On the occurrence of open ocean particle production and growth events


  • Colin O'Dowd,

    1. School of Physics, National University of Ireland Galway, Galway, Ireland
    2. Centre for Climate and Air Pollution Studies, Environmental Change Institute, National University of Ireland Galway, Galway, Ireland
    Search for more papers by this author
  • Ciaran Monahan,

    1. School of Physics, National University of Ireland Galway, Galway, Ireland
    2. Centre for Climate and Air Pollution Studies, Environmental Change Institute, National University of Ireland Galway, Galway, Ireland
    Search for more papers by this author
  • Manuel Dall'Osto

    1. School of Physics, National University of Ireland Galway, Galway, Ireland
    2. Centre for Climate and Air Pollution Studies, Environmental Change Institute, National University of Ireland Galway, Galway, Ireland
    Search for more papers by this author


[1] We present new results from the Mace Head coastal station illustrating that open ocean new particle production and growth events occur frequently during periods of high oceanic productivity over the N.E. Atlantic. For the first time, we report events during which a recently-formed nucleation mode (∼15 nm diameter) is detected and is observed to grow into an Aitken mode (∼50 nm diameter) over periods up to 48 hours. A growth rate of 0.8 nm hour−1 is estimated in a typical case study, pointing to a source region ∼700 km off-shore. The duration of the growth also suggests the particle production events are occurring over large spatial scales. Analysis of seven years of data show occurrence of extended growth events (lasting at least 24 hours) from March to September, with a peak occurrence in May. The events suggest that secondary marine boundary layer aerosol formation contributes to the marine aerosol population.

1. Introduction

[2] Nucleation and subsequent particle production and growth in the marine boundary layer is hypothesized to provide an important source of secondary aerosol which have the potential to increase the marine cloud condensation (CCN) population leading to an important role in climate change [Charlson et al., 1987]. Charlson et al.'s [1987] hypothesis linked plankton activity to new non-sea-salt sulphate aerosol and CCN formation, leading to a negative feedback on climate change. Other studies [e.g., O'Dowd et al., 2002] suggested that condensable iodine vapours may be involved in open ocean particle production. However, despite extensive studies into marine new particle production and various hypothesized species involved, very few studies have reported such events, suggesting that such events occur only under particular or exceptional conditions. One notable event was reported by Clarke et al. [1998] over the Pacific tropical region, and other events have been reported in Antarctic waters [O'Dowd et al., 1997]. The aforementioned studies pointed to marine boundary layer particle production while the event reported by Covert et al. [1996] was attributed to entrainment of free tropospheric air. Other studies [e.g., Hoppel et al., 1994] report new particle production following intense precipitation events and associated aerosol removal.

[3] Modelling studies [Pirjola et al., 2000] evaluated the role of sulphuric acid in marine nucleation and particle production (operationally defined as the production of particles with diameter >3 nm) and found that while sulphuric acid was able to nucleate a significant amount of stable clusters of the order of 1 nm, there was rarely enough sulphuric acid to lead to particle production. Pechtl et al. [2006] modeled OIO particle formation in coastal hotspots while Vuollekoski et al. [2009] further evaluated marine new particle production, examining iodine oxide self nucleation and activation of iodine oxide by sulphuric acid vapours. Vuollekoski et al. [2009] found that for exposure to patchy iodine emitting plankton blooms for up to 30 minutes (corresponding to advection spatial scales of ∼36 km) led to significant nucleation mode particle production; however, they found that iodine oxides alone were unlikely to lead to notable production of particles at CCN sizes. The addition of a low volatility organic vapour was required to produce larger particles. The combined modeling studies promote the concept of a multi-stage process by which marine new particle formation occurs: i.e., either iodine oxide or sulphuric acid, or a combination thereof, is required to enable the nucleation of stable clusters and enable initial cluster growth while low volatility vapours are required to grow the newly-formed particles into Aitken mode (20–80 nm diameter) and accumulation mode particles (80–200 nm diameter), noting that in-cloud aqueous phase oxidation of compounds like SO2 can also grow Aitken mode particles into the accumulation mode [Hoppel et al., 1994]. Mahajan et al. [2010] also estimated that IO concentrations of the order of 8–10 ppt could explain the formation of 103 cm−3 particles of 20 nm diameter size.

[4] New particle production has been most extensively studied in North East Atlantic marine air masses arriving at the Mace Head Atmospheric Research Station. The clean marine air events studied, however, have been exclusively coastal, or tidal-driven events produced as macro algae are exposed to the atmosphere, thus releasing high iodocompound emissions [O'Dowd et al., 1998, 2002]. These events are typically characterized by intense bursts of new particles in the 3–6 nm size range, with concentrations often reaching in excess of 1,000,000 cm−3. This study analyses and probes the Mace Head aerosol size distribution dataset for occurrence of open ocean events as opposed to coastal events.

2. Experiment

[5] The Mace Head coastal Atmospheric Research Station [Jennings et al., 2003; O'Connor et al., 2008] is located on the west coast of Ireland and at the edge of the N.E. Atlantic. The station is subject to both marine and continental air masses [Dall'Osto et al., 2010], with clean marine air prevailing approximately 50% of the time. The marine air masses can be of tropical, polar or Arctic origin, bringing with it different aerosol characteristics [Dall'Osto et al., 2010]. Size distributions were taken using a nano-Scanning Mobility Particle Sizer (n-SMPS) over a 3–20 nm diameter size range and a standard SMPS [Wang and Flagan, 1990] operating over a size range of 20–500 nm diameter. Both instruments sampled dry aerosol sizes. Aerosol sampling was conducted from the community aerosol inlets sampling from a height of 8 m above ground level approximately 100 m from the coast line and has been reported previously [O'Dowd et al., 2002]. Clean air particle production events are defined as those occurring in marine air masses with black carbon mass less than 50 ng m−3 [e.g., Cavalli et al., 2004] and when the number concentration of nucleation mode particles between 3 and 20 nm (N3–20) exceeds 100 cm−3 for an hour average period. Extended growth events are those which last more than 24 hours.

3. Results

[6] A 12 day period during August 2009 is selected to illustrate the occurrence and detection of open ocean particle production and growth events occurring in the N.E. Atlantic marine boundary layer. Figure 1 displays the hour-average number concentration of N3–20 measured during the selected period in clean air. During the selected period, total particle background concentrations are of the order of 300–400 cm−3 representing very clean Atlantic air and even reached below 100 cm−3 briefly at times; however, during particle production events N3–20 concentrations greater than 100 cm−3 over an hour period are encountered and represent particle production events.

Figure 1.

(top) N3–20 concentration in clean marine air at Mace Head, illustrating clean nucleation mode event occurrence. Combined nSMPS and standard SMPS-derived aerosol size distributions (3-500 nm diameter) for a 12 day period in August 2009, corresponding to selected aerosol growth event on (middle) JD 236–238 and (bottom) JD 240–242.

[7] Some of these events are local tidal−related coastal aerosol production events [O'Dowd et al., 2002] while others are classified as open ocean events. The typical coastal nucleation events are seen to occur on Julian Days 234.5, 238.6 and 240.7. These are classified as local or coastal tidal-related production events because the peak in the nucleation mode is at 3–4 nm and no growth to larger sizes (i.e., into the Aitken mode) is observed, reflecting a local line-source plume rather that an event occurring over large spatial scales. By way of contrast, the open ocean events are linked to the more aged or more developed nucleation mode observed at sizes between 10 nm and 20 nm. The attribution of the 10–20 nm nucleation mode to open ocean production is simply based on the unrealistic growth rates of >104 nm hour−1 required to grow particles from 1 nm to 20 nm in a matter of seconds (i.e., the transit time from the shore/tidal region to the sampling point).

[8] Over the selected 12 day period, 10 such events are observed and four of these events are observed to grow over periods in excess of 24 hours. Two such cases are extracted for closer examination and are shown also in Figure 1.

[9] Figure 1 (middle) illustrates a growth event over 48 hours, from JD 236.0 to JD 238. The nucleation mode of ∼15 nm is detected JD 236.0 and continues to grow to ∼50 nm by JD 238.0. It is highlighted that at the end of the presented growth event on JD 238, a younger nucleation mode has appeared, reflecting a second open ocean production event. The growth rate over this period is ∼0.8 nm hour−1. Assuming a constant growth rate, and working backwardly from the initially detected nucleation mode diameter, and taking a wind speed of 10 m s−1, the source region for nucleation clusters of 1 nm is of the order of 675 km off-shore. A similar conclusion can be arrived at for the second growth case over a 48 hour period presented in Figure 1 (bottom) for the period JD 240–242. In the latter case also, a younger nucleation plume emerges on JD 241.1, suggesting that during this period there are multiple large scale nucleation and growth events in the region during the NE Atlantic mid-latitude summer. The growth, or evolution, observed implies that the production events are occurring over large spatial scales. Given that growth from ∼15 nm to ∼50 nms is observed over a 48 hour period, this suggests new particle production events occurring over spatial scales of the order of 1500–1700 km.

[10] For the case on JD 236, size distributions are extracted for three 2.4 hour periods at different stages of the plume evolution, namely for JD periods JD236.1–236.2, JD237.5–237.6, and JD237.9–238 and are shown in Figure 2. At the start of the plume evolution (JD 236.1–236.2), a clear nucleation mode is seen at 15 nm and overshadows the pre-existing Aitken mode, the shoulder of which is visible at sizes from 20–40 nm. The nucleation mode concentration is 3328 cm−3 when first detected. By JD 237.5–237.6 the mode has grown to 29 nm and by JD 237.9–238, the mode has reached 49 nm. By the latter stage of plume evolution, the accumulation mode has also increased in concentration from 150 cm−3 to 425 cm−3 and a new nucleation mode has appeared at 10–15 nm, associated with a second large scale open ocean production event.

Figure 2.

Averaged (2.4 hour) size spectra the growing plume on JD 236–237. 236.1–236.2, 237.5–237.6, and 237.9–238.0.

[11] Figure 3 illustrates a five day back trajectory from Mace Head at noon on JD 236. The trajectories are overlaid on satellite-derived chlorophyll-a concentrations in the surface waters. The trajectories indicate a polar maritime air mass advecting from the NW Atlantic to Mace Head. Although cloud contamination is evident in the satellite data, it is clear that the air mass has advected over biologically rich waters to the west of Mace Head. The estimated source region is illustrated by the 500 km2 box.

Figure 3.

Five-day air mass back trajectories (courtesy of NOAA-HYSPLIT) overlaid on MODIS-derived chlorophyll-a oceanic concentrations for 12:00 on 24th August 2009 (JD 236). A 500 km2 grid box is superimposed on the figure for scaling purposes, noting that the production scale is at least three times this area.

[12] Taking the observed nucleation growth rate of 0.8 nm hour−1, and using the analytic approach of Dal Maso et al. [2002], the required sustained condensable vapour concentration to drive the calculated growth rate is of the order of 107 – 108, and > 109 molecules cm−3 for higher growth rates observed under similar conditions (2–3 nm hour−1). As noted in previous studies, H2SO4 and MSA concentrations are typically below this sustained concentration level [Berresheim et al., 2002], suggesting that organic vapours are driving the growth given that no other condensable inorganic gases achieve higher concentrations in clean marine air.

[13] There are perhaps four possible explanations for the particle production events although none can be evaluated conclusively without measurements of nucleated clusters and the condensing vapours which remain unknown at this stage. The four possibilities are that particle production and growth resulted from: (1) the initial source region of nucleated clusters is 675 km offshore and that over a period of 2 days the source of nucleating vapours continually moves further offshore such that the size of particles measured at Mace Head continually increases as the particles have been in the atmosphere for an increasing time since formation; (2) by mesoscale entrainment where the location of entrainment (of both freshly formed particles and condensable vapour) from the free troposphere (within which large or local scale nucleation is occurring) moves continuously offshore; (3) large scale boundary layer nucleation and continuous growth; or (4) large scale free tropospheric nucleation and entrainment of particles with condensable vapours or their precursors. For hypothesis (1) it is unlikely that the source or emission region changes over time scales of 24 hours as the biological turnover time for plankton is longer than 24 hours. For hypothesis (2), the mesoscale entraining region would not only have to be moving westerly against prevailing wind but would also have to be moving north or south to coincide with the air mass back trajectory. For hypothesis (3), a strong biogenic source of nucleating and condensing vapours would be required which seems likely. Finally, for hypothesis (4), while entrainment of nuclei may promote particle production over large spatial scales, it is difficult to account for the source of condensable vapours driving the growth as only anthropogenic sources from long range transport or a strong free tropospheric source could account for the condensable vapour production rate, the former is inconsistent with soot carbon mass levels encountered during these events, while for the latter, the free troposphere is not regarded as a significant source of natural vapours. For the above reasons, and acknowledging we cannot be conclusive, the most likely scenario is that these events result from boundary layer nucleation and growth from biogenic sources.

[14] A seven year dataset (2002–2007) of aerosol size distributions measured at Mace Head was analysed to quantify the ratio of the average monthly frequency of occurrence of such open ocean production events to the average percentage of marine air encountered at Mace Head. Events had to adhere to clean marine air criteria and exhibit growth extending to 24 hours or more. The average monthly ratio over the seven year period is shown in Figure 4. The peak occurrence is seen in the month of May, with the first events detected as early as March and the last events detected as late as September. Seventeen events occurred during May over the seven-year period.

Figure 4.

(top) Distinguishable nucleation mode monthly average occurrence over the years 2002–2009 divided by the percentage occurrence (bottom) of clean marine air (i.e., black carbon <50 ng m−3). The top graph represents data where a clear mode was observed to appear around 3 nm-20 nm in diameter, and then grow continuously to larger sizes over a period greater than 24 hours.

4. Conclusions

[15] Open ocean particle production and growth events have been detected over the mid-lattitude N.E. Atlantic. Although the events are observed at a coastal station, they cannot be coastal events given the appearance of the nucleation mode at ∼15 nm diameter. Some of these events are observed to last 24–48 hours, during which the nucleation mode grows from 15 nm to 50 nm. An associated growth rate of the order of 0.8 nm hour−1 is estimated for the events presented. At a typical wind speed of 10 m s−1, it is calculated that the source region or region of particle production was at least 675 km off-shore. In addition, the measurement of a growing nucleation mode up to 48 hours in during implies a spatial scale of the order of 1500–1700 km. The events are typically associated with polar marine air masses advecting over biological rich waters to the west and south–west of Mace Head. Analysis of 7 years of size distribution data from Mace Head reveals that extended (i.e., greater than 24 hours in duration) particle production and growth events are not unusual over mid-lattitude N.E. Atlantic waters For the 7 year period, peak occurrence is during May, but such events are detected from March through to September. In addition to the extended events, there are more events of shorter duration. These are the first experimental results reporting frequent open ocean particle production events over N.E. Atlantic waters, the frequency of occurrence of which suggests that marine boundary layer particle production is an important source of secondary marine aerosols.


[16] This work was supported by the European Commission under MAP and EUCAARI and also by Irelands Higher Education Authority under the PRTLI4 programme and Science Foundation Ireland.