Direct and Indirect Shortwave Radiative Effects of Sea Salt Aerosols

July 30, 2008

By Ayash, Tarek Gong, Sunling; Jia, Charles Q

ABSTRACT Sea salt aerosols play a dual role in affecting the atmospheric radiative balance. Directly, sea salt particles scatter the incoming solar radiation and absorb the outgoing terrestrial radiation. By acting as cloud condensation nuclei, sea salt aerosols indirectly modulate the atmospheric radiative budget through their effective contribution to cloud formation. Using the Canadian Aerosol Module (CAM)-Canadian Centre for Climate Modelling and Analysis (CCCma) GCM, version 3 (GCM3) framework, the direct as well as the indirect shortwave (SW) radiative effects of sea salt aerosols are simulated. The model results herein suggest that sea salt aerosols exert a significant direct radiative effect over oceanic regions, with seasonal means in the range from -2 to -3 W m^sup -2^ over the Southern Ocean. Globally, sea salt’s SW indirect effect (annual mean -0.38 W m^sup -2^) is found to be less than its direct effect (annual mean -0.65 W m^sup -2^). However, sea salt’s indirect effect is found to be far stronger over the Southern Hemisphere than over the Northern Hemisphere, especially over the Southern Ocean with seasonal means around -4 W m^sup -2^, which exceed its direct effect. The model results herein suggest that sea salt aerosols significantly modulate the atmospheric radiation budget over oceanic regions and need to be accounted for in global climate models.

(ProQuest: … denotes formulae omitted.)

1. Introduction

Climate on earth is known as a highly dynamic and complex system, and atmospheric aerosols have been increasingly recognized as principal agents in that system. In this regard, the role of aerosols of anthropogenic origin, mainly sulfate and carbonaceous aerosols, within climate change has been of primary interest. However, of equal importance is understanding the role of natural aerosols, dominated by sea salt and soil dust, in modulating the earth-atmosphere radiative balance and, consequently, the dynamics of the climate system. Most global models show that, among all aerosol types, mass emissions are dominated by sea salt aerosol (Textor et al. 2006). However, owing to their high solubility, sea salt particles grow hygroscopically to large sizes and, thus, are quite efficiently removed by both dry and wet deposition. This results in sea salt particles having typically the shortest residence times and dominating atmospheric aerosol mass burdens near their high source regions.

Sea salt aerosols play a dual role in affecting the atmospheric radiative balance. Directly, sea salt particles interact with the incoming solar radiation and the outgoing terrestrial radiation. Unlike the more hydrophobic soil dust aerosol, sea salt particles uptake water readily and, hence, are highly scattered at shortwave (SW; solar) wavelengths with virtually no absorption (e.g., Takemura et al. 2002). However, similar to soil dust aerosol, because of their large sizes, sea salt particles are also efficient absorbers at longwave (LW, terrestrial) wavelengths (Satheesh and Lubin 2003). The highly hygroscopic nature of sea salt particles renders them very efficient cloud condensation nuclei (CCN). Over the remote oceanic regions where the presence of other aerosol species is relatively weak, sea salt aerosols are the dominant contributor to CCN numbers (Quinn et al. 2000). Even in the presence of the efficiently condensing sulfate aerosol, sea salt particles activate to form cloud droplets more readily than sulfate aerosol (O’Dowd et al. 1999) because of their larger size and the lower supersaturation threshold that is required for their activation. However, because CCN concentration depends on supersaturation, the smaller but more abundant sulfate particles often dominate sea salt as CCN at supersaturations exceeding 1%. By acting as CCN, therefore, sea salt aerosols indirectly modulate the atmospheric radiative balance through their effective contribution to cloud formation.

Several model studies provide estimates of the direct radiative effects of sea salt aerosol. General circulation model (GCM) simulations (Dobbie et al. 2003; Grini et al. 2002; Jacobson 2001; Takemura et al. 2002) provide a wide range of estimates of the SW and net (shortwave and longwave) direct radiative effects, which is between -0.15 and -2.2 W m^sup -2^ for the clear-sky global annual mean and between -0.31 and -1.1 W m^sup -2^ for the whole-sky global annual mean. Such a wide range of estimates mainly reflects the large uncertainty underlying sea salt’s emission modeling, which is mainly because sea salt’s generation mechanisms are highly determined by wind speed (e.g., Gong et al. 2002). This is well documented by Haywood et al. (1999), who provide “low” (-1.51 W m^sup -2^) and “high” (-5.03 W m^sup -2^) GCM estimates of the SW sea salt direct effect (annual mean, ocean averaged) by using two different empirical approaches that encompass the considerable range in the effect of wind speed on sea salt aerosol concentration. However, Haywood et al. show that the difference between their GCM- simulated and observed clear-sky irradiance over the oceans is smallest if a high burden of sea salt is assumed. In addition to GCM- based estimates, several estimates based on a combination of models and observational data suggest significant SW and LW direct radiative effects of sea salt at local and regional scales (Lubin and Vogelmann 2004; Podgorny and Ramanathan 2003; Satheesh and Lubin 2003).

Regarding sea salt’s indirect radiative effect (IRE), no GCM- based estimates have yet been made available. By deriving empirical relationships based on shipborne and island-based observational data, Vinoj and Satheesh (2004) estimate that the IRE of sea salt aerosols over the Arabian Sea may be as large as -7 +- 4 W m^sup – 2^ compared to -2 +- 1 W m^sup -2^ for their direct radiative effect (DRE) estimate. Vinoj and Satheesh’s findings, therefore, suggest that the indirect radiative effect resulting from sea salt aerosols may be more prominent than their direct radiative effect. Some observational data provide evidence of the mechanisms underlying the indirect effect of sea salt aerosols. For example, satellite data revealed plumes of reduced cloud particle size and suppressed precipitation originating from major urban and industrial areas; however, precipitation from similarly polluted clouds over oceans appeared to be much less affected (Rosenfeld 2000). One possible explanation is that the giant sea salt nuclei override the precipitation suppression effect of a large number of small pollution nuclei (Rosenfeld et al. 2002). This is in accordance with the findings that the more hygroscopic a CCN, the earlier it will become a droplet as water saturation is achieved, and that, at a given composition, larger CCNs will acquire water first (Kondratyev et al. 2006).

The Canadian Aerosol Module (CAM) was developed for simulating the atmospheric cycling of aerosols, with the Canadian Centre for Climate Modelling and Analysis (CCCma) GCM, version 3 (GCM3) as its climatological driver model. The global budget and distribution of sea salt were simulated by CAM and compared to observations (Gong et al. 2002). Sea salt’s impact on the global budget and distribution of non-sea salt (nss) sulfate aerosols was also simulated (Gong and Barrie 2003). Using this model framework, here we aim at quantifying the direct and indirect SW radiative effects of sea salt aerosols. The model framework and the simulation methodology are described in section 2. Results of model simulations are presented and discussed in section 3, ending with conclusions in section 4.

2. Model framework and simulation methodology

a. CAM-CCCma atmospheric general circulation model, version 3 framework

The atmospheric general circulation model, version 3 (AGCM3) shares many basic features with the AGCM, version 2 (AGCM2; McFarlane et al. 1992). As in AGCM2, the spectral transform method in AGCM3 is used to represent the horizontal spatial structure of the main prognostic variables, while the rectangular finite elements defined for a hybrid vertical coordinate represent the vertical spatial structure (LaPrise and Girard 1990). A semi-Lagrangian transport algorithm also exists in AGCM3 as an option for transport of moisture and other trace constituents. The spectral representation currently used in AGCM3 corresponds to a higher horizontal resolution than that used in AGCM2, composed of a 47- wave triangularly truncated (T47) spherical harmonic expansion. The vertical domain of AGCM3 is deeper than in AGCM2, extending from the surface to the stratopause region (1 hPa), and the vertical resolution is also higher (32 layers). The model time step is 15 min with shortwave radiation calculations being performed every 1 h, while full terrestrial radiation calculations are done every 6 h and partial calculations are made every model time step. Solar radiative heating in AGCM3 is computed using an improved scheme that employs four bands in the visible and near-infrared regions, instead of the two bands in the scheme used in AGCM2. The optical properties of the clouds are parameterized in terms of the liquid water/ice content. The treatment of terrestrial radiation is similar to that used in AGCM2, but with improved treatment of broadband emissivities and the water vapor continuum. The original solar radiation scheme, which uses the two-stream and delta-Eddington approximations for layer reflectance and transmittance computations, has been modified by using the more accurate delta-four-stream approximation (Ayash et al. 2008a). Aerosol processes are simulated by the algorithms of the CAM. CAM (Gong et al. 2003) is a size-segregated multicomponent aerosol module that includes processes that determine the emission, transport, chemical transformation, cloud interaction, and deposition of atmospheric aerosols. The module accounts for the following five aerosol species: sea salt, sulfate, black carbon, organic carbon, and soil dust. Sources include the surface emission rate of both natural and anthropogenic aerosols as follows: black and organic carbon (Cooke et al. 1999; Lavoue et al. 2000; Liousse et al. 1996), sulfur dioxide and sulfate [Global Emissions Inventory Activity (GEIA) level 2], land H2S (Benkovitz and Schwartz 1997), dimethylsulfide [DMS; Kettle et al. (1999) for ocean data; Wanninkhof (1992) for sea-air transfer], and interactive modeling of sea salt (Gong et al. 1997) and soil dust (Marticorena and Bergametti 1995). Production of secondary aerosols, that is, airborne aerosol mass produced by chemical transformation of their precursors, together with particle nucleation, condensation, and coagulation, form the clear-air processes. For sulfate aerosols, an online sulfur chemistry for both clear-sky and cloud droplets using the offline chemical concentrations of OH, O^sub 3^, H^sub 2^O^sub 2^, NO^sub 3^, and NH^sub 3^ is included in the module.

In CAM, aerosol size distribution is treated following the sectional approach method. This representation is employed for its flexibility in treating processes including multicomponent interactions, such as coagulation, condensation, and chemical processes. The aerosol size spectrum is divided into 12 bins, partitioned at multiples-of-two radii between 0.005 and 20.48 [mu]. The hygroscopic growth of soluble aerosol species is accounted for, hence, defining the aerodynamic size of aerosol particles as that in equilibrium with ambient relative humidity. However, for supermicron particles, equilibrium with relative humidity can often produce unrealistically large estimates of water uptake and direct radiative forcing, especially when relative humidity is high in the updrafts. A cloud module with explicit microphysical processes is included to simulate cloud droplet activation, subsequent aerosol-cloud- raindrop interaction, and cloud chemistry. Within a size bin, an internally mixed aerosol is assumed for all types of aerosols except for the freshly emitted (at source grid) insoluble components (black carbon and soil dust), which are assumed to be externally mixed for a fixed amount of time in that grid and layer. The number densities of externally mixed components are calculated for every time step in the source grid, which then can be used in the calculations of externally mixed aerosol activation and radiative forcing.

b. Aerosol optical parameters module

A module that accounts for the aerosol optical parameters needed to compute the aerosol radiative effect in the CAM-CCCma GCM3 modeling framework has been developed and described by Ayash et al. (2008b). The module uses wavelength-specific and band-averaged lookup tables of asymmetry factors and absorption and scattering cross sections over the solar spectrum, and band-averaged absorption cross sections over the terrestrial spectrum that were computed based on the Mie theory, using indices of refraction for five aerosol types (plus water) obtained from Hess et al.’s (1998) Optical Properties of Aerosols and Clouds (OPAC) database (initially compiled by d’Almeida et al. 1991). The asymmetry factor and cross sections are interpolated as a function of the relative humidity value at every model time step, and the latter are used to calculate the optical depth for each aerosol component, size bin, wavelength/ band, and model layer. In computing the optical properties, all aerosol components are assumed to be externally mixed, and soil dust and black carbon are treated as hydrophobic species.

c. Simulation methodology

To simulate the direct and indirect effects of sea salt aerosols, the CAM-CCC GCM3 is run for 5 yr with prescribed climatological- mean sea surface temperatures, and the results are based on the last 3 yr of the model run. By making multiple calls to the SW radiation routine, the DRE of sea salt is then computed as the difference in the resulting net downward fluxes at the top-of-atmosphere (TOA) and the surface (taken as the flux with sea salt optical parameters minus the flux without sea salt optical parameters). As such, sea salt-affected fluxes are not fed back into the model.

By acting as CCN and ice nuclei (IN), aerosols may impact the radiative properties of clouds basically in two ways. First, increasing the number of CCN leads to more, but smaller, cloud droplets in a cloud (whose liquid water content remains constant), which enhances the cloud’s reflection of solar radiation. This is known as the first indirect effect, or the Twomey effect (Twomey 1959). In addition, the more prevalent but smaller cloud droplets reduce the precipitation efficiency and, therefore, enhance the cloud lifetime and, hence, the cloud reflectivity, which is referred to as the cloud lifetime, or second indirect effect (Albrecht 1989).

In CAM, the cloud droplet number concentration (CDNC) is derived from aerosol number concentration in a certain size range according to the empirical formulation (Jones et al. 1994)

, (1)

where N^sub drop^ is cloud droplet number concentration (m^sup – 3^) and Na is total aerosol number concentration (m^sup -3^) in the size range of 0.050 -1.500 [mu]m radius.

Such a relationship was derived based on aircraft data from a wide range of locations by Martin et al. (1994) who derived separate curves for maritime and continental clouds from which Jones et al. (1994) derived Eq. (1) to provide CDNC as a single continuous function of aerosol concentration. Martin et al. (1994) found a good correlation between the aerosol number concentration just below cloud base and the CDNC with little scatter for the maritime air masses, indicating that nearly all aerosols are good CCN over the oceans. In the continental air masses, however, a much larger scatter was found, which relates to the more diverse aerosol chemical characteristics found over continents than over the oceans. Therefore, Eq. (1) can be regarded as a better estimate for relating CDNC to aerosol concentration over oceans than over continents, which would serve our purpose to study the indirect radiative effects of sea salt predominantly over oceanic regions. The precipitation formation, or autoconversion of cloud droplets, is derived from the stochastic collection equation, which describes the time evolution of a droplet spectrum that is changing by collisions among droplets of different sizes (Beheng 1994). The autoconversion rate (Q^sub aut^, kg kg^sup -1^ s^sup -1^) is related to the CDNC, in addition to the liquid water content of clouds, by

… (2)

where n = 10 is the width parameter of the initial cloud droplet spectrum described by a gamma function, pair is airmass density, q^sub cl^ is cloud water mixing ratio in the cloudy part of the grid box, and gamma^sub 1^ = 15 is the microphysical constant, which determines the efficiency of rain formation and the cloud lifetime.

To calculate the indirect effects of sea salt, we follow a variation of the methodology devised by Kristjansson (2002) for calculating the indirect effects of sulfate and black carbon aerosols, which is shown schematically in Fig. 1. The CDNC computed by Eq. (1) is fed to the cloud condensation scheme, which accounts for explicit microphysical processes to treat the cloud-aerosol interactions. In its original formulation (Lohmann and Roeckner 1996), the condensation scheme computes the cloud liquid water and ice contents based on CDNC, which is derived from marine and continental sulfate aerosol mass concentrations. This formulation is modified, therefore, by accounting for all aerosol species; however, variations in aerosol size distribution, composition, and meteorological conditions are not accounted for.

To account for sea salt’s indirect effects, the CDNC is computed twice by Eq. (1): once with Na including all aerosol species, and another with N^sub a^ excluding sea salt. To account for the first indirect effect only, three parallel calls to the clouds’ optical parameters routine are made: one for advancing the model and two diagnostic calls using the two computed CDNC fields, but all use the same cloud water field from a single call of the condensation scheme. To include the second indirect effect, these computations are also performed in another model run making three parallel calls to the condensation scheme: one for advancing the model and two other diagnostic calls. This results in two additional cloud water fields that are then fed, along with the computed CDNC fields, into the multiple calls of the clouds optical parameters routine. In the two model runs, the resulting cloud optical parameters are then fed into two additional parallel calls of the SW radiation routine, and the difference in the resulting TOA SW cloud radiative forcing (whole-sky flux minus clear-sky flux) between these two calls would then give the first and the first-andsecond indirect effects of sea salt aerosols.

By such a methodology, therefore, the calculated first indirect effect would capture sea salt’s modulation of the cloud radiative properties and forcing through its contribution to the number of CCN and, hence, the number of cloud droplets formed, thereby changing the droplets’ sizes that would have otherwise been formed without the presence of sea salt. The combined (first and second) indirect effect would additionally account for sea salt’s effect in changing the clouds’ liquid water content, which results from changes in precipitation release resulting from different CCN amounts and, hence, different CDNCs. Such an approach assumes that the indirect effect occurs only through water but not ice clouds; therefore, clouds in the lower troposphere only will contribute to the estimated indirect effect. Furthermore, in the model’s condensation scheme the fractional cloud cover is an empirical function of relative humidity (Rosenfeld et al. 2002) and, thus, is independent of the cloud water content. Therefore, our approach would not capture changes in cloud coverage, which may lead to an underestimation of the lifetime effect. 3. Model results

a. Direct radiative effect

Global fields of the modeled seasonal mean TOA SW DRE of sea salt aerosols are shown in Fig. 2. In explaining our results here, we also refer to the global distribution of sea salt aerosols simulated using CAM-CCCma GCM3 by Gong et al. (2002). For all seasons, the modeled DRE is strongest over the Southern Ocean, peaking during the Southern Hemisphere (SH) spring and summer with magnitudes between – 2 and -3 W m^sup -2^. This is consistent with the fact that the model predicts the largest sea salt concentrations over this region that strongly correlate with the highest wind speeds over the South Pacific (roaring forties). Similarly, the maxima over the tropical Pacific and the North Atlantic during the Northern Hemisphere (NH) winter would be correlated with the largest sea salt c oncentrations and wind speeds over these regions, but remain within -2 W m^sup – 2^. A strong DRE, exceeding -3 W m^sup -2^, is modeled over the Arabian Sea during the NH summer, which also correlates with the large sea salt surface concentrations modeled for this time and region.

The overall stronger DRE modeled over the Southern Ocean compared to the tropical Pacific and the North Atlantic may be explained, in part, by the regional differences in the dependence of sea salt concentration and loading on the surface wind speed. Gong et al. (2002) find that the sea salt wind dependencies are stronger at the surface than for the column burden. However, in the roaring forties of the South Pacific, they find the highest correlation for both surface concentrations and column loadings. As such, the larger surface concentrations as well as the vertical dispersions over that region would result in a stronger DRE.

Another factor that contributes to the regional differences in the modeled sea salt DRE is the dependence of the aerosol’s scattering efficiency on the particle size. While the total sea salt surface mass concentration is dominated by supermicron particles, the most efficient size range for solar radiation scattering occurs for particle diameters in the 0.2-1.0 [mu]m range, which results in the submicron sea salt aerosols accounting for about 20% of the total light scattering (Quinn et al. 1996). For January, the model predicts the highest surface mixing ratios for the submicron sea salt aerosols over the South Pacific. In April, the highest submicron fraction of sea salt over the Pacific is modeled at the southernmost latitudes. Therefore, the higher submicron fractions over these regions would also contribute to higher magnitudes of the modeled DRE.

To examine the effect of clouds on the DRE of sea salt, global fields of its TOA annual mean in both clear and whole skies are shown in Fig. 3. The effect of clouds in diminishing sea salt DRE is quite noticeable. While in whole skies its magnitude remains globally within -3 W m^sup -2^, in clear skies its magnitude is higher than -3 W m^sup -2^ over large areas of the North Atlantic, and even higher than -5 W m^sup -2^ over large areas of the South Pacific. Because of their partial absorption at solar wavelengths, the presence of soil dust above a cloud layer may result in their TOA radiative effect being more positive due to soil dust’s absorption of multiply scattered solar radiation enhanced by the albedo of the lower cloud layer (Takemura et al. 2002). For a largely scattering aerosol as sea salt, however, their TOA radiative effect would always be negative regardless of the presence of clouds. However, the large reduction of sea salt’s radiative effect by clouds reflects the vertical distribution of the sea salt aerosols relative to the cloud layers. The reduction indicates the overall presence of aerosols below the cloud layers, with the clouds reflecting the solar radiation that would otherwise be reflected by the aerosols. This is consistent with the bulk of sea salt aerosols remaining at lower levels given the dominance of the heavier supermicron particles in their emissions.

b. Indirect radiative effects and comparison to Earth Radiation Budget Experiment data

The modeled contributions of sea salt aerosols to the seasonal mean column-integral CDNC and the cloud liquid water content (CLWC) are shown in Figs. 4 and 5, respectively. Global fields of the modeled seasonal mean TOA SW first indirect radiative effect of sea salt aerosols are shown in Fig. 6, while those for the first-and- second indirect effect are shown in Fig. 7. To gain insight on the significance of the estimated magnitude of the indirect radiative effect of sea salt relative to the overall cloud modulation of the SW radiation budget, the zonal averages of the seasonal mean SW cloud radiative forcing (CRF) were also computed and are shown in Fig. 8. Given that the CRF also reflects the model’s simulation of the clouds’ distribution, the model’s ability to properly simulate cloud distribution would thus be assessed by comparing the modeled CRF to observations. Therefore, the zonal averages of the seasonal mean SW CRF obtained from a 5-yr measurement period (1985-89) of the Earth Radiation Budget Experiment (ERBE) satellites are included in Fig. 8. This dataset is the ERBE S4 Regional, Zonal and Global Average Product, which provides 2.5[degrees] x 2.5[degrees] TOA flux data for both clear and cloudy skies (Barkstrom et al. 1989). In line with Cess et al. (1997), analysis is restricted only to latitudes between 60[degrees]N and 60[degrees]S, given that the ERBE clear-sky scene identification is not reliable over snow and ice, so that high latitudes are better excluded. Furthermore, while the clearsky fluxes in the model simulation are diagnostically evaluated at each grid by setting cloud amount to zero, which is known as method 2 (Cess and Potter 1987), the ERBE clear-sky fluxes were computed from all the pixel values that were flagged as clear during a single month. Therefore, this results in a sampling discrepancy between the model and the ERBE data, which could impact the comparison of the CRF.

Compared to the modeled seasonal averages over southern oceanic regions, the CDNC is generally higher by a factor of between 2 and 3, and the CLWC is higher by a factor of 2 over northern and tropical oceanic regions except for the high June-August averages over subtropical regions west of South America. The modeled sea salt contribution to CDNC and CLWC exhibits a reverse hemispheric pattern, however. Seasonal averages show that sea salt contributes over 10% to CDNC and over 4% to CLWC over much of the southern oceanic regions, which are found to be twice its contributions over northern regions. Both of the modeled first and first-and- second indirect radiative effects of sea salt show marked seasonal variations, and even more so compared to the modeled DRE, peaking over the Southern Ocean during the SH spring and summer seasons. The magnitude of the indirect effects of this region is found to be much higher than that over the North Atlantic and mid-Pacific, especially for the first indirect effect that, except for the NH spring, is found to be negligibly small (within -0.5 W m^sup -2^) over most parts of the northern oceans. While the first indirect effect remains within -3.0 W m^sup -2^ over the Southern Ocean, inclusion of the second indirect effect results in noticeable enhancement of the indirect effect, with magnitudes within -4.5 W m^sup -2^ over most parts of this region. Significant increases are also noticeable over the North Pacific and Atlantic during the NH summer and fall seasons.

The hemispheric discrepancy in the modeled CDNC and CLWC may be linked to the higher concentrations of sulfate aerosol over the NH (Ayash 2007). With less sulfate over the SH, sea salt aerosol would have a larger contribution to CCN and, hence, CDNC and CLWC, all of which would yield higher indirect effects of sea salt over the Southern Ocean. In addition to the reduced competition by sulfate aerosol, with the higher vertical dispersion of sea salt aerosols over the South Ocean, the contribution of sea salt to CCN and, hence, cloud formation, is expected to be larger as more of it reaches higher altitudes. Furthermore, with the CCN in our model scheme being regarded as the aerosols with radii between 0.05 and 1.50 jam [Eq. (I)], the contribution of the smaller (not the larger) aerosols would thus count. With the modeled sea salt submicron fractions being highest over the Southern Ocean, this would result in larger contributions of sea salt to the modeled CCN numbers and, consequently, a stronger indirect effect.

Compared to observations (Fig. 8), the modeled CRF is in reasonable agreement with ERBE data over the tropical and subtropical latitudes for all seasons. Such agreement becomes less apparent at higher latitudes with a general underestimation of the ERBE values by the model, except for southern latitudes during the SH winter. Significant underestimation is found at these latitudes during the SH spring and summer seasons where the observed CRF are highest, which might indicate a possible underestimation of cloudiness by the model. In any case, the modeled indirect effect of sea salt was found to be strongest over the southern latitudes, with maxima between -3 and -4.5 W m^sup -2^ during the SH spring and summer. Over these latitudes and seasons, the modeled CRF is found to be between -40 and -80 W m^sup -2^. Based on these values, the model simulations suggest that sea salt’s contribution to the cloud SW radiative modulation could be at best between 5% and 10%. Such a contribution would be much less over the northern latitudes given that the estimated indirect effect is less and that the land fraction is higher (where the indirect effect is negligible, the zonal average would be even lower). 4. Discussion and conclusions

Annual means of the modeled SW radiative effects of sea salt aerosol at the TOA are listed in Table 1, and other model estimates of the sea salt DRE are included in Table 2. Our model estimates suggest that sea salt aerosols exert an overall stronger direct cooling effect than their indirect cooling effects, with the global average of the former being about one-and-a-half times the latter. While both direct and indirect effects are stronger over the SH than over the NH, there is a marked difference in their hemispheric ratios, with significantly higher SH-to-NH ratios for the indirect effects (about five to more than seven) compared to the DRE (less than two). While sea salt’s indirect effects averaged over land areas are found to be negligible compared to their oceanic averages (being more than an order of magnitude less), the land average of its DRE is nonnegligible, especially for clear skies (less within a factor of 5).

Our model estimates for the sea salt SW DRE is found to be within the range of estimates of other model studies (Table 2). If sea salt’s LW warming is to be included, the resultant net radiative effect would be less than the SW, so our expected estimate for the net effect would still fall within the range of other model estimates. While no direct measurement of sea salt’s DRE is available, Haywood et al. (1999) found that the inclusion of a high sea salt burden would lead to minimal differences between their modeled and ERBEmeasured reflected clear-sky fluxes in the SH, but would lead to an overestimate in the mid- to high-latitude NH oceans. Considering their high estimate of the sea salt SW DRE at – 5.03 W m^sup -2^ (clear sky, ocean averaged), therefore, might suggest that our estimate of -1.95 W m^sup -2^ may be regarded as somewhat conservative.

The clear-sky DRE of sea salt aerosol and of the combined tropospheric aerosols (sea salt, sulfate, black and organic carbon, soil dust) simulated by the same model are compared to satellite- based estimates over two oceanic regions in Table 3. The satellite data are provided by Yu et al. (2005) and include estimates based on measurements made by the Moderate Resolution Imaging Spectroradiometer (MODIS) and Clouds and the Earth’s Radiant Energy System (CERES) instruments. The MODIS-based estimates are derived from MODIS retrievals of aerosol optical depth, single-scattering albedo, and phase function that best matches spectral radiances observed at the TOA. These retrieved aerosol optical properties are then used with the CLIRAD-SW radiative transfer model (Chou 1992) to calculate TOA fluxes that best match the observed radiances from which the aerosol direct radiative effect is calculated. The CERES- based estimates are derived using CERES Terra-measured radiances/ fluxes along with cloud distributions and aerosol properties determined directly from Terra MODIS measurements.

The first oceanic region is the tropical southeast Pacific, which is a pristine region that is characterized by the smallest aerosol DRE without significant seasonal variations. Over this region, sea salt accounts for about half of the modeled aerosol DRE, which is found within 13% of MODIS estimates and 20% of CERES estimates. Over the southern oceanic region, the modeled sea salt DRE accounts for more than two-thirds of the aerosol DRE and exhibits insignificant seasonal variation. Here, the modeled aerosol DRE is within 14% of MODIS estimates and within 5% of CERES estimates, except for its June-August mean, which matches the modeled sea salt DRE. Yu et al. (2005) find that, over most oceanic regions, model simulations of aerosol DRE are generally smaller than measurement-based estimates by 30%-50%. Therefore, we may regard our model results to compare reasonably well with satellitebased estimates. Such comparison, however, should be regarded against the uncertainties in and limitations of both model and measurement-based estimates. For example, Yu et al. (2005) note that MODIS-retrieved optical depths tend to be overestimated by about 10%-15% because of the contamination of thin cirrus and clouds, which would result in a comparable overestimate of the aerosol DRE. On the other hand, as already noted, our modeled DRE is calculated from clear-sky fluxes that are diagnostically evaluated by setting cloud amount to zero (method II). Added to differences between the modeled and satellite- measured cloud fields, this leads to discrepancy in clear-sky sampling that would affect the comparison of clear-sky DRE.

Simulations by CAM-CCCma GCM3 yielded a global sea salt emission that was found to be more than one order of magnitude larger than in the other models (Ayash 2007). Added to the short residence time modeled for sea salt, this was mainly attributed to model disagreement on representation of the particle sizes of sea salt and dust, especially on the choice of the largest particles simulated. With the point made earlier that the finer fractions of sea salt particles contribute more efficiently to scattering, our “conservative” estimate of the sea salt DRE, despite the model’s relatively high emission, might be explained by the dominance of large particles in the modeled size distribution. In any case, our model simulations suggest that sea salt aerosols exert a significant DRE over oceanic regions, with the seasonal means being in the range from -2 to -3 W m^sup -2^. Simulations by the same model framework (Ayash 2007) suggest that soil dust exerts a much stronger cooling effect at its source regions with typical seasonal mean values exceeding -10 W m^sup -2^. However, the magnitude of the global mean SW effect of soil dust was found in the range from -0.63 to -0.85 W m^sup -2^, which is comparable to that of sea salt found here at – 0.65 W m^sup -2^. Therefore, our model findings suggest that, while sea salt’s SW DRE is less prominent than soil dust’s effect at their source regions, these two natural aerosol species would have comparable effects on a global scale and need to be accounted for in global climate models.

Our modeled sea salt indirect radiative effects over the Arabian Sea are not in agreement with the estimates of Vinoj and Satheesh (2004), which suggest a more prominent indirect effect over that region compared to their estimated sea salt DRE. However, as already noted in the introduction, their estimates were derived from empirical relationships based on observational data that are limited both in space and time, which would render any direct comparison of our global, longer-term estimates to theirs inadequate. Disregarding such differences, while our modeled sea salt DRE shows a strong maximum over the Arabian Sea during the NH summer (Fig. 2), our modeled indirect effects over this region are virtually negligible for all seasons (Figs. 6 and 7). Detailed examination of the simulated meteorology, especially wind speed, over this region would be needed to explain such a discrepancy.

Given that a rather simplistic relationship is used to estimate the CDNC from aerosol numbers only [Eq. (1)], our model simulations would not account for a well-established aspect characterizing the formation of CCN and cloud droplets over the marine atmosphere, and that is the competition between sea salt and nss sulfate aerosol species. Using a different parameterization that yields CDNC as a function of sulfate concentration and wind speed, Gong and Barrie’s (2003) simulations with CAM-CCCma GCM3 show that the dominant impact of the sea salt aerosols is to reduce the CDNC through suppression of the peak supersaturation achieved within the cloud. However, Ghan et al. (1998) find that under relatively low sulfate number concentrations and strong updraft velocities, the total number concentration of activated cloud droplets may increase with increasing sea salt concentration because of the activation of accumulation mode sea salt particles, but would otherwise (high sulfate and weak updrafts) decrease because of the reduction in maximum cloud supersaturation caused by competition with coarse mode sea salt particles. Such effects, therefore, would be accounted for only by using a more sophisticated parameterization scheme that provides more realistic estimates of the CDNC.

Another limitation in our model simulations is that of the modeled sea salt particle size. In CAM, a semiempirical formulation (Monahan et al. 1986) is used to obtain the surface emission rate of sea salt aerosol as a function of the surface wind speed and particle size. However, this formulation is only good for particles with a dry radius larger than 0.2 /j,m and would result in uncertainties, especially in the number concentrations, for smaller particle sizes. By observing that around 60% of sea salt particles produced by bubbles from coastal oceanic breaking waves were found to have diameters smaller than 0.1 [mu]m, Clarke et al. (2006) developed a new sea salt source function that would account for these ultrafine sea salt fluxes. Applying this source function, they estimate that in marine regions with little continental impact, the sea salt aerosol flux can contribute around 5%-90% of the marine CCN. Alongside these findings, Pierce and Adams (2006) investigated the impact of sea salt on CCN concentrations relative to simulations where only sulfate aerosols were considered. Their simulations show that the inclusion of sea salt aerosols increased CCN (activated in a cloud of 0.2% supersaturation) over the Southern Ocean by 150%- 500%, depending on the emission estimate used. Furthermore, they found that the highest increases resulted from the simulations that included ultrafine sea salt emissions, which enhanced CCN (0.2%) by more than 50% over both the Southern Ocean and Antarctica. These findings demonstrate that the modeled CCN concentrations can be rather sensitive to ultrafine sea salt emissions, so their inclusion in model simulations needs to be considered. Against such model limitations and the possible resulting uncertainties, our model simulations suggest that, although being less global, the indirect effects of sea salt may be as important as, and even more prominent than, its DRE in modulating the atmospheric radiation budget over oceanic regions. This is specially the case over the Southern Ocean, a remote oceanic region with a typically clean maritime atmosphere where natural aerosol components dominate and, therefore, sea salt’s contribution to cloud formation and, consequently, its indirect effects would be largest. However, relative to the overall SW radiative budget modulation by clouds, our model simulations suggest a relatively small, yet significant, contribution of sea salt aerosols.

In addition to accounting for the direct and indirect radiative effects of sea salt aerosol, the dependence of the indirect effect of anthropogenic aerosol on the natural aerosols adds further weight to the inclusion of sea salt aerosol in climate models. This follows from the fact that cloud susceptibility to additional aerosol is greater if the natural aerosol concentrations are lower, where competition during droplet nucleation is weaker. Thus, regardless of the relative importance of the direct and indirect effects of sea salt, the inclusion of sea salt in climate models is important because it comprises a substantial fraction of the natural CCN concentration. Finally, the relative simplicity of the parameterization used to estimate the numbers of cloud droplets in our model simulations demands further investigation with more sophisticated schemes.

Acknowledgments. This work was generously supported by Environment Canada and by the International Research Centre for Sustainable Materials, Institute of Industrial Science, University of Tokyo.


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

Ayash, T., 2007: Development of an interactive model for studying aerosol-climate interactions using the Canadian Aerosol Module- Canadian Climate Center General Circulation Model modeling framework. Ph.D. thesis, University of Toronto, 205 pp.

_____, S. L. Gong, and C. Q. Jia, 2008a: Implementing the delta- four-stream approximation for solar radiation computations in an atmosphere general circulation model. J. Atmos. Sci., 65, 2448- 2457.

_____, _____, _____, P. Huang, T. L. Zhao, and D. Lavoue, 2008b: Global modeling of multicomponent aerosol species: Aerosol optical parameters. J. Geophys. Res., in press.

Barkstrom, B. R., E. Harrison, G. Smith, R. Green, J. Kibler, R. Cess, and the ERBE Science Team, 1989: Earth Radiation Budget Experiment (ERBE) archival and April 1985 results. Bull. Amer. Meteor. Soc., 70, 1254-1262.

Beheng, K. D., 1994: A parametrization of warm cloud microphysical conversion processes. Atmos. Res., 33, 193-206.

Benkovitz, C. M., and S. E. Schwartz, 1997: Evaluation of modeled sulfate and SO^sub 2^ over North America and Europe for four seasonal months in 1986-1987. J. Geophys. Res., 102, 25 305-25 338.

Cess, R. D., and G. L. Potter, 1987: Exploratory studies of cloud radiative forcing with a general circulation model. Tellus, 39A, 460- 473.

_____, and Coauthors, 1997: Comparison of the seasonal change in cloud-radiative forcing from atmospheric general circulation models and satellite observations. J. Geophys. Res., 102, 16 593-16 603.

Chou, M.-D., 1992: A solar radiation model for use in climate studies. J. Atmos. Sci., 49, 762-772.

Clarke, A. D., S. R. Owens, and J. Zhou, 2006: An ultrafine sea salt flux from breaking waves: Implications for cloud condensation nuclei in the remote marine atmosphere. J. Geophys. Res., 111, D06202, doi:10.1029/2005JD006565.

Cooke, W. F., C. Liousse, H. Cachier, and J. Feichter, 1999: Construction of a 1[degrees] x 1[degrees] fossil fuel emission data set for carbonaceous aerosol and implementation and radiative impact in the ECHAM4 model. J. Geophys. Res., 104, 22 137-22 162.

d’Almeida, G. A., P. Koepke, and E. P. Shettle, 1991: Atmospheric Aerosols: Global Climatology and Radiative Characteristics. A. Deepak Publishing, 561 pp.

Dobbie, S., J. Li, R. Harvey, and P. Chylek, 2003: Sea salt optical properties and GCM forcing at solar wavelengths. Atmos. Res., 65, 211-233.

Ghan, S. J., G. Guzman, and H. Abdul-Razzak, 1998: Competition between sea salt and sulfate particles as cloud condensation nuclei. J. Atmos. Sci., 55, 3340-3347.

Gong, S. L., and L. A. Barrie, 2003: Simulating the impact of sea salt on global nss sulphate aerosols. J. Geophys. Res., 108, 4516, doi:10.1029/2002JD003181.

_____, _____, and J.-P. Blanchet, 1997: Modeling sea salt aerosols in the atmosphere. Part 1: Model development. J. Geophys. Res., 102, 3805-3818.

_____, _____, and M. Lazare, 2002: Canadian Aerosol Module (CAM): A size-segregated simulation of atmospheric aerosol processes for climate and air quality models. Part 2: Global sea salt aerosol and its budgets. J. Geophys. Res., 107, 4779, doi:10.1029/2001JD002004.

_____, and Coauthors, 2003: Canadian Aerosol Module: A size- segregated simulation of atmospheric aerosol processes for climate and air quality models. Part 1: Module development. J. Geophys. Res., 108, 4007, doi:10.1029/2001JD002002.

Grini, A., G. Myhre, J. K. Sundet, and I. S. A. Isaksen, 2002: Modeling the annual cycle of sea salt in the global 3D model Oslo CTM2: Concentrations, fluxes, and radiative impact. J. Climate, 15, 1717-1730.

Haywood, J., V. Ramaswamy, and B. Soden, 1999: Tropospheric aerosol climate forcing in clear-sky satellite observations over the oceans. Science, 283, 1299-1303.

Hess, M., P. Koepke, and I. Schult, 1998: Optical properties of aerosols and clouds: The software package OPAC. Bull. Amer. Meteor. Soc., 79, 831-844.

Jacobson, M. Z., 2001: Global direct radiative forcing due to multicomponent anthropogenic and natural aerosols. J. Geophys. Res., 106, 1551-1568.

Jones, A., D. Roberts, and A. Slingo, 1994: A climate model study of indirect radiative forcing by anthropogenic sulphate aerosols. Nature, 370, 450-453.

Kettle, A. J., and Coauthors, 1999: A global database of sea surface dimethylsulfide (DMS) measurements and a procedure to predict sea surface DMS as a function of latitude, longitude, and month. Global Biogeochem. Cycles, 13, 399-444.

Kondratyev, K. Y., L. S. Ivlev, V. F. Krapivin, and C. A. Varotsos, 2006: Atmospheric Aerosol Properties: Formation, Processes and Impacts. Springer-Verlag, 572 pp.

Kristjansson, J. E., 2002: Studies of the aerosol indirect effect from sulfate and black carbon aerosols. J. Geophys. Res., 107, 4246, doi:10.1029/2001JD000887.

LaPrise, R., and C. Girard, 1990: A spectral general circulation model using a piecewise-constant finite-element representation on a hybrid vertical coordinate system. J. Climate, 3, 32-52.

Lavoue, D., C. Liousse, H. Cachier, B. J. Stocks, and J. G. Goldammer, 2000: Modeling of carbonaceous particles emitted by boreal and temperate wildfires at northern latitudes. J. Geophys. Res., 105, 26 871-26 890.

Liousse, C., J. E. Penner, C. Chuang, J. J. Wallon, H. Eddleman, and H. Cachier, 1996: A global three-dimensional model study of carbonaceous aerosols. J. Geophys. Res., 101, 19 411-19 432.

Lohmann, U., and E. Roeckner, 1996: Design and performance of a new cloud microphysics scheme developed for the ECHAM general circulation model. Climate Dyn., 12, 557-572.

Lubin, D., and A. M. Vogelmann, 2004: Longwave aerosol direct and indirect radiative effects at the NSA site. Proc. 14th ARM Science Team Meeting, Albuquerque, NM, ARM, 1-5.

Marticorena, B., and G. Bergametti, 1995: Modeling the atmospheric dust cycle. Part 1: Design of a soil-derived dust emission scheme. J. Geophys. Res., 100, 16 415-16 430.

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

McFarlane, N. A., G. J. Boer, J.-P. Blanchet, and M. Lazare, 1992: The Canadian Climate Centre second-generation general circulation model and its equilibrium climate. J. Climate, 5, 1013- 1044.

Monahan, E., D. Spiel, and K. Davidson, 1986: Model of marine aerosol generation via whitecaps and wave disruption. Oceanic Whitecaps and Their Role in Air-Sea Exchange Processes, E. C. Monahan and G. Mac Niocaill, Eds., D. Reidel, 167-174.

O’Dowd, C. D., J. A. Lowe, M. H. Smith, and A. D. Kaye, 1999: The relative importance of N^sub ss^-sulphate and sea salt aerosol to the marine CCN population: An improved multicomponent aerosol-cloud droplet parametrization. Quart. J. Roy. Meteor. Soc., 125, 1295- 1313.

Pierce, J. R., and P. J. Adams, 2006: Global evaluation of CCN formation by direct emission of sea salt and growth of ultrafine sea salt. J Geophys. Res., 111, D06203, doi:10.1029/2005JD006186.

Podgorny, I., and V. Ramanathan, 2003: Aerosol-cloud radiative interactions in the longwave domain. Geophysical Research Abstracts, Vol. 5, Abstract 01797. [Available online at http://www.cosis.net/ abstracts/EAE03/01797/EAE03-J-01797.pdf.]

Quinn, P., V. Kapustin, T. Bates, and D. Covert, 1996: Chemical and optical properties of marine boundary layer aerosol particles of the mid-Pacific in relation to sources and meteorological transport. J. Geophys. Res., 101, 6931-6951. _____, and Coauthors, 2000: A comparison of aerosol chemical and optical properties from the 1st and 2nd Aerosol Characterization Experiments. Tellus, 52B, 239-257.

Rosenfeld, D., 2000: Suppression of rain and snow by urban and industrial air pollution. Science, 287, 1793-1796.

_____, R. Lahav, A. Khain, and M. Pinsky, 2002: The role of sea spray in cleansing air pollution over ocean via cloud processes. Science, 297, 1667-1670.

Satheesh, S., and D. Lubin, 2003: Short wave versus long wave radiative forcing by Indian Ocean aerosols: Role of sea-surface winds. Geophys. Res. Lett., 30, 1695, doi:10.1029/2003GL017499.

Takemura, T., T. Nakajima, O. Dubovik, B. N. Holben, and S. Kinne, 2002: Single-scattering albedo and radiative forcing of various aerosol species with a global three-dimensional model. J. Climate, 15, 333-352.

Textor, C., and Coauthors, 2006: Analysis and quantification of the diversities of aerosol life cycles within AeroCom. Atmos. Chem. Phys., 6, 1777-1813.

Twomey, S. A., 1959: The nuclei of natural cloud formation. Part II: The supersaturation in natural clouds and the variation of cloud droplet concentrations. Geofis. Pure Appl., 43, 227-242.

Vinoj, V., and S. Satheesh, 2004: Direct and indirect radiative effects of sea salt aerosols over Arabian Sea. Curr. Sci., 86, 1381- 1390.

Wanninkhof, R., 1992: Relationship between wind speed and gas exchange over the ocean. J. Geophys. Res., 97, 7373-7382.

Yu, H., and Coauthors, 2005: A review of measurement-based assessment of aerosol direct radiative effect and forcing. Atmos. Chem. Phys. Discuss., 5, 7647-7768.


Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario, Canada


Air Quality Research Branch, Meteorological Service of Canada, Toronto, Ontario, Canada


Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario, Canada

(Manuscript received 29 May 2007, in final form 13 November 2007)

Corresponding author address: Tarek Ayash, Dept. of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College Street, Toronto, ON M5S 3E5, Canada.

E-mail: tarek.ayash@utoronto.ca

Copyright American Meteorological Society Jul 1, 2008

(c) 2008 Journal of Climate. Provided by ProQuest Information and Learning. All rights Reserved.

comments powered by Disqus