Emissions of sulphur to the atmosphere over sea areas is as dimethylsulphide (DMS). This subsequently oxidises to sulphate (easily scavenged, with restricted dry deposition velocity). Phytoplankton community composition and biomass are primary determinants of DMS concentrations in seawater, and of its precursor dimethylsulphoniopropionate (DMSP). Whilst there is relatively little DMS production from actively growing phytoplankton (Archer et al., 2002), higher rates of DMS release result from zooplankton grazing pressure (Archer et al., 2001), which encourages breakdown of DMSP. Bacterioplankton are significant agents for this, with viral agents also able increase DMS production notably (Malin et al., 1998 & Simó et al., 2000). Phaeocystis sp.zooplankton herbivory consumes particulate DMSP at rates equal to DMS production, and hence blooms can be sources of high DMS concentrations. The various interrelationships lead to correlations between locations of microzooplankton biomass, phytoplankton, and nearshore waters. Highest biomass is nears hore where zooplankton grazing of phytoplankton crops productivity (Stelfox-Widdicombe et al., 2003).

The ROBUST project has also looked at the mechanisms behind DMS production. Anoxic sediments are proposed as a particular source of DMS, with eutrophic coastal or sedimentary environments favouring Phaeocystis sp., Ulva sp. and Enteromorpha sp. The establishment of high concentrations of DMS in coastal waters is a necessary step for emission to the atmosphere, and one which ELOISE projects have contributed to. The step to identifying the scale of atmospheric release requires consideration of sea-air exchange, and with this of existing air concentrations. This step for sulphur compounds was not taken by ELOISE or IMPACTS projects. Whilst not the purpose of this digest to conduct such research, it is interesting to compare the observed DMS concentrations observed by ELOISE projects in European coastal waters with values used for past estimates of the global sea-air flux of sulphur. Table 4 documents the potential for coastal emissions of dimethylsulphide to the atmosphere.

The quoted values for sea concentrations and for sea to air fluxes from Erickson et al. (1990) are here compared with the observed sea concentrations obtained by the ESCAPE project. The Erickson et al. (1990) values are only an example of those available in literature, and are comparable to the overall global approximation provided by ROBUST of 3nM (Welsh, 2000). Comparison is also made with approximate fluxes obtained by multiplying those concentrations with the sea to air transfer velocities estimated by Erickson et al. (1990). Clearly, the sea concentrations of DMS observed within ESCAPE are sufficient on their own to produce a far greater flux from European coastal waters than some global sea-air estimates have previously.

This remains a step from evaluating the flux to the land surface, there are reasons to believe that the coastal flux of sulphur to the land could be greater than previously estimated. Not only are ELOISE observations revealing greater concentrations of DMS, but they are doing so in the coastal waters, in closer proximity to the receiving land surface. Further, the atmospheric modelling traditionally used to estimate acid loading in Europe does not operate at fine enough spatial scale to reflect near-coast or estuarine sea concentrations. Finally, decreasing anthropogenic sulphur emissions in recent years will have increased the evasion from the sea surface of biogenic sulphur.

Several projects investigated greenhouse gases arising from the fluvial system. EROS-21 picked out the region subject to the Danube river plume as an important source of coastal dissolved CH₄ concentrations, competing for primacy with seeps from geology and with methanogenesis in sediments (Amouroux et al., 2002). More specifically, it was found that bacteria near the river mouth, shallow water bottom sediments and delta regions have greatest significance for atmospheric CH₄ releases. Other gases involved in climate impacts include N₂O which is dependent on the anoxic conditions encountered in the Black Sea.

In the Mediterranean, different projects have pointed to the significance of river plumes for CH₄ and N₂O production. The role of the deltaic sediments is seen in the findings of the PHASE project that biogeochemical cycles below the euphotic zone are pelagically driven. The METROMED project focused on the characteristic of estuaries and coasts that they often have high loads of organic and mineral particles, and of inorganic nitrogen, providing ideal conditions for CH₄ and N₂O production. That such coastal environments represent a key combination of factors behind the generation of these gases is indicated by observed concentrations in the Bay of Lions/Rhone estuary and the Gulf of Thermaikos/Axios estuary, which were 10x and 100x those of open ocean waters. It is to be anticipated that gases may be produced in sediments and thence transported to the surface waters.

The BIOGEST project focused precisely on biogas generation, having phaeocystis sytems at its centre. Taking a coupled estuary ecosystem-atmosphere approach, it seeks to determine and evaluate the atmospheric fluxes of biogases in European estuaries and their impact on the global budget. The biogases CO₂, CH₄, CO, non-methane hydrocarbons, N₂O, NH₃, DMS, COS, halogenated organic compounds and biogenic volatile metals were evaluated in several European estuaries, showing that these environments are important marine sources of a wide range of gases linked to biological systems. Estuaries/coastal zones represent a key coincidence of factors arising from their morphometry, land use, drainage, population, industry, river discharge and climate, which lead typically to factors favouring biogas production (Cabeçadas et al., 1999). These include high organic/nutrient loading to estuaries, sediment oxygen depletion, methylation, and high productivity. The example of CO₂ in the Scheldt shows that whilst river systems might themselves transport appreciable quantities of biogas produced by soil and river respiration, heterotrophic activity in the estuaries dominates, being responsible for 90% of dissolved carbon dioxide (Borges & Frankignoulle, 2002b). This may be lower in more turbid less productive estuaries (e.g. The Gironde, Abril et al., 1999). Further, advective fluxes from the Scheldt to the estuary, and from the estuary to the North Sea were an order of magnitude smaller than estuarine water-air exchange. Thus, the estuary is annually a net source of atmospheric CO₂. Generalising beyond the Scheldt, estuaries are found to be supersaturated with carbon dioxide, and are estimated to be equivalent to 5-10% of European anthropogenic carbon dioxide release (Frankignoulle et al., 1998). This figure understates both the true anthropogenic influence, and the actual potential contribution of estuaries to total releases: not only is most respired labile carbon in polluted European estuaries itself anthropogenic in origin, but for the rest of the world where atmospheric anthropogenic emissions are lower and where significant fluvial organic carbon loading results from overpopulation, the estuarine contribution to the total will be higher.

Beyond investigation of estuarine and coastal carbon dioxide (Abril et al., 2000, Borges & Frankignoulle, 2002a, Borges & Frankignoulle, 2002b) studies were conducted of volatile metal(loid) species (Amouroux et al., 1998), volatilization of organotin compounds (Amouroux et al., 2000 & Tessier et al., 2002b), volatile selenium species and alkyl-iodides (Tessier et al., 2002a), alkyl-metal(loid) species (Tseng et al., 2000), mercury speciation in environmental matrices including estuarine muds (Tseng et al., 1998). Other biogeochemical trace contaminant atmospheric flows which may have regional and climatic roles and which can be influenced by the coastal environment are those of iodine and selenium. As well as local cycling, methyl iodine may operate as a replacement for methyl bromine, and estuaries may account for around 20% of European anthropogenic emissions (EROS-21). Organic substances can also be volatised from coastal environments, e.g. organotins from microbial/chemical methylation in anoxic estuaries. This is a contaminant with global scale cycling.

The land-to-sea transport of mercury has been discussed, and its volatile nature commented. Here the reverse sea-to-land transport is described. Simple observation data on the marine source can be found in the MOE project. Whilst essentially a model evaluation (Schmolke & Petersen, 2003) uses 3 coastal stations for the purpose, allowing areas supplying peak and background levels of observed atmospheric mercury to be estimated. Atlantic and Mediterranean waters are shown to be supersaturated with Hg and thus to emit mercury to the atmosphere (Reschke et al., 2002). This agrees with other opinions that air from the Atlantic and Mediterranean have elevated Hg content, whilst air further from oceans (e.g. over Scandinavia) have low levels (Wängberg et al., 2001). The implication is the potential for sea to land transfer within a coastal strip. Mechanisms are not fully quantified, but a possibility is the over-sea formation of particulate HgCl₂. Not only would this form experience higher rates of dry deposition through its large particle size, but scavenging by precipitation would also not be inconsiderable (Petersen et al., 2001). Coastal meteorology in particular could then be expected to be of consequence. A shift to more turbulent air flow as air masses make landfall would encourage dry deposition, and any increase in precipitation would enhance removal.