Wind-driven influences on aerosol light scattering in north-east Atlantic air



[1] Ten years (2001–2010) of aerosol light-scattering measurements in N.E. Atlantic marine air are analysed to determine wind-speed related influences on scattering properties. The scattering coefficient and the backscattering coefficient dependency on wind speed (U) was determined for the winter (Low Biological Activity-LBA) and the summer seasons (High Biological Activity-HBA), and was found to be dependent on ∼U2. In spite of having a U2dependency, scattering properties for the LBA-period are approximately twice those of the HBA-period. 96% of the LBA-HBA scattering difference can be explained by the combined effects of size distribution and refractive index differences while 70% of the scattering difference can be attributed to a difference in refractive index alone resulting from organic-matter enrichment during the HBA period. The 550 nm scattering coefficient was ∼70 Mm−1 for ∼25 ms−1 wind speeds, which is considerably higher than that encountered under polluted air masses in the same region.

1. Introduction

[2] Sea spray aerosol is perhaps the most significant natural aerosol component with global emissions of the order of 6297 Tg y−1 [Vignati et al., 2010] to 10,120 Tg y−1 [Andreae and Rosenfeld, 2008], compared to 1600 Tg y−1 [Andreae and Rosenfeld, 2008] to 1776 Tg y−1 [Dentener et al., 2006] for soil dust. However, it should be noted that there are still large differences across a wider range of budget estimates [de Leeuw et al., 2011]. O'Dowd et al. [2004]demonstrated that a significant fraction of the sub-μm sea spray mass can be comprised of organic matter depending on ocean biological productivity while Facchini et al. [2008] illustrated that the primary organic matter associated with sea spray is almost exclusively water insoluble. Vignati et al. [2010]calculated the global annual budget for both the sea salt and organic fractions of the sub-μm sea spray and found a global emission of 24 Tg y−1 for sea salt and 8.2 Tg y−1 for primary organics. Given the fact that sea spray is emitted over a dark surface (i.e., the ocean surface), and that the ocean covers approximately 70% of the Earth's surface, aerosol and cloud layers formed on, or modified by, sea spray are likely to have a significant impact on the global radiative budget. Indeed, Mulcahy et al. [2008]reported a high correlation between aerosol optical depth (AOD) and wind-speed with AOD values of 0.3–0.4 at moderately high wind speed. They found a power-law dependency between wind speed and AOD and suggested that sea spray contributed significantly to the direct radiative effect.O'Dowd et al. [1999] highlighted the preferential activation of sea salt nuclei over sulphate nuclei in marine clouds, while Ovadnevaite et al. [2011] found that even the water insoluble organic sea spray plumes have almost a 100% activation efficiency even at a low supersaturation of 0.2%, pointing to a significant role for sea spray in the direct radiative effect.

[3] While there have been many studies of aerosol light-scattering in remote locations and in marine air [Bodhaine, 1996; Parameswaran et al., 1998; Pereira et al., 2011], there have been few studies [e.g., Kleefeld et al., 2002] that have reported wind-dependent scattering dependency. The aim of present study is to establish a wind-speed scattering relationship for clean marine air masses for conditions representative of periods with high organic matter enrichment and periods with low organic matter enrichment.

2. Experiment

[4] Aerosol light scattering coefficient (σscat) and backscattering coefficient (σbscat) measurements were made at three wavelengths (450, 550 and 700 nm) on a continuous basis from 2001–2010 using a commercially available three wavelength Nephelometer instrument [Bodhaine et al., 1991] at the Mace Head Atmospheric Research Station (53°19′N, 9°54′W) [O'Connor et al., 2008] on the west coast of Ireland. The research station, also a Global Atmosphere Watch (GAW) measurement site, is designated as a clean marine background station for atmospheric aerosol research [Jennings et al., 2003]. Air masses arriving at the research station from the wind sector 190°–300° are clean marine in origin with 52% frequency of occurrences [Jennings et al., 2003].

[5] Aerosol measurements at Mace Head are taken from a 10-meter stainless-steel community sampling duct, 100 mm diameter, and with a laminar flow rate of 150 LPM. The Nephelometer is connected to the main duct via a 25 mm diameter tube and operates at a flow rate of 16.6 LPM. The sampling efficiency of the inlet system is reported byKleefeld et al. [2002], where the author also reported the 50% cut-off diameter of the inlet system to be 8μm at wind speeds less than 5 ms−1 and around 4 μm at a wind speed of 10 ms−1. A size cut-off of 4μm or above is considered to be appropriate in the present case as the mean wind speed for the 10 year period is 7 ± 4 ms−1 and the 75th percentile is below 10 ms−1 [Vaishya et al., 2011]. Wind speed and wind direction along with other meteorological parameters are measured on a continuous basis at the research station. Black carbon mass concentration used as an indicator for continental influences [Jennings et al., 1993; O'Dowd et al., 1993] was measured using an Aethalometer instrument [Hansen et al., 1984]. Black carbon data used in this study have inputs from two different Aethalometer models: AE-9 from 2001 up to April 2005 and AE-16 from May, 2005 up to 2010.

3. Data Selection

[6] Scattering measurements were grouped into the winter season, or LBA period, (December, January and February) scattering and the summer season, or HBA period, (June, July and August) scattering. The motivation to classify the scattering measurements in different seasons comes from the studies of O'Dowd et al. [2004] and Yoon et al. [2007]where a seasonal pattern in the physicochemical properties of Northeast Atlantic clean marine atmospheric aerosols has been reported. During the LBA period sea-salt mass concentration shows a maximum whereas it shows a minimum during the HBA period [Yoon et al., 2007]. Organic matter mass fraction in the sub-μm mode increases from 15% during LBA period to 63% during the HBA period [O'Dowd et al., 2004]. Similar seasonal variations in the mass concentration of non-sea-salt (nss) sulphate, water soluble organic carbon (WSOC) and total carbon (TC) have been found [Yoon et al., 2007]. nss sulphate mass concentration in fine mode exhibited lower values during winter season and higher values during summer season.

[7] The following criteria were imposed on the hourly averaged light scattering data to ensure clean marine air conditions: (1) Hourly averaged wind direction was in the designated clean marine sector, i.e., between 190°–300°; and (2) Hourly averaged black carbon mass concentration was <=50 ng/m3. The above mentioned criteria to ensure advection of clean marine air masses over to Mace Head Atmospheric Research Station has been established and reported previously [Cooke et al., 1997; Jennings et al., 1997; O'Dowd et al., 1993].

[8] For the scattering data used in this study, the light scattering measurements are not affected by relative humidity (RH) and are representative of dry scattering coefficient as the average RH within the Nephelometer measurement chamber remained around 34 ± 8% because of the heating caused by the instrument lamp and a warmer laboratory temperature as compared to ambient.

4. Results and Discussion

[9] Hourly averaged clean marine σscat and σbscat data were binned into wind speed bins with a bin width of 1 ms−1. The number of data point (hourly average values) in each wind speed bin is variable between 3 and 594. Ångstöm exponent (α) was calculated from the hourly averaged σscat values using Ångstöm power law for the wavelengths 450 nm and 700 nm. Figures 1 and 2 show the variation of σscat, σbscat, and α with respect to wind speed for the winter (circle) and the summer (square) seasons. Winter in the Northeast Atlantic is characterised by more frequent high wind episodes whereas in the summer, wind activity reduces [Jennings et al., 2003]. Hence in Figures 1 and 2 the highest wind speed bin is 24.5 ms−1 in the winter season and 15.5 ms−1 in the summer season. Both σscat and σbscat at all three wavelengths (450, 550 and 700 nm) are dependent on U through a power law of the form σ = σ0 + aUb. α was also fitted with a power law of the form α = α0 + cUd. The fit parameters σ0, α0, a, b, c, d and the coefficient of determination (R2) for all the wavelengths are listed in Table 1.

Figure 1.

Variation of (a) aerosol σscat at 550 nm with wind speed for winter, spring, summer and autumn. (b) σbscat at 550 nm with wind speed for winter and summer with standard deviations at 1 ms−1 wind speed bins.

Figure 2.

Variation of aerosol α450–700 nm. Circles represent the winter season and squares represent the summer season. Error bars are the standard deviation of the radiative parameter within each wind speed bin: dash-dot line for the winter season and solid line for the summer season.

Table 1. The Dependence of σscat and σbscat and of α on U Through the Relation σ = σ0 + aUb and α = α0 + cUd
Seasonλ (nm)σ0/α0b/da/cR2
Scattering Coefficient (σscat)
Backscattering Coefficient (σbscat)
Ångström Exponent (α)

[10] σscat is found to be dependent on approximate square of U, with the exponent ranging from 1.7–2.3 over the seasons and over the three wavelengths. The determination coefficient (R2) over all power-law fits for scattering coefficient ranged between 0.93–0.96.σbscat is dependent on U in a contrasting manner depending on season with b values ranging from 2.4–2.7 for the winter season and from 1.5–1.7 for the summer seasons, respectively. The higher dependency of σbscat on Uduring the winter season indicates perhaps a larger contribution towards scattering from wind speed generated primary produced sea salt particles relative to a generally lower concentration of secondary particles. During the summer season, when the ocean productivity is high, it leads to high enrichment fractions of organic matter in sub-μm aerosol [O'Dowd et al., 2004] and results in an increased relative contribution of sub-μm particles to σscat [Vaishya et al., 2011]. As wind speed increases, sea salt production increases and hence the particle size increases, thus decreasing the backscatter fraction which might weaken the σbscat dependency on U. σscat and σbscatdependency on U may be somewhat underestimated due to reduced cut-off diameter of the inlet system at higher wind speeds; however, potential losses are also off-set by the Nephelometer non-ideality in not sensing the increased forward-angle scattering contribution by larger particles (to the tune of 30% or more for particles >4μm) [Anderson et al., 1996].

[11] An interesting feature in both σscat and σbscat dependency on U is that the winter season radiative values are twice or more that of the summer season radiative (scattering and backscatter coefficient) values. It is evident also from Figure 2that the winter season is more dominated by super-μm particles at all wind speeds as can be inferred from lower α values [Ahlquist and Charlson, 1969; Smirnov et al., 2002], and the summer season is dominated by sub-μm aerosol particles at lower wind speed. It is also clear that the primary production of aerosols increases with increasing wind speed, which results in decreasing α value.

[12] As stated above, there is an approximate factor of two higher scattering coefficient and backscatter coefficient in the winter/LBA period compared to the summer/HBA period. In-between seasons, with moderate productivity, have scattering values between the LBA and HBA regimes (Figure 1a). In order to elucidate the differences between the absolute scattering coefficients as a function of wind speed, we evaluate the effect of refractive index and the difference in particle size distribution shape on aerosol scattering coefficient.

[13] σscat values for the particle diameter size range from 0.1–0.5 μm were calculated using the Mie code [Mätzler, 2002], based on Bohren and Huffman [1983]. Input parameters were refractive index and the lognormal fitted size distribution for the clean sector at 10 ms−1 wind speed for both HBA and LBA periods. Lognormal parameters were: dmodal = 0.15 μm, σ = 1.64, and N = 124.8 cm−3 for the LBA period and dmodal = 0.18 μm, σ = 1.27, and N = 127.5 cm−3 for the HBA period. The said size range is selected as the organic matter dominates in this size range during HBA and sea salt dominates during LBA [O'Dowd et al., 2004].

[14] Assuming the entire aerosol mass in the sub-μm range is comprised of sea salt in the LBA season, and taking the refractive index of sea salt as 1.5–i1.0e-08 [Kent et al., 1983], we get Σσscat value ∼16.7 Mm−1. For the HBA season, assuming that the organic matter dominates the 0.1–0.5 μm size range, and has a refractive index of 1.4 [Horvath, 1998], we get Σσscat value ∼8.7 Mm−1, which is approximately half that of the LBA season. This indicates that 96% of the scattering difference is explicable by the combined effect of refractive index and size distribution differences between the two seasons. The calculations indicate that, for the same size distribution, whether it be for the LBA or HBA period, refractive index accounts for ∼72% of the difference in scattering coefficient.

[15] Other parameters such as sea surface temperature (SST) are potentially important and may affect the production mechanism of sea spray aerosol particles [Lewis and Schwartz, 2004]. Jaeglé et al. [2011] reported that the ratio of observed to modeled sea salt mass concentration can differ by a factor of 2 at a SST difference of ∼30°C. For the North Atlantic SST range (∼7°), the Jaeglé et al. [2011] SST correction to the sea spray flux and size distribution is rather insensitive; however, small changes of the distribution resulting from temperature or organic surface layer effects cannot be ruled out, but are very difficult to quantify over the above size range.

5. Conclusions

[16] The σscat and σbscat dependency on U was quantified, for the winter/LBA and the summer/HBA seasons, for clean marine Northeast Atlantic Air masses and was found to be function of ∼U2. A high correlation coefficient (0.95) was found for both σscat and σbscat. αvalues indicates the presence of a dominant super-μm mode at all wind speeds during the winter season. In the summer season, αvalues are high at low wind speed and attain lower values as the production of the super-μm mode particles increases, i.e., at higher wind speed. σscat and σbscatvalues for the winter season are approximately twice those of the summer season. Mie modelling studies indicate that both size distribution and refractive index effects combined can account for 96% of the LBA-HBA scattering differences and that refractive index on its own can account for about 70% of the difference.


[17] The work described in this paper was supported by the EU FP6 Integrated Infrastructures Initiatives (I3) Project EUSAAR (European Supersites for Atmospheric Aerosol Research, project FP6-026140). We also acknowledge the support of the European Commission through the GEOmon (Global Earth Observation and Monitoring) Integrated Project under the 6th Framework Programme (contract FP6-2005-Global-4-036677). Ciaran Monahan is gratefully acknowledged for providing the size distribution data. Jurgita Ovadnevaite is also acknowledged for valuable discussion.

[18] The Editor thanks two anonymous reviewers for their assistance in evaluating this paper.