SEARCH

SEARCH BY CITATION

Keywords:

  • sea salt;
  • direct radiative effect;
  • numerical modelling;
  • regional scale

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] This study deals with the direct radiative effect of sea salt on the regional scale, within both the shortwave and longwave ranges. The COSMO-ART model system has been extended and applied for a large part of Europe and adjacent waters within this investigation. For the radiation calculations, we determined the sea salt optical properties based on Mie calculations, giving the optical properties for the three sea salt modes and eight spectral intervals. The simulated sea salt aerosol optical depth is found to show strong dependence on the 10 m wind speed under cloud-free conditions. This relation is best represented by a power law fit and compares well with satellite observations. For clear-sky conditions, the simulated sea salt direct radiative effects on the shortwave and longwave radiative budgets are approximately of the same order of magnitude, but with opposite signs. This causes the net radiative effect to approach zero, which leads to a low impact on the temperature for this area.

1 Introduction

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] Aerosol particles modify the atmospheric radiation budget, directly through the extinction (scattering and absorption) of light and indirectly through the contribution to cloud formation. The aerosol effects on climate are potentially as large as those caused by greenhouse gases, but with more varieties [Intergovernmental Panel on Climate Change IPCC, (2007)] and on the whole probably with an opposite sign. The radiative impact caused by aerosol particles is associated with seasonal dependence and large spatial variations and can become dominant on the regional scale [Tzanis and Varotsos, 2008]. The state of knowledge regarding the aerosol impact on various atmospheric processes on the regional scale is even worse than that on the global scale [Vogel et al., 2009].

[3] To establish a better understanding of the anthropogenic impacts on the atmospheric state, it is of central importance to separate the anthropogenic contribution from the natural one. Sea salt aerosol is produced through primary emissions from the ocean as a result of the wind effect on the water surface. Given that the earth's surface is roughly 70% covered by oceans, sea salt is probably the key aerosol constituent over large parts of the ocean regions. Consequently, sea salt may play a major role for atmospheric processes that govern climate change. This study focuses on the direct radiative effect (DRE) of sea salt, i.e., the extinction of radiation by the particles. The sea salt aerosol is responsible for the majority of aerosol-scattered light in clean marine background air [e.g., Murphy et al., 1998; Quinn et al., 1998; Quinn and Bates, 2005]. Atmospheric models taking the direct sea salt radiative effect into consideration are applied mainly on the global scale [e.g., Grini et al., 2002; Ma et al., 2008], and studies on regional scales are rare. Globally, the sea salt aerosol has a direct cooling effect, because the extinction is dominated by the scattering of sunlight, leading to a decrease in surface temperature [e.g., Lewis and Schwartz, 2004]. Although some previous model studies have dealt with the sea salt direct radiative effect on climate, the magnitude of this impact is still connected to large uncertainties and requires further investigation [Lewis and Schwartz, 2004; Satheesh and Krishna Moorthy, 2005]. Lewis and Schwartz [2004] summarized the global annually averaged direct effect from different studies as ranging from −0.08 Wm−2 to −6.2 Wm−2. For comparison, the direct radiative effect is estimated to be −0.4 ± 0.2 Wm−2 for sulfate, +0.03 ± 0.12 Wm−2 for biomass burning aerosol, −0.05 ± 0.05 Wm−2 for fossil fuel organic carbon, +0.2 ± 0.15 Wm−2 for fossil fuel black carbon aerosol, +0.1 ± 0.1 Wm−2 for nitrate, and −0.1 ± 0.2 Wm−2 for mineral dust [IPCC, 2007].

[4] In addition, previous investigations of the optical characteristics of sea salt aerosol that are important for the direct radiative effect are rare. One of the few studies of this topic is the global model study by Ma et al. [2008]. However, the authors neglected any internal mixing with, e.g., sulfate, which probably also contributes to the extinction of radiation by the particles, and calculated the sea salt mean concentration directly as a function of wind speed, without considering the sea salt emission flux specifically. Ma et al. [2008] applied the optical parameterization by Li et al. [2008], which is one of a few parameterizations describing both the longwave and the shortwave sea salt optical properties. For wet, internally mixed aerosol particles in general, the parameterization of Ghan and Zaveri [2007] can be applied. The evaluation of this parameterization, with respect to sea salt particles and to the behavior in the longwave range is, however, restricted. The optical parameterization of Li et al. [2008] and Ghan and Zaveri [2007] are both based on the volume mixing rule for soluble constituents. The volume mixing rule, however, is not valid for aerosols such as sea salt, which dissolve to liquid droplets at high relative humidity [Liu and Daum, 2008; Irshad et al., 2009]. Any applications with respect to the sea salt direct radiative effect on the regional scale, including shortwave and longwave radiative effects, are to our knowledge not available in the literature.

[5] Simulating the sea salt influence on the state of the atmosphere accurately is a complex task, in part because of the variable state of the particle as a result of ageing and exchange processes with ambient aerosols and gases. Sea salt is often found internally mixed with other chemical compounds and aerosols [e.g., Andreae et al., 1986; Sievering et al., 1992; O'Dowd et al., 1997; Clegg et al., 1998; Katoshevski et al., 1999; Li et al., 2011]. Internal mixtures of sea salt aerosol with non-sea salt sulfate, as a result of the condensation of sulphuric acid or aqueous-phase oxidation, has been found in several studies [e.g., Sievering et al., 1992; O'Dowd et al., 1997; Clegg et al., 1998; Katoshevski et al., 1999]. The hygroscopic nature of the sea salt particles leads to effective growth because of the water uptake, which modifies the aerosol optical properties and alters the aerosol life cycle. However, in previous model studies, the composition of sea salt is often assumed to remain constant during transport, so any exchange with ambient gaseous and aerosol species is neglected [e.g., Foltescu et al., 2005; Grini et al., 2002; Ma et al., 2008]. The feedback between sea salt aerosol and the atmosphere is also seldom taken into account.

[6] In this study, the feedback between sea salt concentrations and the state of the atmosphere on the regional scale, induced by the sea salt direct radiative effect, is considered. The tool used for this study is the regional-scale nonhydrostatic numerical atmospheric model COSMO-ART [Vogel et al., 2009]. COSMO-ART is, within this work, extended with respect to the treatment of sea salt. Internally mixed wet aerosol composed by sodium chloride, sodium sulfate, and liquid water is included in the model. Also, the condensation of sulphuric acid of natural and anthropogenic origin onto sea salt particles is added to the model. The most important natural source for sulfur species is the emissions of oceanic dimethylsulfide (DMS) [Pham et al., 1995]. The emission of DMS is also included in COSMO-ART. For radiation calculations, information about the optical properties of the particles is required. For this reason, a new set of optical properties is determined based on Mie calculations, describing the internally mixed sea salt aerosol optical properties and introduced into COSMO-ART. A case study is performed with the extended version of COSMO-ART for the Mediterranean, Northeast Atlantic, and a large part of Europe.

[7] With the assumption that no loss of sea salt by cloud formation takes place, the main aim of this study is to answer the following questions: 1) How does the marine environment respond to the sea salt aerosol direct effect in the Mediterranean region? 2) How important is the sea salt DRE compared with the DRE of anthropogenic aerosols in this region?

2 The COSMO-ART Model System

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[8] The model system COSMO-ART is designed for applications on the regional scale. The model incorporates sophisticated modules for gaseous chemistry and aerosol dynamics, which are coupled online to the meteorological driver COSMO [Baldauf et al., 2011]. The COSMO-ART model has been described in detail by Vogel et al. [2009]. The capabilities and applications of the model system moreover have been treated by Lundgren [2006], Vogel et al. [2006; 2008], Stanelle et al. [2010], Knote et al. [2011], and Bangert et al. [2011, 2012]. In this study, the chemical composition of the sea salt aerosol in COSMO-ART is extended from a pure sodium chloride (NaCl) aerosol to an internally mixed aerosol containing NaCl, sodium sulfate (Na2SO4), and liquid water. Thus, the sea salt modes are now also coupled to the gas phase in the model. In the following, we focus on the treatment of sea salt.

2.1 Conservation Equations for Sea Salt

[9] The conservation equations of the internally mixed sea salt aerosol are solved for the number density, sulfate content, and content of sea salt. The water content on the aerosol is not transported but locally equilibrated. This may lead to overestimated water content for the coarse sea salt mode, because kinetic effects can delay water uptake, which is not accounted for when assuming equilibrium conditions. The extended equations for sea salt take the form

  • display math(1)

where inline image is the wind speed vector and inline image, inline image, and inline image are the sedimentation velocities for the number density NaCl density, and sulfate mass density of each mode, respectively. The sedimentation velocities are calculated as a function of the total density of the aerosol, the mode median diameter, and the geometric standard deviation of the mode. The turbulent fluxes, inline image, inline image, and inline image, are parameterized by analogy to the turbulent fluxes in the diffusion equation and in the budget equation of the water vapor in the COSMO model [Vogel et al., 2009]. inline image is the change of mass resulting from condensation (see section 2.3), and inline image and inline image are the losses from washout (see section 2.4). inline image, inline image, and inline image are the density-weighted Reynolds averages for the respective components given by inline image, where ρa is the total mass density of humid air. inline image, inline image, and inline image are the normalized number density and the corresponding normalized mass densities, respectively. M0,i is the 0th moment and equals the total number density of the mode. This is a result of the definition of the moment for a log-normal distribution, which for the kth order is defined as inline image, where ni(ln Dp) is the number density distribution, and Dp is the particle diameter. The third moment is proportional to the mass density

  • display math(2)

[10] For the internally mixed aerosol, the total mass density mi,mix is the sum of the mass density of dry sea salt mi,NaCl, the mass density of the sulfate component mi,sulf, and the mass density of water mi,w. The density of the internally mixed aerosol is calculated by ρi,mix = 6mi,mix/(πM3,i,mix). Coagulation is currently not considered for the sea salt particles. This process will be included in future studies with COSMO-ART. However, because of the comparably low concentrations of these particles, any influence of neglecting coagulation is assumed to be low.

2.2 Sea Salt Liquid Water Content

[11] The determination of the aerosol liquid water content, mw, of the aerosol containing H+-Na+-Cl-SO42−-HSO4 is calculated with the use of the Zdanovskii-Stokes-Robinson (ZSR) relation [Stokes and Robinson, 1966]. For these calculations, the concentration and the binary molality of each electrolyte in the solution are required. The electrolyte concentrations are determined from the method presented by Zaveri et al. [2005]. In this study, the concentration of sodium ions predominates over the sulfur ions, i.e. inline image > > nS(VI), and the liquid water calculations are restricted to the “sulfate-poor” domain. The possible aqueous electrolytes are NaCl and Na2SO4. For determining the number of moles of the electrolytes, the equivalent ion fractions of each ion are used [Zaveri et al., 2005; Lundgren, 2010]. The binary molalities, mE0(aw), are determined as a function of the water activity (aw = RH/100), by the polynomial expressions of Tang [1997]. When the relative humidity drops below 58%, no water uptake by Na2SO4 is considered in the model. For conditions in which the relative humidity drops below 47%, the sea salt aerosol liquid water content is neglected. For RH exceeding 99%, the liquid water content is approximated with the corresponding value at 99%.

2.3 Condensational Growth and Formation of Sulphuric Acid from DMS

[12] inline image is the mean growth rate of each mode due to condensation of sulphuric acid. The growth rate is calculated according to Binkowski and Shankar [1995] as function of the available mass of sulphuric acid and the particle size distribution. Sulphuric acid is formed when SO2, which either is emitted directly or formed via chemical reactions in the atmosphere, is oxidized by OH in a three-body reaction. One of the most important sources of SO2 from natural sources is the emission of oceanic DMS. For this reason, the emission of oceanic DMS is implemented into the COSMO-ART model system. The flux of DMS, FDMS, is determined by the product of the parameterized transfer velocity and the DMS water surface concentration, Cw. The parameterization of Nightingale et al. [2000] is applied for calculation of the transfer velocity, giving the emission flux of DMS the form

  • display math(3)

[13] FDMS is the flux of DMS from the ocean surface (kg m−2 s−1), Mw,DMS denotes the molecular weight of DMS (kg mol−1), and U10 is the wind speed at 10 m (m s−1). The concentration Cw (mol L−1) is determined based on the work of Longhurst et al. [1995]. The data of Kettle and Andreae [2000] and the Global Surface Seawater DMS Database (http://saga.pmel.noaa.gov/dms/) are combined into a new DMS seawater climatology in COSMO-ART, as shown in Figure 1 and Table 1. The degradation of DMS through reactions with OH and NO3 is added to the chemistry module RADMKA.

  • display math(R1)
  • display math(R2)
  • display math(R3)
  • display math(R4)
image

Figure 1. DMS sea water provinces in COSMO-ART.

Download figure to PowerPoint

Table 1. DMS Sea Water Concentration in COSMO-ART
ProvinceCw (mol L−1)
NADR8.0 10−9
NAST5.0 10−9
NECS6.0 10−9
MEDI5.0 10−9

[14] The rate constants of the implemented reactions are obtained from Pham et al. [1995]. The dry deposition velocity of DMS is set to zero [Huneeus et al., 2009], and the dry deposition velocity of DMSO is set to 1 cm s−1 over the ocean and zero over the continent [Pham et al., 1995; Huneeus et al., 2009]. The dry deposition of SO2, and other gases, is parameterized according to Bär and Nester [1992]. The fraction of the gas-phase sulphuric acid, which does not take part in the condensation processes, will participate in binary homogeneous nucleation with water. The nucleation rate is calculated as a function of the production rate of sulphuric acid and the loss rate through condensation, based on the parameterization of Kerminen and Wexler [1994].

2.4 Removal Through Washout

[15] The terms inline image, inline image, and inline image denote the losses from washout. The washout terms are parameterized as a function of the particle size distribution and the size distribution of the rain droplets. The droplet number density distribution is described by the Marshall-Palmer distribution. The total droplet number density is determined based on the precipitation classification of Mircea and Stefan [1998]. Here we distinguished between weak, moderate, and strong precipitation. The calculated liquid water content of precipitation of the lowest model layer is used online in the model to determine the current precipitation classification. The parameters of the droplet distribution are kept constant over the complete precipitation column. Thermophoresis and diffusiophoresis are negelected. Moreover, the model does not supply any information on the charge state of the aerosol particles or precipitation droplets. Attraction resulting from electric charge is not taken into account. These considerations result in the following equations for the loss of the number density and mass densities, respectively, due to washout

  • display math(5)
  • display math

where Mk,i is the k:th moment for the mode i, and ρi,mix is the internally mixed particle density for the mode. The parameter ξl is given by inline image for l = 1, 2,… ,8, whereas αl and cl are presented in Table 2. A = 3,34 ⋅ λair and N0 is achieved from the definition of the Gamma function, as a function of e.g., the total droplet number density. For a more extensive description of the washout parameterization, see Rinke [2008].

Table 2. Coefficients αl and cl From the Washout Terms, Equation (2)a
lαlcl
  • a

    μair and μw are the dynamic viscosities of air and water, respectively.

12inline image
2inline imageinline image
3inline imageinline image
4inline imageinline image
5inline image130π
6inline imageinline image
7inline imageinline image
8inline imageinline image

[16] Here the removal through below-cloud scavenging is described. The removal through in-cloud scavenging, however, is not included in this version of the model. The loss through in-cloud scavenging is estimated to contribute 1% of the sea salt removal on average, compared with 28 − 39% and 60 − 70% of washout and dry deposition removal, respectively [Gong et al., 1997]. For this reason we assume that neglecting the removal through in-cloud precipitation scavenging leads only to minor errors.

2.5 Sea Salt Emissions

[17] In the atmosphere, we distinguish between film mode, jet mode, and spume mode particles. The initial geometric median diameter, Dg,ini, with respect to the number density of the freshly emitted sea salt aerosol particles of each mode is 0.2 µm, 2 µm, and 12 µm based on measurements by O'Dowd et al. [1997]. The corresponding standard deviations of the sea salt modes are 1.9, 2.0, and 3.0. A narrower distribution of the largest particles is applied under the assumption that these particles have a short residence time in the atmosphere. Thus, the standard deviation of 1.7 was utilized for the largest mode. The median mode diameters are modified during the transport and chemical transformation processes, the standard deviations are held constant.

[18] The emission fluxes for each mode are calculated online in COSMO-ART and enter equations (1) via the lower boundary conditions. For the emission fluxes, three different parameterizations, one for each of the sea salt modes, are applied. The emission flux of film mode particles is parameterized according to Mårtensson et al. [2003]

  • display math(7)

[19] F0 is the aerosol number flux, and Dp is the dry particle diameter of the freshly emitted particles, ranging between 0.02 and 1 µm. U10 is the horizontal wind speed (m s−1) at 10 m above ground, and the parameter Φ(Dp,Tw) is dependent on the dry particle diameter and the sea surface temperature, Tw, in K

  • display math

q Indicates three different size ranges, and the coefficients c0 − c4 and d0 − d4 for each size range have been given by Mårtensson et al. [2003] and Lundgren [2006]. The emission flux of supermicrometer aerosol particles in the jet mode is parameterized following Monahan et al. [1986].

  • display math(8)

[20] U10 is the wind speed in m s−1, and r80 is the sea salt particle radius in micrometers at the ambient relative humidity of 80%. The size range of the freshly emitted particles that enter the jet mode is 1 − 9 µm Dp. For the largest particles in the spume mode, the emission flux of sea salt particles has been described by Smith et al. [1993]:

  • display math(9)
  • display math

[21] U10 is again the wind speed in m s−1, and r80 is the sea salt particle radius given in micrometers at the ambient relative humidity of 80%. For the spume mode, the size range of dry freshly emitted particles corresponds to 9 − 28 µm Dp. The fluxes 8 and 9 are given for the wet particle radius r80 at ambient relative humidity of 80%. The parameterizations 7 − 9 are brought to the same form by applying the parameterization of [Lewis and Schwartz, 2006], which describes the wet aerosol size as a function of the dry size.

3 Sea Salt Optical Properties

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[22] In COSMO, the radiation scheme GRAALS [Ritter and Geleyn, 1992] is applied to calculate the vertical profiles of the shortwave and longwave radiative fluxes. Eight spectral bands, kb, are utilized in the radiation scheme. The single scattering albedo ω, the asymmetry factor, and the extinction coefficient b, are required for each sea salt mode, i, and all spectral bands, kb, at each grid point and every time step. Because of the enormous computational time of the necessary Mie calculations a new sea salt aerosol optical parameterization is developed based on Mie calculations. This is described below.

3.1 Effective Refractive Index

[23] The real part of the effective refractive index for the multicomponent sea salt aerosol is determined by the molar refraction approach [e.g., Stelson, 1990; Tang, 1997] as

  • display math(14)

[24] ηmix is the real effective refractive index of the solution aerosol at wavelength λ, Rmix,λ is the effective solution molar refraction at wavelength λ, Vmix is the solution volume density, and cn,mix is the total number of moles in the solution. The effective solution molar refraction is calculated by

  • display math(15)

where cnl is the number of moles of the components l in the solution, and Rl,λ is the partial molar refraction of each of the components. The latter take the form

  • display math(16a)
  • display math(16b)
  • display math(16c)

[25] ηl is the refraction index of the components, which is found in the literature (see section 3.2). Mw,l and ρl are the molecular weight and density of each component, respectively. The volume density of the solution is given by the fraction of the solution mass density cm,mix to the solution density ρmix. The aerosol density is derived from Tang [1997] by

  • display math(17)

where cm,l is the mass density of the components in the solution, and inline image and inline image are the densities of the binary solutions, which are expressed by the polynomials [Tang, 1997]

  • display math(18a)
  • display math(18b)

[26] The factor 103 converts the binary densities from g cm−3 to kg m−3, and w is the total solute weight in per cent inline image.

[27] The imaginary part of the refraction index of the internally mixed sea salt aerosol is based on the approach of Jacobson [2001; 2002]. The effective imaginary part of the refractive index, κmix,λ, for the internally mixed aerosol at wavelength λ is given by

  • display math(19)

[28] The product of the effective molar absorption, Amix,λ, and the mole concentration, cn,mix, of the mixed solution is determined from the partial molar absorptions, Al,λ, and the mole concentrations, cn,l , of each constituent, l,

  • display math(20)

[29] The molar and partial molar absorptions are given in m3 mol−1. In assuming that the partial molar absorptions of aqueous electrolytes are equal to those of the corresponding solid electrolytes, κl,λ corresponds to the imaginary part of the refractive index of water and the electrolytes, respectively. For the sea salt constituents, the partial molar absorptions take the form

  • display math(21a)
  • display math(21b)
  • display math(21c)

[30] The effective refractive indices (equations (15) and (20)) serve as input for the Mie calculations.

3.2 Mie Calculations for the Sea Salt Aerosol

[31] The Mie calculations are performed with the code that originally was developed by Bohren and Huffman [1983]. To consider the variability in the chemical composition of the sea salt aerosol, the Mie calculations are performed with pre-simulated aerosol distribution fields over the southwestern Mediterranean as input. The simulated aerosol size distributions and chemical composition are discussed in section 4.

[32] Because of the dominant mass of sea salt close to the surface (Fig. 2a), the direct radiative impact is expected to be as important at low levels. Based on this and the fact that the Mie calculations are highly time consuming, the calculations are carried out for each of the 500 grid cells in the lowest model layer, which has a thickness of ~20 m. By applying the Mie calculations for near-surface grid cells, the impact of varying chemical composition between the grid cells on the optical properties is considered.

image

Figure 2. a: Simulated averaged sea salt mass density for the southwestern Mediterranean at 18 UTC, 24 July 2007. b − d: Volume extinction coefficients for one shortwave interval (blue) and one longwave interval (red) as function of altitude.

Download figure to PowerPoint

[33] The Mie calculations are performed for 134 wavelengths between 250 nm and 30 µm. In the solar range, calculations are carried out for 101 wavelengths between 0.25 and 4 µm, and in the thermal range for 33 wavelengths between 4 and 30 µm. A lack of refractive index data in the longwave range restricts the calculations to 30 µm. For short wavelengths, the optical constants are weighted to the available energy from the solar spectra and averaged bandwise over the three shortwave spectral bands. For the longwave spectra, the optical coefficients are weighted with the Planck function at 275 K for each individual wavelength and averaged bandwise over the five longwave spectral bands.

[34] The real and imaginary parts of the refractive index of water are achieved from Segelstein [1981] as a function of wavelength. The values of bulk NaCl from Eldridge and Palik [1985] as a function of wavelength are utilized. For the solar range, the real part of the refractive index of sodium sulfate was assumed to be the same as the value at 589 nm [Mark Z. Jacobson, 2009, personal communication]. Based on the findings of, e.g., Mogili et al. [2007], the real part of the refractive index of sodium sulfate is achieved by taking the average value of the three components for the different crystal axes of Shannon et al. [2002]. For the thermal spectra, the wavelength dependent values of ammonium sulfate of Toon et al. [1976] are applied as a surrogate to sodium sulfate [Mark Z. Jacobson, 2009, personal communication]. The absorption by sodium sulfate in the solar range is neglected by assuming a value of 10−7 for the imaginary part of the refractive index, and in the thermal range the data for ammonium sulfate from Toon et al. [1976] are applied [Mark Z. Jacobson, 2009, personal communication].

3.2.1 Extinction Coefficient

[35] The volume extinction coefficient, inline image, resulting from the Mie calculations is shown in Figure 3a for the spume mode for all grid cells, for the third spectral band. Further results for the other spectral intervals and the other modes have been given by Lundgren [2010]. The relation between the extinction coefficient and the total wet aerosol mass of each mode shows a linear relation in the form

  • display math(22)

where the volume extinction coefficient form is given in m−1, the specific extinction coefficient, inline image, in m2 g−1, and the wet aerosol mass density, mi, in µg m−3. By applying linear fits, the specific extinction coefficients inline image are obtained as the slopes of each curve (Table 3). Thus, for any arbitrary sea salt aerosol distribution with wet mass mi, the extinction coefficient is obtained for the eight spectral intervals by applying equation (22) and the values of the specific extinction coefficient from Table 3. The dependence on the aerosol mass density that is calculated online in the model ensures the impact of the vertical concentration gradients on the extinction of radiation. The impact on the sea salt optical constants of vertical concentration gradients is treated in section 3.2.4. The total volume extinction coefficient for the whole sea salt aerosol distribution is achieved as the sum of the extinction coefficient for each mode

  • display math(23)
image

Figure 3. a − c: Optical constants as a function of the mass density.

Download figure to PowerPoint

Table 3. Specific Extinction Coefficients (m2 g−1), Single Scattering Albedos, and Asymmetry Parameters Derived from the Mie Calculations Based on Near-Surface Sea Salt Concentrations
kb (Wavelength in µm)
 12345678
(1.53 − 4.64)(0.7 − 1.53)(0.25 − 0.7)(20.0 − 104.5)(12.5 − 20.0)(8.33 − 9.01, 10.31 − 12.5)(9.01 − 10.31)(4.64 − 8.33)
inline image0.98132.18672.69860.14010.24890.10000.06890.1202
inline image0.35200.32320.30030.16980.19400.16550.17410.2704
inline image0.10320.09880.09690.10990.10350.07660.12410.1219
inline image0.92940.99991.00000.01740.02430.11930.16720.4276
inline image0.93980.99961.00000.30580.27980.54790.70440.7959
inline image0.91110.99881.00000.44210.42040.62080.72900.7345
inline image0.69660.76530.77010.10400.16400.27190.29400.3887
inline image0.80170.79170.81550.55030.67370.79270.80000.7999
inline image0.83720.83570.84670.77500.85270.88960.88020.8517

[36] The relation between the aerosol mass density and the extinction coefficients is based on the linear regression method, fitted by using the least squares approach. The correlation coefficients that result from this test exceed 0.98 for all the bands and all the sea salt modes. The deviations around the linear fits are explained by variations in number density, chemical composition, and aerosol size.

3.2.2 Single Scattering Albedo

[37] The total single scattering albedo, ωkb, for all the aerosol modes is achieved by weighting the single scattering albedo, inline image, of each mode i for the band kb, with the volume extinction coefficient

  • display math(24)

[38] The single scattering albedo, inline image, is obtained from the Mie calculations and inline image and inline image from equations (22) and (23), respectively. Since the extinction coefficient is calculated as a function of the aerosol mass density, equation (24) automatically gives a dependence not only on inline image but also on the aerosol mass density and specific extinction coefficient of each sea salt mode and spectral band, respectively. The output from the Mie calculations is illustrated in Figure 3b for the values of the jet mode for the third and fifth bands as a function of the mass densities for all grid cells. Further results for the other spectral intervals and the other sea salt modes have been given by Lundgren [2010]. The single scattering albedo of each mode remains approximately constant within each spectral band. Because of the generally constant values of each mode and band, the single scattering albedo is treated simply by choosing the mean values. This approach is also applied in the longwave range, although deviations from the mean values occur to a grerater extent in this part of the spectrum compared with the solar range. The values of the single scattering albedo for each mode and band are presented in Table 3.

[39] The absorption of sea salt is negligible in the solar range. For these three bands, mass independent constants are good approximations for the single scattering albedo for each of the three modes. The relative importance of absorption to the total extinction increases with increasing wavelength; i.e., the single scattering albedo decreases. The different characteristics of the chemical components may cause scattering about the mean values, depending on the chemical composition.

3.2.3 Asymmetry Parameter

[40] The asymmetry parameter inline image is obtained from the first moment of the phase function for each spectral band, aerosol mode, and grid cell of the model domain. The asymmetry parameter inline image for the whole aerosol distribution, including all three modes, is for each spectral band given by

  • display math(25)

where inline image and inline image are the total scattering coefficient and the scattering coefficient for mode i for each band kb, respectively. Thus, the total asymmetry parameter is calculated by weighting the asymmetry parameter of each mode to the corresponding scattering coefficient of the mode, relative to the total scattering coefficient of all the modes. The scattering coefficients are achieved from the extinction coefficient and the single scattering albedo. Hence, the total asymmetry parameter is determined as a function of the aerosol mass density, specific extinction coefficient, and single scattering albedo of each sea salt mode for each spectral interval. As an example, the calculated asymmetry parameters of the film and spume modes are shown as function of the aerosol mass density for the third spectral interval and each grid cell in Figure 3c. The asymmetry parameters for the other spectral intervals have been given by Lundgren [2010].

[41] The asymmetry parameter is treated by the same approach as the single scattering albedo, using mean values. The mean values of the asymmetry parameter for each mode and interval are summarized in Table 3. Representing the asymmetry parameters by constant mean values is a simplification. Weak deviations from the mean values are results from varying chemical composition. One explanation for the deviations is the change in chemical composition at coastal locations with low relative humidity. Higher wet mass densities are, moreover, more realistic in most marine conditions and are, therefore, of the most interest for climate modelling. Based on this, the mean values are applied, and the spread and deviations at low mass densities are neglected. The impact of the amount of scattered radiation, dependent on the sea salt mass density and specific extinction coefficients at each model layer, is accounted for via equation (25) in the model.

[42] The positive values of the asymmetry parameters for all modes and bands indicate that the scattering of radiation is predominantly in the forward direction. Constant values represent the asymmetry parameters well within the solar range, with only small deviations from the mean values. The asymmetry parameter varies with e.g., particle size, wavelength, and single scattering albedo [Andrews et al., 2006]. Andrews et al. [2006] found increased forward scattering for increased size and increased single scattering albedo, and g decreased with increasing wavelength. The deviations from the mean values are thus explained by altered particle sizes and compositions, which in turn affect the single scattering albedo. The asymmetry parameter increases with larger diameters and is, therefore, largest in the spume mode. As the particle size increases, so does the scattering in the forward direction. The average value for each mode decreases for longer wavelengths. Comparison with the values ranging between 0.75 and 0.85 in the solar range, reported by Winter and Chýlek [1997], reveals reasonable orders of magnitude of the calculated parameters for the spectral intervals in the solar regime.

3.2.4 Evaluation Against Mie Calculations

[43] In sections 3.2.1 − 3.2.3, the sea salt optical properties that are utilized in this study were described. The performance of this parameterization of the optical properties is evaluated against direct Mie output. This evaluation is carried out by applying the parameterization to vertical mean concentration profiles (Fig. 2a) and comparing the results based on the calculated optical properties (method “REF”) with direct Mie output (method “MIE”) for the same vertical mean concentrations. The evaluation is presented for two spectral intervals, one in the solar range (band 3, 0.25 − 0.7 µm) and one in the longwave range (band 5, 12.5 − 20 µm).

[44] The sea salt mass density decreases rapidly as a function of the altitude for all three modes, revealing strong vertical gradients (Fig. 2a). The highest concentrations, with respect to both mass and number (not shown), are found within the lowest 1 − 2 km. These characteristics are reasonable compared with previous findings [e.g., Caffrey et al., 2006; Glantz et al., 2003; 2004; Grini et al., 2002; Gong et al., 2002; Textor et al., 2006].

[45] The volume extinction coefficient (Fig. 2b − d) of all modes decreases rapidly with height because of the dependence on the aerosol mass density. In general, the volume extinction coefficients that results from the calculated specific coefficient (method REF) differ only slightly from the direct Mie output (MIE). The largest difference occurs for the smallest particles, and the smallest absolute deviation occurs for the largest particles. Except for the film mode extinction coefficient in the longwave range, an underestimation of the extinction coefficient, compared with the direct Mie output, is shown. The total calculated extinction coefficient near the surface is ~0.057 km−1 for the shortwave example, based on the surface concentrations. This is comparable to measured extinction coefficients for sea salt particles [e.g., Nishizawa et al., 2010]. The decay of the calculated extinction coefficient with height is, however, less rapid than the extinction coefficient measured by Nishizawa et al. [2010], which decreases to close to zero at a height of 700 m. However, because the aerosol extinction profile may vary rapidly with changes in the atmospheric boundary layer characteristics [Boyouk et al., 2011], it is here of most interest to compare the order of magnitude with the measured sea salt extinction coefficient.

[46] In the radiation calculations, the scattering and absorption optical depths are determined from the extinction coefficient and the single scattering albedo. The scattering, absorption, and total optical depths are calculated for each model layer illustrated as a function of altitude in Figure 4. As expected, scattering dominates the total optical depth in the shortwave range, and absorption dominates the total optical depth in the longwave spectrum. The profiles for the scattering optical depth from the two different approaches show similar features. The same is found for the absorption optical depth. Because the absorption in the solar range is negligible, any errors in this range are, however, unimportant.

image

Figure 4. Vertical profiles of scattering optical depth (a), absorption optical depth (b), total optical depth (c), and asymmetry parameter (d). The parameters are presented for one shortwave (blue) and one longwave (red) interval.

Download figure to PowerPoint

[47] Figure 4d shows the vertical profiles of the total asymmetry parameter from the REF and MIE methods. In the shortwave range, the REF asymmetry parameter is constant (0.77), whereas the Mie output varies between 0.74 and 0.78. In the two different approaches, the total asymmetry parameter is dominated by different particles sizes (not shown). The deviations of the REF method from the direct MIE output are on the order of <0.02. In the longwave range, any uncertainties in the asymmetry parameter have limited effects on the radiative effect, because scattering is much smaller than absorption in this range.

[48] Thus, the optical properties that were achieved by applying the optical coefficients in Table 3 to a vertical mean concentration profile agreed well with the results of direct Mie outputs for the same vertical concentration profile. Any effect on the optical depth was not significant, and the scattering of longwave radiation was small compared with the absorption, which makes the proportion of forward or backward scattering of the longwave radiation comparably unimportant for the calculated radiative effects.

4 Three-Dimensional Case Study

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[49] Three-dimensional simulations are performed with the extended model system COSMO-ART for 24 − 26 July 2007. The model domain is illustrated in Figure 7. The simulations are performed with a time step of 40 seconds, horizontal resolution of 0.25° in both directions, and 40 vertical layers. In the vertical, a pressure-based hybrid coordinate is used; i.e., the coordinate lines follow the terrain to a certain altitude (~11 km), above which it switches back to plane lines following the normal z-system. The depths of the model layers increase with increasing altitude. The lowest 1 km is, in this case, represented by 11 model layers. The top of the model domain is positioned at ~24 km, with a depth of the highest layer on the order of 3 km.

[50] The meteorological boundary and initial data are derived from IFS analyses [White, 2003]. Meteorological boundary data are applied every sixth hour during simulation. Clean air aerosol and gaseous initial and boundary conditions are utilized. This means that very low concentrations are applied at the beginning of the simulation and at the boundaries. A spin-up time of 2 days (22 and 23 July 2007) is applied. After these 2 days, the model is restarted at 20 UTC with the meteorological analysis data from IFS, whereas the simulated gaseous and aerosol concentrations are utilized from the previous spin-up simulation. This approach is applied to minimize the uncertainty connected with the meteorological fields after a few days of simulation and takes into account the chemical composition of the atmosphere based on the simulation of the 2 previous days.

[51] An overview of the individual simulations is provided in Table 4. In the reference run (R), sea salt aerosol containing mixtures of NaCl, sulfate, and water are considered, but no interaction with radiation takes place. In simulations Fssa, Fant, and Fall, the direct radiative effects and their feedback on the state of the atmosphere are considered. In Fssa, the interaction between wet, internally mixed sea salt aerosol and radiation is taken into account. Fant represents the simulation in which anthropogenic aerosol interacts with the state of the atmosphere through the DRE. In run Fall, both sea salt and anthropogenic aerosol direct radiative effects are considered. Anthropogenic gaseous and particulate emissions are derived from the TNO/GEMS (Netherlands) emissions database [Visschedijk et al., 2007]. The emissions of CO, NH3, NOx, SO2, volatile organic compounds (VOCs), and PM2.5 are considered. For the prevailing weather condition (see below), emissions from the European continent and emissions from ships in the model domain are assumed to be of the most importance for the considered area, and no emissions from Northern Africa are considered.

Table 4. Overview of the Simulations Performed With the Extended Model System COSMO-ARTa
RunDirect Radiative Effect
  • a

    The optical properties of the anthropogenic aerosols are treated according to Vogel et al. [2009].

RNo
FssaSea salt
FantAnthropogenic
FallSea salt + anthropogenic

[52] During the simulation period (24 − 26 July 2007), a surface low-level pressure system moved from west of Great Britain toward the northeast (Fig. 5). Southwest to westerly flow was dominating, with stable and warm conditions in the simulation area. Because of a trough approaching from the west, increased southwesterly flow occurred over France and Germany on the 26th (Fig. 5c). Over the Mediterranean Sea and southern Europe, cloud-free and dry conditions without precipitation dominated throughout the three days, 24 − 26 July (Figs. 5, 6). On the contrary, cloudy conditions dominated over central Europe for all 3 days, connected to the low-level pressure systems passing by in the northeastly direction. Over the continent, the precipitation sum over the last 6 hours ranged up to ~10 mm during the simulation period. The wind direction 10 m above the ground is illustrated in Figure 7. Close to the low-level pressure system above the North Atlantic, the wind speed was ~15 m s−1. Over central Europe and the Mediterranean Basin, the wind speed was generally on the order of 5 − 7.5 m s−1, with a maximum of 12 m s−1 south of mainland France, in the beginning of the simulation period. The weather situation during these 3 days does not represent any extreme weather conditions in the area. The synoptic conditions can rather be seen as typical for the summer months for this region. No major dust event over the Mediterranean was reported during this time period. Therefore, the aerosol impacts from marine or anthropogenic aerosol are expected to be the most dominant for these days.

image

Figure 5. a − c: Left: Synoptic situation represented by 500 hPa geopotential in gpdm (black contours), surface pressure in hPa (white contours), and relative topography in gpdm (colors). Right: Sum of precipitation in mm for the last 6 hours. Copyright www.wetter3.de.

Download figure to PowerPoint

image

Figure 6. Cloud cover. Copyright EUMETSAT/NEODAAS/University of Dundee.

Download figure to PowerPoint

image

Figure 7. Simulated 10 m horizontal wind speed and horizontal wind vectors.

Download figure to PowerPoint

4.1 Simulated Sea Salt Aerosol Chemical Composition

[53] An example of the simulated chemical composition is illustrated in Figure 8 for the film mode. Mass densities close to the surface (less than approx. 20 m), averaged over the simulation period, are illustrated for the chemical components. The total mass density is dominated by the liquid water content, generally on the order of 90% of the total mass density. The simulated NaCl content dominates over the sulfate content, which in general contributes to <1% of the total mass density. These simulated orders of magnitudes are representative for all the three sea salt modes. The contribution of the sulfate component is, however, greatest in the film mode compared with the larger sea salt modes.

image

Figure 8. Film mode mass densities of total wet mass (a), NaCl (b), SO42− (c), and H2O (d) for the lowest 20 m above the surface, averaged over 3 days.

Download figure to PowerPoint

4.2 Sea Salt AOD and Wind Speed: Comparison with Satellite Observations

[54] It is well known that an enhancement of the AOD over oceanic regions takes place as a result of the increased contribution of sea salt aerosol with increasing wind speed near the surface [e.g., Satheesh and Krishna Moorthy, 2006; Mulcahy et al., 2008; Glantz et al., 2009; Huang et al., 2010; Mulcahy et al., 2009]. Although many studies have been performed, there is still no consensus on the relation between wind speed and AOD, taking the form as linear exponential and power-law relations [e.g., Smirnov et al., 2003; Jennings et al., 2003; Huang et al., 2010; Lehahn et al., 2010; Satheesh and Krishna Moorthy, 2006; Mulcahy et al., 2008; 2009; Glantz et al., 2009]. Because all the studies are performed for different locations and time periods, the relations are difficult to compare directly. Since the sea salt emission flux is proportional to U103.14 [Monahan et al., 1986; Smith et al., 1993; Mårtensson et al., 2003] and the water absorption by sea salt is exponential with increasing ambient relative humidity [e.g., Tang et al., 1997; Tang, 1997; Lewis and Schwartz, 2004], a linear dependence of AOD and wind speed is in general rather unexpected.

[55] Describing the sea salt AOD as a function of wind speed provides the opportunity to treat the optical depth directly as a function of the accessible wind conditions. This study provides AOD and wind speed data as direct functions of each other since they originate from the same source, the model. This is an advantage compared with studies using data from different sources with individual uncertainties. The data from the simulation are analyzed for 12 UTC at 25 July 2007. Only grid points over the ocean are used, because no correlation over land is expected. Moreover, noncloudy pixels (cloud cover <1%) are selected to ensure the comparability with satellite-based observations. Any effects of boundaries are overcome by neglecting all grid cells that are closer than 1.25° to the edges of the mode domain.

[56] The maximum AOD is 0.3472, and the corresponding 99th percentile of the AOD data is 0.1279. To consider only data that are less well represented, the 99th percentile values were chosen [e.g., Glantz et al., 2009]. The AOD data were classified into 11 bins, with a bin width of 1 m s−1. The average AOD within each bin and the corresponding standard deviation are depicted as a function of the average bin wind speed in Figure 9. The best fit of the data is a power law of the form

  • display math(26)

with nondimensional AOD550 and U10 in m s−1. The power fit according to the 11 mean values is shown in Figure 9 in green. For comparison, the relations from Schinozuka et al., [2004] (red) and Glantz et al. [2009] (blue) are also depicted (Fig. 9) for the wind speed regime in which the relation to this study is valid (0.8 − 10.2 m s−1). In accordance with satellite-based observations, an offset of the sea salt AOD wind dependence is found at wind speeds on the order of 4 m s−1. For example, Lehahn et al. [2010] found an offset of the AOD wind dependence at 3.8 m s−1.

image

Figure 9. Averaged sea salt AOD values with corresponding standard deviations (gray), the power law fit of this study (green), Glantz et al. [2009] (blue), and Schinozuka et al. [2004] (red).

Download figure to PowerPoint

[57] The comparison (Fig. 9) reveals a wind-independent AOD contribution, which causes the AOD not to approach zero for low wind speeds. The magnitude of this contribution from the present study is estimated to be ~0.02, i.e., between the relations found by Schinozuka et al., [2004] and Glantz et al. [2009]. A combination of elevated aerosol layers that show less correlation with surface wind and swelling of the aerosol particles at high relative humidity are thought to be some of the effects behind the wind-independent AOD contribution. Overall, the relation found in this study is within the order of magnitude of those based on satellite retrievals. For this reason, we argue that, for cloud-free situations, the calculated AOD in this work is of a reasonable order of magnitude. Since the formation of cloud droplets is neglected, the mass of sea salt is most probably overestimated in the more cloudy areas. For these regions, the back-scattering is expected to be completely dominated by that of clouds.

4.3 Sea Salt Aerosol Direct Radiative Effect

[58] The sea salt aerosol impact on the radiation fluxes is calculated as the difference in the solar (FSW) and thermal (FLW) radiation budgets at the surface and the top of the atmosphere (TOA), respectively, between the results of run F and run R (see Table 5). The radiation budgets FSW and FLW are defined as the net solar and thermal radiative fluxes, respectively. The difference in the net shortwave radiation has similar properties at the surface and at the TOA (not shown). The sea salt impact on the longwave radiation is most important at the surface, whereas the effects at the TOA are even smaller. The longwave radiation impact is in general lower than the shortwave radiation impact, and with an opposite sign. This is found for areas where the average cloudiness was low. In cloudy areas, the sea salt DRE effect itself plays a minor role. Instead, the back-scattering of clouds is the dominant factor. Since the cloud systems are unstable, they are highly sensitive to small perturbations in the atmospheric state, such as those that are caused by the sea salt radiative effects. The spatial shift in the cloud cover pattern leads to a change in the incoming or outgoing radiation fluxes, i.e., the radiation budget is altered.

Table 5. Summary of the DRE of the Sea Salt Aerosol Mean Changes (run F − run R) in the Solar and Thermal Net Radiative Fluxes in Wm−2a
 TotalLandOcean
Whole SkyClear SkyWhole SkyClear Sky
  • a

    All values (for latitudes north of 35°N) are averaged over the simulation period and presented with the corresponding standard deviation.

ΔFSW(sfc)−0.62 ± 3.71−0.47 ± 4.09−0.06 ± 0.08−0.82 ± 3.08−0.18 ± 0.13
ΔFSW(TOA)−0.47 ± 3.21−0.38 ± 3.59−0.04 ± 0.05−0.61 ± 2.60−0.13 ± 0.08
ΔFLW(sfc)+0.19 ± 1.08+0.18 ± 1.27+0.06 ± 0.18+0.22 ± 0.720.16 ± 0.23
ΔFLW(TOA)+0.05 ± 0.54+0.07 ± 0.67+0.02 ± 0.04+0.03 ± 0.240.03 ± 0.04

[59] To identify the direct radiative effect on radiation budget, clear-sky conditions are considered in the following. The dependence of the radiative budgets on the wet sea salt AOD at 550 nm under cloud-free conditions is illustrated in Figure 10. Here all grid cells with clouds were removed and grid cells closer than 1.25° to the edges of the domain were neglected. The direct model output, without any averaging, is considered. For the shortwave fluxes, data at 8 UTC, 12 UTC, and 16 UTC for each day are considered. For the longwave fluxes, nighttime data are considered in addition to the daytime values. The difference in the shortwave fluxes is depicted in green for 8 UTC, blue for 12 UTC, and gray for 16 UTC. For the longwave range, model outputs at 20 UTC, 00 UTC, and 4 UTC are additionally taken into account. Mean values and standard deviation for each case are depicted in Figure 10. In earlier studies, an almost linear dependence of the radiative budgets on the sea salt loading was found [Li et al., 2008]. In addition, Stanelle et al. [2010] saw clear dependence between the mineral dust loading and differences in net radiation over West Africa at the surface. We find that the solar radiative budgets decrease with increasing sea salt optical depths. Hence, a negative correlation between the solar radiative budgets and the sea salt AOD is simulated. This effect is the most significant at the ocean surface. The opposite occurs for the thermal radiative budgets, which increase with increasing optical depths. The thermal radiative budget changes are, as for the solar range, relatively greater at the surface compared with the effects at the TOA. In comparing the effects over land with those over ocean, the greatest difference is the larger amount of data from over the oceans. This complicates any direct comparison. However, on average, the cooling and warming, respectively, are relatively more significant over the ocean than over land. Even though the effects are small, this indicates the importance of the surface albedo to the sea salt DRE in this study. Moreover, the effect in the shortwave range is seen to be strongest at midday (12 UTC; Fig. 10). The mean values and the corresponding standard deviations over the complete 3 day period (day and night) are given in Table 5 for each case. The average effects on the shortwave radiation and the longwave radiation are for the clear-sky conditions of approximately the same order of magnitude, but with opposite signs, which causes the net direct radiative effect of sea salt to approach zero.

image

Figure 10. a − h: Change (run F − run R) of the solar and thermal radiation budgets under cloud-free conditions as a function of simulated AOD. The differences in the shortwave fluxes are depicted for daytime only (green 8 UTC, blue 12 UTC, gray 16 UTC). The longwave fluxes are depicted for daytime and nighttime.

Download figure to PowerPoint

4.4 Comparison With Simulated Anthropogenic DRE

[60] The relative importance of sea salt aerosol DRE for this region was investigated by comparing the effect with that of anthropogenic aerosol. In simulations Fant and Fall, the feedback between anthropogenic and composite aerosols, respectively, with the atmosphere is considered. Composite aerosol is the sum of anthropogenic and sea salt aerosols. The anthropogenic AOD is on average on the order of up to 0.25 over the European continent (Fig. 11). This is within the typical range of anthropogenic AOD of 0.2 − 0.5 [Mulcahy et al., 2008]. At coastal locations, the sea salt AOD is on the same order as or higher than the anthropogenic AOD. For the radiation calculations, the different optical properties and the simulated mass densities of sea salt and anthropogenic aerosol at each model layer are accounted for. A summary of the mean changes in the radiative budgets and temperature for the run in which sea salt and anthropogenic aerosol are considered is presented in Table 6. The aerosol DRE for this region during 24 − 26 July 2007 was seen to be small. The difference in the radiative budgets when both sea salt and anthropogenic aerosols were considered is larger than when only considering sea salt (see Table 6). This was found for the whole domain and also when looking at the clear-sky land case and the whole-sky ocean case separately. The magnitude of the direct radiative effect is, however, the greatest, on the order of −1.1 Wm−2, averaged over the 3 days.

image

Figure 11. Sea salt (top) and anthropogenic (bottom) AOD, averaged over 3 days.

Download figure to PowerPoint

Table 6. Summary of the DRE of Composite (Anthropogenic + Sea Salt) Aerosola
 TotalLandOcean
Whole SkyClear SkyWhole SkyClear Sky
  • a

    Mean changes (run Fall − run R) in the solar and thermal net radiative fluxes in Wm−2. All values (for latitudes north of 35°N) are averaged over the simulation period and presented with the corresponding standard deviation.

ΔFSW(sfc)−1.07 ± 4.65−1.14 ± 4.89−0.25 ± 0.14−0.98 ± 4.30−0.39 ± 0.16
ΔFSW(TOA)−0.64 ± 4.08−0.70 ± 4.35−0.03 ± 0.10−0.54 ± 3.68−0.15 ± 0.11
ΔFLW(sfc)+0.38 ± 1.64+0.48 ± 1.970.11 ± 0.16+0.23 ± 1.01+0.18 ± 0.19
ΔFLW(TOA)+0.09 ± 0.63+0.13 ± 0.790.04 ± 0.04+0.04 ± 0.28+0.04 ± 0.03

Summary and Conclusions

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[61] The DRE of sea salt aerosol has been investigated on the regional scale for the case of no participation in cloud droplet formation. The nonhydrostatic online–coupled, regional-scale model system COSMO-ART [Vogel et al., 2009] was for this purpose extended with respect to the treatment of the sea salt aerosol. Throughout this study, sea salt was treated as an internal mixture of NaCl, sulfate, and water. Condensation of sulphuric acid from anthropogenic and natural sources onto the sea salt aerosol and the absorption of liquid water were considered. Additionally, the emission of oceanic DMS was implemented in the model within this study.

[62] For the radiation calculations, sea salt optical properties were calculated based on detailed Mie theory. The optical properties are valid for both solar and thermal ranges (0.25 < λ < 30 µm) and include the extinction coefficients, single scattering albedos, and asymmetry factors of wet sea salt aerosol for the three sea salt aerosol modes in both submicrometer and supermicrometer sizes and for eight spectral intervals. The optical properties were applied to a vertical sea salt mean concentration profile, and the results were compared with direct Mie output. The calculated properties revealed only small deviations from the Mie output. The new set of optical properties was implemented in COSMO-ART, and the impact on the radiative budgets was quantified.

[63] Simulations were performed with the extended version of COSMO-ART for a case study in July 2007. The simulated sea salt aerosol optical depth at 550 nm (AOD550) showed a strong wind speed dependence, which was best represented by a power law fit (AOD550 ~ U103.40). In agreement with the satellite observations, the offset of the 10-m wind speed dependence was found at about 4 m s−1. At low wind speeds, the wind-independent contribution was on the order of 0.02. The simulated AOD550 compared well with satellite observations. For cloud-free atmospheres the correlation between the AOD550 and the differences in the radiative fluxes was negative for solar irradiance and positive for thermal fluxes. The mean effect at the surface was greater over the ocean than over land, suggesting a more meaningful effect above surfaces with relatively lower albedo. Because of the interaction with thermal radiation, a warming occurred at the surface, during both night and day. Although a high AOD was simulated, a small temperature effect was found at the surface for clear-sky conditions because of the net radiative effect close to zero. The low net radiative effect at the surface was explained by cooling and warming effects in the solar and thermal spectra, respectively, of about the same order of magnitude. These results were found for conditions with clear skies. Therefore, we conclude that the sea salt DRE still carries large uncertainties and should be further investigated.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[64] We acknowledge Rainer Behrendt and Holger Mahlke. Figure 5 was taken from their website www.wetter3.de. We also thank EUMETSAT/NEODAAS/University of Dundee for providing the satellite images. Christoph Knote, Laboratory for Air Pollution/Environmental Technology, EMPA, and Hugo Denier van der Gon, TNO Environment and Geosciences, are acknowledged for providing the emission data for the simulations. Moreover, we thank David Topping, School of Earth, Atmospheric and Environmental Science, University of Manchester, for fruitful discussions with respect to the thermodynamic properties of sea salt particles.

References

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information
  • Andreae, M.O., R. J. Charlson, F. Bruynseels, H. Storms, R. van Grieken, and W. Maenhaut (1986), Internal mixture of sea salt, silicates, and excess sulfate in marine aerosols, Science, 232, 16201623.
  • Andrews, E., P. J. Sheridan, M. Fiebig, A. McComiskey, J. A. Ogren, P. Arnott, D. Covert, R. Elleman, R. Gasparini, D. Collins, H. Jonsson, B. Schmid, and J. Wang (2006), Comparison of methods for deriving aerosol asymmetry parameter, J. Geophys. Res., 111, D05S04, doi:10.1029/2004JD005734.
  • Baldauf, M., A. Seifert, J. Förstner, D. Majewski, M. Raschendorfer, and T. Reinhardt (2011), Operational convective-scale numerical weather prediction with the COSMO model: description and sensitivities, Mon. Wea. Rev., 139, 38873905, doi:10.1175/MWR-D-10-05013.1.
  • Bangert M., C. Kottmeier, B. Vogel, and H. Vogel (2011), Regional scale effects of the aerosol cloud interaction simulated with an online coupled comprehensive chemistry model, Atmos. Chem. Phys., 11, 44114423.
  • Bangert M., A. Nenes, B. Vogel, H. Vogel, D. Barahona, V. A. Karydis, P. Kumar, C. Kottmeier, and U. Blahak (2012), Saharan dust event impacts on cloud formation and radiation over Western Europe, Atmos. Chem. Phys., 12, 40454063.
  • Bär, M., and K. Nester (1992), Parameterization of trace gas dry deposition velocities for a regional mesoscale diffusion model, Ann. Geophys., 10, 912923.
  • Binkowski, F. S., and U. Shankar (1995), The regional particulate matter model, 1. model description and preliminary results, J. Geophys. Res., 100, 2619126209.
  • Bohren, C. F, and D. R. Huffman (1983), Absorption and Scattering of Light by Small Particles, John Wiley & Sons, New York.
  • Boyouk, N., J.-F. Léon, H. Delbarre, P. Augustin, and M. Fourmentin (2011), Impact of sea breeze on vertical structure of aerosol optical properties in Dunkerque, France, Atmos. Res., 101, 902910.
  • Caffrey, P. F., W. A. Hoppel, and J. J. Shi (2006), A one-dimensional sectional aerosol model integrated with mesoscale meteorological data to study marine boundary layer aerosol dynamics, J. Geophys. Res., 111(D24201), doi:10.1029/2006JD007237.
  • Clegg, N. A., C. D. O'Dowd, M. H. Smith, and J. A. Lowe (1998), Sulphate partitioning in marine aerosol, J. Aerosol. Sci., 29, 881882.
  • Eldridge, J. E., and E. D. Palik (1985), Sodium chloride, in Handbook of Optical Constants of Solids, edited by E. D. Palik, pp. 775793, Academic Press, Orlando.
  • Foltescu, V. L., S. C. Pryor, and C. Bennet (2005), Sea salt generation, dispersion and removal on the regional scale, Atmos. Environ., 39, 21232133.
  • Ghan, S. J., and R. A. Zaveri (2007), Parameterization of optical properties for hydrated internally mixed aerosol, J. Geophys. Res.,112(D10201), 10.
  • Glantz, P., K. J. Noone, and S. R. Osborne (2003), Scavenging efficiencies of aerosol particles in marine stratocumulus and cumulus clouds, Q. J. R. Meteorol. Soc, 129, 13291350.
  • Glantz, P., G. Svensson, K. J. Noone, and S. R. Osborne (2004), Sea-salt aerosols over the north-east Atlantic: Model simulations of the ACE-2 Second Lagrangian experiment, Q. J. R. Meteorol. Soc., 130, 21912215.
  • Glantz, P., E. D. Nilsson, and W. von Hoyningen-Huene (2009), Estimating a relationship between aerosol optical thickness and surface wind speed over the ocean, Atmos. Res., 92, 5868.
  • 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, 17171730.
  • Gong, S. L., L. A. Barrie, J. M. Prospero, D. L. Savoie, G. P. Ayers, J.-P. Blanchet, and L. Spacek (1997), Modeling sea-salt aerosols in the atmosphere 2. Atmospheric concentrations and fluxes, J. Geophys. Res., 102, 38193830.
  • Huang, H., G. E. Thomas, and R. G. Grainger (2010), Relationship between wind speed and aerosol optical depth over remote ocean, Atmos. Chem. Phys., 10, 59435950, doi:10.5194/acp-10-5943-2010.
  • Huneeus, N., O. Boucher, and Chevallier, F. (2009), Simplified aerosol modelling for variational data assimilation, Geosci. Model Dev., 2, 213229.
  • IPCC (2007), Climate change 2007: Physical science basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge.
  • Irshad, R., R. G. Grainger, D. M. Peters, R. A. McPheat, K. M. Smith, and G. Thomas (2009), Laboratory measurements of the optical properties of sea salt aerosol, Atmos. Chem. Phys., 9, 221230.
  • Jacobson, M. Z. (2001), Global direct radiative forcing due to multicomponent anthropogenic and natural aerosols, J. Geophys. Res., 106(D2), 15511568.
  • Jacobson, M. Z. (2002), Analysis of aerosol interactions with numerical techniques for solving coagulation, nucleation, condensation, dissolution, and reversible chemistry among multiple distributions, J. Geophys. Res., 107(D19), 4366.
  • Jennings, S. G., C. Kleefeld, C. D. O'Dowd, C. Junker, T. G. Spain, P. O'Brien, A. F. Roddy, and T. C. O'Connore (2003), Mace Head Atmospheric Research Station—Characterization of aerosol radiative parameters, Boreal Environ. Res., 8, 303314.
  • Katoshevski, D., A. Nenes, and J. H. Seinfeld (1999), A study of processes that govern the maintainance of aerosols in the marine boundary layer, J. Aerosol. Sci., 4, 503532.
  • Kerminen, V.-M., A. S., Wexler (1994), Post-fog nucleation of H2SO4-H2O particles in smog, Atmos. Environ., 28, 23992406.
  • Kettle, A. J., and M. O. Andreae (2000), Flux of dimethylsulfide from the oceans: A comparison of updated data sets and flux models, J. Geophys. Res., 105(D22), 793808.
  • Knote, C., D. Brunner, H. Vogel, J. Allan, A. Asmi, M. Äijälä, S. Carbone, D. van der Gon, J. L. Jimenez, A. Kiendler-Scharr, C. Mohr, L. Poulain, A. S. H. Prévôt, E. Swietlicki, and B. Vogel (2011), Towards an online-coupled chemistry-climate model: evaluation of trace gases and aerosols in COSMO-ART, Geosci. Model Dev., 4, 10771102.
  • Lehahn, Y., I. Koren, E. Boss, Y. Ben-Ami, and O. Altaratz (2010), Estimating the maritime component of aerosol optical depth and its dependency on surface wind speed using satellite data, Atmos. Chem. Phys., 10, 67116720, doi:10.5194/acp-10-6711-2010.
  • Lewis, E. R., and S. E. Schwartz (2006), Comment on “size distribution of sea-salt emissions as a function of relative humidity”, Atmos. Environ., 40, 588590.
  • Lewis, E. R., and S. E. Schwartz (2004), Sea Salt Aerosol Production: Mechanisms, Methods, Measurements and Models: A Critical Review, p. 413, American Geophysical Union, Washington DC.
  • Li, J., X. Ma, K. von Salzen, and S. Dobbie (2008), Parameterization of sea-salt optical properties and physics of the associated radiative forcing, Atmos. Chem. Phys., 8, 47874798.
  • Li W., L. Shao, R. Shen, S. Yang, Z. Wang, and U. Tang (2011), Internally mixed sea salt, soot, and sulfates at Macao, a coastal city in South China, J. Air Waste Manag. Assoc., 11661173.
  • Liu, Y., and P. H. Daum (2008), Relationship of refractive index to mass density and self-consistency of mixing rules for multicomponent mixtures like ambient aerosols, J. Aerosol. Sci., 39, 974986.
  • Longhurst, A., S. Sathyendranath, T. Platt, and C. Caverhill (1995), An estimate of global primary production in the ocean from satellite radiometer data, J. Plankton Res., 17, 12451271.
  • Lundgren, K. (2006), Numerical Simulation of the Spatial and Temporal Distribution of Sea Salt Particles on the Regional Scale, M.S. thesis, Dep. of Met., Stockholm Univ., Sweden.
  • Lundgren, K. (2010), Direct radiative effects of sea salt on the regional scale., PhD thesis,Karlsruhe Institute of Technology, Karlsruhe, Germany.
  • Ma, X., K. von Salzen, and J. Li (2008), Modelling sea salt aerosol and its direct and indirect effects on climate, Atmos. Chem. Phys., 7, 13111327.
  • Mårtensson, M., E. D. Nilsson, G. de Leeuw, and L. H. Cohen (2003), Laboratory simulations and parameterization of the primary marine aerosol production, J. Geophys. Res., 108(D9), 42974308.
  • Mircea, M., and S. Stefan (1998), A theoretical study of the microphysical parameterization of the scavenging coefficient as a function of precipitation type and rate, Atmos. Environ., 32, 29312938.
  • Mogili, P. K., K. H. Yang, M. A. Young, and P. D. Kleiber (2007), Environmental aerosol chamber studies of extinction spectra of mineral dust aerosol components: Broadband IR-UV extinction spectra, J. Geophys. Res., 112, 12.
  • Monahan, E. C., E. Spiel, and K. L. Davidson (1986), A model of marine aerosol generation via whitecaps and wave disruption, in Oceanic Whitecaps, pp. 167174, D. Reidel Publishing Company, Dordrecht, Netherlands.
  • Mulcahy, J. P., C. D. O'Dowd, S. G. Jennings, and D. Ceburnis (2008), Significant enhancement of aerosol optical depth in marine air under high wind conditions, Geophy. Res. Lett., 35(L16810), 14.
  • Mulcahy, J. P., C. D. O'Dowd, and S. G. Jennings (2009), Aerosol optical depth in clean marine and continental northeast Atlantic air, J. Geophys. Res., 114, 10.
  • Murphy, D. M., J. R. Anderson, P. K. Quinn, L. M. McInnes, F. J. Brechtel, S. M. Kreidenweis, A. M. Middlebrook, M. Pósfai, D. S. Thomson, and P. R. Buseck (1998), Influence of sea-salt on aerosol radiative properties in the Southern Ocean marine boundary layer, Nature, 392, 6265.
  • Nightingale, P. D., G. Malin, C. S. Law, A. J. Watson, and P. S. Liss (2000), In situ evaluation of air − sea gas exchange parameterizations using novel conservative and volatile tracers, Glob. Biogeochem. Cycles., 14, 373387.
  • Nishizawa, T., N. Sugimoto, I. Matsui, A. Shimizu, X. Liu, Y. Zhang, R. Li, and J. Liu (2010), Vertical distribution of water-soluble, sea salt, and dust aerosols in the planetary boundary layer estimated from two-wavelength backscatter and one-wavelength polarization lidar measurements in Guangzhou and Beijing, China, Atmos. Res., 96, 602611.
  • O'Dowd, C. D., M. H. Smith, I. E. Consterdine, and J. A. Lowe (1997), Marine aerosol, sea-salt and the marine sulphur cycle: A short review, Atmos. Environ., 31, 7380.
  • Pham, M., J.-F. Müller, G. P. Brasseur, and G. Mégie (1995), A three-dimensional study of the tropospheric sulfur cycle, J. Geophys. Res., 20, 6192.
  • Quinn, P. K., and T. S. Bates (2005), Regional aerosol properties: Comparisons of boundary layer measurements from ACE 1, ACE 2, Aerosols99, INDOEX, ACE Asia, TARFOX, and NEAQS, J. Geophys. Res., 110, 24.
  • Quinn, P. K., D. J. Coffman, V. N. Kapustin, and T. S. Bates (1998), Aerosol optical properties in the marine boundary layer during the first Aerosol Characterization Experiment (ACE 1) and the underlying chemical and physical properties, J. Geophys. Res., 103, 547563.
  • Rinke, R. (2008), Parametrisierung des Auswaschens von Aerosolpartikeln durch Niederschlag, PhD thesis (in German), University of Karlsruhe (TH), Karlsruhe, Germany.
  • Ritter, B,. and J.-F. Geleyn (1992), A comprehensive scheme for numerical weather prediction models with potential applications in climate simulations, Mon. Wea. Rev., 120, 303325.
  • Satheesh, S. K., and K. Krishna Moorthy (2005), Radiative effects of natural aerosols: A review, Atmos. Environ., 39, 20892110.
  • Satheesh, S. K., and K. Krishna Moorthy (2006), Contribution of sea-salt to aerosol optical depth over the Arabian Sea derived from MODIS observations, Geophys. Res. Lett., 33, 4.
  • Segelstein, D. J. (1981), The complex refractive index of water, MS thesis, University of Missouri, Kansas City, Missouri.
  • Shannon, R. D., R. C. Shannon, O. Medenbach, and R. X. Fischer (2002), Refractive index and dispersion of flourides and oxides, J. Phys. Chem. Ref. Data, 31, 931970.
  • Schinozuka, Y., A. D. Clarke, S. G. Howell, V. Kapustin, and B. J. Huebert (2004), Sea-salt vertical profiles over the Southern and tropical Pacific Oceans: Microphysics, optical properties, spatial variability, and variations with wind speed, J. Geophys. Res., 109(D24201), 17.
  • Sievering, H., J. Boatman, E. Gorman, Y. Kim, L. Anderson, G. Ennis, M. Luria, and S. Pandis (1992), Removal of sulphur from the marine boundary layer by ozone oxidation in sea-salt aerosols, Nature, 360, 571573.
  • Smirnov, A., B. N. Holben, O. Dubovik, R. Frouin, T. F. Eck, and I. Slutsker (2003), Maritime component in aerosol optical models derived from Aerosol Robotic Network data, J. Geophys. Res., 108, 11.
  • Smith, M. H., P. M. Park, and I. E. Consterdine (1993), Marine aerosol concentrations and estimated fluxes over the sea, Q. J. R. Meteorol. Soc., 119, 809824.
  • Stanelle, T., B. Vogel, H. Vogel, D. Bäumer, and C. Kottmeier (2010), Feedback between dust particles and atmospheric processes over West Africa during dust episodes in March 2006 and June 2007, Atmos. Chem. Phys., 10, 1077110788.
  • Stelson, A. W. (1990), Urban aerosol refractive index prediction by partial molar refraction approach, Environ. Sci. Technol., 24, 16761679.
  • Stokes, R. H., and R. A. Robinson (1966), Interactions in aqueous nonelectrolyte solutions. I. Solute − solvent equilibria, J. Phys. Chem., 70, 21262130.
  • Tang, I. N. (1997), Thermodynamic and optical properties of mixed-salt aerosols of atmospheric importance, J. Geophys. Res., 102(D2), 18831893.
  • Tang, I. N., A.C. Tridico, and K.H. Fung (1997), Thermodynamic and optical properties of sea salt aerosols, J. Geophys. Res., 102(D19), 269275.
  • Toon, O. B., J. B. Pollock, and B. N. Khare (1976), The optical constants of several atmospheric aerosol species: ammonium sulfate, aluminium oxide and sodium chloride, J. Geophys. Res., 81, 57335748.
  • Textor, C., M. Schulz, S. Guibert, S. Kinne, Y. Balkanski, S. Bauer, T. Berntsen, T. Berglen, O. Boucher, M. Chin, F. Dentener, T. Diehl, R. Easter, H. Feichter, D. Fillmore, S. Ghan, P. Ginoux, S. Gong, A. Grini, J. Hendricks, L. Horowitz, P. Huang, I. Isaksen, T. Iversen, S. Kloster, D. Koch, A. Kirkevåg, J. E. Kristjansson, M. Krol, A. Lauer, J. F. Lamarque, X. Liu, V. Montanaro, G. Myhre, J. Penner, G. Pitari, S. Reddy, Ø. Seland, P. Stier, T. Takemura, and X. Tie (2006), Analysis and quantification of the diversities of aerosol life cycles within AeroCom, Atmos. Chem. Phys., 6, 17771813.
  • Tzanis, C., and C. A. Varotsos (2008), Tropospheric aerosol forcing of climate: A case study for the greater area of Greece, Int. J. Remote Sens., 29, 25072517.
  • Visschedijk, A. J. H., P. Zandveld, and H. A. C. Denier van der Gon (2007), A high resolution gridded European emission database for the EU integrated project GEMS, TNO-report 2007-A-R0233/B, TNO, Apeldoorn, Netherlands.
  • Vogel, B., C. Hoose, H. Vogel, and C. Kottmeier (2006), A model of dust transport applied to the Dead Sea area, Meteorol. Z., 14, 611624.
  • Vogel, H., A. Pauling, and B. Vogel (2008), Numerical simulation of birch pollen dispersion with an operational weather forecast system, Int. J. Biometeorol., 52, 805814.
  • Vogel, B., H. Vogel, D. Bäumer, M. Bangert, K. Lundgren, R. Rinke, and T. Stanelle (2009), The comprehensive model system COSMO-ART—Radiative impact of aerosol on the state of the atmosphere on the regional scale, Atmos. Chem. Phys., 9, 86618680.
  • Whitby, K. T., P. H. McMurray, U. Shankar, and F. S. Binkowski (1991), Modal aerosol dynamics modeling, Technical Report 600/3-91/020, (NTIS PB91-161729/AS Natl. Tech. Inf. Serv. Springfeld, Va.), Atmos. Res. and Exposure Assess. Lab. U.S. Environ. Prot. Agency, Research Triangle Park, N.C.
  • White, P. W. (2003), IFS Documentation Cycle (CY25r1), European Centre for Medium-Range Weather Forecasts, Reading, England.
  • Winter, B., and P. Chýlek (1997), Contribution of sea salt aerosol to the planetary clear-sky albedo, Tellus, 49B, 7279.
  • Zaveri, R. A., R. C. Easter, and A. S. Wexler (2005), A new method for multicomponent activity coefficients of electrolytes in aqueous atmospheric aerosols, J. Geophys. Res., 110(D02201), 23.

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. 1 Introduction
  4. 2 The COSMO-ART Model System
  5. 3 Sea Salt Optical Properties
  6. 4 Three-Dimensional Case Study
  7. Summary and Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

The supporting information may be found in the online version of this article.

FilenameFormatSizeDescription
Figure_Orientation.docWord document6007KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.