The aerosol-driven radiative effects on marine low-level cloud represent a large uncertainty in climate simulations, in particular over the Southern Ocean, which is also an important region for sea spray aerosol production. Observations of sea spray aerosol organic enrichment and the resulting impact on water uptake over the remote Southern Hemisphere are scarce, and therefore the region is under-represented in existing parameterisations. The Surface Ocean Aerosol Production (SOAP) voyage was a 23 d voyage which sampled three phytoplankton blooms in the highly productive water of the Chatham Rise, east of New Zealand. In this study we examined the enrichment of organics to nascent sea spray aerosol and the modifications to sea spray aerosol water uptake using in situ chamber measurements of seawater samples taken during the SOAP voyage.
Primary marine organics contributed up to 23 % of the sea spray mass for particles with diameter less than approximately 1
Nascent 50 nm diameter sea spray aerosol hygroscopic growth factors measured at 90 % relative humidity averaged
Aerosol–cloud interactions represent a large uncertainty in modelled radiative forcing
The inorganic composition of SSA has largely been assumed to mirror the inorganic composition of seawater
Chamber observations of nascent SSA universally indicate the presence of an internally mixed organic component, with the organic contribution varying between studies from approximately 4 % to 80 % by volume, dependent on the method used and size range investigated
The organic fraction of sub-200 nm diameter SSA appears to be comprised of a volatilisable component which evaporates at approximately 150–200
The SSA organic fraction has been associated with the phytoplankton decline due to bacterial grazing and viral infection, which releases fatty acids and polysaccharides predisposed to SSA enrichment
Empirical relationships between the nascent PM
Studies using nascent SSA generation chambers have largely indicated that the presence of primary organics suppresses sub-200 nm diameter SSA hygroscopic growth factors (HGFs) by 4 %–17 % relative to sea salt
Measurements of nascent SSA composition and water uptake taken during the Surface Ocean Aerosol Production (SOAP) research voyage
The SOAP voyage examined air–sea interactions over the biologically productive frontal waters of the Chatham Rise, east of New Zealand in February and March 2012. The Chatham Rise couples pristine marine air masses with high biological activity due to the mixing of warm subtropical water and cool Southern Ocean waters. Subtropical waters have relatively low macronutrient levels, while Southern Ocean waters are depleted in iron but not in macronutrients
Voyage map for SOAP study, coloured by bloom periods.
Seawater samples were collected throughout the voyage for the purpose of generating nascent SSA. Seawater was primarily collected from the ocean surface (approximately 10 cm depth) during workboat operations at a distance from the R/V
Nascent SSA was generated in situ in a 0.45
Experimental schematic of nascent SSA chamber experiments during SOAP voyage. The VH-TDMA (grey) contains an RH-controlled region (blue) used for water uptake measurements.
Size distributions and number concentrations of 10 to 300 nm diameter SSA were measured using a TSI 3080 scanning mobility particle sizer (SMPS), coupled with a 3071 differential mobility analyser and a 3010 condensation particle counter (CPC) (TSI, Shoreview, MN), with an aerosol sample flow rate of 1
Size distributions, volatility and HGFs have also been measured for laboratory sea salt samples, which were generated using the same glass filters, chamber and sample conditioning. A comparison of the sea salt and sea spray volatility (Fig. S3 in the Supplement) was used to calculate the 50 nm SSA organic volume fraction. In addition laboratory sea salt transmission electron microscopy (TEM) samples were collected using a TSI 3089 nanometre aerosol sampler and analysed using X-ray dispersive spectrometry (TEM-EDX). TEM data were collected on a JEOL2100 transmission electron microscope operating at 200 kV coupled with a Gatan high-angle annular dark field (HAADF) detector. TEM images were used to compare the morphology of heated (250
The ultrafine organic tandem differential mobility analyser
SSA generated from 23 ocean water samples was collected on filters for further compositional analysis using transmission Fourier transform infrared (FTIR) and ion beam analysis (IBA). SSA was sampled through a 1
A large number of ocean water measurements were taken, characterising the physical properties, the nutrient concentration, the phytoplankton population, the bloom productivity and the concentration of molecular classes important for SSA, e.g. fatty acids, proteins and carbohydrates. A detailed list of ocean water measurements undertaken during the SOAP voyage is contained in
Nascent SSA size distributions for each water sample were averaged and normalised to their maximum value. Non-linear least square fits of up to four log-normal modes were fitted to each distribution with a random selection of initial values for the geometric mean and standard deviation, constrained to 10–320 nm and less than 2, respectively. The most appropriate fit was determined using the Bayesian information criterion, which is a measure of the error in reconstructing the measured size distribution that applies a penalty based on the number of parameters used and therefore avoids overfitting
All VH-TDMA data were inverted using the TDMAinv algorithm
In this study SSA organic volume fractions were calculated using volatility measurements, by accounting for the presence of sea salt hydrates. The volatility due to hydrates was used as a proxy for the proportion of inorganic sea salt in the natural seawater samples, which in turn provided the proportion of organics. The volume of hydrates is assumed to be a stable proportion of the sea salt volume, and there is assumed to be no contribution to the hydrates from SSA organics. As the organic fraction of internally mixed SSA increases, the sea salt fraction decreases and the hydrate fraction decreases in proportion to the sea salt as illustrated in Fig. S2. The sea salt fraction was computed by comparing natural sea spray volatility profiles and laboratory sea salt volatility profiles. In this study it was assumed that volatility was due to the evaporation of hydrates over the temperature range 200–400
The organic volume fraction was inferred from volatility measurements using the linear model outlined in Eq. (
Laboratory sea salt volatility profiles were measured using three different sea salt samples, a commercially available sea salt (Pro Reef Sea Salt, Tropic Marin, Wartenberg, Germany) and two mixtures of laboratory grade salts, one mimicking Sigma-Aldrich sea salts composition and one mimicking the
The method used to compute the organic volume fraction implicitly assumes that the proportion of hydrates in the sea salt component of SSA is constant; however, observations have shown variability in inorganic composition of SSA can vary
The volume fraction of semi-volatile organics in the nascent SSA generated from natural seawater, which evaporate at temperatures less than 200
The HGF was computed using Eq. (
The PM
Enrichment factors for inorganic elements were calculated with respect to the laboratory-prepared seawater and presented with respect to
The OCEANFILMS model was implemented for the surface and mixed layer nascent SSA experiments with measured water parameters used to represent bulk seawater molecular classes. Lipids were assumed to be equal to the total concentration of fatty acids, the total high-molecular-weight proteins were used to represent the protein molecular class and high-molecular-weight reducing sugars were used to represent polysaccharides. Note that these measurements were not microlayer measurements; the seawater samples were collected via CTDs or on workboats. Missing water composition data were filled using the relationships outlined in
The ZSR approximation was used to compute the ambient temperature nascent SSA HGF, defined as the average of all measurements less than 45
As a counterpoint to the ZSR assumption, which assumes the organic component is dissolved into the bulk, the compressed film model
Chl
Characterisation of biological activity for water samples used to generate SSA. Note that this is a selected subset of all water parameters. Panels show the concentration of Chl
Bloom 1 was dominated by dinoflagellates and displayed the highest average Chl
Bloom 2 was characterised as a coccolithophore bloom
The measured size distributions were broken up into four log-normal modes characterised by geometric mean diameters ranging from 33 to 320 nm, as seen in Fig.
Nascent SSA size distributions from laboratory sea salt measurements
The shape of the nascent SSA size distribution was broadly similar to nascent SSA size distributions observed in previous studies which also used sintered glass filters but shifted to slightly larger mean diameters. For example
Nascent SSA log-normal parameters.
Volatility measurements using the VH-TDMA indicated that the SSA volatile organic fraction made up a relatively consistent proportion of the 50 nm SSA, with a OVF
Summary of nascent SSA properties. Chl
Figure
Comparison of the 50 nm organic volume fraction calculated from volatility measurements (using VH-TDMA) and the PM
Correlations of both the semi-volatile organic volume fraction (OVF
The mass fraction of inorganic species in SSA during SOAP was observed to vary from that of salts in seawater; in particular an enrichment factor of
Inorganic mass fraction
Alcohol functional groups contributed
Ethanol growth factors measured using the UFO-TDMA for preselected 50 nm diameter SSA were
The HGFs observed for SSA generated from both laboratory sea salt and natural seawater samples showed up to three externally mixed HGF modes (Fig. S6). The first natural seawater SSA HGF mode averaged
Ambient HGF
The shape-corrected 50 nm ambient nascent SSA HGF averaged
The measured HGF of 50 nm diameter SSA at ambient temperature
A buffered response of SSA hygroscopicity under supersaturated conditions to OVF has been previously reported
The 50 nm deliquescence relative humidity was measured for the Workboat 9 seawater sample at 69 % RH (Fig.
Deliquescence of 50 nm SSA generated from the SOAP seawater sample (Workboat 9) shown in yellow to blue colour scale. The laboratory sea salt deliquescence curve is shown with circles.
The organic enrichment of SSA was examined using the Chl
PM
The organic macromolecular classes associated with the SSA were determined from the functional group composition using the conversions outlined in
Measured organic composition inferred from functional groups
Over-prediction of alkane-to-hydroxyl ratios for particularly clean marine measurements is a known issue for OCEANFILMS-2, and broadening the model to different saccharides with varying molecular weights has been identified for future research
OCEANFILMS does improve the prediction of organic enrichment from seawater parameters, relative to Chl
The sea spray hydrate volume fraction ranged from 0.05 to 0.16 and is associated with the inorganic sea salt component of SSA. The hydrates make up a larger proportion of the SSA at low OVFs. The inclusion of the hydrates in the ZSR model lowered the HGF by approximately 0.15 at low OVFs and by approximately 0.05 at high OVFs, and it therefore lowers the dependence of SSA HGF on OVF. Reasonable agreement between the modelled and measured HGFs at low OVFs is reached with the ZSR model including hydrates; however, there are still large discrepancies at high OVFs as shown in Fig.
The compressed film model was run to examine the potential impact of organic surface partitioning on SSA water uptake. The proportion of SSA organics at the particle surface was tested assuming that partitioning occurs on the basis of the organic molecular classes as computed from the distribution of functional groups or from the OCEANFILMS-2 model. Four cases were tested assuming the following molecular classes partitioned to the surface:
lipids, lipids and proteins, lipids and polysaccharides, all organics.
The resulting error in the HGF given by the compressed film model is shown in Figs.
Ambient HGF measured minus compressed-film-modelled HGF for lipids partitioned to the surface
A comparison of the HGFs modelled using the compressed film model and those modelled using the ZSR assumption is shown in Fig.
Ambient HGF modelled using the compressed film model (green) as a function of OVF, modelled using the ZSR assumption (light blue) and observed (dark blue). Compressed film model output is for the case where lipid and protein fraction (computed from functional group measurements) can partition to the particle surface. The ZSR model used an organic HGF of 1.15.
The discrepancy between observed and modelled HGFs was particularly apparent for samples with OVFs collected during bloom 1, the most productive phytoplankton bloom. The drivers of the HGF discrepancy are unknown but could potentially relate to short-lived seawater components associated with strong phytoplankton blooms and the interaction between these components. For example, mixtures of surface-active membranes and monolayers (DPPC in particular) with inorganic
Chamber measurements of primary marine aerosol generated from 23 seawater samples collected across three phytoplankton blooms tracked during the 23 d SOAP voyage over the Chatham Rise (east of New Zealand) are examined in this study. The SSA was an internal mixture of sea salt and organics. Volatility measurements at a preselected particle mobility diameter of 50 nm indicated that the SSA had an organic volume fraction of up to 0.79, with an average of
The SSA organic fraction displayed a scattered correlation with chlorophyll
Water uptake measurements revealed that the SSA hygroscopicity was largely invariable with the organic mass fraction, with HGFs averaging
The SSA organics showed consistently low alkane-to-hydroxyl ratios, even in relatively productive waters with high SSA organic fractions, and surface tension effects. These results could indicate that the source and structure of the SSA organics was largely consistent throughout the voyage, for example made up of lipopolysaccharides, which have previously been identified as an important component in primary marine aerosols. These measurements provide valuable comparison with observations for models of SSA organic enrichment and water uptake. Constraints on emissions and process models for this region are of particular importance because existing measurements are sparse and it is a region for which SSA has been observed to make up a large contribution to CCN.
The data are available through the World Data Centre PANGAEA. VH-TDMA data are available at
The supplement related to this article is available online at:
The SOAP campaign was led and coordinated by CSL and MJH. CSL led the ocean biogeochemistry work programme and MJH led the atmospheric work programme. LTC, MDM, PV, MJH and CSL made measurements during the SOAP voyage. GO developed fatty acids and alkane techniques and analysed samples. KS collected workboat samples, conducted Chl
The authors declare that they have no conflict of interest.
This article is part of the special issue “Surface Ocean Aerosol Production (SOAP) (ACP/OS inter-journal SI)”. It is not associated with a conference.
We acknowledge the invaluable assistance of the captain, officers and crew of the R/V
This research has been supported by the Australian Institute of Nuclear Science and Engineering (grant no. ALNGRA13048), the Australian Research Council (grant no. DP150101649), the National Institute of Water and Atmospheric Research (Climate and Atmosphere Research Programme 3 – Role of the oceans, grant no. 2015/16 SCI), the National Science Foundation (grant no. AGS-1013423), the Academy of Finland (grant no. 136841), and the European Cooperation in Science and Technology (Action 735).
This paper was edited by Markus Petters and reviewed by Christopher Oxford and one anonymous referee.