We present the first direct measurements of coastal ozone deposition and ultrafine particle emission fluxes alongside observation of particle growth to sizes where they will be active as cloud condensation nuclei (CCN). Simultaneous ozone deposition and particle emission from exposed macroalgae were observed during daytime low tide. Ozone deposition to seawater at high tide was significantly slower (vd[O3] = 0.302 ± 0.095 mm s−1) than low tidal deposition. Low tide deposition was slower at night (1.00 ± 0.10 mm s−1) than during daytime (2.05 ± 0.16 mm s−1) when ultrafine particle formation results in apparent particle emission. Particle emission fluxes greater than 200,000 cm−2 s−1 were observed during some of the lowest daytime tides and these particles grew large enough to act as CCN. These results provide the first direct evidence of both direct depositional loss and photochemical destruction of ozone in the formation of particles from macroalgal emissions at a coastal location.
 It has long been established that seaweeds accumulate large amounts of iodine in their tissue and brown kelps can accumulate iodine concentrations more than 30,000 times greater than the surrounding seawater [Küpper et al., 1998]. It was shown by Küpper et al.  that the iodine is accumulated as iodide, which acts as an inorganic anti-oxidant. Upon oxidative stress (e.g., when exposed to the oxidising atmosphere at low tide), high levels of iodide are released, which reacts with ozone in the film of water on the seaweed surface to form molecular iodine [Palmer et al., 2005; Küpper et al., 2008]. It has become well-established in recent years that the emission of this molecular iodine [McFiggans et al., 2004; McFiggans, 2005; Saiz-Lopez et al., 2006] from low tidal macroalgal exposure may lead to photochemical formation of ultrafine particles in the coastal atmosphere [Saiz-Lopez et al., 2006]. Growth of such particles may enhance cloud condensation nucleus (CCN) concentrations and hence affect the properties of coastal clouds.
 It has been found in the laboratory that Laminaria sp. macroalgae (those comprising the highest percentage iodine dry weight and likely contributing to the greatest iodine emission on intertidal exposure) must be exposed to ozone in order to produce particles [McFiggans et al., 2004; Palmer et al., 2005; Küpper et al., 2008], though not necessarily to produce molecular iodine. Indeed, the amount of ozone consumed in macroalgal exposure experiments was shown to depend on the initial ozone concentration [Küpper et al., 2008] with a first order loss rate of 0.0115 s−1. This loss rate was used to calculate an apparent ozone deposition velocity to the macroalgal sample of between 2.5 and 10 mm s−1 (assuming a typical range for aerodynamic resistance). This is at least an order of magnitude greater than the value of 0.3 mm s−1 determined for deposition to seawater [Chang et al., 2004], and is also greater than the upper limit of 1.0 mm s−1 estimated by Clifford et al.  based on the reaction of ozone with chlorophyll in the surface marine layer. It may therefore be expected that measurable enhancement of ozone deposition will be exhibited whenever iodine-mediated coastal new particle formation occurs. This was a primary motivation for the current work.
 It is important to note that all possible ozone loss mechanisms are included when calculating apparent deposition velocity. Particle production results from the product of the self-reaction of the iodine monoxide radical, IO. The IO concentration is rapidly established in photostationary state with atomic iodine (I), consuming O3 whenever IO reacts either with itself, HO2 or NO2 rather than being photolysed to release I atoms [McFiggans et al., 2000]. It may therefore be expected that the apparent deposition velocity in laboratory experiments will comprise both direct deposition of ozone to the macroalgae and photochemical ozone consumption in the formation of particles.
 Likewise, any attempts to measure ozone deposition to infra-littoral macroalgal beds by micrometeorological techniques will comprise both direct depositional loss and photochemical destruction occurring between sources of reactive iodine and the measurement location. This study presents such direct measurement of apparent ozone deposition and particle production at a coastal location. Figure 1 illustrates the processes taking place in relation to the measurement location.
2. Site and Instrumentation
 The measurements were made in Roscoff (48°44′N, 3°59′W), a coastal town in Brittany in the northwest of France, during September 2006, as part of the coastal experiment of the Reactive Halogens in the Marine Boundary Layer (RHaMBLe) project. The coast at Roscoff has a very large tidal range (up to 9.60 m), leaving the sea floor exposed for as much as 3 km from the shore at the lowest tides. Instruments were situated at the start of a stone jetty outside the Station Biologique de Roscoff, providing a fetch of at least 800 m, and up to several kilometres, over the inter-tidal zone for a wind direction of between 215° and 005°. Micrometeorological fluxes of sensible heat, momentum, water vapour, ultrafine particles and ozone were measured using eddy covariance with fluxes averaged over 15 minutes. Figure 1 shows a schematic of the instrumental setup and a full methodology description is provided in the auxiliary material.
3.1. Ozone Fluxes
 Mean ozone concentrations of 30 ppb were observed, ranging from 2 ppb to 54 ppb. The mean ozone deposition velocity (vd[O3]) was 0.96 mm s−1 with a standard error of ±0.15 mm s−1 (deposition denoted by a positive value). A vd[O3] time-series is plotted in Figure 2 for a period with a prolonged sea fetch. No relationship between vd[O3] and wind speed was expected nor seen at high tide as, for the majority of the time (about 75%), it remained below 4 m s−1, where Chang et al.  observe very little dependence of vd[O3] on wind speed. On average, vd[O3] was greater during low tide than during high tide; mean low tide vd[O3] was 1.28 mm s−1 for the whole experimental period (standard error ±0.22 mm s−1; number of data points, n = 221), more than four times greater than the high tide value of 0.302 mm s−1 (standard error ±0.095 mm s−1; n = 109). The difference between the low and high tide mean values was significant (p < 0.005), and the two data sets can be compared in the inset of Figure 3. Between 24th to the 28th September, low tide occurred in the middle of the day and the middle of the night. vd[O3] was found to be statistically significantly higher (p < 0.001) during daytime low tide (mean 2.05 ± 0.16 mm s−1 standard error; n = 39) than during night-time low tide (1.00 ± 0.10 mm s−1; n = 31). The night-time deposition velocities were still significantly higher (p < 0.001) than those observed at high tide during the same period. This is expected due to enhanced turbulent transfer of ozone to macroalgal surfaces (and their associated surface films). These data are summarised in the main panel of Figure 3. For comparison with previous studies [e.g., Gallagher et al., 2001], mean surface resistance to seawater from high tide measurements was rs = 1150 ± 150 s m−1 (see Table S1).
3.2. Particle Measurements
 The time-series of number concentrations of particles greater that 3 nm and less than 10 nm diameter (Figure 4a) shows that a number of particle bursts were observed at low tide during the daytime, with durations of around 5 hours. These particle bursts were particularly strong from the 7th to the 11th of September during the period of greatest tidal range. During these events, mean particle number concentrations were as high as 95,000 particles cm−3, reaching peaks of around 2.5 × 105 particles cm−3. The time-series of particles greater than 10 nm does not show corresponding particle bursts, confirming the bursts are of ultrafine particles. Throughout this period of spring tides, the inferred “growth” of the particles from the in situ SMPS size distribution measurements was such that they could be readily expected to behave as cloud condensation nuclei. A time series of the size distributions is shown in Figure 4b. The wind during this period was mostly from the northeast, over the inter-tidal areas but beyond the fetch for flux measurements, so reliable particle and ozone fluxes were not observed. Fortunately, particle bursts of up to 105 particles cm−3 were also observed on the 25th and 26th of September during a period of good fetch and stationary conditions, allowing corresponding fluxes to be analysed. Figure 5 shows ultrafine particle number concentrations (estimated by subtracting the particles greater than 10 nm from particles greater than 3 nm), fluxes of particles greater than 3 nm, and vd[O3]. Apparent net particle emission fluxes of up to 2 × 105 particles cm−2 s−1 can be seen during daytime low tide corresponding with the particle bursts and the highest vd[O3] (up to 3 mm s−1), whilst apparent net particle emission fluxes were not observed during night-time low tides.
 During night-time low tide, direct deposition to the seaweed surface will be a dominant sink of ozone in the coastal atmosphere. If low tide occurs during the daytime, ozone will also react with atmospheric iodine generated by the photolysis of iodine emitted by exposed macroalgae. It is this daytime consumption of ozone which will lead to IO formation and particle production. Measured daytime ozone deposition velocities can therefore be expected to comprise both direct deposition to the macroalgae and photochemical destruction between the measurement height and the sources of reactive iodine, and may be expected to be greater than night-time deposition velocity.
 The results of this study are consistent with the picture described above. The surface resistance to ozone deposition during night-time low tide, when direct deposition to the exposed macroalgae is the dominant loss mechanism, was found to be 650 ± 46 s m−1. This is lower than that observed at high tide (1150 ± 150 s m−1, both daytime and night-time) as the high levels of iodide present at the seaweed surface result in an enhanced uptake. The difference is not an order of magnitude as predicted by Küpper et al. . Küpper et al.  calculated an apparent surface resistance to ozone deposition to exposed Laminaria digitata of between 100–400 s m−1 based on the ozone removal rate observed during chamber measurement results. The removal rate was comparable in both sunlight and in the dark, suggesting the loss mechanism was predominantly deposition to the seaweed surface. This may therefore be considered a lower limit to surface resistance to an exposed seafloor with non-uniformly distributed Laminaria beds amongst other species of seaweed.
 During periods of low tide, vd[O3] is, on average, twice as high during the day as it is at night, illustrating the additional contribution made by photochemical destruction to the removal of ozone from the coastal atmosphere. This is clearly supported by the observation of strong net apparent particle emission fluxes and hence concentrations only during daytime low tide which are fully consistent with observations of particle bursts at other coastal sites [O'Dowd et al., 1998, 2002b] and also in laboratory experiments [O'Dowd et al., 2002a; McFiggans et al., 2004; Palmer et al., 2005]. The particle size spectrum clearly shows that the increase in number following the particle production events on 8th and 9th September is propagated continuously through to larger sizes which persist at the measurement site into the afternoon of the 10th September. The particles are formed at diameters of less than 3 nm, observed only once they have reached 3 nm (by the difference in CPC measurements), and grow to dry sizes of greater than 120 nm in a timescale of tens of hours. It may therefore be inferred that the particle formation must be a widespread phenomenon over the Brittany coastline during the experimental period (see back-trajectories in Figure S5), otherwise the growth would not be observed by single-point in situ measurement. Since regionally, there may be significant additional sources of condensable material, it cannot be claimed that the particle growth results solely from the condensation of oxidation products of macroalgal emissions. However, from the relationships established in the current work it may be concluded that the regional scale enhancement in particle number results from tidally driven and ozone mediated particle formation.
 It should be noted here that it is likely that significant particle nucleation occurs above the sensors, leading to an underestimate in the total apparent particle emission fluxes. The fluxes shown above and in Figure 5 must therefore be lower limits at this location.
 Ozone deposition velocities and ultrafine particle emission fluxes were monitored over an extensive infra-littoral zone on the coast of north-west France during September 2006. Higher vd[O3] was observed during low tide both during daytime and night-time compared to high tide as a result of enhanced uptake to the iodide-rich surfaces of the macroalgae. Furthermore, observed deposition velocities of ozone at low tide were at least twice as high during the day as at night, illustrating the importance of photochemical destruction as a removal mechanism for ozone in the coastal atmosphere. Strong particle bursts (peaking around 2.5 × 105 particles cm−3) were also observed during the daytime low-tides. In good stationarity and fetch conditions, particle fluxes of up to 2 × 105 particles cm−2 s−1 were observed corresponding to the highest observed vd[O3] of 3 mm s−1. These results provide the first direct evidence of both direct depositional loss and photochemical destruction of ozone in the formation of particles from macroalgae beds at a coastal location.
 We would like to acknowledge Philippe Potin and the staff of the Station Biologique de Roscoff for their kind assistance and the interest shown in our work by the people of Roscoff. Thanks also to James Lee of the University of York for providing the ozone concentration data. We gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT transport and dispersion model and READY website (http://www.arl.noaa.gov/ready.html) used in this publication. The RHaMBLe project was funded under a NERC grant NE/D006570/1 within the UK SOLAS programme.