The seaweedsFucus vesiculosusand Ascophyllum nodosumare significant contributors to coastal iodine emissions

Based on the results of a pilot study in 2007, which found high mixing ratios of molecular iodine (I 2) above the intertidal macroalgae (seaweed) beds at Mweenish Bay (Ireland), we extended the study to nine different locations in the vicinity of Mace Head Atmospheric Research Station on the west coast of Ireland during a field campaign in 2009. The mean values of I 2 mixing ratio found above the macroalgae beds at nine different locations ranged from 104 to 393 ppt, implying a high source strength of I 2. Such mixing ratios are sufficient to result in photochemically driven coastal newparticle formation events. Mixing ratios above the Ascophyllum nodosumandFucus vesiculosus beds increased with exposure time: after 6 h exposure to ambient air the mixing ratios were one order of magnitude higher than those initially present. This contrasts with the emission characteristics of Laminaria digitata, where most I 2 was emitted within the first half hour of exposure. Discrete in situ measurements (offline) of I2 emission from ambient air-exposed chamber experiments ofL. digitata, A. nodosumand F. vesiculosus substantially supported the field observations. Further online and time-resolved measurements of the I 2 mission from O3-exposed macroalgal experiments in a chamber confirmed the distinct I 2 emission characteristics of A. nodosumand F. vesiculosuscompared to those of L. digitata. The emission rates of A. nodosumandF. vesiculosuswere comparable to or even higher thanL. digitata after the initial exposure period of∼20–30 min. We suggest that A. nodosumandF. vesiculosusmay provide an important sourc of photolabile iodine in the coastal boundary layer and that their impact on photochemistry and coastal new-particle formation should be reevaluated in light of their longer exposure at low tide and their widespread distribution.


Introduction
Brown algae include kelps of the genus Laminaria, the strongest biological accumulators of iodine currently known.Laminaria species (Laminaria spp.) are a key biogeochemical pump for the transfer of iodine from the sea to the atmosphere (Küpper et al., 2011).Laminaria accumulates iodide for the provision of an extracellular antioxidant.Its reaction with O 3 on the thallus surface results in the release of I 2 (Küpper et al., 2008).However, the iodine metabolism of other brown algae and their role in marine-atmospheric halogen transfer is much less well understood.The current interest in tropospheric iodine chemistry was initiated by observations of iodine oxide (IO) -a product of iodine atom reaction with O 3 -in the marine boundary layer (MBL) at Mace Head, Ireland (Alicke et al., 1999).Reactive iodine can affect the tropospheric oxidizing capacity through catalytic destruction of O 3 , changing the NO 2 / NO and HO 2 / HO ratios, and by reactivating chlorine and bromine from sea salt aerosol (Vogt et al., 1999;McFiggans et al., 2000;Bloss et al., 2005;Saiz-Lopez et al., 2008).The photochemically driven reaction of iodine and O 3 also results in iodine oxide particle (IOP) formation.
Coastal new-particle formation via secondary gas-toparticle conversion is an important process determining the concentration of atmospheric aerosols and, ultimately, the concentration of cloud condensation nuclei (CCN) on the regional scale (O'Dowd and Hoffmann, 2005).The nucleation events generally occur around low tide during daylight and have been known to lead to ultrafine particle number concentrations in excess of 10 6 particles cm −3 (O'Dowd and Hoffmann, 2005).In recent years, numerous studies have shown that the coastal particle bursts are closely linked to iodine emission from low-tidal macroalgae exposure (e.g., McFiggans et al., 2004;Sellegri et al., 2005;Saiz-Lopez et al., 2006a;Huang et al., 2010c;McFiggans et al., 2010).A clear negative correlation between IO (the precursor of IOP) and tidal height, and a positive correlation between IO and solar irradiation, have been observed (Carpenter et al., 2001;Saiz-Lopez et al., 2006b;Huang et al., 2010b;Commane et al., 2011).The emission of iodocarbons such as CH 2 I 2 from macroalgae was first proposed to be the source of photolabile iodine (Hoffmann et al., 2001;Carpenter, 2003).However, recent field measurements and laboratory experiments show that the emission of molecular iodine (I 2 ) is the dominant source of iodine and is responsible for the observed iodine chemistry in the coastal MBL (McFiggans et al., 2004;Saiz-Lopez and Plane, 2004;Sellegri et al., 2005;Huang et al., 2010b, c;Monahan et al., 2012).
Measurements of I 2 mixing ratios have so far been reported at four different coastal locations: Mace Head and vicinity, Ireland (Saiz-Lopez and Plane, 2004;Saiz-Lopez et al., 2006a, b;Huang et al., 2010b, c), Roscoff, France (Leigh et al., 2010;McFiggans et al., 2010), O Grove, Galicia, Spain (Mahajan et al., 2011), and La Jolla, California (Finley and Saltzman, 2008).The observations of high concentrations of I 2 at Roscoff and O Grove are thought to be a consequence of large I 2 emissions from Laminaria spp.such as L. digitata andL.hyperborea, which are the dominant species at these measurement sites.
Since the daytime reaction cycle of iodine in the coastal MBL is initiated by rapid photolysis of I 2 , measurement of iodine close to its source is required for understanding the contribution of macroalgal iodine emissions to local atmospheric processes.However, such field observations are scarce because measurement of I 2 directly above the algal beds is challenging.Efforts have therefore been made to study the I 2 emission profiles of macroalgae through laboratory incubation experiments (Dixneuf et al., 2009;Ball et al., 2010;Nitschke et al., 2011;Kundel et al., 2012b;Ashu-Ayem et al., 2012) in order to understand the emission mechanism and to better estimate the flux of I 2 from macroalgae.These studies have mainly focused on L. digitata as this species accumulates iodine at up to around 5 % of its dry weight (Küpper et al., 1998;Gall et al., 2004) and emits large amounts of I 2 when exposed to ambient air; the results showed a high variability of I 2 emission rates, with values ranging from 3 to 2500 pmol min −1 g fresh weight −1 (FW −1 ) (Bale et al., 2008;Ball et al., 2010;Ashu-Ayem et al., 2012).The I 2 emission profiles of L. digitata are characterized by an intense initial burst when first exposed to air, followed by an approximately exponential decay over a short period of about 20-30 min (Bale et al., 2008;Dixneuf et al., 2009;Ball et al., 2010;Nitschke et al., 2011;Ashu-Ayem et al., 2012).
During a field campaign at Mweenish Bay, Ireland, in August/September 2007, a pilot study of I 2 emissions was carried out directly above macroalgal beds of Ascophyllum nodosum and Fucus vesiculosus located in the mid-littoral zone.Elevated I 2 mixing ratios of up to 302 ppt were observed after the algae had been exposed to air for several hours (Huang et al., 2010b).This behavior differs from the emission characteristics of L. digitata found in laboratory incubation experiments, where I 2 emissions are small or stop completely after prolonged exposure (Dixneuf et al., 2009;Ball et al., 2010;Nitschke et al., 2011).We hypothesized that the temporal behavior of I 2 emission of A. nodosum and F. vesiculosus would likely be different and of longer duration than that of L. digitata.Both A. nodosum and F. vesiculosus are widely distributed throughout the world and exposed to air even at moderately low tidal levels since they inhabit the midto upper littoral regions.If the emission indeed increases with time, the impact of these two macroalgae on the local atmospheric iodine chemistry could be more significant than currently believed.This finding would likely apply to numerous other brown algae (Phaeophyceae) species, implying greater macroalgal sea-air transfer of iodine than expected from Laminaria emissions only.Moreover, different emission characteristics compared to Laminaria spp.may provide an explanation for the frequently observed new-particle formation events at the west coast of Ireland, considering the large population of A. nodosum and F. vesiculosus distributed there.For example, during a field campaign in 2007 enhanced nucleation events were observed in 14 out of 23 days of measurements at Mweenish Bay, where A. nodosum and F. vesiculosus are dominant species.The ultrafine particle bursts typically lasted for about 4-6 h, which was closely   related to the diurnal variation of the exposure period of these two species (Huang et al., 2010c).
In this paper, we present results of a 2-week field campaign carried out on the west coast of Ireland in August 2009 with a special focus on the investigation of mixing ratios and emission characteristics of I 2 above the macroalgae beds in the field.The results derived from chamber experiments of L. digitata, A. nodosum and F. vesiculosus exposed to ambient air and to ozone are also presented.

Sampling sites
Measurements were carried out at nine different sites in the vicinity of the Mace Head Atmospheric Research Station on the west coast of Ireland (see Fig. 1) from 17 August to 28 August 2009.This area is characterized by a high abundance (but discrete zonation) of brown macroalgae along the coastline and is representative of typical North Atlantic rocky seashores in terms of the macroalgal exposure time.A. nodosum and F. vesiculosus are dominant algal species at sites #1, #2, and #4-7 and can be exposed to the air for up to 6-8 h during low tides (average exposure ∼4 h at normal low tides).No Laminaria spp.were observed in the immediate vicinity of these sites.In contrast, most L. digitata algae at sites #3 and #8 are extensively exposed only at around lowest water (spring low tides) and for a much shorter period (∼20-30 min).Measurements at site #9 (dominated by A. nodosum and F. vesiculosus) were carried out when Laminaria spp. in the nearby site #8 were submerged in seawater.Note that at all sampling sites samples were taken when the wind came from the sea (wind direction 179 • -287 • ).The wind speed ranged between 5-11 m s −1 .

Denuder sampling with GC-MS quantification
Molecular iodine was measured using a diffusion denuder system in combination with a gas chromatography-mass spectrometry (GC-MS) method, which provides "point" in situ concentrations of I 2 at the sampling site.Details of this denuder/GC-MS method are given in Huang and Hoffmann (2009), and it will be only briefly described here.Ambient I 2 samples were collected at a flow rate of 500 mL min −1 for 5-30 min in brown denuder tubes (6 mm i.d., 50 cm length) which were uniformly coated with ∼11 mg α-cyclodextrin (α-CD) and trace 129 I − .The potential interference iodine species such as ICl and HOI were removed by coupling a 1,3,5-trimethoxybenzene-coated denuder upstream of the α-CD/ 129 I − coated denuder (data not shown here).The inlet of the denuder was set up very close to the algal beds (∼5-10 cm) during sampling to minimize potential photolysis of I 2 .Although Saiz-Lopez et al. ( 2004) calculated a lifetime of I 2 of about 8 s under noontime clear sky at Mace Head for this time of year, the relatively low solar flux (as indicated in Fig. 3) for most measurements implies a lifetime that is several times longer (i.e., the low irradiation (< 200 W m −2 ) during most of our measurements implies a much longer I 2 lifetime of over 40 s, while in four of the measurements the I 2 lifetime is shorter and around 12-21 s).After sampling, the open ends of the denuders were sealed with PP caps and kept under refrigeration until analysis.In the laboratory, the samples were eluted with five 2.0 mL portions of ultrapure water into a 25 mL flask to which 500 µL of phosphate buffer (pH 6.4), 100 µL of 2,4,6-tribromoaniline (2.5 mg L −1 , internal standard), 400 µL of sodium 2-iodosobenzoate, and 300 µL of N,N -dimethylaniline were added.The solution was shaken at room temperature for about 120 min, leading to >98 % conversion of I 2 into 4-iodo-N ,N-dimethylaniline (Huang et al., 2010a).Finally, the solution was extracted with 100 µL of cyclohexane.One (1.0) µL of the extraction solution was injected into a GC-MS system (Agilent 6850 GC interfaced to a 5973N MSD, Agilent Technologies, Santa Clara, CA).A Rtx-5MS fused-silica capillary column (Restek Co., Bad Homburg, Germany) was used for chromatographic separation, and the MS was run in selected ion monitoring (SIM) mode to enhance the sensitivity.The detection limit was generally below 1.0 ppt during this campaign, and the collection efficiency was greater than 98 %.The precision of the method for the determination of I 2 is < ± 10 %.The reported values in this study have been corrected for the mean value of blanks, which were based on the analysis of denuders that had been sealed throughout the campaign.

Ambient air chamber experiments
L. digitata specimens were collected at Mweenish Bay during spring low tide on 24 August 2009.Since thalli were tightly attached to the rock through their rhizoids, rock and thalli were taken together to avoid injuring the alga.Laminaria thalli were kept in seawater during a short transport time of ∼10 min and then stored in running seawater in a transparent tank placed outside the Martin Ryan Institute, Carna, (MRI-Carna) building, where macroalgae were exposed to the natural diurnal light and temperature cycle.The seawater was freshly pumped from the sea and close to the sea temperature in the area.Algal thalli were used within 4 days of collection.When required for the incubation experiments, the whole alga together with its anchoring rock was removed from the tank and placed into a 35 L translucent polyethylene chamber.Ambient air (O 3 mixing ratio 35-40 ppb) was then drawn through the chamber, and the emitted I 2 was collected using the denuder at a flow rate of 500 mL min −1 .After the first exposure of 20 min and air sample collection, the alga was removed from the chamber and exposed to outside ambient air for a specified duration of up to several hours (similar to what it would experience during tidal exposure).Afterwards the alga was placed back into the chamber and its emission was measured again.Whole, submerged A. nodosum and F. vesiculosus algae were taken directly from the intertidal zone outside the MRI-Carna building and used immediately.Again, algal thalli were usually taken together with small rocks attached to their holdfasts.The transit time from collection in the intertidal zone to beginning an experiment was around 5-10 min, during which algal thalli were kept in seawater.The alga and its anchoring rock were weighed before and after the experiments, and then the alga mass alone was weighed after detaching the rock.About 300-500 g algae were used in individual experiments.When air samples in the chamber were taken, the I 2 mixing ratios in the ambient air outside the chamber were measured simultaneously to correct for background I 2 .

I 2 mixing ratios on the west coast of Ireland
At the Mace Head station intense bursts of new-particles have been frequently observed, with concentrations often reaching in excess of 10 6 cm −3 (O' Dowd and Hoffmann et al., 2005).Airborne measurements further reveal that such particle bursts are almost ubiquitous along the coastline in the vicinity of Mace Head (O'Dowd et al., 2007).Close to the plume head the growth rate of nucleated particles can be as high as several hundred nanometers per hour, indicating the presence of a high concentration of precursor gases.Table 1 shows the denuder results from nine different sampling sites in the vicinity of Mace Head station.It is evident that elevated levels of I 2 were indeed present over the local source regions (macroalgae beds).Based on model study predictions that 80-100 ppt I 2 is required for iodine oxide particle bursts (Saiz-Lopez et al., 2006a), the mixing ratios of I 2 , ranging from 104 ppt to 393 ppt (see Table 1), observed in all sampling sites would be sufficient to result in photochemically driven new-particle formation events.Note that the measurements given in Table 1 simply show elevated levels of I 2 when macroalgae beds are exposed to ambient air.They were collected under specific conditions and should not be taken as being more generally representative.
The I 2 mixing ratio of 547 ppt (Fig. 2) measured over a short time period of 5 min immediately after exposure of the L. digitata beds is one of the highest observations reported to date.Similar "point" in situ techniques at other coastal locations have also found high mixing ratios at sites close to Laminaria spp.belts: a daytime maximum of 350 ± 100 ppt was reported at O Grove (Mahajan et al., 2011) and 50 ppt was observed at Roscoff (McFiggans et al., 2010).Although it is difficult to compare observations made at different geographical locations, in particular where observations were obtained at different distances downstream of highly localized emission sources, considering the dilution effect, the short photolysis lifetime and potential chemical recycling of I 2 , these measurements are consistent with the current consensus that Laminaria spp.are very strong emitters of I 2 .What the observations reported in Table 1 also indicate, however, is that the I 2 mixing ratios observed above the A. nodosum and F. vesiculosus mixed beds are a significant proportion of those above the L. digitata beds (with an average of 134 ppt versus 301 ppt, respectively).This observation is somehow different from the macroalgae incubation experiments of Ball et al. (2010) which showed that I 2 emissions from A. nodosum and F. vesiculosus were several orders of magnitude lower than those from Laminaria spp.We attribute the enhanced I 2 emissions of A. nodosum and F. vesiculosus to the longer exposure period in our study and the distinct time-dependent emission characteristics of A. nodosum and F. vesiculosus discussed below.We note that, in the vicinity of Mace Head, A. nodosum and F. vesiculosus are extensively exposed to the air during most low tides, whereas the exposure of Laminaria beds is limited to extremely low-water (spring low tide) events.The different exposure profiles, together with the elevated mixing ratios observed, leads us to suggest that A. nodosum and F. vesiculosus could be the main sources of I 2 in the vicinity of Mace Head (see Fig. 1) during most low tides, whereas Laminaria spp.would be the major contributor to I 2 emissions at spring low tides.This suggestion is supported by the observation that new-particle formation events occur on more than half of the days at Mace Head and are therefore not limited to spring low tides.Moreover, the new-particle formation events observed at Mace Head station are often characterized by number size distributions with "apple" (i.e., the shape of the number concentration vs. time plot resembles an apple), "bump" (i.e., particles do not usually grow larger than 10 nm in diameter, and the shape of the number concentration vs. time plot resembles a bump) or "mixed" (i.e., the formation of intermediate air ions is clearly observed, but the events cannot be classified as banana-, apple-or bump-type) shapes (Vana et al., 2008).Ehn et al. (2010) explained these particle formation characteristics on the basis of an inhomogeneous distribution of precursor gases.Our findings of high I 2 mixing ratios above the A. nodosum and F. vesiculosus beds, when taken together with the inhomogeneous distribution of these two macroalgae species (whose habitat is restricted to a number of small areas around Mace Head, e.g., Mweenish Bay, Glinsk and Roundstone), are consistent with the suggestion of Ehn et al. (2010).

Field observations
The emission characteristic is one of the most important factors determining iodine flux from macroalgae to the atmosphere.However, relevant studies are scarce and have mainly concentrated on L. digitata (Dixneuf et al., 2009;Ball et al., 2010;Nitschke et al., 2011;Ashu-Ayem et al., 2012).Laboratory incubation experiments have revealed that I 2 emis-16 Fig. 2. I 2 mixing ratio against the exposure time of L. digitata beds o the first sample (from 0-5 min) was taken when L. digitata was just an interval of 10 min for setting up the second sampler, the second was taken.The data set is limited to two measurements because accessible only at a single spring low tide at Mweenish Bay, which bars in this Figure.sions from this species are intense within the first few minutes when subjected to air exposure but decrease strikingly afterwards (Dixneuf et al., 2009;Ball et al., 2010;Nitschke et al., 2011;Ashu-Ayem et al., 2012).This phenomenon was also observed in our field observations.As shown in Fig. 2, the average I 2 mixing ratio measured above the L. digitata beds at Mweenish Bay was 547 ppt over the first 5 min of exposure, but decreased to 239 ppt over the interval of 15-20 min.
Figure 3 shows the 2-day temporal profile of I 2 emission (averaged over a 30 min sampling time) at sampling site #1, dominated by A. nodosum and F. vesiculosus, at Mweenish Bay.A lower mixing ratio was observed at the beginning of the ebbing tide when the macroalgae were just exposed to the ambient air.However, the mixing ratio increased gradually with exposure time, reaching a value of one order of magnitude higher than the initial value after about 6 h (1/2 tidal cycle).Note that the contribution of nearby macroalgae source to the I 2 level at the sampling site could be rather small due to the rapid photochemical destruction of I 2 , the efficient dilution during transport and the short distance (5-10 cm) between the denuder inlet and the macroalgae bed.This emission profile is markedly different from that of L. digitata observed in the present and previous studies (e.g., Ashu-Ayem et al., 2012).It may at first appear that the increased emissions of I 2 from A. nodosum and F. vesiculosus at Mweenish Bay are a consequence of elevated O 3 concentration in the ambient air, which we observed in our previous field studies (Huang et al., 2010b).However, this effect is unlikely to be the case as the O 3 concentrations were relatively an interval of 10 min for setting up the second sampler, the second sample (from 15-20 min) was taken.The data set is limited to two measurements because the L. digitata beds were accessible only at a single spring low tide at Mweenish Bay, which explains the lack of error bars in this Figure.stable (35-39 ppb, measured by an ozone analyzer, Model 1008-RS, Dasibi Environmental Corp., Glendale, USA) over two consecutive days (20-21 August 2009).The data measured at a similar O 3 level (36-39 ppb) but from a previous study in 2007 also fit well into the emission profile.The intensity of solar irradiation had no noticeable impact on the emission of I 2 (see Fig. 3).

Chamber studies
During the campaign, I 2 emissions from L. digitata, A. nodosum and F. vesiculosus were also investigated in a flow chamber at Mweenish Bay.The algal samples were exposed to the ambient air that flowed into the chamber to simulate the natural process of exposure.The results, after correction for background concentration in the ambient air, show that the emission rates from A. nodosum and F. vesiculosus rose significantly upon prolonged exposure (see Fig. 4a and b) whereas L. digitata emission rates decreased strikingly over time (see Fig. 4c).This temporal behavior was observed in both daytime and nighttime experiments, providing substantial support for the field observations reported above.Figure 4 also shows that the initial emission rates of A. nodosum and of F. vesiculosus were relatively constant although different specimens were used.The relative standard deviations (RSDs) are 12 % (n=3) for A. nodosum and 18.8 % (n=3) for F. vesiculosus, respectively.In contrast, the initial I 2 emissions by L. digitata differed significantly between specimens (RSD 74 %).Variable emissions between plants have been found in a relatively large sample size 17 the symbol, solid circle represents data from this work and solid diamond repres from the 2007 campaign (Huang et al., 2010b).The O 3 concentrations in the surrou were 35-39 ppb over the course of measurements.2012) and may arise from the inhomogeneous accumulation of iodine in individual algae.Additionally, iodine emission differs across parts of the thallus, and Nitschke et al. (2011) have shown that the stipe of L. digitata emits up to 19 times more I 2 compared to the distal blade.The L. digitata #1 studied here had a bigger and longer stipe than L. digitata #2 (16 cm length, ∼1.5 cm i.d.versus 10 cm length, ∼1.2 cm i.d.) and would be expected to emit more I 2 after correction for its grams fresh weight.
The time profile of I 2 emissions from L. digitata, A. nodosum and F. vesiculosus was further investigated in a simulation chamber using a recently developed time-offlight aerosol mass spectrometer in combination with a gaseous compound trapping in artificially generated particle (GTRAP-AMS) method (Kundel et al., 2012a).The measurement system was calibrated using a capillary-based diffusion device as an I 2 test gas source (Huang and Hoffmann, 2010).Algae in chamber experiments were directly exposed to 50 ppb O 3 .Details of this set of chamber experiments are given in Kundel et al. (2012b).This online GTRAP-AMS provided a much higher time resolution compared to the offline denuder method and therefore enables us to look at a more detailed profile of I 2 emissions from macroalgae.The results (see Fig. 5) provide strong evidence that the I 2 emission rate from A. nodosum and F. vesiculosus rises with increasing exposure time over a period of hours, while L. digitata displays the opposite behavior, emitting a strong, short pulse immediately on exposure.These measurements further support our findings from field measurements and ambient air-exposed chamber studies.

Potential contribution of Fucus and Ascophyllum seaweeds to coastal iodine level
It is beyond the scope of the present work to investigate the biochemical mechanism governing the distinct I 2 emission features between different macroalgal species, but it is tempting to hypothesize that it is linked to different physiological adaptations of Ascophyllum, Fucus and Laminaria to their differing positions in the littoral zone.Ascophyllum and Fucus are intertidal species and get exposed at every low tide, while Laminaria is a mostly submerged-living species, only getting exposed during stronger spring tides.The location of their habitat in the littoral zone governs the frequency and duration of exposure to air at low tide and therefore the extent of exposure to ozone and desiccation stress.It is clear that L. digitata is able to emit much larger amounts of I 2 within the first exposure of ∼20-30 min than A. nodosum and F. vesiculosus.However, a major finding of this study is that the emission rates of A. nodosum and F. vesiculosus are comparable to or even higher than L. digitata after the initial exposure period (see Figs.   tion studies, taken from Ball et al. ( 2010)) for A. nodosum and F. vesiculosus.Nevertheless, these authors also showed that the I 2 mixing ratios calculated from the emissions of Laminaria spp.were lower than observed values, and suggested the presence of an additional source to account for the discrepancy.Our results show that upon exposure to 50 ppb O 3 in a synthetic air stream the first 10 min integrated I 2 emission rates of A. nodosum (0.014 pmol min −1 g FW −1 ) and F. vesiculosus (0.049 pmol min −1 gFW −1 ) are indeed much lower than that of L. digitata (4.23 pmol min −1 gFW −1 ).However, the first 1 h integrated I 2 emission rates are 159, 69, 19 pmol h −1 g FW −1 for L. digitata, A. nodosum and F. vesiculosus, respectively.Note that the emission rate is calculated based on one or two samples of each species.
It is expected that the sum of I 2 emitted by A. nodosum and F. vesiculosus could be comparable to or even larger than that from L. digitata under natural conditions, considering their distinct time-dependent emission trends and the longer exposure time of A. nodosum and F. vesiculosus in comparison to L. digitata (the former two inhabit upper the littoral zone, while the latter the sublittoral zone).It should be noted that A. nodosum and F. vesiculosus (both Fucales) are commonly found on the coasts of the North Atlantic Ocean and that Fucus species are widely distributed along rocky coasts throughout the world.Therefore, Fucales and other large, morphologically complex brown algae may provide an important source for the observed tropospheric iodine level, and their impact on the photochemistry and coastal new-particle formation should be reevaluated.

Summary
The mixing ratio of I 2 above the macroalgae beds at nine different locations on the west coast of Ireland has been measured using diffusion denuders in combination with a gas chromatography-mass spectrometry (GC-MS) method.
The results show the occurrence of elevated I 2 levels above macroalgae beds, ranging from 104 ppt to 393 ppt, which is in line with a previous pilot study (Huang et al., 2010b).Most importantly, it is found that the mixing ratio above the A. nodosum and F. vesiculosus beds correlates positively with their exposure time, reaching a value of one order of magnitude higher than the initial emission after exposure to ambient air for ∼6 h.In contrast, the mixing ratio above the L. digitata beds decreases with increasing exposure time, as observed in the present and previous studies.This feature can be attributed to the distinct time-dependent I 2 emission characteristic of macroalgae confirmed in two sets of chamber experiments.A particularly interesting aspect is the different emission profiles over time -of Fucus and Ascophyllum on the one hand, and Laminaria on the other.The results derived from these chamber experiments indicate that the emission rates of A. nodosum and F. vesiculosus are comparable to or even higher than that of L. digitata after the initial exposure period of ∼20-30 min.Given the longer low-tide exposure time of A. nodosum and F. vesiculosus (as they usually inhabit the upper littoral zone and are therefore easily exposed to air) as well as their large distribution on the coasts of Atlantic and Pacific Oceans, we suggest that A. nodosum, F. vesiculosus and possibly other Fucales may provide an important source for the observed tropospheric iodine level and that their impact on the photochemistry and coastal newparticle formation should be reevaluated.Furthermore, more studies are needed to determine whether other macroalgae that are often exposed to ambient air during low tide are also significant contributors to coastal emissions of I 2 .

.
Fig. 1.The denuder sampling sites on the west coast of Ireland (revised from Google earth).

Fig. 1 .
Fig. 1.The denuder sampling sites on the west coast of Ireland (revised from Google earth).

Fig. 3 .
Fig. 3.I 2 mixing ratio above the A. nodosum and F. vesiculosus sampling #1 at Mweenish Bay, Ireland as a function of algal e

Fig. 2 .
Fig.2.I 2 mixing ratio against the exposure time of L. digitata beds observed at site #3.Here, the first sample (from 0-5 min) was taken when L. digitata was just exposed to air.Following an interval of 10 min for setting up the second sampler, the second sample (from 15-20 min) was taken.The data set is limited to two measurements because the L. digitata beds were accessible only at a single spring low tide at Mweenish Bay, which explains the lack of error bars in this figure.

Fig. 3 .
Fig. 3.I 2 mixing ratio above the A. nodosum and F. vesiculosus mixed beds observed at sampling #1 at Mweenish Bay, Ireland as a function of algal exposure time and solar Fig. 3.I 2 mixing ratio above the A. nodosum and F. vesiculosus mixed beds observed at sampling site #1 at Mweenish Bay, Ireland, as a function of algal exposure time and solar irradiation.Each data point represents the mean value over a sampling period of 30 min.Note the symbol solid circle represents data from this work, and solid diamond represents data from the 2007 campaign (Huang et al., 2010b).The O 3 concentrations in the surrounding air were 35-39 ppb over the course of measurements.

Fig. 4 .
Fig. 4. The time-dependent I 2 emission rates of A. nodosum (a), F. vesiculosus (b) and L. digitata (c) at daytime and nighttime when subjected to ambient air exposure in a flow chamber at Mweenish Bay.Note the difference in scale of L. digitata emissions and the lack of error bars in (c) as only one sample was taken for individual experiments.
4 and 5) and are sustained over a period of several hours.Due to a relatively limited dataset, we have not yet been able to reconstruct detailed emission profiles of these three macroalgae under natural conditions, nor do we compare their overall emission rates over an integrated lowtide period.Leigh et al. (2010) concluded that, in comparison to Laminaria spp., the contributions from A. nodosum and F. vesiculosus to the total I 2 emissions were relatively small in the coastal region around Roscoff by assuming a lower and constant emission rate throughout the low-tide period (derived from the first exposure period of ∼10 min in incuba-18 of error bars in (c) as only one sample was taken for individual experiments.
at daytime and nighttime when subjected to ambient air exposure in chamber at Mweenish Bay.Note the difference in scale of L. digitata emissions and of error bars in (c) as only one sample was taken for individual experiments.

Fig. 5 .
Fig. 5.A typical time-dependent I 2 emission rate of A. nodosum (a), F. vesiculosus (b) and L. digitata (c) when exposed to 50 ppb O 3 in synthetic air in a simulation chamber (modified from Kundel et al., 2012).

Table 1 .
The mixing ratio of I 2 at different sites on the west coast of Ireland.