1. Introduction
Tropical convection is important to many of the interactions and feedback mechanisms of the climate system (Arakawa 2004; Stephens 2005), playing a crucial role in the production of precipitation and the maintenance of the earth’s energy budget. The importance of convection in regulating the large-scale circulations, thermal structure, and heat balance of the tropical atmosphere has long been recognized (Riehl and Malkus 1958) and continues to receive attention (Fierro et al. 2009). Any changes to tropical convection through external forcing may therefore have significant implications for the tropical atmosphere and globe as a whole. Enhanced aerosol concentrations, such as those associated with Saharan or Asian dust intrusions, represent one such potential forcing mechanism.
Aerosols, whether natural or anthropogenic, affect the climate both directly through the absorption and scattering of solar radiation (the direct aerosol effect) and indirectly by modifying cloud properties and precipitation processes, which in turn influence cloud radiative feedbacks [aerosol indirect effects (AIEs)]. Quantitative estimates of these effects are highly uncertain, and AIEs are still regarded as one of the most uncertain aspects of climate change (Houghton et al. 2001; Solomon et al. 2007). If we are to represent both convective processes and aerosol indirect forcing (AIF) more effectively in regional and global climate models, we need to enhance our understanding of AIEs on tropical convection. The overarching goal of the research presented here is therefore to investigate the impacts of AIF associated with cloud condensation nuclei (CCN) on the dynamical and microphysical properties of tropical convective clouds through the use of a radiative–convective equilibrium (RCE) framework.
a. Tropical clouds
Initial studies of tropical cloud distributions suggested a bimodal distribution consisting of shallow trade wind cumulus clouds associated with the trade wind inversion of the descending branches of the Hadley and Walker circulations, and deep cumulonimbus clouds (“hot towers”) in the intertropical convergence zone (ITCZ) (e.g., Riehl and Malkus 1958). A trimodal distribution of clouds consisting of shallow cumulus, cumulus congestus, and deep convection was subsequently identified within regions of active convection in the tropics (Johnson et al. 1999). The cloud tops of these three cloud types appear to be in close proximity to prominent stable layers that inhibit cloud growth and promote cloud detrainment (Johnson et al. 1999; Posselt et al. 2008). Shallow cumulus occur over most of the tropical oceans beyond convergence zones and are important in maintaining the trades and in serving as a moisture source for the boundary layer that supplies deep convection (Riehl and Malkus 1958; Augstein et al. 1973; Betts 1973; Nitta and Esbensen 1974). Cumulus congestus frequently form part of multiple-layered systems (Stephens and Wood 2007) and appear to play a significant role in moistening the middle troposphere prior to deep convection (Johnson and Lin 1997; Johnson et al. 1999). The deep convective mode maintains the temperature and moisture structure of the convective atmosphere through latent heat release and convective entrainment, and it affects the general circulation through the transport of momentum, moisture, and energy (Riehl and Malkus 1958; Fierro et al. 2009). Each cloud mode therefore appears to play an important, unique role in the tropics. In spite of this previous research, much is still not understood about these tropical cloud modes including how they might respond to AIF and the resultant response that this may have on the moisture and energy budgets of the tropics.
b. Aerosols
A number of observational studies have demonstrated the first (Twomey 1974) and second (Albrecht 1989) aerosol indirect effects on clouds and precipitation, including smaller cloud droplets, optically thicker clouds, and the suppression of precipitation (e.g., Warner and Twomey 1967; Borys et al. 1998; Rosenfeld and Lensky 1998; Rosenfeld 1999, 2000; Heymsfield and McFarquhar 2001; Hudson and Yum 2001; Andreae et al. 2004; Jirak and Cotton 2006). However, while remote sensing techniques have been used to identify AIEs on clouds (e.g., Kaufman and Nakajima 1993; Feingold et al. 2003; Berg et al. 2008), there is still great uncertainty regarding how to estimate these effects using such data and methods (Feingold et al. 2003; Berg et al. 2008). That changes in aerosol concentrations are often associated with variations in meteorological conditions further complicates the assessment of AIEs (Matsui et al. 2004, 2006; Kaufman et al. 2005), and observations have shown that not only can the LWP increase with enhanced aerosol concentrations, but it can also decrease or remain constant (Platnick et al. 2000; Coakley and Walsh 2002; Han et al. 2002).
Recent modeling studies have also demonstrated that the relationships between cloud processes and aerosols may differ from those proposed in the original aerosol indirect hypotheses. Enhanced CCN concentrations may reduce rather than increase the LWP in stratocumulus through drizzle and humidity effects (Jiang et al. 2002; Ackerman et al. 2004; Lu and Seinfeld 2005) and may cause a decrease in the cloud fraction of warm trade cumulus through enhanced entrainment (Xue and Feingold 2006; Jiang et al. 2006). It has also been found that the cloud fraction of shallow cumulus may only increase with enhanced aerosol concentrations when the background concentrations are relatively low (Xue et al. 2008) and that while less precipitation is produced by polluted cumulus clouds, the rain drops may be larger because of a more efficient collection process (Altaratz et al. 2007; Berg et al. 2008; Saleeby et al. 2010; Storer et al. 2010).
Simulations of deep convective storms (e.g., Fridland et al. 2004; Khain et al. 2004, 2005; Zhang et al. 2005; van den Heever et al. 2006; Carrió et al. 2007; van den Heever and Cotton 2007; Storer et al. 2010) have demonstrated that not only do aerosols influence the microphysical properties of these clouds, but they also have an important effect on the storm dynamics and subsequent convective development. AIEs have also been found to vary based on the cloud or storm type under consideration (Seifert and Beheng 2006; Khain et al. 2008), the stage in the storm life cycle (van den Heever et al. 2006; van den Heever and Cotton 2007; Berg et al. 2008), and the background aerosol concentrations (van den Heever and Cotton 2007; Xue et al. 2008). Khain et al. (2008) found that increased aerosol concentrations produce increases in both the generation and the loss of condensate mass, the dominance of which was environmentally dependent. They conclude that the many discrepancies arising from aerosol indirect studies reported in the literature may be attributed to differences in the environmental conditions of the observations or numerical simulations, as well as to the different types of cloud systems being analyzed. Thus it is clear that there is still much we do not understand about aerosol indirect effects on various cloud types.
c. Radiative–convective equilibrium
One way to consider the connection between convection and the energy budget of the earth is through the concept of RCE. The tropical atmosphere is never far from a state of RCE (Stephens et al. 1994, 2004a). In this equilibrium state, the loss of energy from the atmosphere that occurs as a result of the emission of radiation from the atmosphere exceeding the absorption of radiation by the atmosphere is balanced by the heat transferred from the surface by convective processes and large-scale winds. The RCE framework has been used in a number of experiments focusing on the feedbacks among radiation, clouds, water vapor, and precipitation in the tropics (Held et al. 1993; Randall et al. 1994; Sui et al. 1994; Xu and Randall 1999; Tompkins and Craig 1998a,b; Su et al. 2000; Grabowski and Moncrieff 2004; Stephens et al. 2004b; Bretherton et al. 2005; Grabowski 2006); however, few experiments have been performed to examine AIEs on tropical convection, with the possible exception of Grabowski (2006). He found from his experiments that AIF is dominated by the first indirect effect, that there was little if any impact on the hydrological cycle, and that while precipitation develops more readily in clouds developing in clean environments, the difference was insignificant when averaged over many clouds and over extended time periods. Grabowski did suggest that his results be viewed as preliminary because several model simplifications were used and only shallow convective clouds were represented.
d. Goals
The primary goal of the research described here was to investigate the impacts of enhanced CCN concentrations associated with dust intrusions on the microphysical, dynamical, and precipitation characteristics of the range of tropical cloud regimes developing within an environment that is in RCE. More specifically, the following questions were investigated:
What is the effect of enhanced CCN concentrations on the state of RCE and on the large-scale organization of convection?
What influence do CCN exert on the surface precipitation produced by the entire spectrum of tropical cloud modes, as well as the precipitation produced by each cloud mode?
What are the impacts of CCN on the microphysical characteristics of the modes of the tropical convection? In particular, how do CCN affect the liquid water and ice species, and the partitioning between these species?
What are the impacts of CCN on the dynamical characteristics of the modes of the tropical convection? More particularly, what impact do CCN exert on the strength, organization, and frequency of the updrafts?
2. Model and experiment setup
a. CRM description
One of the reasons that AIEs are such an uncertain aspect of climate modeling is that current global climate models have to make use of parameterization schemes to represent convective and cloud processes. While such schemes produce many key characteristics of clouds, the interactions among cloud physics, cloud dynamics, cloud radiative transfer, and surface forcing are not well represented. Also, aerosols are not explicitly represented in GCM microphysical schemes and even within many CRM schemes. AIEs are often simply represented by techniques such as altering the autoconversion rate. Several studies have shown, however, that vertical velocity may be as important as the concentration of aerosols when considering aerosol effects (Nenes and Seinfeld 2003), and the number of activated CCN has been observed to increase with increasing vertical velocity (Saleeby and Cotton 2004). A CRM that eliminates the need for a convective parameterization scheme and that explicitly includes the effects of aerosol would therefore be most suitable for this research. The Regional Atmospheric Modeling System (RAMS; Pielke et al. 1992; Cotton et al. 2003) is one such model. RAMS is a nonhydrostatic CRM that has been successfully used to simulate many different atmospheric systems on a range of space and time scales.
The RAMS microphysics scheme (Walko et al. 1995; Meyers et al. 1997; Saleeby and Cotton 2004) in particular makes this CRM highly suitable for these simulations. This microphysical scheme has also been introduced into other CRMs such as the Goddard Cumulus Ensemble (GCE) model (Lee et al. 2009) and has been found to compare well with observations when conducting case study simulations (e.g., Cheng et al. 2009; Saleeby et al. 2009). For the experiments conducted here the two-moment scheme, which predicts both hydrometeor mixing ratios and number concentration, was utilized, and the following hydrometeor species were activated: pristine ice, snow, aggregates, graupel, hail, cloud water, and rain. The hydrometeor size distribution spectra are represented using a generalized gamma distribution function. While a two-moment bulk scheme is used for this research, this scheme in RAMS does not follow the methodology spearheaded by Kessler (1969) but rather attempts to represent the essence of bin-resolved microphysics models while still maintaining the cost effectiveness of a bulk scheme. For example, the activation of CCN is achieved through the use of look-up tables previously generated offline using a detailed bin-resolving parcel model (Feingold and Heymsfield 1992). Cloud droplet collection is simulated using stochastic collection equation solutions (Feingold et al. 1988), also facilitated by look-up tables rather than by continuous accretion approximations, and the philosophy of bin representation of collection is extended to calculations of drop sedimentation (Feingold et al. 1998).
b. Model configuration
Based on the need for a sufficiently large model domain to capture the large-scale circulation between convecting and nonconvecting regions, as well as sufficiently fine grid resolution to represent tropical convective processes, the first suite of simulations performed for this research was conducted in two dimensions. While dimensionality can be expected to influence the scale, strength, and dimensionality of the simulated convection (Tompkins 2000; Stephens et al. 2008), these simulations do capture many of the characteristics of tropical convection, as will be seen below, and are thus felt to provide a simple yet useful framework with which to examine AIF on the wide range of tropical cloud types that develop. Three-dimensional aerosol experiments with grid setups similar to those described in Stephens et al. (2008) are currently being performed, although these simulations represent a compromise between dimensionality and grid spacing. The two-dimensional simulations were run using a grid spacing of 1 km and 10 000 points in the zonal direction. The stretched vertical grid setup consisted of 36 levels, eight of which are found within the first 1-km AGL. The model top extended to approximately 26 km AGL. Periodic lateral boundary conditions were employed, while the top boundary was a rigid lid with four Rayleigh absorbing layers to prevent the reflection of gravity waves, and the lower boundary was a fixed oceanic surface with a SST of 300 K. A time step of 10 s was utilized and the simulations were run for 100 days. Coriolis force was turned off. A fixed solar zenith angle of 50° was employed, and thus the diurnal cycle is not represented in the model. The radiation scheme was updated every 5 min. While the radiation interacts fully with the microphysical processes in this model, it does not directly interact with aerosol species. Hence while AIF is accounted for, the direct forcing effects of aerosols are not represented here.
All of the simulations were initialized using the 0000 UTC 5 December 1992 Tropical Ocean and Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE) with initial zero mean wind in the vertical. The wind field evolves in time because of localized turbulence and broader-scale circulations that develop as the simulations progress. Convection was initiated by randomly and weakly perturbing the potential temperature field across the entire domain below 2 km AGL and, given the 1-km grid spacing, no convection parameterization scheme was necessary. While much of the setup is similar to that described in Stephens et al. (2008), differences do exist in that the two-moment as opposed to single-moment microphysics parameterization scheme is used here, and the grid spacing is reduced from 2.4 to 1 km. Other relevant model options are summarized in Table 1. This paper therefore describes the results of simple RCE experiments in which equilibrium is achieved over a lower boundary of fixed SST for a range of tropical cloud regimes in which there is no large-scale dynamical forcing. As such, this means that convection is driven solely by surface heat and moisture fluxes and radiational cooling, and that no large-scale convergence, mean wind or vertical wind shear were imposed, being similar in this respect to numerous previous RCE studies (e.g., Tompkins and Craig 1998a; Pakula and Stephens 2009).
Parameterization schemes and settings for the CRM simulations.
c. Experiment design
A series of experiments was conducted in which the number of aerosol available to act as CCN was varied from pristine to polluted concentrations. While it is recognized that aerosols such as salt and dust can also serve as giant CCN and IN, the focus here is on CCN. The CONTROL simulation was initialized horizontally homogeneously with pristine surface CCN concentrations of 25 cm−3 that decrease with height as a function of air density. A threshold is imposed in the model that prevents the aerosol concentrations dropping below 20 cm−3. The CONTROL simulation was run until RCE was achieved, a process that takes about 60 simulation days.
Sensitivity tests were then conducted in which the CONTROL simulation was restarted from day 60 and a layer of enhanced aerosol concentrations was introduced between 2 and 4 km AGL. This idealized model setup can be thought to be representative of the situation that occurs in the tropics when a polluted layer, such as a Saharan or Asian dust layer, is advected over the tropical Atlantic or Pacific Oceans, respectively. Saharan dust tends to reside in a layer between 2 and 4 km AGL (e.g., Prospero and Carlson 1981). For each consecutive experiment the number concentrations of the aerosol available to serve as CCN between 2 and 4 km AGL were doubled, starting with 100 cm−3. Similar aerosol concentrations have been used previously (e.g., Xue and Feingold 2006; van den Heever et al. 2006), and concentrations between 300 and 1200 cm−3 at 0.7% supersaturation were observed in association with Saharan dust during the Cirrus Regional Study of Tropical Anvils and Cirrus Layers–Florida-Area Cirrus Experiment (CRYSTAL-FACE) field campaign (van den Heever et al. 2006). The details and naming conventions of the sensitivity experiments that were conducted are shown in Table 2. The sensitivity experiments were then run for another 40 days. As these experiments were performed over such a long time period the aerosol concentrations within the 2–4-km layer were updated each time step, thus providing a continuous source of aerosol and thereby being representative of a dust layer advecting over an oceanic region. This was the only source of aerosol allowed. The aerosols are removed from the model domain following processes such as activation and precipitation.
The naming convention and aerosol concentrations within the 2–4-km layer AGL of the sensitivity experiments described in the text.
It is important to note that the aerosols available for activation are advected around the model domain by the model winds and are not therefore confined to the 2–4-km layer but are found throughout the domain in varying concentrations, even within the lowest levels of the model. The vertical motion associated with the development of cumulus, congestus, and deep convective clouds well above the height of the trade inversion in the disturbed regions, as well as occasional penetrations of the trade wind inversion by shallow cumulus in the regions of broad descent, are examples of processes that influence the CCN concentrations in the lower levels, which may be locally high at times. While the focus of this investigation is on the impacts of an elevated dust source of high aerosol concentrations on tropical convection, scenarios in which aerosol concentrations are greatest at the surface and decrease with height could also be expected influence tropical cloud distributions and are suggested as future research.
As it will be seen below, a wide range of typical tropical cloud systems develop under the environmental conditions that arise in these simulations. These experiments therefore allow for the assessment of AIEs not only on the tropical cloud spectrum as a whole, but also on the individual cloud types making up this spectrum, all within a self-consistent modeling framework.
3. Results
a. RCE
A time series of the domain-averaged column-integrated water vapor or precipitable water (PW), the thermal (latent plus sensible) heat flux (THF), the column net radiative flux divergence (QRAD), and the precipitation rate (PR) for the CONTROL simulation are shown in Fig. 1. Multiple time scales are evident in these fields as equilibrium is achieved in the model. Such time scales have been previously observed (e.g., Robe and Emanuel 1996; Tompkins 2000). On the longer time scales, the PW field demonstrates that equilibrium is achieved after ∼60 days (Fig. 1a), while shorter times scales are evident in the precipitation rate (Fig. 1c). The radiative flux divergence and the thermal heat flux reach balance between 113.5 and 116.1 W m−2, and the mean precipitation rate is approximately 3.4 mm day−1, thus comparing relatively well with the global average precipitation rates, which are on the order of 3 mm day−1 (Kiehl and Trenberth 1997). At equilibrium, the domain-averaged PW values for the CONTROL and aerosol experiments range between 31.3 and 33.5 mm depending on the experiment (Table 3). Stephens (1990) showed that for SSTs of 300 K, PW values average 41 ± 4 mm. The model PW values are thus 4–5 mm (∼10%) drier than observations over 300-K waters. This can in part be attributed to the two-dimensionality of the grid that artificially enlarges the areas of subsidence (Bretherton and Smolarkiewicz 1989; Tompkins 2000; Pakula and Stephens 2009). If the PW is averaged over what may be referred to as the convectively disturbed regions, defined after Stephens et al. (2008) to be those regions of the grid where the vertically integrated condensate is greater than 0.01 kg m−2 and the outgoing longwave radiation (OLR) is less than 260 W m−2, the values range from 44.1 to 45.7 mm (experiment DIST-PW in Table 3), which is more in keeping with observations of deep convective regions in the tropics.
A time series of the domain-averaged (a) precipitable water (mm), (b) thermal heat flux (black) and radiative flux divergence (gray) (W m−2), and (c) precipitation rate (mm day−1) for the CONTROL simulation for the entire 100-day simulation.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
A time series of the domain-averaged (a) precipitable water (mm), (b) thermal heat flux (black) and radiative flux divergence (gray) (W m−2), and (c) precipitation rate (mm day−1) for the CONTROL simulation for the entire 100-day simulation.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
A time series of the domain-averaged (a) precipitable water (mm), (b) thermal heat flux (black) and radiative flux divergence (gray) (W m−2), and (c) precipitation rate (mm day−1) for the CONTROL simulation for the entire 100-day simulation.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Domain-averaged precipitable water (PW), surface precipitation rate (PR), thermal heat flux (THF), radiative flux divergence (QRAD), and the difference between THF and QRAD averaged over days 60–100 for the CONTROL and aerosol experiments. Also shown is the PW averaged over disturbed regions (DIST-PW).
b. CONTROL convection characteristics
As the CONTROL simulation progresses it is evident that the water vapor field becomes organized into coherent moist and dry bands that tend to remain relatively stationary with time over the 100-day simulation period (Fig. 2) and are maintained by the circulations established between them. Such organization has been previously observed (e.g., Held et al. 1993; Tompkins 2001; Grabowski and Moncrieff 2004; Stephens et al. 2008). Self-aggregation of convection occurs preferentially within the regions of greatest available PW (Fig. 2). The scale of the band organization differs from that described in Stephens et al. (2008), which points to the role played by microphysical processes, as well as to the impact of better resolving the convection.
Hovmöller plots of precipitable water (mm) for the entire CONTROL simulation.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Hovmöller plots of precipitable water (mm) for the entire CONTROL simulation.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Hovmöller plots of precipitable water (mm) for the entire CONTROL simulation.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Temporally averaged zonal–height cross sections through these moist and dry bands are shown in Fig. 3. Many of the features typical of tropical convection (e.g., Simpson 1992; Johnson et al. 1999) are evident in this figure. The broad regions of ascent (descent) (Fig. 3a) are associated with regions of high (low) relative humidity (Fig. 3b). Three relatively moist layers are centered around ∼2, 5, and 11 km (Fig. 3b), arising as a result of detrainment associated with the stable or weakly unstable layers of the trade wind inversion, the freezing level (∼5 km), and the tropopause (∼15 km) (Fig. 3c). Both the regions of detrainment and the stable layers compare favorably with previous observations (Johnson et al. 1999). They are associated with three overturning circulations, the roles of which have previously been described (Posselt et al. 2008). Trade wind cumuli are evident below the strong trade wind inversion in the regions of large-scale subsidence. Also evident is the overshooting of these cumulus turrets above the inversion layer and the deepening of the trade cumulus layer as the disturbed regions are approached (Simpson 1992). Within the regions of deep ascent, the trade wind inversion does not exist as a true temperature inversion but rather a region of weakly decreasing temperature, and two populations of cumuli are evident, appearing to represent the forced and active trade cumulus described previously (e.g., Randall 1980; Johnson and Lin 1997). The condensate maxima at midlevels are associated with midlevel detrainment from congestus clouds, although detrainment from the sides of deep convection can also contribute to this detrainment (Johnson et al. 1999). This detrainment (Fig. 3b) is associated with enhanced cloud populations (Fig. 3d; see also below) and is supported by previous observational (Malkus and Riehl 1964; Mapes and Zuidema 1996) and modeling studies (Liu and Moncrieff 1998). As it will be seen below, congestus clouds may overshoot the 0°C stable layer. Should these clouds overshoot deep enough, the additional buoyancy obtained from glaciation will allow them to rise to the tropopause (Zuidema 1998). Finally, the deep layer of steep lapse rates in the upper troposphere is collocated with deep convective systems and the associated condensate maxima.
Zonal–height cross section showing temporally averaged (a) vertical velocity (contour interval 0.2 m s−1, with red indicating rising motion and blue sinking motion) and the 0°C isotherm (horizontal black line); (b) relative humidity (shading) and wind vectors (m s−1), where the vertical scale has been multiplied by 100); (c) dT/dz (K km−1); and (d) total condensate mixing ratio (g kg−1). Note that (b)–(d) are focused on the western disturbed region of (a).
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Zonal–height cross section showing temporally averaged (a) vertical velocity (contour interval 0.2 m s−1, with red indicating rising motion and blue sinking motion) and the 0°C isotherm (horizontal black line); (b) relative humidity (shading) and wind vectors (m s−1), where the vertical scale has been multiplied by 100); (c) dT/dz (K km−1); and (d) total condensate mixing ratio (g kg−1). Note that (b)–(d) are focused on the western disturbed region of (a).
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Zonal–height cross section showing temporally averaged (a) vertical velocity (contour interval 0.2 m s−1, with red indicating rising motion and blue sinking motion) and the 0°C isotherm (horizontal black line); (b) relative humidity (shading) and wind vectors (m s−1), where the vertical scale has been multiplied by 100); (c) dT/dz (K km−1); and (d) total condensate mixing ratio (g kg−1). Note that (b)–(d) are focused on the western disturbed region of (a).
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
The cloud fraction (defined where the total condensate is greater than 0.01 g kg−1) as a function of height is shown in Fig. 4. The trimodal distribution of convective clouds is once again apparent in this figure, with the modes being separated at heights of approximately 4 and 7 km. The shallow cloud mode, which takes into account both the trade wind cumulus of the subsidence regions and the trade-like cumuli of the ascending regions, is the most predominant mode, as is typically seen in observational studies (Rossow and Schiffer 1999; Sassen and Wang 2008). The middle or congestus mode, which peaks around 5 km, has the lowest cloud fractions of the three modes and will also include those clouds associated with the midlevel detrainment of moisture. Finally, the cloud fractions of the deep convective mode peak at ∼10 km and are slightly greater than those of the middle or congestus mode. Congestus clouds may also form part of the deep convective mode following glaciation.
Cloud fraction (total condensate mixing ratio > 0.01 g kg−1) as a function of height for all of the aerosol sensitivity tests.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Cloud fraction (total condensate mixing ratio > 0.01 g kg−1) as a function of height for all of the aerosol sensitivity tests.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Cloud fraction (total condensate mixing ratio > 0.01 g kg−1) as a function of height for all of the aerosol sensitivity tests.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Thus it is apparent that these simulations capture many of the cloud distribution characteristics of tropical convection within both the ascending and descending regions of the tropics. While the wet and dry bands of this simple model setup can be thought of as being representative, to some degree, of the ascending and descending branches of the Hadley or Walker circulations, strong caution must be exercised when making comparisons of such simple model setups (fixed SST, 2D, no Coriolis force) with reality. These simple simulations do, however, provide us with a self-consistent manner in which to investigate the impacts of aerosol indirect forcing on the range of different cloud types that do develop within the ascending and descending regions of the tropics. It is to these aerosol effects that we now turn our attention.
c. AIEs on RCE
The impacts of varying aerosol concentrations on the state of equilibrium may be deduced from Table 3, which shows various quantities averaged over days 60–100 of the simulation, the time period from when the aerosol layer was introduced to the end of the simulation. It is apparent that enhanced aerosol concentrations tend to result in increasing trends in PW, weak mixed trends in the DIST-PW, decreasing trends in the PR, decreases in the THF, and a somewhat mixed response in QRAD. The differences between THF and QRAD, which represents the state of RCE, tend to increase in magnitude with increasing aerosol concentrations, ranging from −1.1 W m−2 at low aerosol concentrations to about −3.4 W m−2 at high concentrations. However, the changes in all of these fields as a result of variations in aerosol concentrations are on the order of ∼5%. It would thus appear that the enhanced availability of aerosols to act as CCN does not significantly affect the state of RCE.
d. AIEs on convective organization
The impact of enhanced aerosol concentrations on the convective organization is evident in Fig. 5, which shows Hovmöller plots of PW for the various aerosol experiments. While there are some differences, the overall large-scale organization of the moist and dry regions remains relatively similar. These results therefore suggest that AIEs have a relatively small impact on the large-scale organization and self-aggregation of convection, and that the large-scale controls associated with radiational cooling and surface heat and moisture fluxes are more significant. However, AIF on the cloud fraction at each level is more significant, as is evident in Fig. 4. Enhanced aerosol concentrations result in a decrease in the cloud fraction of the shallow mode (although there is a weak increase in cloud fraction between ∼3 and ∼4 km, an increase in the middle or congestus mode, and a weak increase in the high-level clouds. The maximum changes in cloud fraction due to the greatest enhancements in available aerosol concentrations (between CCN-100 and CCN-1600) are on the order of approximately 35% for the shallow mode, 45% for the middle mode, and about 10% for the deep mode. AIEs therefore exert a relatively significant effect on low and middle cloud fractions but a weaker influence on the fraction of high clouds. The causes for this will be discussed below.
Hovmöller plots of precipitable water (mm) for the (a) CONTROL simulation and (b)–(f) all of the CCN experiments from when the aerosol layer was introduced.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Hovmöller plots of precipitable water (mm) for the (a) CONTROL simulation and (b)–(f) all of the CCN experiments from when the aerosol layer was introduced.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Hovmöller plots of precipitable water (mm) for the (a) CONTROL simulation and (b)–(f) all of the CCN experiments from when the aerosol layer was introduced.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
It was shown in Stephens et al. (2008) that the large-scale organization of the moist and dry bands was strongly dependent on the upper-tropospheric radiative heating by high clouds. While the impacts of enhanced CCN concentrations have a significant effect on the cloud fraction at low and middle levels, this effect is reduced for high clouds. It is thus reasonable to expect that the response of the large-scale convective organization to changes in CCN concentrations is weak, given the small response in high cloud fraction to enhanced aerosol concentrations. The impacts of AIEs on the large-scale organization may be significantly greater when considering aerosol species that serve as IN (Carrió et al. 2007).
e. Dynamic response to AIF
Variations in the cloud microphysics through enhanced aerosol concentrations have previously been observed to have an impact on the storm dynamics (van den Heever and Cotton 2004; Khain et al. 2005; Seifert and Beheng 2006; van den Heever and Cotton 2007). The average updraft velocities as a function of height, expressed as a difference from the CCN-100 simulation, are shown for all the aerosol experiments in Fig. 6. The averages are calculated for all of the model domain points where the updraft velocities are greater than 0, 1, and 5 m s−1, respectively. It is apparent from Fig. 6 that increased aerosol concentrations result in general in increased updraft strengths, being particularly obvious for the higher aerosol cases. There is a mixed response for the middle clouds for the lower updraft thresholds. The maximum differences between the aerosol experiments and the CCN-100 case are on the order of ∼10% to ∼15% of the mean updraft velocities for all three velocity thresholds of the CCN-100 case. Accompanying the enhanced updraft strengths is an increase in the domain fraction covered by these stronger updrafts (Fig. 7), although the domain fraction is less for the CCN-1600 case in the upper levels for the updrafts greater than 5 m s−1. The maximum fractional domain differences are approximately 2%, 15%–20%, and 50% (reaching close to 100% for some levels) of the fractional means for the 0, 1, and 5 m s−1 velocity thresholds in the CCN-100 case, respectively. These simulations therefore demonstrate that enhanced aerosol concentrations are associated with stronger updrafts, and these stronger updrafts cover a greater fraction of the domain more frequently, a trend that becomes more evident when focusing on those updrafts larger in magnitude. It is interesting to note in these simulations that the impacts of AIEs are seen not only in the upper levels where the suppression of the warm rain process results in greater amounts of cloud water being available for lofting, freezing, and latent heat release in the upper levels of the cloud, but also in the middle mode and the upper regions of lower mode where the formation of greater amounts of cloud water due to enhanced CCN concentrations appears to produce a similar effect.
Average updraft velocity expressed as a difference from the CCN-100 field for (a) w > 0, (b) w > 1, and (c) w > 5 m s−1.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Average updraft velocity expressed as a difference from the CCN-100 field for (a) w > 0, (b) w > 1, and (c) w > 5 m s−1.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Average updraft velocity expressed as a difference from the CCN-100 field for (a) w > 0, (b) w > 1, and (c) w > 5 m s−1.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Fraction (%) of the domain covered by updrafts that are (a) >0, (b) >1, and (c) >5 m s−1 expressed as a difference from the CCN-100 field.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Fraction (%) of the domain covered by updrafts that are (a) >0, (b) >1, and (c) >5 m s−1 expressed as a difference from the CCN-100 field.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Fraction (%) of the domain covered by updrafts that are (a) >0, (b) >1, and (c) >5 m s−1 expressed as a difference from the CCN-100 field.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
f. AIEs on precipitation rates
The precipitation rates, averaged over the entire domain for the last 40 days of the simulation, are shown in Table 3. It is apparent from this table that the grid domain averaged precipitation rates decrease with increasing aerosol concentrations, although there is an increase in these rates between CCN-100 and CCN-200. The decreasing trend in the domain-averaged precipitation is in keeping with the first AIE. The range in response of the domain-averaged precipitation between CCN-200 and CCN-1600 is weak, being only approximately 7% for a 16 times increase in aerosol concentrations. However, when the precipitation is averaged only over those points that are precipitating, the precipitation rates increase with increasing CCN (Fig. 8a), being on the order of ∼13% and 15% for the 1 and 5 mm day−1 thresholds, respectively. This is associated with a decrease in fractional domain coverage of approximately 10%, 17%, and 16% for these cases (not shown). A frequency distribution plot for rainfall rates (mm h−1) for the CCN-200 and CCN-800 cases demonstrates that the frequency of those systems producing rainfall rates less than ∼7 mm h−1 are reduced in the presence of enhanced CCN concentrations, whereas the frequency of heavier rainfall producing systems is enhanced under polluted conditions, a fact that is supported below when examining the frequency and precipitation contributions of congestus and deep convective systems. Such trends can be attributed to the more frequent, stronger updrafts associated with polluted conditions.
(a) Average precipitation rates for all those points where precipitation rates are >0.24, >1, and >5 mm day−1; and (b) precipitation rate (mm h−1) frequency distribution plot for the CCN200 and CCN800 cases. The frequencies are expressed in a relative sense where, for example, 50% implies that the frequency for some precipitation rate is the same in both cases.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
(a) Average precipitation rates for all those points where precipitation rates are >0.24, >1, and >5 mm day−1; and (b) precipitation rate (mm h−1) frequency distribution plot for the CCN200 and CCN800 cases. The frequencies are expressed in a relative sense where, for example, 50% implies that the frequency for some precipitation rate is the same in both cases.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
(a) Average precipitation rates for all those points where precipitation rates are >0.24, >1, and >5 mm day−1; and (b) precipitation rate (mm h−1) frequency distribution plot for the CCN200 and CCN800 cases. The frequencies are expressed in a relative sense where, for example, 50% implies that the frequency for some precipitation rate is the same in both cases.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
g. AIEs on water and ice characteristics
The temporally and spatially averaged vertically integrated hydrometeor species are shown in Fig. 9. Enhanced aerosol concentrations result in an overall increase in the amount of cloud water and a decrease in the amount of rainwater produced (Fig. 9a), which is in keeping with AIF theory and the suppression of the warm rain process. The increase in cloud water over the range of experiments is ∼22%, while the decrease in rain is ∼27%. Overall, the amount of liquid water within the atmosphere increases (∼7%), with increasing aerosol concentrations being dominated by the cloud water trends. Increases in both the amount of cloud ice (pristine ice + snow + aggregates) (15%) and in graupel and hail (36%) also occur as the CCN concentrations are increased (Fig. 9b), resulting in a total ice mass increase of ∼23%. With the suppression of the warm rain process and the resultant increase in cloud water, more cloud water is available to be vertically transported within the storm, and hence is available for participation in ice formation processes. Increased temperatures and subsequent changes in supersaturation as a result of the latent heat release also contribute to the ice content. The ratio of the ice mass to that of liquid water increases by ∼18% across the range of aerosol experiments (not shown), which can have significant implications for the earth’s radiation budget given their different scattering properties.
Temporally and spatially averaged vertically integrated fields of (a) cloud, rain, and total liquid, and (b) PSA (pristine ice + snow + aggregates), GH (graupel + hail), and total ice for the various aerosol experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Temporally and spatially averaged vertically integrated fields of (a) cloud, rain, and total liquid, and (b) PSA (pristine ice + snow + aggregates), GH (graupel + hail), and total ice for the various aerosol experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Temporally and spatially averaged vertically integrated fields of (a) cloud, rain, and total liquid, and (b) PSA (pristine ice + snow + aggregates), GH (graupel + hail), and total ice for the various aerosol experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
h. AIF of cloud mode characteristics
As we saw from Fig. 4, the convection within these simulations is distributed in a trimodal manner, with maxima in cloud fraction being separated around 4 and 7 km AGL. The characteristics of these three cloud modes will now be discussed in more detail. For the analysis presented here, low clouds are defined to have cloud tops below 4 km, middle clouds have cloud tops between 4 and 7 km AGL, and high clouds have tops above 7 km AGL. Cloud top was determined by starting at the top of each model column and sampling downward until reaching the level where the total condensate was greater than 0.01 g kg−1. This sampling procedure has weaknesses as discussed below. Such weaknesses are also encountered with satellite data.
Frequency distributions of low, middle, and high cloud modes for each of the aerosol experiments are shown in Fig. 10. The total cloud fraction is approximately 40%. From Fig. 10 it is apparent that the low cloud mode is predominant, with a cloud fraction of between 20% and 25%; the middle mode has cloud fractions of between 3% and 5% and the high mode between 12% and 15%. Enhanced CCN concentrations result in a decrease in low cloud frequency on the order of magnitude of 22% across the aerosol experiments, but an increase in the middle and high level cloud modes on the orders of 51% and 6%, respectively. The enhanced CCN concentrations in these large-scale simulations therefore do not result in greater low cloud fractions, as suggested by the second AIE, and are more in keeping with the findings of Xue and Feingold (2006) and Jiang et al. (2006). While this has been attributed to greater entrainment rates in the more polluted cases (Xue and Feingold 2006; Altaratz et al. 2007), this cannot be assessed with the grid dimensions utilized here. The increased cloud frequencies in the upper levels with enhanced CCN concentrations are associated with more frequent, stronger updrafts in these cases. The reasons for the increased frequency of the middle-level clouds appear to be related to the enhanced buoyancy associated with ice processes as these clouds penetrate the 0°C stable layer. This is further evident when examining cloud echo-top height–precipitation echo-top height (CETH–PETH) histograms (Masunaga et al. 2005; Stephens and Wood 2007) produced using Quickbeam, a CloudSat simulator (Haynes et al. 2007) that generates a radar reflectivity field using model output (Fig. 11). The CETH and PETH are determined at the level of the −30- and 10-dBZ returns, respectively, and the fields shown in Fig. 11 are difference plots from CCN-100. It is evident from this plot that enhanced CCN concentrations produce an increase (decrease) in the frequency of the population with cloud tops above (below) the freezing level. This appears to occur as a result of the suppression of warm rain processes in the lower population, which results in more cloud water being available for lofting above the 0°C, upon which freezing releases latent heat thereby enhancing the buoyancy. This is supported by the trends in liquid (LWP) and ice water path (IWP) for the congestus modes (shown in Tables 6 and 7 and discussed below). The trends in the other cloud types evident in Fig. 11 are in keeping with those previously discussed.
Frequency occurrence (%) of low (<4 km), middle (4–7 km), and high (>7 km) cloud for each of the aerosol experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Frequency occurrence (%) of low (<4 km), middle (4–7 km), and high (>7 km) cloud for each of the aerosol experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Frequency occurrence (%) of low (<4 km), middle (4–7 km), and high (>7 km) cloud for each of the aerosol experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Normalized frequency distribution plots of the cloud echo-top height vs the precipitation echo-top height expressed as a difference from the CCN-100 case for (a) CCN-200, (b) CCN-400, (c) CCN-800, and (d) CCN-1600 experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Normalized frequency distribution plots of the cloud echo-top height vs the precipitation echo-top height expressed as a difference from the CCN-100 case for (a) CCN-200, (b) CCN-400, (c) CCN-800, and (d) CCN-1600 experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
Normalized frequency distribution plots of the cloud echo-top height vs the precipitation echo-top height expressed as a difference from the CCN-100 case for (a) CCN-200, (b) CCN-400, (c) CCN-800, and (d) CCN-1600 experiments.
Citation: Journal of the Atmospheric Sciences 68, 4; 10.1175/2010JAS3603.1
The weaknesses of classifying clouds into low, middle, and high clouds by sampling down a column from the top of the atmosphere include not correctly accounting for multilevel cloud systems, undercounting those cloud types that are located under higher cloud types, and grouping together cloud types (such as anvil cirrus and deep convection) that have very different contributions to processes such as surface precipitation rates. A powerful aspect of a CRM is that clouds can be more accurately classified into various cloud types rather than simply into low, middle, and high clouds. The output from all the aerosol experiments was reclassified using the scheme shown in Table 4. In this table, M1 refers to those clouds below 4 km AGL, M2 to those clouds located between 4 and 7 km AGL, and M3 to those clouds with cloud tops above 7 km AGL. The arrows in this scheme, such as for M1→M3, imply that cloud extends from M1 to M3 (below 4 km to above 7 km in this example), while the ampersand symbol in the scheme, such as in “M1&M3,” indicates that two discontinuous layers of cloud exist (one below 4 km AGL and one above 7 km AGL in this example). Only 2% of all the cloudy points within these simulations do not fall into one of the types of this classification scheme.
Table showing the cloud classification schemes referred to in the text. The ovals represent the presence of cloud (defined to exist wherever the total condensate is greater than 0.01 g kg−1). The surface is indicated by thick horizontal lines, while the 4- and 7-km AGL levels are shown with single horizontal lines. The name of each element of the scheme is given below each figure in the table. The arrow indicates that the cloud extends from one level to the next, while the ampersand implies the existence of more than one layer of cloud, but these clouds are not joined (e.g., MM1&MM3 indicates the presence of multiple layers below 4 km and above 7 km AGL, respectively).
Using this cloud classification scheme, the overall cloud fraction for the CCN-100 case is ∼42%, and while this does vary as a result of the aerosol concentrations, the variation is not significant. Single cloud layers contribute ∼71% to the cloud cover, while double and triple cloud layers make up ∼21% and 6%, respectively. Multilayered cloud systems therefore make up ∼29% of the systems occurring within these simulations. Stephens and Wood (2007) emphasized the importance of multilayered structures in their radar analysis of tropical convection, finding that precipitating systems were multilayered between 45% and 53% of the time. For these simulations, surface precipitation was associated with multilayered systems ∼35% of the time, thus also demonstrating the importance of multilayered systems in the tropical precipitation.
The cloud fraction of each cloud type, as well as the contributions made by each cloud type to the LWP, IWP, and the surface precipitation, all as a function of aerosol concentration, are shown in Tables 5–8, respectively. The cloud types are grouped into low, middle, and high clouds by their cloud tops, but their contributions are kept distinct. The cloud type with the greatest frequency of occurrence is M1 (e.g., trade wind cumulus), followed by M3 (e.g., cirrus or upper-level anvil clouds). It is apparent from Table 5 that the trends in modal frequency are consistent with those discussed previously, and that the relative changes in frequency of middle clouds is more susceptible to AIF than the other two modes, although their initial frequencies are low. This represents the importance of AIF on those cloud systems that vary between being warm-phase and mixed-phase systems, with AIEs assisting in the transition between these two phases through buoyancy effects.
Fractional coverage (as a percentage of the entire domain) of the various cloud modes described in the text for the different aerosol experiments. The magnitude of the aerosol response (%) is shown in the last column (Aero effect) and is given by [1 − (minimum value/maximum value)] × 100.
As in Table 5, but for the fractional contributions by the various cloud modes to the LWP for the different aerosol experiments.
As in Table 5, but for the fractional contributions (%) of the different cloud types to the IWP for the various aerosol experiments.
As in Table 5, but for the fractional contributions by the various cloud modes to the surface precipitation rates for the different aerosol experiments.
The greatest contributors to the LWP are low-level cumulus/stratocumulus (M1 and MM1), deep convection (M1→M3), congestus below upper-level cirrus (M1→M2&M3), and cirrus overlaying low-level cumulus–stratocumulus (M1&M3). All of the other cloud types contribute less than 7% to the LWP, although the congestus contributions are close to this. Enhanced CCN concentrations have the greatest impacts on the LWP contributions of the middle clouds, leading to an increase of ∼30%, while the high cloud contributions to the LWP increase by only ∼4.4%. Enhanced aerosol concentrations result in a decrease in the LWP of ∼20% for low clouds. That the LWP decreases with enhanced CCN concentrations for these low-level clouds is not in keeping with the traditional second AIE, but rather is in keeping with other studies (Jiang et al. 2006; Xue and Feingold 2006; Altaratz et al. 2007). Deep convection is by far the most significant contributor to the IWP of all the cloud types contributing on the order of 79%–80% of the ice mass (Table 7). Contributions to the IWP also come from those cloud types involving M3 (upper-level cirrus), as well as those cloud types in which M1→M2 (congestus) forms a part. While the contributions from the high clouds including deep convection vary little in response to AIF (∼1%), the middle clouds, in particular congestus, show a much greater variation (76.5%), although their overall contribution to the total IWP is small.
As with the IWP, the most significant contributor to the surface precipitation (Table 8) is the deep convective mode (M1→M3), contributing between 60 and 70%, depending on the aerosol concentration. Lesser, although still significant contributors include congestus under upper-level cirrus (M1→M2&M3), low-level cumulus–stratocumulus (M1), and congestus (M1→M2). Enhanced aerosol concentrations cause a reduction of ∼60% in the contributions to surface precipitation made by the low-level clouds over the range of aerosols considered here as a result of the suppression of the warm rain process. A mixed precipitation response in the middle mode and a ∼8% high cloud enhancement are evident. Precipitation contributions from the deep convective clouds, those clouds contributing most significantly to the total precipitation, are increased by ∼14% across the range of aerosol concentrations, while congestus contributions are increased by ∼19% and are associated with the aerosol dynamic effect. These results therefore suggest that the impacts of enhanced CCN concentrations are to decrease the cloud fraction, LWP, and precipitation contributions from low-level clouds and to increase the cloud fraction, LWP, and precipitation contributions from congestus and deep convective clouds.
4. Discussion and conclusions
The impacts of enhanced CCN concentrations on the characteristics of tropical convection have been investigated through the use of large-domain, fine-resolution, long-duration idealized CRM simulations conducted under a RCE framework. These simulations replicated numerous features typical of tropical convection. Aerosol concentrations ranging from 100 to 1600 cm−3 were released between 2 and 4 km AGL, being representative of an outbreak of dust over oceanic regions. The scale and duration of these simulations provide a unique opportunity to investigate the impacts of aerosol indirect forcing on the wide range of cloud distributions that develop within a consistent model framework.
The CRM model results demonstrate that the large-scale organization of convection is only weakly influenced by variations in CCN concentrations, with the large-scale dynamics exerting the far more dominant control. Also evident is that decreases in the domain-averaged precipitation (∼7%) and domain total cloud fraction (∼6%) associated with enhanced CCN concentrations are small in magnitude. Such weak, large-scale responses in precipitation and cloud fraction are in keeping with findings of Grabowski (2006). While the large-scale, domain-averaged responses to increased CCN concentrations tend to be weak, both modal and local responses to aerosol indirect forcing tend to be larger in magnitude. Cloud frequencies vary by approximately 22%, 51%, and 6% for the shallow, middle, and high modes of the trimodal distribution in response to enhanced CCN concentrations. The precipitation contribution from shallow clouds decreases on the order of 60%, shows a mixed response for the middle mode, and increases by 8% for the high cloud mode, with contributions from the analysis of specific clouds types being even greater (14% for deep convection and 19% for congestus clouds). The LWP contributions from shallow clouds are found to decrease, but they increase for the middle and high clouds modes, and the IWP contributions vary most for the middle mode, although their contributions to the total IWP are small. The different responses to aerosol indirect forcing based on the cloud type or cloud mode evident in this study provide support of other previous studies in which the importance of cloud type is stressed (e.g., Seifert and Beheng 2006; Khain et al. 2008). A dynamic response to increased aerosol concentrations is evident, with increases both in updraft speed and frequency. The influence of aerosol indirect forcing on the total cloud water (22% increase), rain (27% decrease), cloud ice (15% increase), graupel and hail (36% increase), and the ratio of ice to liquid water (18% increase) are also not insignificant and tend to be in keeping with aerosol indirect forcing theories.
It should be noted that the domain-averaged responses in precipitation and water vapor are largely constrained by the fixed SST assumption of these RCE simulations, and hence it may not be unexpected that changes in the domain-averaged responses are small. However, aspects associated with enhanced CCN concentrations (such as greater surface evaporation rates driven by stronger storms, variations in evaporation rates associated with changes in the drop size distribution, and differences in radiative forcing due to the changes in cloud fraction, type, and vertical distribution) will also all impact the latent and sensible heat fluxes, thus indicating that fixed SSTs are not the only constraint on the system.
In summary, the RCE simulations presented here demonstrate that for tropical environments, which are never far from a state of RCE, the large-scale organization, domain-averaged precipitation, and total cloud fraction responses to enhanced CCN concentrations associated with dust intrusions are relatively weak. However, the role of aerosol indirect forcing may be quite substantial when examining localized effects or the influences pertaining to a specific cloud type within the tropics. Aerosol indirect forcing associated with the three tropical cloud modes was found to be quite significant in magnitude and often different in sign, a fact which appears to contribute to the weaker domain-averaged response. These simulations therefore suggest that aerosol indirect effects associated with shallow clouds in the tropics may offset or compensate for the aerosol indirect effects associated with congestus and deep convection systems and vice versa, thus producing a more moderate domain-wide response to aerosol indirect forcing.
Acknowledgments
This research was supported by the National Science Foundation under Grant ATM-0820557, with some of the initial investigation being funded by the National Oceanic and Atmospheric Administration under Grant NA17RJ1228. The authors are very grateful for the helpful comments and suggestions of the anonymous reviewers, which significantly improved this manuscript.
REFERENCES
Ackerman, A. S., M. P. Kirkpatrick, D. E. Stevens, and O. B. Toon, 2004: The impact of humidity above stratiform clouds on indirect aerosol climate forcing. Nature, 432, 1014–1017.
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, 2007: Aerosols’ influence on the interplay between condensation, evaporation and rain in warm cumulus cloud. Atmos. Chem. Phys. Discuss., 7, 12 687–12 714.
Andreae, M. O., D. Rosenfeld, P. Artaxo, A. A. Costa, G. P. Frank, K. M. Longo, and M. A. F. Silva-Dias, 2004: Smoking rain clouds over the Amazon. Science, 303, 1337–1342.
Arakawa, A., 2004: The cumulus parameterization problem: Past, present and future. J. Climate, 17, 2493–2525.
Augstein, E., H. Riehl, R. Ostapoff, and V. Wagner, 1973: Mass and energy transports in an undisturbed Atlantic trade wind flow. Mon. Wea. Rev., 101, 101–111.
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.
Betts, A., 1973: Non-precipitating cumulus convection and its parameterization. Quart. J. Roy. Meteor. Soc., 99, 178–196.
Borys, R. D., D. H. Lowenthal, M. A. Wetzel, F. Herrera, A. Gonzalez, and J. Harris, 1998: Chemical and microphysical properties of marine stratiform cloud in the North Atlantic. J. Geophys. Res., 103, 22 073–22 085.
Bretherton, C. S., and P. K. Smolarkiewicz, 1989: Gravity waves, compensating subsidence and detrainment around cumulus clouds. J. Atmos. Sci., 46, 740–759.
Bretherton, C. S., P. N. Blossey, and M. Khairoutdinov, 2005: An energy-balance analysis of deep convective self-aggregation above uniform SST. J. Atmos. Sci., 62, 4273–4292.
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.
Cheng, W. Y. Y., G. G. Carrió, W. R. Cotton, and S. M. Saleeby, 2009: Influence of cloud condensation and giant cloud condensation nuclei on the development or precipitating trade wind cumuli in a large eddy simulation. J. Geophys. Res., 114, D08201, doi:10.1029/2008JD011011.
Coakley, J. A., and C. D. Walsh, 2002: Limits to the aerosol indirect radiative effect derived from observations of ship tracks. J. Atmos. Sci., 59, 668–680.
Cotton, W. R., and Coauthors, 2003: RAMS 2001: Current status and future directions. Meteor. Atmos. Phys., 82, 5–29.
Feingold, G., and A. J. Heymsfield, 1992: Parameterizations of condensational growth of droplets for use in general circulation models. J. Atmos. Sci., 49, 2325–2342.
Feingold, G., S. Tzivion, and Z. Levin, 1988: Evolution of raindrop spectra. Part I: Solution to the stochastic collection/breakup equation using the method of moments. J. Atmos. Sci., 45, 3387–3399.
Feingold, G., R. L. Walko, B. Stevens, and W. R. Cotton, 1998: Simulations of marine stratocumulus using a new microphysical parameterization scheme. Atmos. Res., 47–48, 505–258.
Feingold, G., W. L. Eberhard, D. E. Veron, and M. Previdi, 2003: First measurements of the Twomey indirect effect using ground-based remote sensors. Geophys. Res. Lett., 30, 1287, doi:10.1029/2002GL016633.
Fierro, A. O., J. Simpson, M. A. LeMone, J. M. Straka, and B. F. Smull, 2009: On how hot towers fuel the Hadley Cell: An observational and modeling study of line-organized convection in the equatorial trough from TOGA COARE. J. Atmos. Sci., 66, 2730–2746.
Fridland, A. M., and Coauthors, 2004: Evidence for the predominance of mid-tropospheric aerosols as subtropical anvil cloud nuclei. Science, 304, 718–722.
Grabowski, W. W., 2006: Indirect impact of atmospheric aerosols in idealized simulations of convective–radiative quasi equilibrium. J. Climate, 19, 4664–4682.
Grabowski, W. W., and M. W. Moncrieff, 2004: Moisture–convection feedback in the tropics. Quart. J. Roy. Meteor. Soc., 130, 3081–3104.
Han, Q., W. B. Rossow, J. Zeng, and R. Welch, 2002: Three different behaviors of liquid water path of water clouds in aerosol–cloud interactions. J. Atmos. Sci., 59, 726–735.
Harrington, J. Y., 1997: The effects of radiative and microphysical processes on simulated warm and transition season Arctic stratus. Ph.D. dissertation, Colorado State University, 289 pp. [Available from Department of Atmospheric Science, Colorado State University, Fort Collins, CO 80523.]
Haynes, J. M., R. T. Marchand, Z. Luo, A. Bodas-Salcedo, and G. L. Stephens, 2007: A multi-purpose radar simulation package: QuickBeam. Bull. Amer. Meteor. Soc., 88, 1723–1727.
Held, I. M., R. S. Hemler, and V. Ramaswamy, 1993: Radiative–convective equilibrium with explicit two-dimensional moist convection. J. Atmos. Sci., 50, 3909–3927.
Heymsfield, A. J., and G. M. McFarquhar, 2001: Microphysics of INDOEX clean and polluted trade cumulus clouds. J. Geophys. Res., 106, 28 653–28 673.
Hill, G. E., 1974: Factors controlling the size and spacing of cumulus clouds as revealed by numerical experiments. J. Atmos. Sci., 31, 646–673.
Houghton, J. T., Y. Ding, D. J. Griggs, M. Noguer, P. J. van den Linden, X. Dai, K. Maskell, and C. A. Johnson, 2001: Climate Change 2001: The Scientific Basis. Cambridge University Press, 881 pp.
Hudson, J. G., and S. S. Yum, 2001: Maritime–continental drizzle contrasts in small cumuli. J. Atmos. Sci., 58, 915–926.
Jiang, H., G. Feingold, and W. R. Cotton, 2002: Simulations of aerosol-cloud-dynamical feedbacks resulting from entrainment of aerosol into the marine boundary layer during the Atlantic Stratocumulus Transition Experiment. J. Geophys. Res., 107, 4813, doi:10.1029/2001JD001502.
Jiang, H., H. Xue, A. Teller, G. Feingold, and Z. Levin, 2006: Aerosol effects on the lifetime of shallow cumulus. Geophys. Res. Lett., 33, L14806, doi:10.1029/2006GL026024.
Jirak, I. L., and W. R. Cotton, 2006: Effect of air pollution on precipitation along the Front Range of the Rocky Mountains. J. Appl. Meteor. Climatol., 45, 236–245.
Johnson, R. H., and X. Lin, 1997: Episodic trade wind regimes over the western Pacific warm pool. J. Atmos. Sci., 54, 2020–2034.
Johnson, R. H., T. M. Rickenbach, S. A. Rutledge, P. E. Ciesielski, and W. H. Schubert, 1999: Trimodal characteristics of tropical convection. J. Climate, 12, 2397–2418.
Kaufman, Y. J., and T. Nakajima, 1993: Effect of Amazon smoke on cloud microphysics and albedo analysis from satellite imagery. J. Appl. Meteor., 32, 729–744.
Kaufman, Y. J., I. Koren, L. A. Remer, D. Rosenfeld, and Y. Rudich, 2005: The effect of smoke, dust, and pollution aerosol on shallow cloud development over the Atlantic Ocean. Proc. Natl. Acad. Sci. USA, 102, 11 207–11 212.
Kessler, E., 1969: On the Distribution and Continuity of Water Substance in Atmospheric Circulation. Meteor. Monogr., No. 32, Amer. Meteor. Soc., 84 pp.
Khain, A. P., A. Pokrovsky, M. Pinsky, and A. Seifert, V. Phillips, 2004: Simulation of effects of atmospheric aerosols on deep turbulent convective clouds using a spectral microphysics mixed-phase cumulus cloud model. Part I: Model description and possible applications. J. Atmos. Sci., 61, 2963–2982.
Khain, A. P., D. Rosenfeld, and A. Pokrovsky, 2005: Aerosol impact on the dynamics and microphysics of deep convective clouds. Quart. J. Roy. Meteor. Soc., 131, 2639–2663.
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.
Kiehl, J. T., and K. E. Trenberth, 1997: Earth’s annual global mean energy budget. Bull. Amer. Meteor. Soc., 78, 197–208.
Lee, S. S., J. E. Penner, and S. M. Saleeby, 2009: Aerosol effect on liquid-water path of thin stratocumulus clouds. J. Geophys. Res., 114, D07204, doi:10.1029/2008JD010513.
Lilly, D. K., 1962: On the numerical simulation of buoyant convection. Tellus, 14, 148–172.
Liu, C., and M. W. Moncrieff, 1998: A numerical study of the diurnal cycle of tropical oceanic convection. J. Atmos. Sci., 55, 2329–2344.
Lu, M.-L., and J. H. Seinfeld, 2005: Study of the aerosol indirect effect by large-eddy simulation of marine stratocumulus. J. Atmos. Sci., 62, 3909–3932.
Malkus, J. S., and H. Riehl, 1964: Cloud Structure and Distributions over the Tropical Pacific Ocean. University of California Press, 229 pp.
Mapes, B. E., and P. Zuidema, 1996: Radiative-dynamical consequences of dry tongues in the tropical atmosphere. J. Atmos. Sci., 53, 620–638.
Masunaga, H., T. L’Ecuyer, and C. D. Kummerow, 2005: Variability in the characteristics of tropical precipitation systems. Part I: Spatial structure. J. Climate, 18, 823–840.
Matsui, T., S. M. Kreidenweis, R. A. Pielke Sr., B. Schichtel, H. Yu, M. Chin, D. A. Chu, and D. Niyogi, 2004: Regional comparison and assimilation of GOCART and MODIS aerosol optical depth across the eastern U.S. Geophys. Res. Lett., 31, L21101, doi:10.1029/2004GL021017.
Matsui, T., H. Masunaga, S. M. Kreidenweis, R. A. Pielke Sr., W.-K. Tao, M. Chin, and Y. Kaufman, 2006: Satellite-based assessment of global warm cloud properties associated with aerosol, atmospheric stability, and diurnal cycle. J. Geophys. Res., 111, D17204, doi:10.1029/2005JD006097.
Mesinger, F., and A. Arakawa, 1976: Numerical methods used in atmospheric models. GARP Publication Series 14, WMO/ICSU Joint Organizing Committee, 64 pp.
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.
Nenes, A., and J. H. Seinfeld, 2003: Parameterization of cloud droplet formation in global climate models. J. Geophys. Res., 108, 4415, doi:10.1029/2002JD002911.
Nitta, T., and S. Esbensen, 1974: Heat and moisture budget analyses using BOMEX data. Mon. Wea. Rev., 102, 17–28.
Pakula, L., and G. L. Stephens, 2009: The role of radiation in influencing tropical cloud distributions in a radiative–convective equilibrium cloud-resolving model. J. Atmos. Sci., 66, 62–76.
Pielke, R. A., and Coauthors, 1992: A comprehensive meteorological modeling system—RAMS. Meteor. Atmos. Phys., 49, 69–91.
Platnick, S., and Coauthors, 2000: The role of background cloud microphysics in the radiative formation of ship tracks. J. Atmos. Sci., 57, 2607–2624.
Posselt, D. J., S. C. van den Heever, and G. L. Stephens, 2008: Trimodal cloudiness and tropical stable layers in simulations of radiative convective equilibrium. Geophys. Res. Lett., 35, L08802, doi:10.1029/2007GL033029.
Prospero, J. M., and T. N. Carlson, 1981: Saharan air outbreaks over the tropical North Atlantic. Pure Appl. Geophys., 119, 677–691.
Randall, D. A., 1980: Conditional instability of the first kind upside-down. J. Atmos. Sci., 37, 125–130.
Randall, D. A., Q. Hu, K.-M. Xu, and S. K. Krueger, 1994: Radiative-convective disequilibrium. Atmos. Res., 31, 315–327.
Riehl, H., and J. S. Malkus, 1958: On the heat balance in the equatorial trough zone. Geophysica, 6, 503–538.
Robe, F. R., and K. A. Emanuel, 1996: Dependence of tropical convection on radiative forcing. J. Atmos. Sci., 53, 3265–3275.
Rosenfeld, D., 1999: TRMM observed first direct evidence of smoke from forest fires inhibiting rainfall. Geophys. Res. Lett., 26, 3105–3108.
Rosenfeld, D., 2000: Suppression of rain and snow by urban and industrial air pollution. Science, 287, 1793–1796.
Rosenfeld, D., and I. M. Lensky, 1998: Satellite-based insights into precipitation formation processes in continental and maritime convective clouds. Bull. Amer. Meteor. Soc., 79, 2457–2476.
Rossow, W. B., and R. A. Schiffer, 1999: Advances in understanding clouds from ISCCP. Bull. Amer. Meteor. Soc., 80, 2261–2287.
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. Meteor., 43, 182–195.
Saleeby, S. M., W. R. Cotton, D. Lowenthal, R. D. Borys, and M. A. Wetzel, 2009: Influence of cloud condensation nuclei on orographic snowfall. J. Appl. Meteor. Climatol., 48, 903–922.
Saleeby, S. M., W. Berg, S. C. van den Heever, and T. L’Ecuyer, 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.
Sassen, K., and Z. Wang, 2008: Classifying clouds around the globe with the CloudSat radar: 1 year of results. Geophys. Res. Lett., 35, L04805, doi:10.1029/2007GL032591.
Seifert, A., and K. D. Beheng, 2006: A two-moment cloud microphysics parameterization for mixed-phase clouds. Part 2: Maritime vs continental deep convective storms. Meteor. Atmos. Phys., 92, 67–82.
Simpson, J., 1992: Global circulation and tropical cloud activity. The Global Role of Tropical Rainfall, J. S. Theon et al., Eds., A. Deepak, 77–92.
Smagorinsky, J., 1963: General circulation experiments with the primitive equations. I. The basic experiment. Mon. Wea. Rev., 91, 99–164.
Solomon, S., and Coauthors, 2007: Technical summary. Climate Change 2007: The Physical Science Basis, S. Solomon et al., Eds., Cambridge University Press, 19–91.
Stephens, G. L., 1990: On the relationship between water vapor over the oceans and sea surface temperature. J. Climate, 3, 634–645.
Stephens, G. L., 2005: Cloud feedback in the climate system: A critical review. J. Climate, 18, 237–273.
Stephens, G. L., and N. B. Wood, 2007: Properties of tropical convection observed by millimeter-wave radar systems. Mon. Wea. Rev., 135, 821–842.
Stephens, G. L., A. Slingo, M. J. Webb, P. J. Minnett, P. H. Daum, L. Kleinman, I. Wittmeyer, and D. A. Randall, 1994: Observations of the Earth’s Radiation Budget in relation to atmospheric hydrology. 4. Atmospheric column radiative cooling over the world’s oceans. J. Geophys. Res., 99, 18 585–18 604.
Stephens, G. L., P. J. Webster, R. H. Johnson, R. Engelen, and T. L’Ecuyer, 2004a: Observational evidence for the mutual regulation of the tropical hydrological cycle and tropical sea surface temperatures. J. Climate, 17, 2213–2224.
Stephens, G. L., N. B. Wood, and L. A. Pakula, 2004b: On the radiative effects of dust on tropical convection. Geophys. Res. Lett., 31, L23112, doi:10.1029/2004GL021342.
Stephens, G. L., S. C. van den Heever, and L. A. Pakula, 2008: Radiative–convective feedbacks in idealized states of radiative–convective equilibrium. J. Atmos. Sci., 65, 3899–3916.
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.
Su, H., C. Bretherton, and S. S. Chen, 2000: Self-aggregation and large-scale control of tropical deep convection: A modeling study. J. Atmos. Sci., 57, 1797–1816.
Sui, C. H., K. M. Lau, W.-K. Tao, and J. Simpson, 1994: The tropical water and energy cycles in a cumulus ensemble model. Part I: Equilibrium climate. J. Atmos. Sci., 51, 711–728.
Tompkins, A. M., 2000: The impact of dimensionality on long-term cloud-resolving model simulations. Mon. Wea. Rev., 128, 1521–1535.
Tompkins, A. M., 2001: Organization of tropical convection in low vertical wind shears: The role of water vapor. J. Atmos. Sci., 58, 529–545.
Tompkins, A. M., and G. C. Craig, 1998a: Radiative–convective equilibrium in a three-dimensional cloud-ensemble model. Quart. J. Roy. Meteor. Soc., 124, 2073–2097.
Tompkins, A. M., and G. C. Craig, 1998b: Time-scales of adjustment to radiative-convective equilibrium in the tropical atmosphere. Quart. J. Roy. Meteor. Soc., 124, 2693–2713.
Twomey, S., 1974: Pollution and the planetary albedo. Atmos. Environ., 8, 1251–1256.
van den Heever, S. C., and W. R. Cotton, 2004: The impact of hail size on simulated supercell storms. J. Atmos. Sci., 61, 1596–1609.
van den Heever, S. C., and W. R. Cotton, 2007: Urban aerosol impacts on downwind convective storms. J. Appl. Meteor. Climatol., 46, 828–850.