Atmospheric input of nitrogen into the North Sea: ANICE project overview

Continental Shelf Research 21 (2001) 2073–2094

Review article

Atmospheric input of nitrogen into the North Sea:ANICE project overview

Gerrit de Leeuwa,*, Leo Cohena, Lise Marie Frohnb, Gary Geernaertb, Ole Hertelb,Bjarne Jensenb, Tim Jickellsc, Laura Kleind, Gerard J. Kunza, Soren Lunde,Marcel Moermana, Frank M .ullerd, Britta Pedersenb, Knut von Salzend,f,

K. Heinke Schl .unzend, Michael Schulzg, Carsten A. Skj�thb, Lise-Lotte Sorensene,Lucinda Spokesc, Susanne Tammg, Elisabetta Vignatib,e,h

aTNO Physics and Electronics Lab., P.O. Box 96864, 2509 JG The Hague, NetherlandsbNational Environmental Research Institute, P.O. Box 358, 4000 Roskilde, DenmarkcSchool of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, UK

dMeteorologisches Institut, Universitat Hamburg, Bundesstr. 55, 20146 Hamburg, GermanyeRisø National Laboratory, P.O. Box 49, 4000 Roskilde, Denmark

fCanadian Centre for Climate Modelling and Analysis, University of Victoria, P.O. Box 1700, STNCSC, Victoria, BC,

Canada V8W 2Y2g IAAC, Universitat Hamburg, Martin-Luther-King Platz 6, 20146 Hamburg, Germany

hJoint Research Centre, Environment Institute, TP 460, 21020 Ispra, Italy

Received 11 April 2000; accepted 29 January 2001

Abstract

The aim of the atmospheric nitrogen inputs into the coastal ecosystem (ANICE) project is to improvetransport–chemistry models that estimate nitrogen deposition to the sea. To achieve this, experimental andmodelling work is being conducted which aims to improve understanding of the processes involved in thechemical transformation, transport and deposition of atmospheric nitrogen compounds. Of particularemphasis within ANICE is the influence of coastal zone processes. Both short episodes with high depositionand chronic nitrogen inputs are considered in the project. The improved transport–chemistry models will beused to assess the atmospheric inputs of nitrogen compounds into the European regional seas (the NorthSea is studied as a prototype) and evaluate the impact of various emission reduction strategies on theatmospheric nitrogen loads. Assessment of the impact of atmospheric nitrogen on coastal ecosystems willbe based on comparisons of phytoplankton nitrogen requirements, other external nitrogen inputs to theANICE area of interest and the direct nitrogen fluxes provided by ANICE. Selected results from both theexperimental and modelling components are presented here. The experimental results show the large spatialand temporal variability in the concentrations of gaseous nitrogen compounds, and their influences on

*Corresponding author. Tel.: +31-70-374-0462; fax: +31-70-374-0654.

E-mail address: [email protected] (G. de Leeuw).

0278-4343/01/$ - see front matter r 2001 Elsevier Science Ltd. All rights reserved.

PII: S 0 2 7 8 - 4 3 4 3 ( 0 1 ) 0 0 0 4 3 - 7

fluxes. Model calculations show the strong variation of both concentrations and gradients of nitric acid atfetches of up to 25 km. Aerosol concentrations also show high temporal variability and experimentalevidence for the reaction between nitric acid and sea salt aerosol is provided by size-segregated aerosolcomposition measured at both sides of the North Sea. In several occasions throughout the experimentalperiod, air mass back trajectory analysis showed connected flow between the two sampling sites (theWeybourne Atmospheric Observatory on the North Norfolk coast of the UK and Meetpost Noordwijk, aresearch tower at 9 km off the Dutch coast). Results from the METRAS/SEMA mesoscale chemistrytransport model system for one of these cases are presented. Measurements of aerosol and rain chemicalcomposition, using equipment mounted on a commercial ferry, show variations in composition across theNorth Sea. These measurements have been compared to results obtained with the transport–chemistrymodel ACDEP which calculates the atmospheric inputs into the whole North Sea area. Finally, the resultswill be made available for the assessment of the impact of atmospheric nitrogen on coastal ecosystems.r 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Atmospheric nitrogen; Aerosols; Gases; Assessment of atmospheric nitrogen loads; North Sea nutrientfluxes; Source–receptor relationships

1. Introduction

The European coastal zones are areas of great concern because of growing problems associatedwith increasing inputs of nutrients. These have resulted in a higher incidence of harmful algalblooms and other eutrophication phenomena and caused deleterious impacts on fisheries andtourism (Lancelot et al., 1989). To understand the impacts of increased nutrient additions to thecoastal area, the full magnitude and seasonality of nutrient inputs must be described, including theeffect of extreme events such as high atmospheric deposition episodes (Spokes et al., 1993, 2000)which, while small in overall annual budget terms, may be able to trigger algal blooms undernutrient depleted conditions in summer and early autumn. These blooms will be followed byoxygen depletion due to decay of the algae when the growth season is over.A significant fraction of the nutrients entering the coastal system consist of nitrogen

compounds which are introduced both as inorganic species and as dissolved organic nitrogen(DON). Biological effects of nitrogen inputs to marine ecosystems are extremely difficult topredict. The productivity of marine planktonic ecosystems is generally considered to be limited bythe availability of nitrogen (Kronvang et al., 1993). Terrestrial inputs of nitrogen are oftendominated by riverine transport, especially in the winter (Sanders et al., 1997). However,atmospheric deposition of nitrogen compounds has been shown to be a significant terrestrial inputto the coastal seas, especially in summer and fall. As an example, atmospheric nitrogen inputs intothe North Sea are dominated by nitrate and ammonium (Rendell et al., 1993). These inputs areoverwhelmingly anthropogenic and comparable in size to riverine inputs. The atmosphericcontribution to the total land based nitrogen input has been reported to be on the order of 30%for the total North Sea (North Sea Task Force, 1993). It is in the same order even in coastal,river dominated areas like the German Bight (Beddig et al., 1997) and still 5%, if the riverand run-off dominated Waddensee area is considered (Schl .unzen, 1994). Calculationspresented by Asman et al. (1994) show that ammonia deposition to the southern bight ofthe North Sea amounts to 7.6 kton Nyr�1. The atmospheric nitrogen load to the Danish part of

G. de Leeuw et al. / Continental Shelf Research 21 (2001) 2073–20942074

the North Sea has been estimated to about 46 kton Nyr�1 in 1998 or 0.93 ton Nkm�2 yr�1 (Skovet al., 1999).In addition, the atmospheric nitrogen load is directly available for algal growth, which makes

this contribution more significant because a large fraction of the river run off contributes nitrogenfixed to biological material and hence is not directly available. The contribution of atmosphericinputs to coastal eutrophication problems is now beginning to be recognised in a number of areasaround the world (Paerl, 1995). Given that atmospheric emissions of nitrogen compounds arecontinuing to increase world wide, this problem can be expected to grow.In the polluted atmosphere, nitrate exists as nitric acid and as fine mode (o1mm diameter)

ammonium nitrate aerosol. On mixing with marine air in the coastal zone, ammonium nitrateaerosols dissociate to liberate the precursor gases HNO3 and NH3. HNO3 is also produced fromNO2, a gas phase product from combustion processes that is converted into nitric acid at the rateof a few percent per hour. Hence HNO3 is both produced in the marine atmosphere and removedby reaction with sea salt aerosol:

HNO3ðgÞ þNaClðsÞ-NaNO3ðsÞ þHClðgÞ:

Sea salt aerosol is produced from bubbles that are entrained in the water during wave breakingand burst after rising back to the water surface. Sea spray mass (or surface area, which is animportant parameter for this reaction) occurs predominantly in the coarse aerosol mode (>1mmdiameter). Hence also the NaNO3 particles occur in the coarse mode and due to the describedprocesses nitrate shifts in the marine atmosphere from the fine mode to the coarse mode. This seasalt displacement reaction not only influences dry deposition of nitrate through competitionbetween gravitational settling of large particles and gaseous deposition, but also increases theefficiency of precipitational scavenging via inertial impaction, a process strongly dependent onparticle size. An important sea spray source is the surf zone, where sea spray concentrations maybe enhanced by 1–2 orders of magnitude with respect to those at open sea, even in off-shore winds(De Leeuw et al., 2000a, b). Reactions involving sea salt aerosol will therefore be focussed in thecoastal areas.Estimates of nitrogen inputs are often based on measurements over land and sometimes on

measurements at open sea. Coastal effects on the total nitrogen depositions are included in asimplistic way, if at all. However, the largest changes in both the physical and chemical propertiesof an air mass advected from land over sea, and therefore also in the resulting processes, areexpected in coastal regions. Since continental sources of atmospheric nitrogen species dominate,sharp gradients in concentrations and associated fluxes are expected across the coastal zone.Reduction of ammonia concentrations in the German Bight by a factor of 4 after 70 km transportout to the sea has been observed (Plate et al., 1995), while 40% of the sea salt is transformed tosodium nitrate (Schulz, 1996). Little is known about the relative importance of the competingeffects of dilution and entrainment, deposition, chemical reactions and re-emission from the seasurface for the total atmospheric input to regional seas. These processes may significantly changewith distance from the coast due to changes in surface properties such as wave properties androughness length, effects of solar radiation on heat transfer, etc., which in turn change theatmospheric mixing properties and air–sea exchange processes.The sudden change in the surface properties at the coastline leads to a non-uniform situation.

As a consequence, the horizontal homogeneity condition on which current description of the

G. de Leeuw et al. / Continental Shelf Research 21 (2001) 2073–2094 2075

boundary layer physics is based, does not strictly apply. As a semi-enclosed basin, the North Seacontains surface waves that are governed by fetch-limited conditions, and strong step changes inroughness and stability. Most atmospheric models assume local horizontal homogeneity (e.g.,Schl .unzen and Pahl, 1992). A new model has been constructed that takes local inhomogeneitiesinto account (Geernaert and Astrup, 1999). Because the physical processes involved in gasexchange are related to the same physical processes governing momentum exchange, similarity inthe range dependent patterns of momentum and gas fluxes can be anticipated. However, whilemomentum and heat fluxes can generally be regarded as ‘‘conservative quantities’’, gaseousnitrogen compounds are often so reactive that source and sink terms must be taken into accountto evaluate their fluxes. Source and sink terms depend on governing chemical reactions, which inturn are dependent on temperature, humidity, and the reaction time scales in relation to thehorizontal and vertical mixing rates. Transport issues that may be important are, for example, theformation and development of an internal boundary layer and associated near-surface mixing,and changes in the boundary layer structure and their influence on transport of gases and aerosolsboth throughout the boundary layer and across the inversion at the top.

2. The ANICE project

Atmospheric Nitrogen Inputs into the Coastal Ecosystem are addressed in the ANICE project.ANICE focuses on quantifying the deposition of atmospheric inputs of nitrogen compounds(HNO3, NO3

�, NH3 and NH4+) into the sea, both near the coast and in open water, and the

governing processes. A small effort aims at the determination of the DON in rain and aerosol. Thesouthern North Sea is studied as a prototype. Because the physical and chemical processes aredescribed, as opposed to empirical relations, the results can also be applied to other regional seaslike the Mediterranean, the North Atlantic continental shelf area and the Baltic.The aim of the ANICE project is to improve transport–chemistry models that estimate nitrogen

deposition to the sea. To achieve this, experimental and modelling work is being conducted whichaims to improve understanding of the processes involved in the chemical transformation,transport and deposition of atmospheric nitrogen compounds. Of particular emphasis withinANICE is the influence of coastal zone processes. Most current models use grids that are toocoarse to describe the governing processes with sufficient accuracy, particularly in coastal regions.The changes in concentrations and fluxes across the coastal zone must be taken into accountbecause of the short residence time of gaseous species such as NH3 and HNO3. Due to theirreactivity and due to scavenging by sea salt aerosol, a significant fraction of these gases may beremoved from the atmosphere in the first 10–20 km. Not only does this affect the atmosphericinputs of nitrogen near the coast, but the rapid removal of the gaseous species will also result indecreasing concentrations with distance off the coast, and thus the atmospheric inputs to openwaters.Large deposition of nitrogen may occur during short periods. Both such short lived extreme

events and chronic nitrogen input are considered in the project. Assessment of the impact ofatmospheric nitrogen on coastal ecosystems will be based on comparisons of phytoplanktonnitrogen requirements, other external nitrogen inputs to the ANICE area of interest and the directnitrogen fluxes provided by ANICE.


Experimental work within ANICE consists of a long-term observational programme usingscientific equipment mounted on commercial ferries, complemented by two intensive fieldexperiments that focus on process studies. The experimental results provide data to evaluatemodel results through case studies with connected air flow between the UK coastal site and theNorth Sea scientific platform near the Dutch coast (a true Lagrangian situation). They also,through the long-term monitoring from ferrys, provide data from the open North Sea. Theintensive field experiments were designed to study relevant processes that should be included inthe models and provide information on the concentrations in air and water and their spatial andtemporal variability including the variation of atmospheric concentrations with fetch.An interesting feature of the ANICE project is the use of two atmospheric chemistry transport

models, ACDEP (Hertel et al., 1995) and METRAS (Schl .unzen, 1990; Schl .unzen et al., 1996),both with their respective aerosol modules, as described below. ACDEP is used to estimateatmospheric inputs of nitrogen to the whole North Sea, integrated over periods varying from 6hto a year. ACDEP also provides the ‘large’ scale lateral boundary conditions for the calculationsin the coastal model domain using METRAS. METRAS in turn, will be used for studyingscenarios for specific days, to calculate the atmospheric nitrogen input to coastal waters with ahigh resolution in space (some 100m) and time (minutes). The combined modelling effort isexpected to lead to a major improvement in the estimate of atmospheric inputs into the NorthSea, which can subsequently be used in effect studies. The models are complementary because ofthe different scales, the different mixing schemes and the different initialisations.

3. Modelling

ACDEP is a Lagrangian transport–chemistry model that is used for a coarse gritted mapping ofnitrogen deposition to the North Sea. The model was developed within the Danish MarineResearch Programme with the aim of describing deposition to Danish waters (Hertel et al., 1995).The chemical module is a modified Carbon Bond Mechanism IV with 37 compounds and 80chemical reactions. The model accounts for emissions, advection, diffusion, dry and wetdeposition. A special module is developed to describe dry deposition to water surfaces. The resultsinclude concentrations, dry and wet deposition of gaseous as well as particulate nitrogencompounds, with a time resolution of 6 h. The calculations are presently based on EMEPmeteorological data. Compounds in the aerosol phase are treated as gases except for thedescription of dry and wet deposition. For these processes all compounds in the aerosol phase areassumed to be associated with aerosol particles with a diameter of 0.8mm. One of the goals of theANICE project is the improvement of the description of aerosols and heterogeneous processes inACDEP. This is accomplished through the development of an aerosol module in which theparticles are discretised in 10 size classes, and the aerosol is assumed to consist of internally mixedparticles containing the relevant ions. The starting point is a 1-D model for the coastal zone(Vignati, 1999). The aerosol module accounts for emission, diffusion and deposition of theaerosols. Equilibrium with water vapour is assumed. Condensation/evaporation of gaseouscompounds will be included in the module.High spatial and temporal resolution emission data used in the ACDEP calculations were

obtained from the EUROTRAC GENEMIS programme. Data are given on a 16.67� 16.67 km2


grid for the EMEP area, as a true sub-grid of the EMEP grid of 50� 50 km2. The ACDEPmodel will provide results for a receptor net with a 30� 30 km2 grid mesh for the years 1998and 1999. Furthermore the model will be used for studies of source–receptor relationships toevaluate the impact of various source regions and the effect of potential emission reductionstrategies.The mesoscale transport and chemistry model METRAS is applied to simulate coastal

atmospheric phenomena including nitrogen transport and the composition of coastal aerosol.METRAS is a multi-layer fluid dynamics, dispersion and chemistry model which can generatethree-dimensional meteorology fields. This model is coupled with the chemical transport model(MECTM) which includes the complex RACM gas phase chemistry (Stockwell et al., 1997) andthe aerosol model SEMA (Von Salzen, 1997, Von Salzen and Schl .unzen, 1999a, b). SEMAcombines a thermodynamic equilibrium approach with a kinetic approach to predict thecondensation and evaporation at the aerosol surface. SEMA uses a sectional algorithm approachto solve the resulting equations (Von Salzen and Schl .unzen, 2000).METRAS is initialised and forced at the lateral boundaries using the output of the German

Weather Forecast Services’s model, the Deutschland Model (DM) (DWD, 1998). Emission datafor the model domain are provided by the EUROTRAC2 project GENEMIS on an 8� 8 km2 gridfor the area of the southern North Sea. The initial data for concentrations and the lateralboundary values are taken from ACDEP results.

4. ANICE experiments

The ANICE experimental efforts consisted of long-term monitoring spanning a total periodof about 1.5 yr, including several extended down periods, and two intensive observationalperiods (IOPs) involving sites on the UK coast and near the Dutch coast. The second IOP alsoinvolved water sampling in the area between Meetpost Noordwijk (see below) and the Dutchcoast.

4.1. Long-term observations

A system for automatic sampling of rain water and aerosols was developed for deploymenton a commercial ferry travelling between Hamburg and Harwich and in summer also betweenHamburg and Newcastle. The system consists of an automated rain sampler and analyserfor major ions, pH and conductivity to obtain an accurate estimate of the wet depositioncontribution across the North Sea. Concentrations of gaseous NH3 and HNO3 were continuouslymeasured, while simultaneously aerosols were sampled on filter packs to determine nitrateand ammonia concentrations. An optical particle counter was used to continuously measureaerosol particle size distributions, which are important parameters for aerosol transport,deposition and heterogeneous processes involving aerosols. Concentrations of NOx and NOy werealso continuously measured. Support measurement included a meteorological station providingwind speed and wind direction, and air temperature. GPS was used to determine samplinglocations.


4.2. Intensive observational periods: process studies and model validity assessment

Intensive 3-week field campaigns were undertaken in June 1998 (IOP1, focussing on HNO3/NO3

� relations) and in August 1999 (IOP2, with emphasis on processes related to NH3). Data werecollected at the Weybourne Atmospheric Observatory (WAO) in North Norfolk, at the east coastof the UK, and on Meetpost Noordwijk (MPN), a research tower at 9 km from the Dutch coast.During each IOP, concentrations of HNO3 and NH3 were measured at both stations, together

with aerosol chemical composition using filter packs and concentrations of relevant ions in rainwater. Size segregated aerosols were sampled using a HiVolume aerosol sampler with sizesegregation using Sierra impactors. The aerosol and rain water samples were analysed for majorions including NO3

�, NO2�, SO4

2�, Cl�, NH4+, Na+, K+, Ca2+ and Mg2+. By analysing both

aerosol and rainwater, scavenging ratios can be estimated which, although controversial, do offerone mechanism of characterising the efficiency of atmospheric wet deposition. Also other relevantgases, aerosol particle size distributions, and a suite of meteorological and oceanographicparameters were measured.

4.2.1. UK coastal site, Weybourne Atmospheric Observatory, North NorfolkThe UK coastal site is the WAO (521 570 2500 N, 0011 070 4000 E), immediately on the coast of

Northeast Norfolk (UK). The observatory is specifically equipped as a specialist atmosphericsampling site with routine sampling of meteorological and some atmospheric chemistryparameters. WAO is maintained by the University of East Anglia.In addition to the analyses indicated above, selected samples collected at the WAO site were

also collected for DON determination and for stable nitrogen isotope analysis. This data shouldallow various nitrogen sources to be distinguished since, for example, marine emissions ofammonia have a distinctly different nitrogen isotope signature to ammonia arising from animalwastes. Such knowledge of sources is of importance because it will be necessary to determinewhether changes occurring in an air mass travelling over the southern North Sea reflect onlyinternal atmospheric processes or addition of nitrogen species from other sources. The results ofthese studies will be presented elsewhere.

4.2.2. Meetpost NoordwijkMeetpost Noordwijk (MPN) is a research tower in the North Sea (position 521 160 25.900 N;

041 170 45.800 E), at 9 km from the Dutch coast near the village Noordwijk. The tower is equippedwith a 20-m long boom at the west side of the tower, which allows for unperturbed (by the tower)turbulence measurements. This boom was used for measurements of fluxes of HNO3 (IOP1) andNH3 (IOP2) using the eddy accumulation technique. Fluxes of CO2 were measured as an inerttracer, i.e. CO2 fluxes are not influenced by reactions with sea spray aerosol. Also concentrationsof O3, NO and NO2 were measured.In addition to the measurements indicated above, particle size distributions were measured

using optical particle counters, and the vertical variation of particles larger than about 13 mm wasmeasured using Rotorod impaction samplers. Together, these data allow for determination of thescavenging of HNO3 and NH3 by heterogeneous or homogeneous reactions. By combination withsimultaneously measured bubble size distributions, information was obtained on the aerosolsource function and thus the availability of fresh aerosol surface.


LIght Detection And Ranging (Lidar) measurements provided information on the mixing layerdepth, the boundary layer structure and on mixing processes in the marine atmospheric boundarylayer. In support, meteorological parameters were measured with local sensors and data fromnearby monitoring stations were also collected.

5. Results

Experimental and model results show the distributions of nitrogen compounds across thesouthern North Sea, including the gradients near the coast and the changes in gaseous andparticulate concentrations in an air mass travelling across the North Sea. Model results providethe deposition into the entire North Sea. Selected results are presented below.

5.1. Gas transfer

The air–sea exchange processes are illustrated in Fig. 1, where the concentrations of NHx in airand in water are plotted for 23–27 August 1999. The measurements were made on and aroundMPN. In easterly winds with a short fetch to the Dutch coast, the concentrations were often high;in westerly winds with long fetch the concentrations were low. The concentrations in the waterand the atmosphere exhibited large temporal and spatial variations, with values from 0.05mgm�3

to more than 10mgm�3 in the atmosphere and water concentrations between 100 and 900 nMol

Fig. 1. Concentrations of NHx(aq) in water in nMol (1 nMol=14mgm�3) and concentrations of NH3(g) in theatmosphere in mgm�3. NHx(aq)CAL is the equilibrium concentration in water calculated from the atmosphericconcentration (in equilibrium there will be no vertical flux of NH3). The horizontal scale gives the time, starting on 23

August at 00:00, ending on 27 August at 24:00; each tic (and vertical line) indicates the start of a new day.


(1 nMol=1nmol l�1=14mgm�3). Also shown are the concentrations in water that would havebeen observed if the atmospheric and water concentrations would have been in equilibrium. Inthat case, there would be no net gas transfer. In all other cases, a net flux of ammonia would beestablished depending on the partial pressure difference between both phases. The flux depends onthe solubility of the gas, as well as on the horizontal homogeneity of the concentrations in air andin water. To illustrate the effect of these two parameters, results are presented from calculationsmade using the data of 26 August 1999. The atmospheric concentrations during this day were lowdue to westerly winds (long fetch) and also the water concentrations varied. Neglecting thesevariations and assuming total solubility for NH3 in water, the mean flux integrated over the daywould have been –0.217mg NH3m

�2 day�1. With the actual water concentrations the flux wouldbe 40% smaller, accounting for horizontal gradients as well as a reduction in the flux by 60%.

5.2. Near-coastal variations and fetch effects

An example of the variation of atmospheric concentrations of NH3 across the North Sea,showing the spatial variability on a larger scale, is shown in Fig. 2. The measurements were madeon the ferry sailing from Hamburg to Harwich and back, from the 30th of August until the 2nd ofSeptember 1999. The NH3 concentrations are very high close to the coast, especially nearHamburg which is situated about 80 km inland. Here concentrations of up to 5 mgm�3 wereobserved. The concentrations decreased rapidly as the ship crossed the North Sea. The influenceof the Westfrisian Islands is clearly visible, but at open sea the concentrations were similar tothose measured at MPN in on shore winds. Near the UK coast the concentrations increased

Fig. 2. Concentrations of atmospheric NH3 over the North Sea measured on the ferry sailing from Hamburg toHarwich and back.


again. During the return to Hamburg, similar observations were made, although this time noinfluence of the Westfrisian islands was observed since they were downwind from the ship.The data on the ferry were collected along fixed ferry routes, i.e. usually not along the direction

of the wind. Observed variations often result from localised emissions ashore upwind from theferry route, rather than from transport phenomena and heterogeneous processes alone.Nevertheless, the results in Fig. 2, as well as observed variations in the gas concentrations inchanging wind directions on MPN, suggest a strong dependence on fetch. A dedicated researchvessel was not available for ANICE, and detailed measurements of the fetch dependence of gas oraerosol concentrations close to the coast line could not be made. Therefore the effect of fetch canonly be illustrated with model results. Calculations were made with the coastal aerosol model(CAT) (Vignati et al., 2001), over fetches of up to 25 km. CAT was initialised with sea sprayconcentrations generated in the surf zone that were assumed to be emitted in a boundary layerwith a homogeneous distribution of continental background aerosol. During transport over seathe aerosol concentrations changed dynamically due to dry deposition, vertical mixing and surfacegeneration of sea spray. To calculate the nitric acid concentrations as a function of fetch, CATwas extended with a chemical scheme describing the formation of sea spray, the transport and drydeposition of nitric acid, and the reaction between nitric acid and sea spray. At the coastline thenitric acid profile was initialised with a constant value of 200 ppt (0.5 mgm�3). Results presented inFig. 3 show the variations in both the HNO3 concentrations and their gradients with fetch. Thehigh sea spray concentrations over the surf zone cause an immediate effect when the air masspasses the coast line. In the further evolution of the gas profile the effect of uptake by sea sprayaerosols is most evident at the lower levels where the concentrations of the particles are highest.Diffusion distributes the particles homogeneously in the vertical, causing a homogeneous uptakein the whole column. At a fetch of 25 km the profile becomes almost uniform except very close tothe surface.

Fig. 3. Modelled concentration profiles of nitric acid for various distances from the coast.


5.3. Aerosol processes during transport across the North Sea: connecting air flow

The effect of transport over longer distances across the North Sea is illustrated withmeasurements of aerosol chemical composition at the Weybourne site and at MPN, which arefurther simulated with model calculations. Fig. 4 shows the measured concentrations of nitrateand ammonium in the particulate phase, measured during IOP1 at both the sites in June 1998. Theconcentrations of both species fluctuate in time by about two orders of magnitude. Temporalpatterns have been interpreted using air mass backward trajectories. The very low levels of bothnitrate and ammonium seen on the 12th of June are associated with very clean northerly airoriginating in the Arctic region. This is confirmed by the observation of high concentrations ofmarine derived sodium, chloride, magnesium and sea salt sulphate during this period. The highconcentration event centred on the 20th and 21st of June is associated with flow off the Europeancontinent with air passing over the Netherlands prior to arrival at the MPN platform and thenacross the North Sea to Weybourne on the UK coast. High levels of nitrate and ammonium areseen in the same air mass as the source region is both highly populated and industrialised as wellas being intensively farmed. Furthermore, the concentrations of both species are significantlyreduced during passage over the North Sea. The reduction is largest for nitrate which, as a resultof its significant coarse mode component (>1 mm diameter), is more efficiently dry deposited thanammonium which occurs mainly in the fine aerosol fraction (o1 mm diameter).By looking at the size distribution of nitrate (Fig. 5), evidence can be seen for the sea salt

displacement reaction. Under the polluted conditions around the 21st of June, nitrate shows acharacteristic bimodal distribution withB45% of the nitrate in the coarse mode fraction. The finemode nitrate size distribution correlates well with the ammonium size distribution suggesting that

Fig. 4. Comparison of aerosol nitrate and ammonium concentrations at WAO and MPN during the first ANICE fieldcampaign.


ammonium nitrate is a significant fraction of the fine mode nitrogen budget. For the coarse mode,nitrate and sodium size distributions correlate well suggesting that sodium nitrate is a componentof the coarse mode aerosol. However, there is insufficient sodium for this to be the only coarsemode nitrate species. Pakkanen (1996) showed that in continental air masses nitric acid can reactwith soil derived coarse mode calcium carbonate:

2HNO3ðgÞ þ CaCO3ðsÞ-CaðNO3Þ2ðsÞ þH2OðaqÞ þ CO2ðgÞ

again shifting nitrate from the fine to the coarse aerosol mode. Insufficient coarse mode calcium is,however, present in these samples to fully explain the presence of coarse mode nitrate. Previousstudies in the coastal zone have shown the presence of significant quantities of coarse modeammonium (Spokes et al., 2000). To account for its presence, Yeatman (2000) proposed thatdissolution and coagulation reactions involving NH4NO3 and sea salt result in the formation ofboth coarse ammonium and nitrate. Sufficient coarse mode ammonium is present in these samplesto account for the excess coarse nitrate over that formed through the sea salt displacementreaction. However, in the present case of the 12th of June 1998, with clean marine conditions, theconcentrations are too low for significant coagulation and the sea salt displacement reactionresults in almost complete conversion of nitrate from the fine to the coarse aerosol mode withexcess coarse sodium remaining.Another episode of connected air flow occurred on the 16th of June 1998, in this case from

WAO to MPN (Fig. 6). Indeed, concentrations of ammonium at MPN were significantly lowerthan those at WAO, but for nitrate the concentrations were rather similar. This situation has beensimulated with the SEMA model, using a Lagrangian approach. The model was initialised withconcentrations observed at WAO (Fig. 7a). The concentrations at the downwind station MPNwere then prognostically calculated (Fig. 7b), by taking into account condensation and

Fig. 5. Comparison of aerosol nitrate, sodium and ammonium size distributions (represented as the massconcentrations measured in each size fraction) under clean marine and polluted continental conditions during thefirst ANICE field campaign.


evaporation and diffusive growing of aerosol particles. An empirical deposition/sea saltgeneration function was included to simulate what may have been happening over water. Thereis no gas phase chemistry in the box-model. The meteorological parameters were held constantthroughout the model run.The box-model was integrated for 12 h, a period similar to the air mass travel time

between monitoring stations. The resulting concentrations and size dependent compositionsare presented in Fig. 7c. Both the SEMA results and the data measured at MPN (Fig. 7b)show the bi-modal distribution, but with slightly increased intensity in the modelresults. Comparing the general trends in the reactions between the measured and modelled,it appears that SEMA box-model handles the chloride reactions quite well. The modelresults show more Na+ than measured, mainly because the only cat ion represented isNa+, the missing Mg2+, Ca2+ and K+ are replaced by Na+. NO3

� is underestimated bythe model, most likely because photochemical production of NO3 is not included as indicatedabove.

Fig. 6. Backward wind trajectories for 16 June 1998 showing connecting flow between the two monitoring stations. The24 h trajectories are plotted every 3 h for both surface (solid lines) and geostrophic (dashed lines) winds. The length of

the trajectory represents the distance the air parcel has travelled during the 24 h previous to the times identified in theupper left hand corner of the figure.


5.4. Spatial distributions in the southern North Sea

Concentrations of gaseous species NH3 and HNO3 and their particulate forms NH4+ and NO3

�

across the North Sea were experimentally determined on the ferry. The ferry measurements wereintegrated over the five transects indicated in Fig. 8, where the two transects of the German Bight(DTB) were taken together into a single data point. At the transition from one transect to

Fig. 7. (a) Measured concentrations at WAO, from 8:02 to 16:02 h 16 June 1998; these measurements were redistributedto create an input for the SEMA box model. (b) Measured concentrations at MPN, from 21:38 h 16 June 1998 to12:06 h 17 June 1998; these measurements should have captured some of the air mass measured at WAO assuming

connected flow and are used as comparison data set for SEMA model results. (c) SEMA model results for MPN withinput data from (a), using a relative humidity of 95%, temperature of 151C and completing the ion budget with Na+.


another, as indicated by a GPS receiver, valves were automatically switched to start a new sample.Fig. 8 shows an example of the results, i.e. the distributions of NH3/NH4

+ and HNO3/NO3� in the

five sectors during two sampling periods indicated by P1 and P2 (see legend for identification).The concentrations were estimated from the filterpack sampling and are in good agreement withthose from simultaneous measurements in WAO and on MPN (see Fig. 4). P1 shows higherconcentrations especially for the two southern sectors. During P2 the concentrations measured inthe two southern sectors were very low in comparison with P1.Fig. 9 shows corresponding results from 14 five sector measurements (for the designation of the

sectors see Fig. 8). The distributions of the concentrations are given as notch boxes with the shorthorizontal line in the middle of the box representing the median, the highest and lowest horizontallines showing extreme values and finally the upper and lower part of the box representing the 75%and 25% percentiles. These sampling periods span the time from 26 May 1999 to 13 September1999. Ammonia and nitric acid concentrations were obtained with denuders. The salts have beencalculated from the particulate matter sampled on a Teflon filter. It is remarkable that theaveraged ammonia concentrations in the middle of the North Sea (MIN) are only slightly lowerthan the concentrations in the more coastal influenced northwest sector (NWN). For the small(NH4)2SO4 particles we find an increase of concentration from the coast to the open sea(NWN>MIN), which can be explained with long-range transport of these particles.

Fig. 8. Filter pack and denuder results from the ferry sampling during IOP 1 (1998). The sampling periods P1 and P2are indicated in the legend.


Fig. 9. Results from 14 ferry measurements during 1999. The ammonia and nitric acid were determined using denuders

and the salt concentrations were calculated from measurements of the particulate matter using filter packs.


Formation of nitric acid takes place throughout the column from the NOx-pool. The slowdegradation of the NOx-pool is probably responsible for the high HNO3 concentration in opensea areas. In contrast, the main source for ammonia is from farming activities over land. Thus, asindicated in Fig. 2, ammonium concentrations in the air over sea are high at short fetchesdownwind from land masses, and decrease fast with fetch. This could be a reason for the differentspatial patterns of nitric acid and ammonia. Over the North Sea there is in general not asignificant source of ammonia over sea. Rather, the North Sea is a sink due to efficient deposition.It is noted that in general the production of ammonia cannot be neglected. Emission of ammoniafrom the oceans contributes 15% of the total (anthropogenic plus natural) and almost 50% of thenatural sources (Seinfeld and Pandis, 1998).

5.5. Calculation of nitrogen deposition to the entire North Sea

One of the final results from the ANICE project will be an estimate of the atmospheric inputsinto the North Sea using the modified ACDEP model. A first estimate of the annual nitrogenloads of the North Sea was calculated with ACDEP, using the simplified aerosol descriptionindicated in Section 3. The calculations were performed for 1117 receptor points in the30� 30 km2 receptor net. Computed annual mean ammonia concentrations (not shown),reflecting mainly the emission areas over the Netherlands and Great Britain, are qualitativelyin reasonable agreement with the 1999 ferry measurements. Generally the concentrations rangebetween 0.1 and 1mgm�3 with highest concentrations in the coastal regions of England and theNetherlands/Germany. The annual nitrogen deposition was calculated from the loads and ispresented in Fig. 10. The deposition ranges between 0.3 and 1.3 ton Nkm�2. The local maximaclose to the Norwegian coast are caused by relatively large amounts of precipitation in the area,causing wet deposition.

6. Discussion

The presented brief overview of the initial results of the ANICE project supports the primaryhypothesis on the importance of coastal effects on nitrogen inputs to the regional seas. Gaseousnitrogen compounds are primarily produced over land and very high concentrations wereobserved close to the coast in off-shore winds. The ANICE experimental results indicate that theseconcentrations decrease rapidly with increasing fetch resulting in a reduction to ‘background’levels when the air mass is transported across the North Sea, over a distance of only about200 km. The gases are highly soluble and are therefore either directly deposited to the surface ortaken up by aerosols where they are accommodated through chemical reactions. Evidence for thelatter effect is presented. The reactions with aerosols obviously affect the gaseous air–sea fluxes asevidenced from the shape of the gas concentration profiles (Geernaert et al., 1998). The aerosoldry deposition flux is in part due to different physical processes and depends on particle size, andthus the dry deposition velocities for gases and particulate nitrogen compounds are different.Moreover, the direct gas fluxes are determined by the partial pressure difference of the gaseousspecies in the water and in the air directly above the water, both of which have been observedduring the ANICE experiments to vary strongly in both space and time. In model calculations


Fig. 10. Annual nitrogen deposition (ton Nkm�2) to the North Sea computed for the year 1998. Baseline calculation isperformed with ACDEP.


such variations, especially those in the sea, are usually not taken into account. Examples presentedin Section 5.1 show that neglecting spatial variations may lead to significant overestimation of drydeposition of NH3.Both NH3 and HNO3 play a role in heterogeneous chemistry, as discussed in Section 1. In

particular HNO3 reacts directly with sea spray aerosol and evidence of the sea spray replace-ment reaction is presented. Current atmospheric chemistry transport models do not includesuch reactions. Work is underway in several groups to implement heterogeneous chemistryinvolving HNO3. In ANICE this occurs through the implementation of heterogeneousreaction schemes in both METRAS, using the SEMA model, and in ACDEP using a 1-Dmodel. The latter is in part based on the model of Vignati (1999), which was applied to show thelarge influence of the reaction between nitric acid and sea spray on both the concentrations andthe gradients of HNO3, over a fetch of only 25 km. The strong variations imply that a coarse gridmodel will not be able to correctly predict the effect of coastal processes at very short fetches onthe nitrogen input, unless these processes are included through a coastal sub-grid or a properparameterisation.A detailed discussion of other effects of coastal processes, briefly indicated in Section 1, has not

been presented in this brief overview paper. In the ANICE project, several efforts have beenundertaken to study such effects, both experimentally and through modelling (De Leeuw et al.,2000a, b). The experimental efforts included simultaneous measurements of bubbles generated inthe water by breaking waves and aerosols just above the water surface. The comparison of therespective size distributions leads to conclusions on the bubble-mediated aerosol source functionand the amount of fresh sea spray available for heterogeneous reactions. These small scalemeasurements are supported by lidar measurements on atmospheric boundary layer propertiesthat influence the atmospheric mixing properties of both gases and aerosols, as well as aerosolplumes generated by waves breaking at the sea surface or in the surf zone. In off-shore winds,plumes of surf generated sea spray aerosol were observed to be transported out over the sea andwere visible over distances of at least 5 km, in good agreement with model calculations using CAT(Vignati, 1999). This confirms that also in off-shore wind large amounts of sea spray are availablefor heterogeneous reaction, at least in the study area along the Dutch coast. Other experimentalefforts included micro-meteorological studies on fluxes of NH3, HNO3, CO2, as well asmomentum, heat and water vapour, which are used in studies on air–sea exchange processes. Themodelling effort includes the construction of a model that accounts for effects of non-homogeneityacross the coast line (Geernaert and Astrup, 1999), as discussed in Section 1.Effects on regional scales were addressed through measurements at the WAO site on the UK

coast and at MPN near the Dutch coast, as well as through ferry measurements. The largeconcentration differences measured on the ferry during different periods show the necessity oflong-term measurements to establish a representative data set on nitrogen inputs into the southernNorth Sea. The analysis of cases with connecting air flow between the WAO site and MPN showsthe change in aerosol particle size distributions and chemical properties. One of these cases wasmodelled using the SEMA box model in a Lagrangian sense. The encouraging results giveconfidence on the use of SEMA in METRAS for fine gritted estimates of the nitrogen inputs intothe southern part of the North Sea. METRAS prognostically calculates the three-dimensionalfields of wind, temperature, humidity, cloud and rain water content and derives exchangecoefficients from first order closure theory. This scheme has been thoroughly tested with


favourable results. Finally, the ACDEP model extended with a detailed aerosol module will beused to calculate the nitrogen inputs for the whole North Sea. The model results will be evaluatedusing the ferry experimental data. An example of the expected output is presented in Fig. 10,which shows a baseline calculation performed with the model before implementation of aerosolmodule and fine resolution emissions.The updated atmospheric chemistry transport models will be used for the assessment of the

impact of atmospheric nitrogen on coastal ecosystems. Both episodic and chronic nitrogen inputswill be considered and the assessment will be based on comparisons of phytoplankton nitrogenrequirements, other external nitrogen inputs to the area of interest to ANICE and the directnitrogen fluxes provided by ANICE.In the study presented above, only inorganic nitrogen has been considered. A limited effort is

made in the ANICE project to characterise DON in rain and in aerosols. DON species areubiquitous in rain and aerosols and comprise between B20% and 80% of total dissolvablenitrogen. Further analysis is currently underway to quantify these species. This data will bothincrease our knowledge of the size distribution of this, largely chemically uncharacterised,material and improve our estimates of the total flux of combined nitrogen species to the surfacewaters of the North Sea.

Acknowledgements

ANICE is part of ELOISE and is supported by EC DG XII, contract ENV4-CT97-0526, as wellas by internal funding of participating institutes. Meetpost Noordwijk is owned and maintainedby the Dutch Ministry of Public Works. The co-operation of Rijkswaterstaat, DirectoraatNoordzee for making available Meetpost Noordwijk, and in particular the co-operation of theplatform managers, Guus Goossens and Jaap van der Horn, and their crews is very muchappreciated. The Weybourne Atmospheric Observatory is run by the University of East Angliaand the help of Prof. Stuart Penkett at UEA is acknowledged here. Scandinavian Seaways kindlygave the permission to mount scientific equipment on board the ferry ‘‘Prince of Scandinavia’’ forthe long-term measurements that started in May 1998 and ended early 2000. The NoordwijkseReddingsbrigade provided their life boat as well as able seamen for water sampling in the areabetween the Dutch coast and MPN. Dr. Uwe Schwarz, Stefan Reis and Dr. Burkhard Wickert atStuttgart University provided the high resolution emission data that are used in the ACDEP andthe METRAS models. The meteorological data used in ACDEP were obtained from Egil St�renfrom the EMEP MSC-W at the Norwegian Meteorological Institute and the data for METRASfrom the German Weather Service, Offenbach.

References

Asman, W.A.H., Harrison, R.M., Ottley, C.J., 1994. Estimation of the net air–sea flux of ammonia over the southern

bight of the North sea. Atmospheric Environment 28, 3647–3654.Beddig, S., Brockmann, U., Dannecker, W., K .orner, D., Pohlmann, T., Puls, W., Radach, G., Rebers, A., Rick, H.-J.,

Schatzmann, M., Schl .unzen, K.H., Schulz, M., 1997. Nitrogen fluxes in the German Bight. Marine Pollution

Bulletin 34, 382–394.


De Leeuw, G., Kunz, G.J., Moerman, M., Cohen, L.H., Schl .unzen, K.H., Klein, L., M .uller, F., von Salzen, K., Lenz,C.-J., Schulz, M., Tamm, S., Plate, E., Geernaert, G., Hertel, O., Vignati, E., Frohn, L., Pedersen, B., Jensen, B.,

S�rensen, L.L., Lund, S., Jickells, T., Spokes, L., 2000a. Atmospheric nitrogen inputs into the coastal ecosystem(ANICE) ENV4-CT97-0594. Second Annual Report (February 1, 1999–January 31, 2000), Report FEL-00-C125,TNO Physics and Electronics Laboratory.

De Leeuw, G., Neele, F.P., Hill, M., Smith, M.H., Vignati, E., 2000b. Sea spray aerosol production by waves breakingin the surf zone. Journal of Geophysical Research 105 (D2), 29,397–29,409.

DWD, 1998. Quarterly report of operational NWPFmodel of the Deutscher Wetterdienst, Business Area Research

and Development, Deutscher Wetterdienst, Offenbach a.M., Germany, 62pp.Geernaert, G.L., Astrup, P., 1999. Wind profile, drag coefficient, and radar cross section in the coastal zone for quasi-

homogeneous conditions. In: Banner, M. (Ed.), The Wind Driven Air–Sea Interface. University of New SouthWales, pp. 305–316.

Geernaert, L.L., Geernaert, G.L., Granby, K., Asman, W.A.H., 1998. Fluxes of soluble gases in the marineatmospheric surface layer. Tellus 50B, 111–127.

Hertel, O., Christensen, J., Runge, E.H., Asman, W.A.H., Berkowicz, R., Hovmand, M.F., Hov, Ø., 1995.

Development and testing of a new variable scale air pollution modelFACDEP. Atmospheric Environment 29 (11),1267–1290.

Kronvang, B., Ærtebjerg, G., Grant, R., Kristensen, P., Hovmand, M., Kirkegaard, J., 1993. Nationwide monitoring of

nutrients and their ecological effects: state of the Danish aquatic environment. Ambio 22, 176–187.Lancelot, C., Billen, G., Barth, H. (Eds.), 1989. Eutrophication and Algal Blooms in the North Sea, the Baltic and

Adjacent Areas: Predictions and Assessment of Preventive Actions. Report No. 12. Water Pollution ResearchProgress Series, EC, Brussels.

Paerl, H.W., 1995. Coastal eutrophication in relation to atmospheric nitrogen deposition: current perspectives. Ophelia41, 237–259.

Pakkanen, T.A., 1996. Study of formation of coarse particle nitrate aerosol. Atmospheric Environment 30, 2475–2482.

Plate, E., Stahlschmidt, T., Schulz, M., Dannecker, W., 1995. The influence of air–sea-exchange on the partitioningof N-species during transport over sea. In: Jaehne, B., Monahan, E.C. (Eds.), Third International Symposium onAir–Water Gas Transfer. AEON-Verlag, Hanau, pp. 735–744.

North Sea Task Force, 1993. North Sea Quality Status Report 1993. Oslo and Paris Commissions, London. Olsen &Osen, Fredensborg, Denmark.

Rendell, A.R., Ottley, C.J., Jickells, T.D., Harrison, R.M., 1993. The atmospheric input of nitrogen species to the

North Sea. Tellus 45B, 53–63.Sanders, R.J., Jickells, T., Malcolm, S., Brown, J., Kirkwood, D., Reeve, A., Taylor, J., Horrobin, T., Ashcroft, C.,

1997. Nutrient fluxes through the Humber estuary. Journal of Sea Research 37, 3–23.Schl .unzen, K.H., 1990. Numerical studies on the inland penetration of sea breeze fronts at a coastline with tidally

flooded mudflats. Contributions to Atmospheric Physics 63, 243–256.Schl .unzen, K.H., 1994. Atmosph.arische Eintr.age von N.ahr- und Schadstoffen. In: Loz!an, J.L., Rachor, E., Reise, K.,

v. Westernhagen, H., Lenz, W. (Eds.), Warnsignale aus dem WattenmeerFWissenschaftliche Fakten. Blackwell

Wissenschafts-Verlag, Berlin, pp. 45–48.Schl .unzen, K.H., Bigalke, K., L .upkes, C., Niemeier, U., Von Salzen, K., 1996. Concept and realization of the

mesoscale transport- and fluid-model ‘METRAS’. METRAS Technical Report 5, Meteorologisches Institut,

Universit.at Hamburg, 156pp.Schl .unzen, K.H., Pahl, S., 1992. Modification of dry deposition in a developing sea-breeze circulationFa numerical

case study. Atmospheric Environment 26A, 51–61.Schulz, M., 1996. Fluxes across marine surfaces. In: Borrell, P.M., Borrell, P., Kelly, K., Cvita$s, T., Seiler, W. (Eds.),

Transport and Transformation of Pollutants in the Troposphere, Vol. 2: Emissions, Deposition, Laboratory Workand Instrumentation. Computational Mechanics Publications, Southampton, UK, pp. 45–49.

Seinfeld, J.H., Pandis, S.N., 1998. Atmospheric Chemistry and Physics. Wiley, New York.

Skov, H., Hertel, O., Ellermann, T., Ambelas Skj�th, C., Heidam, N.Z., 1999. NOVA 2003, Atmospheric deposition ofnitrogen 1998 (in Danish: NOVA 2003 1998, Atmosfærisk deposition af kvælstof 1998). Technical Report No. 289,NERI, 102pp.


Spokes, L., Jickells, T., Rendell, A., Schulz, M., Rebers, A., Dannecker, W., Kr .uger, O., Leermakers, M., Baeyens, W.,1993. High atmospheric nitrogen deposition events over the North Sea. Marine Pollution Bulletin 26, 698–703.

Spokes, L.J., Yeatman, S.G., Cornell, S.E., Jickells, T.D., 2000. Nitrogen deposition to the eastern Atlantic Ocean: theimportance of south-easterly flow. Tellus 52, 37–49.

Stockwell, W.R., Kirchner, F., Kuhn, M., Seefeld, S., 1997. A new mechanism for regional atmospheric chemistry

modeling. Journal of Geophysical Research 102, 25,847–25,879.Vignati, E., 1999. Modelling interactions between aerosols and gaseous compounds in the polluted marine atmosphere.

Ph.D. Thesis, Ris� National Laboratory, Report No. Ris�-R-1163(EN), 133pp.

Vignati, E., de Leeuw, G., Berkowitcz, R., 2001. Modelling coastal aerosol transport and effects of surf-producedaerosols on processes in the marine atmospheric boundary layer. JGR-Atmospheres, in press.

Von Salzen, K., 1997. Entwicklung und Anwendung eines Modells f .ur die Dynamik und Zusammensetzung dessekund.aren und marinen Aerosols. Ber. aus dem Zentrum f. Meeres- und Klimaforschung, Meteorologisches

Institut, Universit.at Hamburg 30, 167pp.Von Salzen, K., Schl .unzen, K.H., 1999a. A prognostic physico-chemical model of secondary and marine inorganic

multicomponent aerosols: Part I. Model description. Atmospheric Environment 33, 567–576.

Von Salzen, K., Schl .unzen, K.H., 1999b. A prognostic physico-chemical model of secondary and marine inorganicmulticomponent aerosols: Part II. Model tests. Atmospheric Environment 33, 1543–1552.

Von Salzen, K., Schl .unzen, K.H., 2000. Simulation of the dynamics and composition of secondary and marine

inorganic aerosols in the coastal atmosphere. Journal of Geophysical Research 104, 30,201–30,217.Yeatman, S.G., 2000. Major ion and isotopic studies of aerosol nitrogen species in the marine atmosphere. Ph.D.

Thesis, University of East Anglia, Norwich, UK.


Atmospheric input of nitrogen into the North Sea: ANICE project overview

Documents

Transcript of Atmospheric input of nitrogen into the North Sea: ANICE project overview