Primary and secondary marine organic aerosols over the North Atlantic Ocean during the MAP experiment
2011; American Geophysical Union; Volume: 116; Issue: D22 Linguagem: Inglês
10.1029/2011jd016204
ISSN2156-2202
AutoresStefano Decesari, E. Finessi, Matteo Rinaldi, Marco Paglione, S. Fuzzi, Euripides G. Stephanou, T. Tziaras, Apostolos Spyros, Darius Čeburnis, Colin O’Dowd, Manuel Dall’Osto, Roy M. Harrison, J. D. Allan, Hugh Coe, M. C. Facchini,
Tópico(s)Toxic Organic Pollutants Impact
ResumoJournal of Geophysical Research: AtmospheresVolume 116, Issue D22 Aerosol and CloudsFree Access Primary and secondary marine organic aerosols over the North Atlantic Ocean during the MAP experiment S. Decesari, S. Decesari [email protected] Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorE. Finessi, E. Finessi Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorM. Rinaldi, M. Rinaldi Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorM. Paglione, M. Paglione Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorS. Fuzzi, S. Fuzzi Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorE. G. Stephanou, E. G. Stephanou Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, Iraklio, GreeceSearch for more papers by this authorT. Tziaras, T. Tziaras Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, Iraklio, GreeceSearch for more papers by this authorA. Spyros, A. Spyros Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, Iraklio, GreeceSearch for more papers by this authorD. Ceburnis, D. Ceburnis School of Physics and Centre for Climate and Air Pollution Studies, Ryan Institute, National University of Ireland, Galway, IrelandSearch for more papers by this authorC. O'Dowd, C. O'Dowd School of Physics and Centre for Climate and Air Pollution Studies, Ryan Institute, National University of Ireland, Galway, IrelandSearch for more papers by this authorM. Dall'Osto, M. Dall'Osto National Centre for Atmospheric Science, School of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham, UK Now at Institute of Environmental Assessment and Water Research, Consejo Superior de Investigaciones Científicas, Barcelona, Spain.Search for more papers by this authorR. M. Harrison, R. M. Harrison National Centre for Atmospheric Science, School of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham, UKSearch for more papers by this authorJ. Allan, J. Allan School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester, UK National Centre for Atmospheric Science, University of Manchester, Manchester, UKSearch for more papers by this authorH. Coe, H. Coe School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester, UKSearch for more papers by this authorM. C. Facchini, M. C. Facchini Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this author S. Decesari, S. Decesari [email protected] Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorE. Finessi, E. Finessi Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorM. Rinaldi, M. Rinaldi Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorM. Paglione, M. Paglione Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorS. Fuzzi, S. Fuzzi Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this authorE. G. Stephanou, E. G. Stephanou Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, Iraklio, GreeceSearch for more papers by this authorT. Tziaras, T. Tziaras Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, Iraklio, GreeceSearch for more papers by this authorA. Spyros, A. Spyros Environmental Chemical Processes Laboratory, Department of Chemistry, University of Crete, Iraklio, GreeceSearch for more papers by this authorD. Ceburnis, D. Ceburnis School of Physics and Centre for Climate and Air Pollution Studies, Ryan Institute, National University of Ireland, Galway, IrelandSearch for more papers by this authorC. O'Dowd, C. O'Dowd School of Physics and Centre for Climate and Air Pollution Studies, Ryan Institute, National University of Ireland, Galway, IrelandSearch for more papers by this authorM. Dall'Osto, M. Dall'Osto National Centre for Atmospheric Science, School of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham, UK Now at Institute of Environmental Assessment and Water Research, Consejo Superior de Investigaciones Científicas, Barcelona, Spain.Search for more papers by this authorR. M. Harrison, R. M. Harrison National Centre for Atmospheric Science, School of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham, UKSearch for more papers by this authorJ. Allan, J. Allan School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester, UK National Centre for Atmospheric Science, University of Manchester, Manchester, UKSearch for more papers by this authorH. Coe, H. Coe School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester, UKSearch for more papers by this authorM. C. Facchini, M. C. Facchini Institute of Atmospheric and Climate Sciences, National Research Council of Italy, Bologna, ItalySearch for more papers by this author First published: 24 November 2011 https://doi.org/10.1029/2011JD016204Citations: 82AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Abstract [1] The organic chemical composition of atmospheric submicron particles in the marine boundary layer was characterized over the northeast Atlantic Ocean in summer 2006, during the season of phytoplankton blooms, in the frame of the Marine Aerosol Production (MAP) experiment. First measurements of water-insoluble organic carbon (WIOC) in marine aerosol particles by nuclear magnetic resonance (NMR) spectroscopy showed that it is structurally similar to lipids, resembling the organic fraction of sea spray formed during bubble-bursting experiments. The composition of the water-soluble organic carbon (WSOC) fraction was investigated by liquid chromatography – mass spectrometry and by 1D- and 2D-NMR spectroscopy, and showed a less hydrophilic fraction containing traces of fatty acids and rich of alkanoic acids formed by lipid degradation, and a more hydrophilic fraction, containing more functionalized species encompassing short-chain aliphatic acids and sulfate esters of hydroxyl-carboxylic acids. The more oxidized fraction of WSOC accounts for the oxidized organic aerosol components, which can form by either gas-to-particle conversion or extensive chemical aging of lipid-containing primary particles, as also suggested by the parallel measurements using online mass spectrometric techniques (presented in a companion paper), showing oxidized organic substances internally mixed with sea salt particles. These measurements are also compared with online measurements using an Aerosol Time-Of-Flight Mass Spectrometer (ATOFMS) and Aerodyne Aerosol Mass Spectrometer (AMS). Given the large variability in the chemical composition of marine organic aerosol particles, a multitechnique approach is recommended to reduce method-dependent categorizations and oversimplifications and to improve the comparability with the results obtained in different oceanic areas. Key Points Marine organic aerosol is a mixture of primary and secondary components Water-insoluble organics in marine aerosol are related to primary production Aging of primary organic aerosol is as important as gas-to-particle conversion 1. Introduction [2] Submicron marine aerosols are an important controlling factor in the atmospheric radiative budget by dictating the number of cloud condensation nuclei (CCN) in the marine atmosphere [O'Dowd and de Leeuw, 2007]. The chemical nature and origin of the fine particulate matter over oceanic regions, and especially of its organic fraction, are still largely unknown, because of the insufficient spatial and temporal coverage of in situ measurements; the large variability in the emission from the ocean resulting from changing biological activity in seawater; from the complex transfer of gases and particles at the air-sea interface; and finally due to the low concentrations, typically ≤1 μg/m3, which can be affected by the transport of aerosol particles from distant continental sources. Past studies on marine organic aerosols have focused on specific mechanisms of production, which are reviewed by the recent paper by Rinaldi et al. [2010], but quantitative methods for source apportionment of marine organic aerosols are still missing. Submicron particles can reside for weeks in the marine atmosphere, thus integrating the contributions from multiple secondary and primary sources of organic compounds in a complex manner. At the same time, aging of the particles in the atmosphere inevitably leads to the progressive loss of molecular markers and to their replacement by complex mixtures of oxidized compounds [e.g., McFiggans et al., 2005; Jimenez et al., 2009]. Investigating the nature of the oxidized fraction of particulate organic carbon (OC) by looking at integral chemical properties (mass fragmentation patterns, N:O:C ratio, functional group distribution, OC isotopic enrichment) of the unresolved mixtures of compounds is therefore a priority for supporting molecular tracer analysis with additional “distinguishing features” in elucidating the sources of marine organic particles [Fuzzi et al., 2006]. Techniques, such as aerosol mass spectrometry (AMS), Fourier transform infrared spectroscopy (FTIR) and nuclear magnetic resonance spectroscopy (NMR) have progressively gained importance in source identification and source apportionment of airborne organic particles [e.g., Sullivan and Prather, 2005; Hersey et al., 2009; Dall'Osto et al., 2006; Russell et al., 2010; Decesari et al., 2007]. The inherent complexity and variability of spectral data of ambient aerosol samples can be rationalized using techniques of data clustering and factor analysis. Factor analysis in particular, allows extracting correlated variables, providing simplified patterns, which can be used to identify sources, processing or transport characteristics of atmospheric particulate organic compounds. Positive matrix factorization (PMF), for instance, is widely used for aerosol mass spectrometric data analysis in many environments [Ulbrich et al., 2009], and it has also been applied to FTIR data for identification and quantification of marine phenols in airborne particles in the marine boundary layer [Bahadur et al., 2010]. [3] The “Marine Aerosol Production (MAP)” project provides an integrated assessment of the sources of MBL organic aerosols in the North Atlantic using molecular marker analysis and state-of-the-art online and off-line spectroscopic techniques. The experiment was conducted at the Mace Head research station with continuous measurements of the aerosol physical properties and semi-continuous measurements of the aerosol chemical properties throughout 2006 [Rinaldi et al., 2010]. Between June and July of the same year, a ship campaign was conducted during which sophisticated physical and chemical aerosol measurements were made in the marine boundary layer off the west coast of Ireland. During the cruise, online aerosol chemical measurements were performed using state-of-the-art mass spectrometric techniques (AMS, Aerodyne “Aerosol Mass Spectrometry,” and ATOFMS, “Aerosol Time-Of-Flight Mass Spectrometry”), while time-integrated filter samples were collected for off-line chemical analysis with analytical techniques showing a high recovery for the polar organic fraction of the aerosol and including liquid-chromatography-mass-spectrometry (LC/MS) and nuclear magnetic resonance spectroscopy (NMR). In a companion paper (M. Dall'Osto et al., Open ocean aerosol measurements during the MAP experiment, submitted to Journal of Geophysical Research, 2011) the results of the online mass spectrometric techniques are discussed thoroughly, while here we focus on the off-line methods and we also provide a synthesis of the main findings of the campaign. In parallel with ambient aerosol measurements, laboratory experiments were conducted onboard the research ship for simulating the production of sea spray aerosol, and the results, reported in detail by Facchini et al. [2008a], are quoted in this study with the aim of comparing the composition of ambient marine aerosols with that of marine primary organic aerosol (POA) representative for the northeastern North Atlantic region. 2. Experimental Setup Ship Campaign [4] Aerosol measurements were carried out on board the research ship Celtic Explorer (CE) from 12 to 20 June 2006 (1st leg) and from 25 June to 5 July 2006 (2nd leg) [Rinaldi et al., 2009]. The CE crossed oceanic regions rich of chlorophyll NW of Ireland between 14 and 18 June, and N of Ireland between 29 June and 3 July. In the same days, an increase of DOC and POC in seawater was also observed (auxiliary material Figure S1). Other, more distant blooms were identified by MODIS (http://oceancolor.gsfc.nasa.gov/) in the north oceanic basins (Denmark Strait, Greenland and Norwegian seas) or, at lower latitudes (45°–55° N), around midway between Europe and Canada [Rinaldi et al., 2009]. [5] Winds were more sustained during the first leg of the cruise, peaking on 20th June, when the CE encountered a summer storm. During the second leg, wind intensity was variable with some days with prevalently calm conditions (auxiliary material Figure S1). The analysis of back trajectories for the periods when the aerosol samplers were operative (see following section) shows that different types of maritime air masses were encountered during the campaign (auxiliary material Figure S2). During the first days of the first leg (12 – 15 June), an almost permanent cyclonic circulation brought polar air masses from the Denmark Strait and Greenland Sea toward the Irish coast. On 15 June some trajectories had a partial southwesterly component. In the following days (auxiliary material Figure S2b), the air mass origin was quite variables and the trajectories had alternatively a northerly or a southwesterly component. Finally, more stagnant conditions were encountered during the second leg (25 June – 4 July) with most of the air masses spending 48 h over an oceanic sector enclosed between 45° and 55°N and between −20° W and the Irish coast before reaching the research ship. In this period (auxiliary material Figure S2c), the air masses were characterized by a shorter range transport compared to the first days of the first leg (auxiliary material Figure S2a). Consequently, also the potential source regions for marine biogenic particles were probably different between these periods of the campaign. Further information about of air mass types and origin can be found in the parallel paper by Dall'Osto et al. (submitted manuscript, 2011). Aerosol Sampling and OC Analysis [6] Sampling at Mace Head during MAP for chemical analysis is discussed in the paper by Rinaldi et al. [2010]. Briefly, the station is equipped with an automated sectored sampling system which provided high volume (HiVol) filter samples for organic analysis differentiating between clean and polluted air masses, where the “clean” sector was defined by: wind direction from 180° to 300°, total particle number below 700 cm−3, and black carbon (BC) concentrations lower than 50 ng m−3. Onboard the Celtic Explorer, the collection of aerosol samples selectively in clean air masses was performed based on weather forecast and trajectory forecast plus active (based on CPC) though manual control. Sampling was also interrupted when the CE was oriented with the aerosol inlets downwind the funnels of the ship. Samples were collected using an 8-stage Berner impactor for the analysis of inorganic ions and of water-soluble organic carbon (WSOC) [Rinaldi et al., 2009] and in parallel using two Sierra Anderson HiVol samplers equipped with slotted impactors for size-segregated sampling (hereafter referred as HiVol1 and HiVol2) for detailed organic analysis (Table 1). Due to the necessity of collecting sufficient amounts of samples for the subsequent detailed chemical analyses, long (∼50 h) time-integrated samplings were performed. The filter samples, provided by the two HiVol systems, were processed for organic chemical characterization according to two distinct protocols (Figure 1). Figure 1Open in figure viewerPowerPoint Scheme of the off-line analysis of aerosol organic composition. The analytical protocols adopted for HiVol1 and HiVol2 samples are indicated by blue and red lines, respectively. Table 1. Sampling Times of HiVol1 and HiVol2 Filter Samplesa Sample Size Cut TSP TC OC EC WSOC MSA MBAS Leg 1 HiVol1 12–15/06 PM1.5 n.a. 0.24 n.a. n.a. 0.15 0.23 n.a. PM1.5–10 n.a. 0.12 n.a. n.a. 0.082 0.075 n.a. 16–20/06 PM1.5 n.a. 0.18 n.a. n.a. 0.10 0.11 n.a. PM1.5–10 n.a. 0.080 n.a. n.a. 0.062 0.050 n.a. HiVol2 12–15/06 TSP 23.59 0.50 0.40 0.10 0.64 n.a. 1.33 (0.38) 16–20/06 Size segregated >14.11 0.19 n.a. n.a. n.a. 0.02 n.a. 1.67 (0.48) 5.90–14.11 ND n.a. n.a. n.a. 0.19 n.a. 0.78 (0.23) 2.93–5.90 1.26 n.a. n.a. n.a. 1.60 n.a. 0.55 (0.16) 1.90–2.93 1.12 n.a. n.a. n.a. 0.34 n.a. ND 1.02–1.90 ND n.a. n.a. n.a. 0.12 n.a. 2.14 (0.62) <1.02 28.15 n.a. n.a. n.a. 0.83 n.a. 18.27 (5.27) Leg 2 HiVol1 26/06–04/07 PM1.5 n.a. 0.29 n.a. n.a. 0.20 0.16 n.a. PM1.5–10 n.a. 0.16 n.a. n.a. 0.11 0.12 n.a. HiVol2 26/06–02/07 TSP 16.76 0.37 0.36 0.01 0.19 n.a. 1.84 (0.53) 02–05/07 TSP 16.92 0.44 0.44 ND 0.21 n.a. 1.40 (0.40) a Total carbon (TC), organic carbon (OC), elemental carbon (EC), water-soluble organic carbon (WSOC), methansulphonate (MSA), methylene-blue-active substances (MBAS) concentrations. All units are μg m−3 except for MBAS that is expressed both in pmol m−3 and ng m−3 (in brackets). Abbreviations: n.a. = not analyzed; ND = not detected. 2.2.1. HiVol 1 [7] The WSOC and TC analyses were performed using a Multi N/C 2100 elemental analyzer (Analytik Jena, Germany), equipped with a furnace solids module. Water-soluble organic compounds were extracted from fine particles with ultrapure Milli-Q water by sonication for 1 h. Generally 100 mL of Milli-Q water were used for one half filter. Water extracts were filtered on quartz-fiber filters (pore size 0.45 μm) using a Büchner flask in order to remove suspended particles. After filtration, the water extract was analyzed by ion chromatography [Cavalli et al., 2004] for the speciation of low-molecular weight organic acids and of inorganic ions, and by proton-Nuclear Magnetic Resonance (1H-NMR) spectroscopy for WSOC functional group analysis. After the extraction with water, the filters were consequently extracted by trifluoroacetic acid (TFA), a solvent suitable for dissolving less polar substances and for hydrolyzing and solubilizing polymeric water-insoluble oxygenated compounds (such as polysaccharides and organo-silicon compounds), and proved to be effective for marine macromolecular biological material [Kovac et al., 2002]. About 30 mL TFA was used for each filter and a sonication time of 30 min was chosen. The TFA solution was first separated from the filter quartz pulp using a concentrator (Eppendorf 5301) and prepared for 1H-NMR analysis of water-insoluble organic carbon (WIOC). 2.2.2. HiVol 2 [8] The organic carbon (OC) and elemental carbon (EC) content of the samples was determined using the OC/EC analyzer of Sunset Laboratory Inc. and quantified straight from a small piece (1,5 cm2) of the quartz microfiber filters. Water-soluble organic carbon (WSOC) has been measured with a Shimadzu TOC VCSH instrument. The filters were then cut into small pieces and extracted with 120–150 ml of nanopure water by the use of the accelerated solvent extraction method with a DIONEX (ASE 300) device. The oven temperature was set at 100°C. Twenty ml of the solution was used for surface activity determination with methylene blue reagent according to Latif and Brimblecombe [2004]. The concentration of anionic aerosol surfactants was determined as methylene blue active substances (MBAS). OASIS HLB extraction cartridges (60 mg) were used for the isolation of the water-soluble organic fraction from the extract, which was first acidified to pH 2 by the addition of HCl. The retained compounds were eluted from the column with 4 ml methanol, 2 ml are used for LC/MS analysis and the rest 2 ml was evaporated to dryness by means of a nitrogen stream and kept for NMR analysis. Speciation of Aerosol Organic Compounds: Liquid Chromatography–Mass Spectrometry (LC/MS) [9] LC/MS analysis of HiVol2 samples was carried out with Thermo Finnigan's TSQ Quantum, equipped with a triple quadrupole for MS2 analysis. The HPLC conditions were: Injection loop 50 μL; column C8, 250 × 4.6 mm, 5 μm (MZ-PerfectSil 300 C8); methanol isocratic flow 500 μL/min. The analysis time was 30 min, although all compounds have been eluted by the first 8 min. Electro Spray (ES) ionization in the negative mode was used for the analysis of all samples: Spray Voltage 4.5 kV, Sheath Gas Pressure 50 psi, Aux Valve Flow 25 psi, Capillary Temperature 310°C, Source CID 0 eV, Tube Lens Offset −143 V, Lens 0 Offset 1.3 V. MS2-CID spectra were measured with collision energies varying from 10 up to 30 eV. Additionally, with the field samples blank filters were also analyzed following the same analytical protocol to determine possible contamination during the transport, storage or laboratory. Between consecutive analyses the LC/MS system was sufficiently rinsed with methanol. Blank runs were performed to exclude memory effects between samples. For comparison purpose MS2-CID spectra of authentic standards of linear alkylbenzene sulfonates (supplied by Lamda Detergents, Athens, Greece) and carboxylic acids (tetradecanoic acid and heptadecanoic acid puriss. supplied by Fluka, Germany) were also measured. Speciation of Aerosol Organic Compounds: Nuclear Magnetic Resonance (NMR) Spectroscopy [10] 1H-NMR spectroscopy was exploited for functional group characterization of WSOC and WIOC. HiVol1 samples for NMR analysis of WSOC were prepared by freeze-drying an aliquot of the water-extracts in a rotary-evaporation and then re-dissolving it with 650 μL of deuterated water (D2O). 50 μL D2O solution containing 0.05% by weight of sodium 3-(trimethylsilyl)-2,2,3,3-d4-propionate (TSP-d4) were added as internal reference standard. 1H-NMR spectra were acquired with a Varian Mercury 400 spectrometer in a 5mm probe. Presaturation of HDO (mono-deuterated water) was performed, nevertheless residual signals of HDO can interfere in the region between 4.5 and 6.0 ppm [Decesari et al., 2007]. The TFA extract containing WIOC material was also freeze-dried by rotary-evaporation. Finally 650 μL of deuterated TFA solution (TFA-d) with TSP-d4 as internal reference standard were use to re-dissolve the WIOC for 1H-NMR spectroscopy. [11] The dried water-extracts of the HiVol2 samples after the chemical workup, described in 2.2.2, were redissolved in a solution of 120 μL D2O containing 3.22 mM trimesic acid (TMA). TMA was used as the internal standard for quantitative analysis, and it was calibrated via a sample of levoglucosan of known concentration.1H-NMR spectra were acquired on a Bruker AMX-500 spectrometer using a 5mm probe. Gradient homonuclear 1H-1H COSY 2D NMR spectra were acquired using a modified water suppression pulse sequence and used for further functional group analysis of the samples. Factor Analysis of NMR Spectra [12] The 1D 1H-NMR spectra of WSOC extracted from the three HiVol1 PM1.5 samples collected onboard the CE and the ten samples collected at Mace Head during MAP were aggregated to other eight spectra from PM1.5 samples collected during past experiments at the same station (Table 2), and the resulting spectral data set was subjected to cluster analysis and factor analysis aiming to summarize and represent concisely the information contained in the spectra and its variability between samples. The algorithms were applied to blank-subtracted spectra and after binning to 200 or 400 points. The spectral regions containing only sparse signals (δH < 0.5 ppm; 4.5 < δH < 6.5 ppm; and δH > 8.5 ppm) were omitted from the data set. Table 2. Synopsis of the Aerosol Samples Providing NMR Data for Cluster and Factor Analysisa Sample Identification Start Date Stop Date MH Sector Na+ nssSO42− NO3− TC WSOC BC Past Mace Head campaigns MH01 05/03/02 12/03/02 marine 0.27 0.28 0.009 NA NA MH02 19/03/02 26/03/02 marine 0.14 0.15 0.016 NA 0.063 MH03 02/04/02 09/04/02 marine 0.31 0.35 0.076 NA 0.23 NA MH04 02/10/02 09/10/02 marine 0.71 0.27 0.036 NA NA 18.3 MH05 23/10/02 30/10/02 marine 5.7 NA <DL 0.23 0.039 20.7 MH06 31/05/02 06/06/02 modified marine 0.18 0.99 0.044 NA 0.29 96.0 MH07 31/05/02 06/06/02 modified marine 0.19 0.30 0.018 0.98 0.17 70.0 MH08 18/08/04 29/08/04 marine 0.064 0.35 0.005 0.19 0.11 26.5 MAP campaign at MH MH09 11/01/06 18/01/06 marine 0.25 0.27 0.009 0.37 0.072 15.0 MH10 10/02/06 20/02/06 marine 0.27 0.11 0.016 0.13 0.071 212 MH11 12/04/06 26/04/06 marine 0.11 0.61 0.007 0.17 0.11 34.9 MH12 03/05/06 10/05/06 continental 0.075 0.37 0.052 0.60 0.36 286 MH13 12/06/06 19/06/06 marine 0.055 0.47 <DL 0.26 0.18 30.7 MH14 19/06/06 28/06/06 marine 0.024 0.31 0.005 0.18 0.12 82.9 MH15 28/06/06 05/07/06 marine 0.032 0.41 <DL 0.32 0.25 55.0 MH16 05/07/06 12/07/06 marine NA NA NA 0.35 0.25 41.9 MH17 05/10/06 11/10/06 marine 0.16 0.17 0.008 0.071 0.051 9.7 MH18 11/10/06 18/10/06 continental 0.062 2.2 1.4 1.57 0.68 761 Celtic Explorer samples CE01 12/06/06 15/06/06 0.071 0.60 <DL 0.35 0.23 CE02 16/06/06 20/06/06 0.023 0.27 <DL 0.26 0.14 CE03 26/06/06 04/07/06 0.048 0.48 0.004 0.43 0.29 a Concentrations are in μg/m3 except for BC (from aethalometer) in ng/m3. Dates are given as dd/mm/yy. [13] Cluster analysis and principle component analysis (PCA) were used to analyze the similarity between spectra and to identify main sources of variability in the shape of the spectra. To this aim, spectra were normalized for their total intensity before analysis. Hierarchical cluster analysis based on Pearson distances between samples provided a categorization scheme, which was consistent with the subjective groupings extracted from PCA score plots. [14] Common factor analysis was performed using PMF (“Positive Matrix Factorization,” EPA v3.0), NMF (“Non-negative Matrix Factorization”) and MCR (“Multivariate Curve Resolution”). Factor analysis was performed on spectra normalized to total NMR-detectable WSOC concentrations (μmol H m−3). PMF [Paatero and Tapper, 1994] is by far the most widespread tool for AMS spectral data analysis and it is here applied exploratively to aerosol NMR spectral data. The uncertainties of the measurements were estimated for the 200- (or 400-) points binned spectra from the NMR detection limit calculated as three times the baseline noise (peak-to-peak noise in the region 6.0 – 6.5 ppm). NMF and MCR comprise the most common NMR spectral unmixing techniques in many chemometric applications [Viant et al., 2009]. Two different algorithms were used for NMF, employing a projected gradient bound-constrained optimization [Lin, 2007], or a multiplicative update approach [Lee and Seung, 2001]. MCR was run according to two different approaches: the alternating least square (MCR-ALS [Tauler, 1995; Jaumot et al., 2005]) and weighted alternating least square (MCR-WALS [Wentzell et al., 2006]) methods. 3. Results Main Carbonaceous Species and Surfactants [15] Table 1 reports the concentrations of the main carbonaceous fractions (TC, OC, EC, WSOC) and of methanesulphonate (MSA) during the ship campaign. Concentration ranges are comparable to those found in other oceanic regions in the northern hemisphere during the season of high biological productivity [Miyazaki et al., 2010]. MSA always accounted for a significant fraction of the water-soluble organic carbon, both in the submicron and in the supermicron fractions (14 ± 5% and 12 ± 2%, respectively), with a maximum (19%) during the first sampling period (12 to 15 June). The average WSOC fraction of TC was 63%, well above the water-soluble fraction of POA (ca. 10%) formed by bubble bursting experiments onboard the CE [Facchini et al., 2008a]. The enrichment of WSOC in ambient aerosol particles can be explained by condensation of secondary organic aerosol (SOA) compounds, or by the progressive oxidation of POA over time scales not amenable by standard bubble bursting experiments. Aged material and gas-to-particle conversion products
Referência(s)