvil region and an increase in
maximum heating in the lower anvil region with higher aerosol concentrations. These changes are shown more
clearly by the radiative heating difference profiles (Figures 10d and 10i).
The net radiative effect of aerosol modulation of the cirrus anvil ice can be quantified by examining the difference relative to CLE in the net column radiative flux (net TOA upward flux net surface upward flux),
which is shown in Figures 10e and 10j. The aerosol indirect effects tend to reduce the net radiative flux.
Thus, it appears that aerosol loading in deep convection with large, long-lived anvils may act to offset the
cooling effect of cloud radiative forcing by changing the ice spectra, cloud top albedo, cloud opacity, and
anvil cloud top elevation.
4. Summary and Conclusions
In this study the indirect effects of cloud-droplet nucleating aerosols on the mean vertical distributions of
liquid and ice condensate and cirrus anvil microphysical and radiative characteristics of midlatitude squall
lines were examined through the use of cloud-resolving model simulations. The RAMS model was used to
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,895
Journal of Geophysical Research: Atmospheres
2
10.1002/2016JD025082
1
Figure 10. (a, f) Cloud top albedo (fraction), (b, g) TOA-OLR (W m ), (c, h) radiative heating rate (K d ), (d, i) difference in
1
radiative heating rate (K d ), and (e, j) percent change in net column radiative flux for the 20 May (Figures 10a–10e) and 23
May (Figures 10f–10j) cases.
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,896
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
simulate two squall line MCS events that occurred on 20 and 23–24 May 2011 during the MC3E field project.
To investigate potential aerosol indirect effects, four sensitivity experiments were performed for each case
study, with initial surface aerosol concentrations of 600, 2000, 3000, and 4000 cm 3, named, respectively,
as CLE, POL2K, POL3K, and POL4K. The aerosol concentrations were based on CCN observations from 22 to
24 May 2011 at the DOE ARM-SGP site, and the vertical profiles of concentrations decreased exponentially
with height.
The primary goal of this study was to improve our understanding of aerosol indirect effects on the vertical
distribution of liquid and ice condensate and the anvil properties of squall lines from an overarching mean
perspective. By examining the aerosol impacts via spatial and temporal means over the full squall line system,
a greater understanding was gained on the overall regional radiative impact of MCSs, and the predominant
sign and magnitude of the cloud microphysical and radiative responses were assessed. The aerosol-induced
competition between cloud water generation at lower cloud levels and the removal by riming at middlecloud levels is key to understanding how aerosol concentration impacts the cloud water mixing ratio and droplet size distributions in deep convection. The influence of cloud droplet spectra on anvil characteristics is
then communicated via homogenous freezing of lofted cloud droplets.
In this study, the sign of the response in microphysical quantities to aerosol loading was variable. Cloud
droplet and cloud ice number concentrations increased monotonically with increased aerosol loading, while
cloud water mixing ratio exhibited nonmonotonic trends that varied with height. In both MCS events, cloud
water mixing ratio in the lower levels (up to ~6 km) exhibited monotonic increases with aerosol loading. At
greater aerosol concentrations, more cloud droplets were nucleated, readily grew by condensation,
consumed the abundant supersaturation, and generated more cloud water. In the upper portions of the
liquid cloud (from ~6 to 10 km), droplets were readily rimed because of their relatively large size, while
at lower cloud levels, droplets were smaller and scavenging was less efficient. A balance between an
increase in the cloud water available to rime and a decrease in rime collection efficiency in response
to aerosol loading led to a variable reduction in cloud mixing ratio above ~6 km. An aerosol-induced
decrease in cloud water above ~6 km relative to CLE was referred to in this study as a trend reversal,
and the greatest trend reversal in cloud water aloft occurred in POL2K.
Since increasing aerosol concentrations led to increases in the cloud droplet mixing ratio (below ~6 km) and
number concentration, while droplet sizes were only moderately reduced, total riming rates generally
increased as aerosol concentration increased. Compared to the change in total riming from CLE to POL2K,
there was still an increase in riming between POL2K and POL4K in spite of reduced collection efficiency of
smaller cloud droplets. This occurred because there was more cloud water and more numerous droplets
available to rime in the more polluted simulations. However, the change in total riming was less from
POL3K to POL4K compared to the change from POL2K to POL3K. This occurred because the riming efficiency
of smaller droplets decreased enough to sufficiently offset the contribution to total riming from increased
generation of more cloud water and number. In and above the cloud water trend reversal layer, there was
less cloud water in POL2K compared to POL4K in spite of greater total riming in POL4K. This occurred because
POL2K simply had less cloud water generated than POL4K and because the riming efficiency is greater than
in POL4K.
The variable profiles in cloud water directly impacted anvil ice characteristics. Increased removal of cloud
water via riming limited the amount of cloud water that was lofted to the upper anvil levels and frozen homogeneously. Thus, there tended to be reduced cloud ice mixing ratio in the upper anvils for an increase in aerosol concentration. The magnitude and depth over which cloud ice was reduced in the anvil tended to scale
with the relative amount of cloud water removed by riming. These trends varied slightly between the 20 and
23 May events. This is to be expected since the response to a specific aerosol loading depends on the amount
of condensed water and ice, which, in turn, is controlled by the environment and the strength of the
squall line.
Regardless of the variability in cloud water and cloud ice profiles, the increase in aerosol concentration led to
monotonic increases in cloud droplet and cloud ice number concentration and subsequent monotonic
decreases in particle size. Smaller ice particles have smaller collection efficiencies and smaller fall velocities.
As such, their residence times were longer, and their horizontal advection in the divergent anvil led to greater
anvil cloud coverage at certain elevations. In the 20 and 23 May case studies, the upper level (>12 km) anvil
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,897
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
areal coverage in the polluted simulations increased by up to 50% and 150%, respectively, while the lower
anvil region (<12 km) showed little change. The total squall line cloud coverage increased only modestly
up to 1–2%. The addition of numerous small cloud ice particles led to enhanced cloud top albedo by up to
4–6%. Further, there was up to a 6–8% reduction in the TOA-OLR due to reduced emission by the upper level
anvil and increased opacity resulting from persistence of more numerous small ice particles at the
uppermost levels.
The aerosol effects on simulated squall lines discussed herein indicate that a series of complex and competing processes determine the vertical distribution of water and ice and the microphysical, macrophysical, and
radiative properties of the cirrus anvil clouds. While some microphysical trends, such as hydrometeor number
and size, are monotonic, others, such as hydrometeor mixing ratios, are not. The concentration of aerosols,
strength of convection, and availability of water vapor are key controlling factors in dictating monotonicity.
In spite of nonmonotonic trends and reduced magnitudes in upper level cloud water and cloud ice mixing
ratios in the more polluted simulations, there were clear monotonic trends in cloud top radiative effects in
response to aerosol loading. The cloud top albedo was enhanced, TOA-OLR was reduced, and the net column
radiative flux was reduced, thus acting to offset the radiative cooling effects of squall lines. This indicates a
positive radiative forcing (warming effect) from increased aerosol loading on squall line clouds resulting from
an increase in anvil ice number concentration, a decrease in ice size, an increase in ice particle residence time,
and a shift to higher-altitude anvil clouds. This radiative response occurs regardless of the decrease in anvil
condensate mixing ratio in more polluted conditions, thus, emphasizing that the radiative forcing of anvils
may be more sensitive to changes in hydrometeor size spectra than mass mixing ratio.
Results shown herein reveal the potential impact that aerosols can have on clouds that persist at the top of
the troposphere and are created through deep convection. The imposed changes to anvil ice clouds and their
radiative properties are consistent between these two completely different MCSs that developed in different
environments. For future work, it would be useful, however, to test the robustness of our conclusions by
performing a similar study over a spectrum of MCS intensities and structures.
Acknowledgments
This work was supported by the
Department of Energy under Grant DESC0010569. The simulation data are
available upon request from Stephen
Saleeby (Stephen.Saleeby@colostate.
edu). The surface CCN concentration
data were accessed from the ARM Data
Archive, http://www.archive.arm.gov.
Satellite imagery was provided courtesy
of University of Wisconsin-Madison
Space Science and Engineering Center.
SALEEBY ET AL.
References
Albrecht, B. (1989), Aerosols, cloud microphysics, and fractional cloudiness, Science, 245, 1227–1230.
Altaratz, O., I. Koren, T. Reisin, A. Kostinski, G. Feingold, Z. Levin, and Y. Yin (2008), Aerosols' influence on the interplay between condensation,
evaporation and rain in warm cumulus cloud, Atmos. Chem. Phys., 8, 15–24.
Berg, W., T. L'Ecuyer, and S. C. van den Heever (2008), Evidence for the impact of aerosols on the onset and microphysical properties of
rainfall from a combination of satellite observations and cloud-resolving model simulations, J. Geophys. Res., 113, D14S23, doi:10.1029/
2007JD009649.
Carrió, G. G., S. C. van den Heever, and W. R. Cotton (2007), Impacts of nucleating aerosol on anvil-cirrus clouds: A modeling study, Atmos.
Res., 84, 111–131.
Carrió, G. G., W. R. Cotton, and W. Y. Y. Cheng (2010), Urban growth and aerosol effects on convection over Houston. Part I: The August 2000
case, Atmos. Res., 96, 560–574.
Chen, T., W. B. Rossow, and Y. Zhang (2000), Radiative effects of cloud-type variations, J. Clim., 13, 264–286, doi:10.1175/1520-0442(2000)
013<0264:REOCTV>2.0.CO;2.
Chou, M.-D., R. S. Lindzen, and A. Y. Hou (2002), Comments on the iris hypothesis: A negative or positive cloud feedback?, J. Clim., 15,
2713–2715.
Churchill, D. D., and R. A. Houze Jr. (1984), Development and structure of winter monsoon cloud clusters on 10 December 1978, J. Atmos. Sci.,
41, 933–960.
Cotton, W. R., G. D. Alexander, R. Hertensetin, R. L. Walko, R. L. McAnelly, and M. Nichols (1995), Cloud venting—A review and some new
global annual estimates, Earth Sci. Rev., 39, 169–206.
Cotton, W. R., et al. (2003), RAMS 2001: Current status and future directions, Meteorol. Atmos. Phys., 82(1), 5–29, doi:10.1007/s00703-001-0584-9.
Davies, H. C. (1983), Limitations of some common lateral boundary schemes used in regional NWP models, Mon. Weather Rev., 111,
1002–1012.
Del Genio, A. D., J. Wu, and Y. Chen (2012), Characteristics of mesoscale organization in WRF simulations of convection during TWP-ICE,
J. Clim., 25(17), 5666–5688, doi:10.1175/JCLI-D-11-00422.1.
DeMott, P. J., M. P. Meyers, and W. R. Cotton (1994), Parameterization and impact of ice initiation processes relevant to numerical model
simulations of cirrus clouds, J. Atmos. Sci., 51, 77–90.
DeMott, P. J., A. J. Prenni, X. Liu, S. M. Kreidenweis, M. D. Petters, C. H. Twohy, M. S. Richardson, T. Eidhammer, and D. C. Rogers (2010),
Predicting global atmospheric ice nuclei distributions and their impacts on climate, Proc. Natl. Acad. Sci. U.S.A., 107, 11,217–11,222.
Fan, J., R. Zhang, G. Li, and W.-K. Tao (2007), Effects of aerosols and relative humidity on cumulus clouds, J. Geophys. Res., 112, D14204,
doi:10.1029/2006JD008136.
Fan, J., J. M. Comstock, and M. Ovchinnikov (2010a), The cloud condensation nuclei and ice nuclei effects on tropical anvil characteristics and
water vapor of the tropical tropopause layer, Environ. Res. Lett., 5(4), 044005, doi:10.1088/1748-9326/5/4/044005.
Fan, J., J. M. Comstock, M. Ovchinnikov, S. A. McFarlane, G. McFarquhar, and G. Allen (2010b), Tropical anvil characteristics and water vapor of
the tropical tropopause layer: Impact of heterogeneous and homogeneous freezing parameterizations, J. Geophys. Res., 115, D12201,
doi:10.1029/2009JD012696.
AEROSOL EFFECTS ON MCS ANVILS
10,898
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
Fan, J., Y. Liu, K. Xu, K. North, S. Collis, X. Dong, G. J. Zhang, Q. Chen, P. Kollias, and S. J. Ghan (2015), Improving representation of convective
transport for scale-aware parameterization: 1. Convection and cloud properties simulated with spectral bin and bulk microphysics: CRM
model evaluation, J. Geophys. Res. Atmos., 120, 3485–3509, doi:10.1002/2014JD022142.
Feingold, G., R. L. Walko, B. Stevens, and W. R. Cotton (1998), Simulations of marine stratocumulus using a new microphysical scheme, Atmos.
Res., 47–48, 505–528.
Feng, Z., X. Dong, B. Xi, C. Schumacher, P. Minnis, and M. Khaiyer (2011), Top-of-atmosphere radiation budget of convective core/stratiform
rain and anvil clouds from deep convective systems, J. Geophys. Res., 116, D23202, doi:10.1029/2011JD016451.
Ferrare, R., G. Feingold, S. Ghan, J. Ogren, B. Schmid, S. E. Schwartz, and P. Sheridan (2006), Preface to special section: Atmospheric Radiation
Measurement Program May 2003 Intensive Operations Period examining aerosol properties and radiative influences, J. Geophys. Res., 111,
D05S01, doi:10.1029/2005JD006908.
Fowler, L. D., and D. A. Randall (1994), A global radiative convective feedback, Geophys. Res. Lett., 21, 2035–2038, doi:10.1029/94GL01711.
Fritsch, J. M., R. J. Kane, and C. H. Chelius (1986), The contribution of mesoscale convective weather systems to the warm season precipitation
in the United States, J. Clim. Appl. Meteorol., 25, 1333–1345.
Futyan, J. M., J. E. Russell, and J. E. Harries (2005), Determining cloud forcing by cloud type from geostationary satellite data, Geophys. Res.
Lett., 32, L08807, doi:10.1029/2004GL022275.
Gebhart, K. A., S. M. Kreidenweis, and W. C. Malm (2001), Back-trajectory analyses of fine particulate matter measured at Big Bend National
Park in the historical database and the 1996 scoping study, Sci. Total Environ., 276, 185–204.
Grant, L. D., and S. C. van den Heever (2015), Cold pool and precipitation responses to aerosol loading: Modulation by dry layers, J. Atmos. Sci.,
72, 1398–1408.
Hallett, J., and S. C. Mossop (1974), Production of secondary ice crystals during the riming process, Nature, 249, 26–28.
Hand, J. (2011), IMPROVE: Spatial and seasonal patterns and temporal variability of haze and its constituents in the United States, Rep. V.
[Available at http://vista.cira.colostate.edu/improve/Publications/Reports/2011/2011.htm.]
Harrington, J. Y. (1997), The effects of radiative and microphysical processes on simulated warm and transition season Arctic stratus, PhD
dissertation, 289 pp., Colo. State Univ., Atmos. Sci. Pap. 637.
Heymsfield, A. J., L. M. Miloshevich, C. Schmitt, A. Bansemer, C. Twohy, M. R. Poellot, A. Fridlind, and H. Gerber (2005), Homogeneous ice
nucleation in subtropical and tropical convection and its influence on cirrus anvil microphysics, J. Atmos. Sci., 62, 41–64, doi:10.1175/
JAS-3360.1.
Jensen, E. J., S. Kinne, and O. B. Toon (1994), Tropical cirrus cloud radiative forcing: Sensitivity studies, Geophys. Res. Lett., 21, 2023–2026,
doi:10.1029/94GL01358.
Jensen, M. P., T. Toto, D. Troyan, P. E. Ciesielski, D. Holdridge, J. Kyrouac, J. Schatz, Y. Zhang, and S. Xie (2016), The Midlatitude Continental
Convective Clouds Experiment (MC3E) sounding network: Operations, processing and analysis, Atmos. Meas. Tech., 8, 421–434,
doi:10.5194/amt-8-421-2015.
Jirak, I. L., and W. R. Cotton (2007), Observational analysis of the predictability of mesoscale convective systems, Weather Forecasting, 22,
813–838.
Jirak, I. L., W. R. Cotton, and R. L. McAnelly (2003), Satellite and radar survey of mesoscale convective system development, Mon. Weather Rev.,
131, 2428–2449.
Kain, J. S., and J. M. Fritsch (1993), Convective parameterization for mesoscale models: The Kain-Fritsch scheme, in The Representation of
Cumulus Convection in Numerical Models, Meteorol. Monogr., vol. 24, edited by K. Emanuel and D. Raymond, pp. 165–170, Am. Meteorol.
Soc., Boston, Mass.
Khain, A., D. Rosenfeld, and A. Pokrovsky (2005), Aerosol impact on the dynamics and microphysics of deep convective clouds, Q. J. R.
Meteorol. Soc., 131, 2639–2663, doi:10.1256/qj.04.62.
Khain, A. P., N. BenMoshe, and A. Pokrovsky (2008), Factors determining the impact of aerosols on surface precipitation from clouds: An
attempt at classification, J. Atmos. Sci., 65, 1721–1748.
Koren, I., Y. J. Kaufman, D. Rosenfeld, L. A. Remer, and Y. Rudich (2005), Aerosol invigoration and restructuring of Atlantic convective clouds,
Geophys. Res. Lett., 32, L14828, doi:10.1029/2005GL023187.
Koren, I., L. A. Remer, O. Altaratz, J. V. Martins, and A. Davidi (2010), Aerosol-induced changes of convective cloud anvils produce climate
warming, Atmos. Chem. Phys., 10, 5001–5010, doi:10.5194/acp-10-5001-2010.
Kreidenweis, S. M., L. A. Remer, R. Bruintjes, and O. Dubovik (2001), Smoke aerosol from biomass burning in Mexico: Hygroscopic smoke
optical model, J. Geophys. Res., 106, 4831–4844, doi:10.1029/2000JD900488.
Lang, S. E., W.-K. Tao, J. Simpson, and B. Ferrier (2003), Modeling of convective–stratiform precipitation processes: Sensitivity to partitioning
methods, J. Appl. Meteorol., 42, 505–527, doi:10.1175/1520-0450(2003)042<0505:MOCSPP>2.0.CO;2.
Lang, S. E., W.-K. Tao, J.-D. Chern, D. Wu, and X. Li (2014), Benefits of a fourth ice class in the simulated radar reflectivities of convective
systems using a bulk microphysics scheme, J. Atmos. Sci., 71, 3583–3612, doi:10.1175/JAS-D-13-0330.1.
Lebo, Z. J. (2014), The sensitivity of a numerically simulated idealized squall line to the vertical distribution of aerosols, J. Atmos. Sci., 71,
4581–4596.
Lee, S. S., L. J. Donner, and V. T. J. Phillips (2009), Sensitivity of aerosol and cloud effects on radiation to cloud types: Comparison between
deep convective clouds and warm stratiform clouds over one-day period, Atmos. Chem. Phys., 9, 2555–2575, doi:10.5194/acp9-2555-2009.
Lensky, I. M., and D. Rosenfeld (2006), The time-space exchangeability of satellite retrieved relations between cloud top temperature and
particle effective radius, Atmos. Chem. Phys., 6, 2887–2894.
Li, G., Y. Wang, and R. Zhang (2008), Implementation of a two-moment bulk microphysics scheme to the WRF model to investigate aerosol–
cloud interaction, J. Geophys. Res., 113, D15211, doi:10.1029/2007JD009361.
Li, G., Y. Wang, K.-H. Lee, Y. Diao, and R. Zhang (2009), Impacts of aerosols on the development and precipitation of a mesoscale squall line,
J. Geophys. Res., 114, D17205, doi:10.1029/2008JD011581.
Lin, B., B. Wielicki, L. Chambers, Y. Hu, and K.-M. Xu (2002), The iris hypothesis: A negative or positive cloud feedback?, J. Clim., 15, 3–7.
Lin, J. C., T. Matsui, R. A. Pielke Sr., and C. Kummerow (2006), Effects of biomass burning-derived aerosols on precipitation and clouds in the
Amazon Basin: A satellite-based empirical study, J. Geophys. Res., 111, D19204, doi:10.1029/2005JD006884.
Loftus, A. M., and W. R. Cotton (2014), Examination of CCN impacts on hail in a simulated supercell with triple-moment hail bulk microphysics, Atmos. Res., 147–148, 183–204.
Lynn, B. H., A. P. Khain, J. Dudhia, D. Rosenfeld, A. Pokrovsky, and A. Seifert (2005), Spectral (Bin) microphysics doupled with a mesoscale
model (MM5). Part II: Simulation of a CaPE rain event with a squall line, Mon. Weather Rev., 133, 59–71, doi:10.1175/MWR-2841.1.
Maddox, R. (1980), Meoscale convective complexes, Bull. Am. Meteorol. Soc., 61, 1374–1387.
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,899
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
Marinescu, P. J., S. C. van den Heever, S. M. Saleeby, and S. M. Kreidenweis (2016), The microphysical contributions to and evolution of latent
heating profiles in two MC3E MCSs, J. Geophys. Res. Atmos., 121, 7913–7935, doi:10.1002/2016JD024762.
May, P. T., V. N. Bringi, and M. Thurai (2011), Do we observe aerosol impacts on DSDs in strongly forced tropical thunderstorms?, J. Atmos. Sci.,
68, 1902–1910.
Mellor, G. L., and T. Yamada (1974), A hierarchy of turbulence closure models for planetary boundary layers, J. Atmos. Sci., 31(7), 1791–1806,
doi:10.1175/1520-0469(1974)031<1791:AHOTCM>2.0.CO;2.
Mesinger, F., and A. Arakawa (1976), Numerical Methods Used in Atmospheric Models, GARP Publ. Ser. 17, vol. 1, 66 pp., WMO, Geneva, Switzerland.
Meyers, M. P., P. J. DeMott, and W. R. Cotton (1992), New primary ice nucleation parameterizations in an explicit cloud model, J. Appl.
Meteorol., 31, 708–721.
Meyers, M. P., R. L. Walko, J. Y. Harrington, and W. R. Cotton (1997), New RAMS cloud microphysics parameterization: Part II. The two-moment
scheme, Atmos. Res., 45, 3–39.
Morrison, H. (2012), On the robustness of aerosol effects on an idealized supercell storm simulated with a cloud system-resolving model,
Atmos. Chem. Phys., 12, 7689–7705, doi:10.5194/acp-12-7689-2012.
Morrison, H., and W. W. Grabowski (2011), Cloud-system resolving model simulations of aerosol effects on tropical deep convection and its
thermodynamic environment, Atmos. Chem. Phys., 11, 10,503–10,523.
Parker, M. D., and R. H. Johnson (2000), Organizational modes of midlatitude mesoscale convective systems, Mon. Weather Rev., 128,
3413–3436.
Platt, C. M. R. (1989), The role of cloud microphysics in high-cloud feedback effects on climate change, Nature, 341, 428–429, doi:10.1029/
JD093iD09p11051.
Platt, C. M. R., and Harshvardhan (1988), Temperature dependence of cirrus extinction: Implications for climate feedback, J. Geophys. Res., 93,
11,051–11,058, doi:10.1029/JD093iD09p11051.
Ramanathan, V., and W. D. Collins (1991), Thermodynamic regulation of the ocean warming by cirrus clouds deduced from observations of
the 1987 El Niño, Nature, 351, 27–32.
Rotunno, R., J. B. Klemp, and M. L. Weisman (1988), A theory for strong, long-lived squall lines, J. Atmos. Sci., 45, 463–485.
Saleeby, S. M., and W. R. Cotton (2004), A large-droplet mode and prognostic number concentration of cloud droplets in the Colorado State
University Regional Atmospheric Modeling System (RAMS). Part I: Module descriptions and supercell test simulations, J. Appl. Meteorol., 43,
182–195.
Saleeby, S. M., and W. R. Cotton (2008), A binned approach to cloud-droplet riming implemented in a bulk microphysics model, J. Appl.
Meteorol. Climatol., 47, 694–703.
Saleeby, S. M., and S. C. van den Heever (2013), Developments in the CSU-RAMS aerosol model: Emissions, nucleation, regeneration,
deposition, and radiation, J. Appl. Meteorol. Climatol., 52(12), 2601–2622, doi:10.1175/JAMC-D-12-0312.1.
Saleeby, S. M., W. Berg, T. L'Ecuyer, and S. C. van den Heever (2010), Impact of cloud-nucleating aerosols in cloud-resolving model simulations
of warm-rain precipitation in the East China Sea, J. Atmos. Sci., 67, 3916–3930.
Saleeby, S. M., S. R. Herbener, S. C. van den Heever, and T. L'Ecuyer (2015), Impacts of cloud droplet–nucleating aerosols on shallow tropical
convection, J. Atmos. Sci., 72, 1369–1385, doi:10.1175/JAS-D-14-0153.1.
Seigel, R. B., S. C. van den Heever, and S. M. Saleeby (2013), Mineral dust indirect effects and cloud radiative feedbacks of a simulated
idealized nocturnal squall line, Atmos. Chem. Phys., 13, 4467–4485, doi:10.5194/acp-13-4467-2013.
Sheffield, A. M., S. M. Saleeby, and S. C. van den Heever (2015), Aerosol-induced mechanisms for cumulus congestus growth, J. Geophys. Res.
Atmos., 120, 8941–8952, doi:10.1002/2015JD023743.
Smagorinsky, J. (1963), General circulation experiments with the primitive equations, Mon. Weather Rev., 91(3), 99–164, doi:10.1175/15200493(1963)091<0099:GCEWTP>2.3.CO;2.
Solomon, S., D. Qin, M. Manning, M. M. K. Averyt, M. M. B. Tignor, H. L. Miller, and Z. Chen (2007), Climate Change 2007: The Physical Basis, IPCC,
Cambridge Univ. Press, Cambridge, U. K.
Stephens, G. L. (2005), Cloud feedbacks in the climate system: A critical review, J. Clim., 18, 237–273.
Stephens, G. L., S. C. van den Heever, and L. Pakula (2008), Radiative-convective feedbacks in idealized states of radiative convective equilibrium, J. Atmos. Sci., 65, 3899–3916.
Storer, R. L., and S. C. van den Heever (2013), Microphysical processes evident in aerosol forcing of tropical deep convective clouds, J. Atmos.
Sci., 70, 430–446.
Storer, R. L., S. C. van den Heever, and G. L. Stephens (2010), Modeling aerosol impacts on convective storms in different environments,
J. Atmos. Sci., 67, 3904–3915.
Tao, W.-K., X. Li, A. Khain, T. Matsui, S. Lang, and J. Simpson (2007), Role of atmospheric aerosol concentration on deep convective precipitation: Cloud-resolving model simulations, J. Geophys. Res., 112, D24S18, doi:10.1029/2007JD008728.
Tao, W.-K., J.-P. Chen, Z. Li, C. Wang, and C. Zhang (2012), Impact of aerosols on convective clouds and precipitation, Rev. Geophys., 50,
RG2001, doi:10.1029/2011RG000369.
Tao, W.-K., D. Wu, T. Matsui, C. Peters-Lidard, S. Lang, A. Hou, M. Rienecker, W. Petersen, and M. Jensen (2013), Precipitation intensity and
variation during MC3E: A numerical modeling study, J. Geophys. Res. Atmos., 118, 7199–7218, doi:10.1002/jgrd.50410.
Tao, W.-K., D. Wu, S. Lang, J.-D. Chern, C. Peters-Lidard, A. Fridlind, and T. Matsui (2016), High-resolution NU-WRF simulations of a deep
convective-precipitation system during MC3E: Further improvements and comparisons between Goddard microphysics schemes and
observations, J. Geophys. Res. Atmos., 121, 1278–1305, doi:10.1002/2015JD023986.
Twomey, S. (1977), The influence of pollution on the shortwave albedo of clouds, J. Atmos. Sci., 34, 1149–1152.
Tzivion, S., G. Feingold, and Z. Levin (1987), An efficient numerical solution to the stochastic collection equation, J. Atmos. Sci., 44, 3139–3149.
van den Heever, S. C., and W. R. Cotton (2007), Urban aerosol impacts on downwind convective storms, J. Appl. Meteorol. Climatol., 46,
828–850.
van den Heever, S. C., G. G. Carrió, W. R. Cotton, P. J. DeMott, and A. J. Prenni (2006), Impacts of nucleating aerosol on Florida storms. Part I:
Mesoscale simulations, J. Atmos. Sci., 63, 1752–1775.
van den Heever, S. C., G. L. Stephens, and N. B. Wood (2011), Aerosol indirect effects on tropical convection characteristics under conditions
of radiative-convective equilibrium, J. Atmos. Sci., 68, 699–718.
Walko, R. L., W. R. Cotton, M. P. Meyers, and J. Y. Harrington (1995), New RAMS cloud microphysics parameterization: Part I. The singlemoment scheme, Atmos. Res., 38, 29–62.
Walko, R. L., W. R. Cotton, M. P. Meyers, and J. Y. Harrington (2000a), Efficient computation of vapor and heat diffusion between hydrometeors in a numerical model, Atmos. Res., 53, 171–183.
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,900
Journal of Geophysical Research: Atmospheres
10.1002/2016JD025082
Walko, R. L., et al. (2000b), Coupled atmosphere–biophysics–hydrology models for environmental modeling, J. Appl. Meteorol., 39, 931–944,
doi:10.1175/1520-0450(2000)039<0931:CABHMF>2.0.CO;2.
Wang, P. K., and W. Ji (2000), Collision efficiencies of ice crystals at low–intermediate Reynolds numbers colliding with supercooled cloud
droplets: A numerical study, J. Atmos. Sci., 57, 1001–1009, doi:10.1175/1520-0469(2000)057<1001:CEOICA>2.0.CO;2.
Webster, P. J., and G. L. Stephens (1980), Tropical upper troposphere extended clouds: Inferences from Winter MONEX, J. Atmos. Sci., 37,
1521–1541.
SALEEBY ET AL.
AEROSOL EFFECTS ON MCS ANVILS
10,901