carbon–nitrogen interactions in european forests...

Biogeosciences 17 1583ndash1620 2020httpsdoiorg105194bg-17-1583-2020copy Author(s) 2020 This work is distributed underthe Creative Commons Attribution 40 License

Carbonndashnitrogen interactions in European forests and semi-naturalvegetation ndash Part 1 Fluxes and budgets of carbon nitrogen andgreenhouse gases from ecosystem monitoring and modellingChris R Flechard1 Andreas Ibrom2 Ute M Skiba3 Wim de Vries4 Marcel van Oijen3 David R Cameron3Nancy B Dise3 Janne F J Korhonen56 Nina Buchmann7 Arnaud Legout8 David Simpson910 Maria J Sanz11Marc Aubinet12 Denis Loustau13 Leonardo Montagnani1415 Johan Neirynck16 Ivan A Janssens17 Mari Pihlatie56Ralf Kiese18 Jan Siemens19 Andreacute-Jean Francez20 Juumlrgen Augustin21 Andrej Varlagin22 Janusz Olejnik2324Radosław Juszczak25 Mika Aurela26 Daniel Berveiller27 Bogdan H Chojnicki25 Ulrich Daumlmmgen28Nicolas Delpierre27 Vesna Djuricic29 Julia Drewer3 Eric Dufrecircne27 Werner Eugster7 Yannick Fauvel1David Fowler3 Arnoud Frumau30 Andreacute Granier31 Patrick Gross31 Yannick Hamon1 Carole Helfter3Arjan Hensen30 Laacuteszloacute Horvaacuteth32 Barbara Kitzler33 Bart Kruijt34 Werner L Kutsch35 Raquel Lobo-do-Vale36Annalea Lohila3726 Bernard Longdoz38 Michal V Marek39 Giorgio Matteucci40 Marta Mitosinkova41Virginie Moreaux1342 Albrecht Neftel43 Jean-Marc Ourcival44 Kim Pilegaard2 Gabriel Pita45 Francisco Sanz46Jan K Schjoerring47 Maria-Teresa Sebastiagrave4849 Y Sim Tang3 Hilde Uggerud50 Marek Urbaniak23 Netty van Dijk3Timo Vesala376 Sonja Vidic29 Caroline Vincke51 Tamaacutes Weidinger52 Sophie Zechmeister-Boltenstern53Klaus Butterbach-Bahl18 Eiko Nemitz3 and Mark A Sutton3

1Institut National de la Recherche en Agriculture Alimentation et Environnement (INRAE) UMR 1069 SAS65 rue de Saint-Brieuc 35042 Rennes France2Department of Environmental Engineering Technical University of DenmarkBygningstorvet 2800 Kgs Lyngby Denmark3UK Centre for Ecology and Hydrology (UK CEH) Bush Estate Penicuik EH26 0QB UK4Wageningen University and Research Environmental Systems Analysis Group PO Box 476700 AA Wageningen the Netherlands5Department of Agricultural Sciences Faculty of Agriculture and Forestry Environmental Soil SciencePO Box 56 00014 University of Helsinki Helsinki Finland6Institute for Atmospheric and Earth System ResearchForest Sciences Faculty of Agriculture and ForestryPO Box 27 00014 University of Helsinki Helsinki Finland7Department of Environmental Systems Science Institute of Agricultural Sciences ETH Zurich LFW C56Universitatstr 2 8092 Zurich Switzerland8Institut National de la Recherche en Agriculture Alimentation et Environnement (INRAE) BEF 54000 Nancy France9EMEP MSC-W Norwegian Meteorological Institute Oslo Norway10Department of Space Earth and Environment Chalmers University of Technology Gothenburg Sweden11Ikerbasque Foundation and Basque Centre for Climate Change Sede Building 1 Scientific Campus of the University of theBasque Country 48940 Leioa Biscay Spain12TERRA Teaching and Research Centre Gembloux Agro-Bio Tech University of Liegravege Gembloux Belgium13Bordeaux Sciences Agro Institut National de la Recherche en Agriculture Alimentation et Environnement (INRAE) UMRISPA Villenave drsquoOrnon 33140 France14Forest Services Autonomous Province of Bolzano Via Brennero 6 39100 Bolzano Italy15Faculty of Science and Technology Free University of Bolzano Piazza Universitagrave 5 39100 Bolzano Italy16Environment and Climate Research Institute for Nature and Forest (INBO) Gaverstraat 35 9500 Geraardsbergen Belgium17Department of Biology Centre of Excellence PLECO (Plant and Vegetation Ecology)University of Antwerp 2610 Wilrijk Belgium

Published by Copernicus Publications on behalf of the European Geosciences Union

1584 C R Flechard et al Carbonndashnitrogen interactions in European ecosystems ndash Part 1

18Karlsruhe Institute of Technology (KIT) Institute of Meteorology and Climate Research AtmosphericEnvironmental Research (IMK-IFU) Kreuzeckbahnstr 19 82467 Garmisch-Partenkirchen Germany19Institute of Soil Science and Soil Conservation iFZ Research Centre for Biosystems Land Use and NutritionJustus Liebig University Giessen Heinrich-Buff-Ring 26ndash32 35392 Giessen Germany20University of Rennes CNRS UMR6553 ECOBIO Campus de Beaulieu263 avenue du Geacuteneacuteral Leclerc 35042 Rennes France21Leibniz Centre for Agricultural Landscape Research (ZALF) Eberswalder Straszlige 8415374 Muumlncheberg Germany22AN Severtsov Institute of Ecology and Evolution Russian Academy of Sciences119071 Leninsky pr33 Moscow Russia23Department of Meteorology Poznan University of Life Sciences Piatkowska 94 60-649 Poznan Poland24Department of Matter and Energy Fluxes Global Change Research Centre AS CRvvi Belidla 9864a 603 00 Brno Czech Republic25Department of Ecology and Environmental Protection Laboratory of BioclimatologyPoznan University of Life Sciences Piatkowska 94 60-649 Poznan Poland26Finnish Meteorological Institute Climate System Research PL 503 00101 Helsinki Finland27Ecologie Systeacutematique Evolution Univ Paris-Sud CNRS AgroParisTechUniversiteacute Paris-Saclay 91400 Orsay France28Weststrasse 5 38162 Weddel Germany29Air Quality Department Meteorological and Hydrological Service Gric 3 10000 Zagreb Croatia30TNO Environmental Modelling Sensing and Analysis Petten the Netherlands31Institut National de la Recherche en Agriculture Alimentation et Environnement (INRAE) UMR1434 Silva Site de NancyRue drsquoAmance 54280 Champenoux France32Greengrass ndash Atmospheric Environment Expert Ltd fellowship Korneacutelia utca 14a 2030 Eacuterd Hungary33Federal Research and Training Centre for Forests Natural Hazards and LandscapeSeckendorff-Gudent-Weg 8 1131 Vienna Austria34Wageningen University and Research PO Box 47 6700AA Wageningen the Netherlands35Integrated Carbon Observation System (ICOS ERIC) Head Office Erik Palmeacutenin aukio 1 00560 Helsinki Finland36Centro de Estudos Florestais Instituto Superior de Agronomia Universidade de LisboaTapada da Ajuda 1349-017 Lisbon Portugal37Institute for Atmospheric and Earth System ResearchPhysics Faculty of SciencePO Box 68 00014 University of Helsinki Helsinki Finland38Gembloux Agro-Bio Tech Axe Echanges Ecosystegravemes Atmosphegravere 8 Avenue de la Faculteacute 5030 Gembloux Belgium39Global Change Research Institute Academy of Sciences Belidla 4a 603 00 Brno Czech Republic40National Research Council of Italy Institute for Agriculture and Forestry Systems in the Mediterranean(CNR-ISAFOM) Via Patacca 85 80056 Ercolano (NA) Italy41Department of Air Quality Slovak Hydrometeorological Institute Jeseniova 17 83315 Bratislava Slovakia42Institute for Geosciences and Environmental research (IGE) UMR 5001 Universiteacute Grenoble Alpes CNRS IRDGrenoble Institute of Technology 38000 Grenoble France43NRE Oberwohlenstrasse 27 3033 Wohlen bei Bern Switzerland44CEFE CNRS Univ Montpellier Univ Paul Valeacutery Montpellier 3 EPHE IRD Montpellier France45Mechanical Engineering Department Instituto Superior Teacutecnico (Technical University of Lisbon)Ave Rovisco Pais IST 1049-001 Lisbon Portugal46Fundacioacuten CEAM CCharles R Darwin 46980 Paterna (Valencia) Spain47Department of Plant and Environmental Sciences Faculty of Science University of CopenhagenThorvaldsensvej 40 1871 Frederiksberg C Denmark48Laboratory of Functional Ecology and Global Change (ECOFUN) Forest Science and Technology Centreof Catalonia (CTFC) Carretera de Sant Llorenccedil de Morunys 25280 Solsona Spain49Group GAMES amp Department of Horticulture Botany and Landscaping School of Agrifood and Forestry Scienceand Engineering University of Lleida Av Rovira Roure 191 25198 Lleida Spain50Norsk institutt for luftforskning Postboks 100 2027 Kjeller Norway51Earth and Life Institute (Environmental sciences) Universiteacute catholique de Louvain Louvain-la-Neuve Belgium52Department of Meteorology Eoumltvoumls Loraacutend University 1117 Budapest Paacutezmaacuteny Peacuteter s 1A Hungary53Department of Forest and Soil Sciences Institute of Soil Research University of Natural Resourcesand Life Sciences Vienna Peter Jordan Str 82 1190 Vienna Austria

Biogeosciences 17 1583ndash1620 2020 wwwbiogeosciencesnet1715832020

C R Flechard et al Carbonndashnitrogen interactions in European ecosystems ndash Part 1 1585

Correspondence Chris R Flechard (christopheflechardinraefr)

Received 22 August 2019 ndash Discussion started 11 September 2019Revised 11 December 2019 ndash Accepted 10 February 2020 ndash Published 26 March 2020

Abstract The impact of atmospheric reactive nitrogen (Nr)deposition on carbon (C) sequestration in soils and biomassof unfertilized natural semi-natural and forest ecosystemshas been much debated Many previous results of this dCdNresponse were based on changes in carbon stocks from pe-riodical soil and ecosystem inventories associated with es-timates of Nr deposition obtained from large-scale chem-ical transport models This study and a companion paper(Flechard et al 2020) strive to reduce uncertainties of N ef-fects on C sequestration by linking multi-annual gross andnet ecosystem productivity estimates from 40 eddy covari-ance flux towers across Europe to local measurement-basedestimates of dry and wet Nr deposition from a dedicated col-located monitoring network To identify possible ecologicaldrivers and processes affecting the interplay between C andNr inputs and losses these data were also combined with insitu flux measurements of NO N2O and CH4 fluxes soilNOminus3 leaching sampling and results of soil incubation ex-periments for N and greenhouse gas (GHG) emissions aswell as surveys of available data from online databases andfrom the literature together with forest ecosystem (BAS-FOR) modelling

Multi-year averages of net ecosystem productivity (NEP)in forests ranged from minus70 to 826 g C mminus2 yrminus1 at totalwet+ dry inorganic Nr deposition rates (Ndep) of 03 to43 g N mminus2 yrminus1 and from minus4 to 361 g C mminus2 yrminus1 at Ndeprates of 01 to 31 g N mminus2 yrminus1 in short semi-natural veg-etation (moorlands wetlands and unfertilized extensivelymanaged grasslands) The GHG budgets of the forests werestrongly dominated by CO2 exchange while CH4 and N2Oexchange comprised a larger proportion of the GHG balancein short semi-natural vegetation Uncertainties in elementalbudgets were much larger for nitrogen than carbon espe-cially at sites with elevated Ndep where Nr leaching losseswere also very large and compounded by the lack of reli-able data on organic nitrogen and N2 losses by denitrifica-tion Nitrogen losses in the form of NO N2O and especiallyNOminus3 were on average 27 (range 6 ndash54 ) ofNdep at siteswith Ndep lt 1 g N mminus2 yrminus1 versus 65 (range 35 ndash85 )for Ndep gt 3 g N mminus2 yrminus1 Such large levels of Nr loss likelyindicate that different stages of N saturation occurred at anumber of sites The joint analysis of the C and N budgetsprovided further hints that N saturation could be detected inaltered patterns of forest growth Net ecosystem productiv-ity increased with Nr deposition up to 2ndash25 g N mminus2 yrminus1with large scatter associated with a wide range in carbon se-questration efficiency (CSE defined as the NEP GPP ra-tio) At elevated Ndep levels (gt 25 g N mminus2 yrminus1) where in-

organic Nr losses were also increasingly large NEP levelledoff and then decreased The apparent increase in NEP at lowto intermediateNdep levels was partly the result of geograph-ical cross-correlations between Ndep and climate indicatingthat the actual mean dCdN response at individual sites wassignificantly lower than would be suggested by a simplestraightforward regression of NEP vs Ndep

1 Introduction

The global terrestrial net sink for atmospheric carbon diox-ide (CO2) is approximately 17 Pg C yrminus1 ie roughly one-fifth of global CO2-C emissions by fossil fuel combustionand industry (94plusmn 05 Pg C yrminus1) This corresponds to theland-based carbon (C) uptake of 32plusmn 08 Pg C yrminus1 minusemissions from deforestation and other land-use changes of15plusmn 07 Pg C yrminus1 The ocean sink is of the same order(24plusmn 05 Pg C yrminus1) while twice as much CO2-C (47plusmn002 Pg C yrminus1) is added yearly to the atmosphere (Le Queacutereacuteet al 2018) Data from atmospheric CO2 inversion meth-ods (eg Bousquet et al 1999 Ciais et al 2010) from na-tional to global forest C inventory approaches (Goodale et al2002 Pan et al 2011) and from eddy covariance (EC) fluxnetworks (Luyssaert et al 2007) have suggested that a dom-inant part of this terrestrial CO2 sink is currently occurring inforests and especially in boreal and temperate forests of theNorthern Hemisphere (Ciais et al 2010 Pan et al 2011)Tropical forest areas are believed to be closer to carbon neu-tral (Pan et al 2011) or even a net C source globally (Bac-cini et al 2017) due to emissions from deforestation forestdegradation and land-use change offsetting their sink poten-tial However others (Stephens et al 2007) have argued thatthe tropical land CO2 sink may be stronger ndash and the north-ern hemispheric land CO2 sink weaker ndash than was generallybelieved At the European scale Schulze et al (2010) calcu-lated that the net biome productivity (NBP the mean long-term carbon sink at a large spatial scale) of temperate andboreal forests was 81 of the total continental-scale landsink

The large European and North American CO2 sinks havebeen attributed to a combination of factors including af-forestation of abandoned land and formerly cut forests re-duced forest harvest CO2 fertilization changes in manage-ment and age structure legacy effects in Europe (Vileacuten etal 2016) and atmospheric reactive nitrogen (Nr) deposition(Reay et al 2008 Ciais et al 2013 and references therein

wwwbiogeosciencesnet1715832020 Biogeosciences 17 1583ndash1620 2020


De Vries et al 2017) However some studies (Nadelhofferet al 1999 Gundale et al 2014 Fernaacutendez-Martiacutenez et al2017) have questioned the widespread theory that elevatedNr deposition boosts forest C sequestration and the magni-tude of the N fertilization effect on forest C sequestration hasbeen a matter of much debate (Magnani et al 2007 2008Houmlgberg 2007 De Schrijver et al 2008 de Vries et al2008 Sutton et al 2008 Dezi et al 2010 Binkley and Houmlg-berg 2016) A better understanding of the impact of nitrogendeposition on natural and semi-natural ecosystems in partic-ular over forests and the impact on the carbon and nitrogencycles as an indirect effect resulting from anthropogenic ac-tivities (Canadell et al 2007) remains key to improving theforecast of regional (de Vries et al 2017) and global (Du andde Vries 2018) models

The relevance of Nr deposition for the global C seques-tration potential or more explicitly the dCdN response(change in C storage with change in Nr deposition) hasbeen estimated typically through meta-analyses of Nr addi-tion experiments (eg Schulte-Uebbing and de Vries 2018)or by combining forest growth inventories together with esti-mates of Nr deposition obtained from large-scale forest mon-itoring plots (Solberg et al 2009 Laubhann et al 2009De Vries et al 2008) Both methods have many sourcesof uncertainty One key difficulty in the latter approach liesin estimating total (wet+ dry) Nr deposition (Ndep) espe-cially dry deposition which is highly variable spatially verychallenging to measure and consequently hard to parame-terize in regional-scale chemical transport models (CTMs)(Flechard et al 2011 Simpson et al 2014 Schwede et al2018) The annual or long-term dry deposition componentof Ndep to forests in all the diversity of N-containing forms(gaseous vs aerosol reduced vs oxidized inorganic vs or-ganic eg Zhang et al 2009) has been actually measured(by micrometeorological methods) in very few forests world-wide (Neirynck et al 2007 Erisman et al 1996) Due tothe large diversity of atmospheric compounds that contributeto total Nr and the complexity of the measurement tech-niques required for each compound (Flechard et al 2011)it is even debatable that complete measurements of all Nrdeposition terms have ever been achieved anywhere Thusvirtually all studies of the forest dCdN response so far haverelied on modelled atmospheric Nr deposition estimates atleast for the dry and occult deposition fractions and furtherthe Nr deposition data being used were systematically pro-vided by the outputs of large-scale regional (eg Sutton etal 2008 Fernaacutendez-Martiacutenez et al 2017) or even global(Fleischer et al 2013) models with resolutions of typically10 kmtimes 10 km or 1times 1 respectively Grid averaging insuch large-scale models introduces a large uncertainty in lo-cal (ecosystem-scale)Nr dry deposition rates (Schwede et al2018) particularly when the forest sites are located near agri-cultural or industrial Nr sources (Loubet et al 2009 Fowleret al 1998)

Additionally nitrogen losses may significantly offset at-mosphericNr inputs at eutrophicated and acidified sites withthe consequence that dCdN may correlate better with netrather than gross atmospheric Nr inputs Depending espe-cially on the extent of ecosystem N saturation (De Schri-jver et al 2008) substantial N losses may occur in the formof nitrate (NOminus3 ) leaching (Dise et al 2009) nitric oxide(NO) and nitrous oxide (N2O) emissions (Pilegaard et al2006) ammonia (NH3) bidirectional exchange (Hansen etal 2013) and emissions of di-nitrogen (N2) from total den-itrification (Butterbach-Bahl et al 2002) (Fig 1) The im-plication is that the carbon response to Ndep would be non-linear with larger dCdN at low Ndep rates and a lower-ing of dCdN as Ndep increases as suggested in the reviewby Butterbach-Bahl and Gundersen (2011) and further elab-orated in De Vries et al (2014) The latter authors showin their review that above a certain N deposition level thedCdN response declines due to adverse effects of excessNr deposition and high soil ammonium (NH+4 ) concentra-tion and nitrification (eg acidification nutrient base cationlosses aluminium mobility) which are known to reduce soilfertility and affect ecosystem health and functioning (Aber1992)

Carbon losses through dissolved organic carbon (DOC)and biogenic dissolved inorganic carbon (DIC) leaching canalso be significant especially for wetlands (Dinsmore et al2010) and also grassland and cropland ecosystems (Kindleret al 2011 Gielen et al 2011) This is relevant for the netecosystem carbon balance (NECB) or the net biome produc-tivity (NBP) estimates obtained on the basis of EC flux sys-tems and needs to be accounted for as a part of the net ecosys-tem productivity (NEP) that is not actually stored in the sys-tem (Chapin et al 2006 Schulze et al 2010) (Fig 1) Dis-solved andor emitted methane (CH4) may further representa significant loss from organic soils (Hendriks et al 2007)while CH4 oxidation which is often observed in well-aeratedsoils and can be suppressed by Nr addition especially NH+4(Steudler et al 1989) may affect the net greenhouse gas(GHG) budget Nitrogen-deposition-induced N2O emissionsfrom the forest floor (Pilegaard et al 2006 Liu and Greaver2009) or from denitrification triggered by deposited NOminus3 inpeatland (Francez et al 2011) can also offset the gain in theecosystem GHG balance resulting from a hypothetical nitro-gen fertilization effect

Nitrogen deposition or addition is known to affect soilmicrobial C cycling in many different ways for examplehigh-level N enrichment generally leading to reduced micro-bial biomass and suppressed soil CO2 respiration (Treseder2008) a reduction of basal respiration without significant de-cline in total microbial biomass following N addition to in-cubated peat cores (Francez et al 2011) and added NOminus3 al-tering directly the oxidative enzyme production by microbialcommunities and hence controlling extracellular enzyme ac-tivity (Waldrop and Zak 2006) Nitrate addition can lead toa reduction in CH4 emissions from wetlands and peatlands



Figure 1 Flux terms and boundaries of the carbon (a) and nitrogen (b) budgets discussed in this paper Net ecosystem productivityNEP=GPPminusReco (asympNPPminusRhet) based on multi-annual eddy covariance CO2 flux data The net ecosystem carbon balance (NECB)includes in addition other C loss fluxes such as DICDOC CH4 and VOC as well as harvest thinning or other disturbances (eg fire) In-organic reactive nitrogen (Nr) budget=NdepminusDINleachminusNOminusN2O The total N budget includes in addition organic nitrogen deposition(WSON) and leaching (DON) as well as N2 inputs and losses from biological fixation and denitrification respectively CLBS CSOM CRCLITT carbon stocks in leaves branches and stems in soil organic matter in roots and in litter layers respectively Terms highlighted inred indicate that direct or measurement-based estimates were not available for some or all sites in our datasets (see also Table 2 for a list ofacronyms Table 3 for a summary of methods and Table S6 for data availability)

(Francez et al 2011) since in anaerobic conditions and inthe presence of NOminus3 as an electron acceptor denitrifyingbacteria can oxidize organic C substrates (eg acetate) andthus outcompete methanogenic communities (Boone 1991)However if chronic N enrichment of peatland ecosystemsleads to floristic changes especially an increase in vascularplants at the expense of bryophytes the net effect may bean increase in CH4 emissions (Nykaumlnen et al 2002) as theaerenchyma of tracheophytes provides a direct diffusion pathto the atmosphere for soil-produced CH4 bypassing oxida-tion in the peat by methanotrophs Excess-nitrogen-inducedvegetation composition changes in Sphagnum moss peatlandare believed to reduce C sequestration potentials and the ef-fect is likely to be exacerbated by climate change (Limpenset al 2011)

This complex web of interactions between the C and N cy-cles and losses shows the need for integrated approaches forstudying the impacts ofNr deposition on C sequestration and

net GHG budgets Ideally all C and N gain and loss pathways(including infrequently or rarely measured fluxes such as Nrdry deposition organic C and N leaching fluxes and GHGfluxes see Fig 1) should be quantified at long-term exper-imental sites to improve and calibrate process-based mod-els Closing the C and N budgets experimentally at each siteof large (eg FLUXNET) monitoring networks is unlikelyto occur in the near future but realistic and cost-effectivemeasurement approaches can be used to progressively re-duce the uncertainties for the large terms of the budgetsSuch approaches were tested and implemented in this studyas part of a large-scale effort within the NitroEurope Inte-grated Project (NEU 2013 Sutton and Reis 2011) to quan-tify Nr deposition and N losses from ecosystems in paral-lel and coordinated with the CarboEurope Integrated Project(CEIP 2011) to estimate the net C and GHG balance for for-est and semi-natural ecosystems in Europe



A main objective of this paper is to build tentative C Nand GHG budgets as well as analyse CndashN interactions em-pirically for a wide range of European monitoring sites byusing measurements or observation-based data as far as pos-sible complemented by modelling Important methodologi-cal goals are to critically examine uncertainties in measure-ment methods and elemental budgets to identify knowledgeand data gaps and to assess the current state of process un-derstanding as encoded in models To this end we compiledthe C N and GHG flux data from NEU CEIP and other com-plementary datasets using a combination of in situ measure-ments empirical relationships ecosystem modelling litera-ture and database surveys at the scale of the CEIP and NEUflux monitoring networks This study presents the method-ologies and discusses the different terms of the budgets in-cluding atmospheric deposition from gas aerosol and pre-cipitation Nr concentration monitoring soil NOminus3 leachingmeasurements and modelling GHG and Nr emission es-timates from chamber measurements and laboratory-basedsoil bioassays EC-tower-based C budgets and historicalpublished data Forest ecosystem modelling (BASFOR) isused to simulate C N and GHG fluxes with the double ob-jective to compare with actual measurements and to fill somegaps in the datasets Wherever possible alternative measure-ments datasets or modelled data are shown alongside the pri-mary data in order to provide an estimate of the uncertainty inthe different terms In the companion paper (Flechard et al2020) the response of C sequestration to Ndep is quantifiedusing the same datasets

2 Materials and methods

21 Monitoring sites

The study comprised 40 terrestrial ecosystem-scale car-bon and nitrogen flux monitoring sites including 31 forests(F) and nine natural or semi-natural (SN) short vegeta-tion ecosystems primarily moorlands wetlands and exten-sively managed unfertilized grasslands (Table 1) The sitesspanned a European geographical and climatic gradient fromthe Mediterranean to the Arctic and from the Atlantic towestern Russia (Fig S1 in the Supplement) an elevationrange ofminus2 m to 1765 m amsl a mean annual temperature(MAT) range of minus10 to 176 C and a mean annual precip-itation (MAP) range of 500 to 1365 mm Selected referencesare provided for each site in Table S1 in the Supplement Alist of the main acronyms and abbreviations used in the paperis provided in Table 2

The forest sites of the study ranged from very young (lt 10years old) to mature (gt 150 years old) and can be broadlyclassified into four plant functional types (PFTs) or five dom-inant tree categories (Table 1) deciduous broadleaf (DB) ev-ergreen needleleaf (EN comprising mostly spruce and pinespecies) mixed deciduousndashconiferous (MF) and Mediter-

ranean evergreen broadleaf (EB) Forest species composi-tion stand characteristics C and N contents of differentecosystem compartments (leaves wood soil) soil physi-cal properties and micro-climatological characteristics aredescribed in Tables S2ndashS5 Semi-natural short vegetationecosystems included unimproved (mountainous and semi-arid) grasslands wetlands and peatlands they are includedin the study as unfertilized C-rich soil systems providing acontrast with forests where storage also occurs above ground(thus with different CN ratios) Among the 40 EC-CO2 fluxmeasurement stations most sites (36) were part of the CEIPCO2 flux network A further three CO2 flux sites were op-erated as part of the NEU network (EN2 EN16 and SN3)and one site (DB4) was included from the French F-ORE-Tobservation network (F-ORE-T 2012) Table S6 provides anoverview of the available C N and GHG flux measurementsdetailed hereafter

22 Nitrogen fluxes

Input and output fluxes of the ecosystem nitrogen and carbonbudgets are represented schematically in Fig 1 The follow-ing sections describe the methods used to quantify the differ-ent terms summarized in Table 3

221 Atmospheric deposition

To obtain realistic estimates of total (dry+wet) Nr deposi-tion at the 40 sites of the network it was necessary to mea-sure ambient air concentrations of the main N-containingchemical species at each location due to the large spatial het-erogeneity in gas-phase concentrations especially for NH3The requirement for local measurements of wet depositionwas relaxed because this is much less spatially variable Forboth dry and wet components measurements had to be com-plemented by models either to calculate fluxes based on lo-cal concentration data at each site or to obtain local estimatesfrom a large-scale CTM when data were missing

Atmospheric inorganic Nr concentrations available fromthe NEU (2013) database were measured monthly for 2ndash4 years in the gas phase (NH3 HNO3 HONO) and inthe aerosol phase (NH+4 NOminus3 ) using DEnuder for Long-Term Atmospheric sampling (DELTA) systems (Sutton et al2001 Tang et al 2009) Concentrations of nitrogen diox-ide (NO2) not covered by DELTA sampling were mea-sured by chemiluminescence at a few sites only and wereotherwise taken from gridded concentration outputs of theEuropean-scale EMEP CTM (details given below) The Nrdata initially reported in Flechard et al (2011) covered thefirst 2 years of the NEU project (2007ndash2008) here the datafrom the entire 4-year NEU monitoring period (2007ndash2010)were used and averaged to provide a more robust long-term4-year estimate of Nr dry deposition The inferential mod-elling method was used to calculate dry deposition for N-containing gas and aerosol species whereby measured am-



Tabl

e1

Ove

rvie

wof

ecos

yste

man

dcl

imat

icch

arac

teri

stic

san

din

ter-

annu

alm

ean

ecos

yste

mndasha

tmos

pher

eex

chan

geflu

xes

forf

ores

tand

sem

i-na

tura

lsho

rtve

geta

tion

site

s

Site

Site

nam

ePF

Ta

Dom

inan

tFo

rest

age

Hm

axb

LA

I max

cL

at

Lon

gE

leva

tiond

MA

Te

MA

PfN

depg

GPP

hR

ecoi

NE

Pj

acro

nym

coun

try

shor

tnam

eve

geta

tion

(201

0)(m

)(m

2mminus

2 )(

N)

(E

)(m

as

l)(

C)

(mm

)(g

Nmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )

DE

-Hai

Hai

nich

Ger

man

yD

B1

Fagu

ssy

lvat

ica

142

234

051

079

104

5243

08

477

52

315

5310

7447

9D

K-S

orSo

roslashD

enm

ark

DB

2Fa

gus

sylv

atic

a91

314

655

487

116

4640

89

730

22

1883

1581

301

FR-F

onFo

ntai

nebl

eau-

Bar

beau

Fra

nce

DB

3Q

uerc

uspe

trae

a11

128

51

484

762

780

9211

069

01

718

5011

8566

5FR

-Fgs

Foug

egraveres

Fra

nce

DB

4Fa

gus

sylv

atic

a41

206

048

383

minus1

185

140

103

900

24

1725

1316

409

FR-H

esH

esse

Fra

nce

DB

5Fa

gus

sylv

atic

a45

166

748

674

706

630

010

297

51

716

3411

8744

6IT

-Col

Col

lelo

ngo

Ital

yD

B6

Fagu

ssy

lvat

ica

120

225

741

849

135

8815

607

211

401

214

2577

665

0C

Z-B

K1

Bily

Kri

zC

zech

Rep

E

N1

Pic

eaab

ies

3313

98

495

0318

538

908

78

1200

21

1548

767

781

DE

-Hoe

Houmlg

lwal

dG

erm

any

EN

2P

icea

abie

s10

435

63

483

0011

100

540

89

870

32

1856

1229

627

DE

-Tha

Tha

rand

tG

erm

any

EN

3P

icea

abie

s12

027

67

509

6413

567

380

88

820

23

1997

1396

601

DE

-Wet

Wet

zste

inG

erm

any

EN

4P

icea

abie

s56

227

150

453

114

5878

56

695

02

218

0917

6743

IT-R

enR

enon

Ita

lyE

N5

Pic

eaab

ies

111

295

146

588

114

3517

304

610

101

313

5352

882

6R

U-F

yoFy

odor

ovsk

oye

Rus

sia

EN

6P

icea

abie

s19

021

28

564

6232

922

265

53

711

10

1488

1559

minus70

UK

-Gri

Gri

ffin

UK

EN

7P

icea

sitc

hens

is29

126

556

617

minus3

800

340

77

1200

07

989

677

311

BE

-Bra

Bra

ssch

aat

Bel

gium

EN

8P

inus

sylv

estr

is82

211

951

309

452

116

108

850

41

1272

1149

123

ES-

ES1

ElS

aler

Spa

inE

N9

Pin

usha

lepe

nsis

111

102

639

346

minus0

319

517

655

12

115

5296

059

3FI

-Hyy

Hyy

tiaumllauml

Fin

land

EN

10P

inus

sylv

estr

is48

183

461

848

242

9518

13

870

90

511

1484

526

8FI

-Sod

Soda

nkyl

aumlFi

nlan

dE

N11

Pin

ussy

lves

tris

100

131

267

362

266

3818

0minus

04

527

03

551

598

minus47

FR-B

ilB

ilos

Fran

ceE

N12

Pin

uspi

nast

er9

40

544

522

minus0

896

5012

493

00

811

7898

918

9FR

-LB

rL

eB

ray

Fran

ceE

N13

Pin

uspi

nast

er41

221

944

717

minus0

769

6112

997

21

619

0614

7942

7IT

-SR

oSa

nR

osso

reI

taly

EN

14P

inus

pina

ster

6118

40

437

2810

284

414

992

01

622

5617

0255

4N

L-L

ooL

oobo

sN

ethe

rlan

dsE

N15

Pin

ussy

lves

tris

101

181

552

168

574

425

100

786

42

1617

1141

476

NL

-Spe

Speu

lder

bos

Net

herl

ands

EN

16P

seud

otsu

gam

enzi

esii

5132

75

522

525

691

5210

083

44

314

1610

1540

1SE

-Nor

Nor

unda

Sw

eden

EN

17P

inus

sylv

estr

is11

228

46

600

8317

467

456

852

70

614

1413

5658

SE-S

k2Sk

ytto

rpS

wed

enE

N18

Pin

ussy

lves

tris

3916

32

601

2917

840

557

452

70

512

3595

328

2E

S-L

Ma

Las

Maj

adas

Spa

inE

B1

Que

rcus

ilex

111

80

639

941

minus5

773

258

161

528

09

1091

958

133

FR-P

uePu

eacutecha

bon

Fran

ceE

B2

Que

rcus

ilex

696

29

437

413

596

270

137

872

11

1309

1030

279

IT-R

o2R

occa

resp

ampa

niI

taly

EB

3Q

uerc

usce

rris

2116

38

423

9011

921

224

157

876

18

1707

886

821

PT-E

spE

spir

raP

ortu

gal

EB

4E

ucal

yptu

sgl

obul

us25

202

738

639

minus8

602

9516

170

91

214

7311

6331

1PT

-Mi1

Mitr

aPo

rtug

alE

B5

Que

rcus

ilex

Que

rcus

sube

r91

83

438

541

minus8

000

264

145

665

09

870

817

53B

E-V

ieV

iels

alm

Bel

gium

MF1

Fagu

ssy

lvat

ica

Pse

udot

suga

men

zies

ii86

305

150

305

599

745

08

110

001

717

9212

4754

5C

H-L

aeL

aumlger

enS

witz

erla

ndM

F2Fa

gus

sylv

atic

aP

icea

abie

s11

130

36

474

788

365

689

77

1100

22

1448

757

692

DE

-Meh

Meh

rste

dtG

erm

any

SN1

Aff

ores

ted

gras

slan

dn

a0

52

951

276

106

5729

39

154

71

511

7111

75minus

4E

S-V

DA

Val

ldrsquoA

linya

Spa

inSN

2U

plan

dgr

assl

and

na

01

14

421

521

448

1765

64

1064

12

669

528

140

FI-L

omL

ompo

lojauml

nkkauml

Fin

land

SN3

Peat

land

na

04

10

679

9824

209

269

minus1

052

10

137

734

532

HU

-Bug

Bug

acH

unga

rySN

4Se

mi-

arid

gras

slan

dn

a0

54

746

692

196

0211

110

750

01

410

4491

812

6IT

-Am

pA

mpl

ero

Ital

ySN

5U

plan

dgr

assl

and

na

04

25

419

0413

605

884

98

1365

09

1241

1028

213

IT-M

Bo

Mon

teB

ondo

neI

taly

SN6

Upl

and

gras

slan

dn

a0

32

546

029

110

8315

505

111

891

714

3513

4789

NL

-Hor

Hor

stem

eer

Net

herl

ands

SN7

Peat

land

na

25

69

520

295

068

minus2

108

800

31

1584

1224

361

PL-w

etPO

LWE

TR

zeci

nPo

land

SN8

Wet

land

(ree

dss

edge

sm

osse

s)n

a2

14

952

762

163

0954

85

550

14

937

642

295

UK

-AM

oA

uche

ncor

thM

oss

UK

SN9

Peat

land

na

06

21

557

92minus

323

927

07

611

650

878

670

581

aPF

T(p

lant

func

tiona

ltyp

es)

DB

ndashde

cidu

ous

broa

dlea

ffor

est

EN

ndashev

ergr

een

need

lele

afco

nife

rous

fore

stE

Bndash

ever

gree

nbr

oadl

eafM

edite

rran

ean

fore

stM

Fndash

mix

edde

cidu

ousndash

coni

fero

usfo

rest

SN

ndashsh

orts

emi-

natu

ral

incl

udin

gm

oorl

and

peat

land

shr

ubla

ndan

dun

impr

oved

upl

and

gras

slan

db

Max

imum

cano

pyhe

ight

cM

axim

umle

afar

eain

dex

defin

edas

one-

side

dor

half

ofth

eto

tal

dA

bove

mea

nse

ale

vel

eM

ean

annu

alte

mpe

ratu

ref

Mea

nan

nual

prec

ipita

tion

gN

itrog

ende

posi

tion

hG

ross

prim

ary

prod

uctiv

ityi

Eco

syst

emre

spir

atio

nj

Net

ecos

yste

mpr

oduc

tivity

na

not

appl

icab

le



Table 2 Main acronyms and abbreviations used in the study

Carbon fluxes and stocks

NEE Net ecosystem exchangeGPP Gross primary productivityNPP Net primary productivityNEP Net ecosystem productivityNECB Net ecosystem carbon balanceNBP Net biome productivityReco Ecosystem respirationRaut Autotrophic respirationRhet Heterotrophic respirationRsoil Soil (heterotrophic and rhizospheric) respirationSCE Soil CO2 efflux measured by chamber methodsCSEobs CSEmod Carbon sequestration efficiency calculated from EC observations or by modellingSOM Soil organic matterCSOM Carbon stock in soil organic matterCR Carbon stock in rootsCLITT Carbon stock in litter layers of the forest floorCLBS Carbon stock in leaves branches and stemsLeafC Leaf carbon contentDIC DOC Dissolved inorganic or organic carbondCdN dNEPdNdep Response (slope) of ecosystem C productivity versus atmospheric Nr deposition

Nitrogen fluxes and stocks

Ndep Total (wet+ dry) atmospheric reactive nitrogen depositionNr Reactive nitrogenNmin Norg Mineral or organic reactive nitrogen formsLeafN Leaf nitrogen contentDIN DON Dissolved inorganic or organic nitrogenDINTF Throughfall inorganic Nr depositionWSON Wet deposition of water-soluble organic nitrogen

Water budget terms

SWC Soil water contentWFPS Water-filled pore spaceET Evapotranspiration

Ecosystem characteristics

PFT Plant functional typeENF Evergreen needleleaf forestDBF Deciduous broadleaf forestMF Mixed (needleleafndashbroadleaf) forestEBF Evergreen broadleaf forestSN Short semi-natural vegetationH Canopy heightDBH Tree diameter at breast height (forests)LAI Leaf area indexSD Stand density (forests) number of trees per unit areaMAT Mean annual temperatureMAP Mean annual precipitation

Methods and general terminology

EC Eddy covarianceDELTA DEnuder for Long-Term Atmospheric samplingBASFOR BASic FORest ecosystem modelCTM Chemical transport modelEMEP European Monitoring and Evaluation Programme (httpwwwemepint last access 22 August 2019)GHG Greenhouse gasGWP Global warming potentialCEIP CarboEurope Integrated ProjectNEU NitroEurope Integrated ProjectFLUXNET Worldwide carbon flux monitoring network



Table 3 Summary of the main methods used to quantify carbon nitrogen and greenhouse gas fluxes and budgets for the 31 forests and nineshort semi-natural vegetation sites included in this study Horizontal bars (green forests blue short semi-natural vegetation) indicate thepercentage of study sites with available data (filled bars) or without available data (open bars) See also Tables S6ndashS7 for details at individualsites

1 Aubinet et al (2000) 2 Daumlmmgen (2006) 3 Dinsmore et al (2010) 4 Dise et al (2009) 5 Flechard et al (2011) 6 Gielen et al (2011) 7 Hendriks et al (2007) 8 Ilvesniemi etal (2009) 9 Kindler et al (2011) 10 Kowalska et al (2013) 11 Legout et al (2016) 12 Luo et al (2012) 13 Pilegaard et al (2006) 14 REddyProc (2019) 15 Schaufler et al (2010)16 Simpson et al (2012) 17 Tang et al (2009) 18 van Oijen et al (2005) 19 See Table S7

bient Nr concentrations were multiplied by a vegetation- meteorology- and chemical-species-dependent depositionvelocity (Vd) (Flechard et al 2011 2013 Bertolini et al2016 Thimonier et al 2018) In the case of NH3 a canopycompensation point scheme was applied in some models al-lowing bidirectional exchange between the surface and theatmosphere Considering notoriously large uncertainties indeposition velocities and large discrepancies between thesurface exchange schemes currently used in different CTMswe tried here to minimize such uncertainties by using theensemble average dry deposition predicted by four differentmodels as in Flechard et al (2011)

The dry deposition of atmospheric organic Nr (ON)species not accounted for by the EMEP model (eg aminesurea) and not included in DELTA measurements can con-tribute a fraction of totalNr deposition However Kanakidouet al (2016) suggest that particulate ON largely dominatesthe atmospheric ON load and for particles the main atmo-spheric removal mechanism is through precipitation Thusdry deposition of ON is expected to be much smaller than

wet deposition of water-soluble organic compounds (see be-low)

For wet deposition several sources of data were used andthe final wet deposition estimate was derived from the arith-metic mean of the different sources where available Firstwithin the NEU project a survey was made of the availablenational andor transnational (eg EMEP 2013 ICP 2019)wet deposition monitoring network concentration data for in-organic N (NH+4 NOminus3 ) in the different European countrieshosting one or several CEIPNEU flux sites These data werechecked for consistency and outliers harmonized and thenspatially interpolated by kriging to provide measurement-based estimates of solute concentrations in rainfall for eachof the 40 sites of this study Wet deposition was then calcu-lated as the product of interpolated concentration times mea-sured precipitation at each site

Next 13 sites (DB1 DB3 DB4 EN4 EN9 EN13 EN14EB2 EB3 MF1 MF2 SN3 SN8) were identified as lack-ing local or nearby wet deposition measurements These siteswere equipped for three years (2008ndash2010) with bulk (open



funnel) precipitation samplers (Model B Rotenkamp Ger-many Daumlmmgen 2006) mounted above the canopy or insidea clearing for some of the forest sites with monthly samplechange and analysis The precipitation samples were stabi-lized by addition of thymol at the beginning of each expo-sure period and were analysed subsequently for inorganic Nr(NH+4 and NOminus3 ) as well as SO2minus

4 Clminus PO3minus4 base cations

(Mg2+ Ca2+ K+ Na+) and pH A few other sites (EN2EN8 EN10 EN16 DB2 SN9) were already equipped withwet-only or bulk precipitation collectors No correction wasapplied to the bulk deposition estimates to account for a pos-sible contribution by dry deposition within the sampler glassfunnel (eg Daumlmmgen et al 2005) since there did not ap-pear to be any systematic overestimation compared with wetdeposition estimates from the monitoring networks or EMEPdata (see Results and Fig S2) even if a more significant biasmay be expected in dry (Mediterranean) regions

In addition to inorganic nitrogen the wet deposition ofwater-soluble organic Nr (WSON) compounds was also in-vestigated in precipitation samples at 16 sites (Cape et al2012) However since WSON data were not available forall sites and the measurements were subject to considerableuncertainties (Cape et al 2012) and also because the contri-bution of WSON to total Nr deposition was on average lessthan 5 WSON was not included in the final estimates oftotal Nr deposition

The last data source was the ca 50 kmtimes 50 km griddedmodelled wet inorganic Nr deposition (also NO2 concen-trations discussed above) simulated by the European-scaleEMEP CTM (Simpson et al 2006a b 2012 2014) for theyears 2007ndash2010 available from EMEP (2013) The datawere downloaded in 2013 and it should be noted that in thisdata series different model versions were used for the differ-ent years This leads to some uncertainty especially in thedry deposition estimates but it is hard to say which modelversion is the most realistic Evaluation of the model againstmeasurements over this period has shown quite consistent re-sults for the wet-deposited components and NO2 concentra-tions but the dry deposition rates cannot be evaluated versusactual measurements at the European scale We chose there-fore to make use of all versions and years giving a smallensemble of simulations

222 Soil gaseous and leaching losses

Nitrogen losses to the atmosphere (gaseous emissions) and togroundwater (N leaching) are especially hard to quantify andthus typically cause large uncertainties in ecosystem N bud-gets These Nr losses were estimated by direct flux measure-ments or by indirect empirical methods Soil NO and N2Oemissions were measured in the field using closed static anddynamic chamber methods as part of NEU (eg EN2 EN10EN16 DB2 SN3 SN8 SN9) andor collected from the liter-ature (eg EN2 EN10 EN14 EN16 DB2 Pilegaard et al2006 long-term data at EN2 in Luo et al 2012) Such data

were available for N2O at seven forest sites and four semi-natural sites as well as at five forest sites for NO (Table S6)Manual static chamber N2O measurements were made man-ually at a typically fortnightly (growing season) or monthly(winter half-year) frequency at many sites Automatic cham-ber systems allowing continuous N2O measurements at afrequency of four times per day were deployed at EN2EN10 DB2 and SN3 Fluxes of NO were only measured byautomatic dynamic (open) chambers Measured fluxes werescaled up to yearly values by linear interpolation or usingthe arithmetic mean of all flux measurements There may beconsiderable uncertainty in the annual flux if gap-filling isbased on linear interpolation between discrete values whenflux measurements are made manually and are therefore dis-continuous and infrequent (Parkin 2008) This is due to theepisodic nature and log-normal distribution of NO and N2Oemissions observed particularly in fertilized croplands andgrasslands However this episodicity is less pronounced insemi-natural ecosystems or at least the magnitude of theepisodic fluxes is generally much smaller than in fertilizedagro-systems (Barton et al 2015) The uncertainty in annualemissions estimated in our study from manual chamber mea-surements is related to the observation frequency (fortnightlyor monthly) and is likely larger than in the case of automatic(continuous) chamber measurements

Direct in situ Nr and non-CO2 GHG gas flux measure-ments were unavailable at many sites These soil N2O andNO (and also CH4) fluxes were therefore also estimated aspart of NEU from empirical temperature and moisture re-sponses of soils These responses were established in a seriesof factorial soil incubation experiments in controlled condi-tions with four levels of temperature (5ndash20 C) and water-filled pore space (20ndash80 WFPS ) following the protocoldescribed in Schaufler et al (2010) Twenty-four undisturbedsoil cores (top 5 cm of the mineral soil Ah horizon) weretaken from each of 27 forests and eight semi-natural sites inspring after soils had warmed up above 8 C for 1 week inorder to guarantee phenological comparability of the differ-ent climatic zones Sampling was conducted in 2008 2009and 2010 and cores were sent to a common laboratory atthe Federal Research and Training Centre for Forests (BFWVienna Austria) for the controlled environment bioassayswhich were carried out straight away The 5 cm topsoil layerwas selected as it represents the highest microbial activ-ity and correspondingly high GHG productionconsumptionrates although processes in deeper soil layers should notbe neglected (Schaufler et al 2010) Site-specific empiricalbivariate (temperature WFPS) relationships describing soilfluxes for CO2 N2O NO and CH4 were derived from the in-cubation results and then applied to multi-annual time seriesof soil temperature and moisture measured at the sites mim-icking field conditions and providing scaled-up estimates ofpotential annual trace gas emissions

Leaching of dissolved inorganic nitrogen (DIN=NH+4 +NOminus3 ) was measured using lysimeter setups or estimated



from a combination of suction cup measurements (typicallysim 1 m soil depth) and a hydrological drainage model at a fewsites during the NEU monitoring period (EN2 EN4 EN10EN15 EN16 DB1 DB2) and as part of parallel projects(EN8 DB4) One-dimensional (1-D) drainage models werebased on the soil water balance equation using evapotran-spiration observed precipitation and changes in soil watercontent (Kindler et al 2011 Gielen et al 2011) For theforest sites where no leaching measurements were availablethe empirical algorithm by Dise et al (2009) was appliedto predict DIN leaching based on key variables (through-fall inorganic Nr deposition DINTF organic horizon CNratios MAT) The algorithm developed from the extensiveIndicators of Forest Ecosystem Functioning (IFEF) database(gt 300 European forest sites) simulates the non-linearity ofDIN leaching with respect to DINTF and soil CN ratiowith critical thresholds for the onset of leaching of DINTF =

08 g N mminus2 yrminus1 and CN= 23 respectively Since the al-gorithm requires DINTF as input as opposed to total (abovecanopy) Ndep in the present study we applied a reductionfactor of 085 from Ndep to DINTF (ie a canopy retentionof 15 of atmospheric N) which was calculated as the av-erage of all available individual DINTF Ndep ratios in theIFEF database A comparison with values of DINTF Ndepratios actually measured at the EN2 EN8 EN10 EN16and DB2 sites (071 080 029 085 and 111 respectivelymeanplusmnSD 075plusmn 030) shows that the applied ratio of 085is plausible but also that much variability in canopy reten-tionleaching may be expected between sites

23 Carbon fluxes

231 Ecosystemndashatmosphere CO2 exchange

Half-hourly rates of net ecosystemndashatmosphere CO2 ex-change (NEE) were measured over several years (on aver-age 5 years see Table S6) by the eddy covariance (EC) tech-nique at all sites The long-term net ecosystem productivityis defined following Chapin et al (2006) as the difference be-tween gross primary production (GPP) and ecosystem respi-ration (Reco) and is thus calculated as the straightforward an-nual sum of NEE fluxes (with opposite sign) The net ecosys-tem carbon balance may differ from the NEP if C fluxesother than assimilation and respiration such as DICDOCleaching CH4 and other volatile organic compound (VOC)emissions as well as lateral fluxes (harvest thinning) andother disturbances (fire) are significant over the long term(Chapin et al 2006) For convenience in this paper we usethe following sign convention for CO2 fluxes GPP and Recoare both positive while NEP is positive for a net sink (aC gain from an ecosystem perspective) and negative for a netsource Previous studies have normalized C flux data throughthe carbon use efficiency (CUE) commonly defined from aplantrsquos perspective as the ratio of net to gross primary pro-ductivity (NPP GPP) or the biomass production efficiency

(BPE=BP GPP Vicca et al 2012) which is a CUE proxyBy analogy we define here an ecosystem-scale medium-term indicator of carbon sequestration efficiency (CSE) asthe NEP GPP ratio calculated from measurable fluxes overthe CEIPNEU project observation periods

The EC technique is based on fast-response (samplingrates typically 10ndash20 Hz) open-path or closed-path infraredgas analyser (IRGA) measurements of turbulent fluctuationsin CO2 concentration (c) in the surface layer above theecosystem coupled with ultrasonic anemometer measure-ments of the three components of wind (u v w) and tem-perature The NEE flux is calculated as the average productof c and w fluctuations ie the covariance (Swinbank 1951Lee et al 2004)

The EC-CO2 flux measurements reported here followedthe protocols established during the CEIP project largelybased on the EUROFLUX methodology (Aubinet et al2000) Briefly post-processing of the raw high-frequency ECdata included typically de-spiking to remove outliers 2-Drotation of the coordinate system time lag optimization bymaximization of the covariance between CO2 concentrationand the vertical component of wind speed (w) and block-averaging over the flux-averaging interval of 30 min Cor-rections were applied for various methodological artefactsincluding notably (i) flux losses at the different frequenciesof flux-carrying eddies caused eg by attenuationdampingin the inlettubing system (Ibrom et al 2007 Fratini etal 2012) path averaging sensor separation analyser re-sponse time and high- and low-pass filtering (ii) effectsof temperature fluctuations and dilution by water vapouron measured fluctuations in concentrations of CO2 (WebbndashPearmanndashLeuning corrections Webb et al 1980) and (iii)CO2 storage below sensor height Quality assurance andquality control procedures were further developed and agreedupon within CEIP including statistical tests non-stationarityintegral turbulence characteristics (Foken et al 2004) andfootprint evaluation (Goumlckede et al 2008) Friction velocity(ulowast) threshold filtering was implemented using the movingpoint test according to Papale et al (2006) and as describedin REddyProc (2019) in order to discard flux data from pe-riods of low turbulence

Different EC post-processing software was used at the dif-ferent sites within the project such that the data were notevaluated in exactly the same way across the CEIP networkbut a reasonably good overall agreement was found amongthe different software within 5 ndash10 difference for 30 minCO2 flux values (Mauder et al 2008 Mammarella et al2016) Similarly for the gap-filling of the 30 min flux timeseries during periods of instrument malfunction or unsuit-able measurement conditions (low turbulence insufficientfetch etc) and for the partitioning of NEP into GPP andReco a number of alternative algorithms have been devel-oped in the past based on different sets of principles (Falgeet al 2001 Barr et al 2004 Reichstein et al 2005 Lass-lop et al 2010) The gap-filling and partitioning algorithm



used by default in this study was the generic online REd-dyProc (2019) software implemented also in the EuropeanFluxes Database Cluster REddyProc was based on (i) Re-ichstein et al (2005) for the filling of gaps in the NEE fluxdata on the basis of information from environmental con-ditions (ii) Reichstein et al (2005) for the night-time-data-based Reco parameterization (using an Arrhenius-type func-tion of temperature) and (iii) on Lasslop et al (2010) for thedaytime-data-based GPP evaluation (using a rectangular hy-perbolic light response curve for NEE and including a tem-perature sensitivity of respiration and limitation of GPP byvapour pressure deficit)

In this study for all CEIP flux sites we have retrieved thefully analysed and validated half-hourly (level-3) and dailyto annual (level-4) CO2 flux (NEP GPP Reco) data as avail-able initially from the CEIP database and later from the Eu-ropean Fluxes Database Cluster (2012) or from the GHG-Europe portal (GHG-Europe 2012) For these data althoughthe evaluation methods were not necessarily harmonized be-tween sites we hold that the data available in the databasewere obtained using the best possible state-of-the-art evalu-ation methods at the time of retrieval For the four non-CEIPflux sites flux evaluation closely followed CEIP protocolsin the case of DB4 the EddyPro (v62) software was usedwhich was based on a synthesis of calculation and correc-tion methods from CEIP and other FLUXNET flux networksaround the globe

The EC-CO2 flux measurements used in this study mostlyspanned the 5-year period of CEIP (2004ndash2008) except fora dozen sites where measurements continued until 2010 iethe end of NEU and of atmospheric Nr sampling Older ECdata (since the mid-late 1990s) were also available at DB5EN6 and EN13 Data collection started and ended later atDB4 at which both EC-CO2 flux and DELTA-Nr measure-ments spanned the 7-year period 2009ndash2015 Data analysespresented in the paper based on inter-annual mean CO2 bud-gets and mean Nr deposition assume that five or more yearsof monitoring yield reasonably robust estimates of long-termfluxes for the different sites and that the small time shift be-tween the CEIP and NEU project periods (2ndash3 year overlap)does not affect the results significantly At some sites such asDB2 long-term NEE measurements showed multi-decadalvariations (Pilegaard et al 2011 Wu et al 2013) thus itwas essential to use the years overlapping with NEU

232 Soil CO2 and CH4 fluxes

In situ soil CO2 efflux (SCE) measurements by opaque (staticor dynamic) manual chambers were carried out at 25 of theforest sites with typically weekly to monthly sampling fre-quency with fluxes being measured continuously (hourly) byautomated chambers at a few sites (eg EN2) The SCE isusually considered a proxy for CO2 production by soil respi-ration (Rsoil) though the two may not be equal as part of theCO2 production is dissolved into pore water and may reach

the atmosphere only later either on-site or even off-site if dis-solved CO2 (DIC) leaches to groundwater Annual Rsoil datascaled up from SCE measurements are available for 19 for-est sites and were collected from the CEIP or GHG-Europedatabases andor from various peer-reviewed publications forthe different sites (see Table S7) The ratio of heterotrophicrespiration (Rhet) to Rsoil was determined on an annual scaleat 16 sites by different techniques (root-exclusion meshestrenching experiments radiocarbon or stable isotope tracingtree girdling eg Subke et al 2006) (Table S7)

Methane fluxes were measured by chamber methods oreddy covariance at six forest sites and five semi-natural (peat-land wetland) sites (Hendriks et al 2007 Skiba et al 2009Drewer et al 2010 Shvaleva et al 2011 Luo et al 2012Kowalska et al 2013 Juszczak and Augustin 2013) (Ta-ble S6) These data were complemented by bioassay mea-surements of CH4 emission or uptake (net oxidation) by thelaboratory soil cores as described previously for NO andN2O estimates (Schauffler et al 2010)

233 Dissolved carbon losses

Dissolved inorganic (excluding CO2 from weathering of car-bonate rocks) and organic carbon (DICDOC) fluxes weremeasured at six forest sites (DB1 DB2 EN4 EN8 EN10EN15) using suction cups for sampling soil water and com-bined with soil drainage data or by monitoring water runoffthrough weirs as part of CEIP NEU and other projects (Il-vesniemi et al 2009 Kindler et al 2011 Gielen et al 2011Verstraeten et al 2014) Data were also available for peat-land at SN7 with DIC DOC and also dissolved CH4 con-centrations in pore water of the clayey peat in groundwaterfrom the sand aquifer and in ditch water as described in Hen-driks et al (2007) For the peatland within SN9 Dinsmoreet al (2010) measured stream concentrations and export ofDIC DOC and particulate organic carbon (POC) and theyalso estimated stream evasion of CO2 CH4 and N2O in ad-dition to the land-based flux (EC chamber) measurements inthe tower footprint

24 Ecosystem greenhouse gas balance

Net GHG budgets were constructed from inter-annual meanEC-based NEP combined with measured and scaled-up N2Oand CH4 fluxes wherever available (nine and six sites re-spectively) or with bioassay-derived fluxes (most sites) ormodelled data (BASFOR forestsN2O only) using 100-year global warming potentials (GWPs) of 265 and 28 forN2O and CH4 respectively (Fifth Assessment Report IPCC2013) The sign convention for non-CO2 GHG fluxes and forthe net ecosystem GHG balance in this paper adopts an at-mospheric warming perspective ie positive fluxes for emis-sions toward the atmosphere (warming) and negative for up-take by the surface (cooling)



25 Ancillary soil plant and ecosystem measurements

Ancillary data were collected mainly for the purpose of as-sembling input parameters and calibration datasets for for-est ecosystem (BASFOR) modelling (see below) Texture( clay sand silt) pH soil organic carbon (SOC) con-centration and CN ratios were measured in soils of 35 sitesas part of the bioassay experiments described previously butwere otherwise also documented in the CEIP database andin papers previously published for the majority of sites (Ta-ble S1) For the forest sites ecosystem data for soil watercontent (SWC) porosity saturation water content (8SAT)field capacity (8FC) and wilting point (8WP) as well as forcanopy height (H) leaf area index (LAI) diameter at breastheight (DBH) basal area (BA) number of trees per unit areaor stand density (SD) and thinning events were obtainedfrom CEIP and other project (eg FLUXNET) databases andcomplemented by various publications (Tables S2ndashS5) Suchwas also the case for ecosystem carbon stocks in soil organicmatter (CSOM) and in roots (CR) stems (CS) branches(CB) leaves (CL) and litter layers (CLITT) for which theglobal database assembled by Luyssaert et al (2007) pro-vided additional data At sites for which published values of8FC and 8WP were not available default estimates were in-ferred from soil texture by means of van Genuchten (1980)pedo-transfer functions using tabulated values from the Ger-man soil description handbook (Eckelmann et al 2005)

Foliar C and N contents (LeafC LeafN) were measuredas part of NEU for EN1 EN2 EN5 EN8 EN10 EN15EN16 DB2 (Wang et al 2013) DB4 SN3 SN4 SN8 andSN9 or were otherwise taken from CEIP GHG-Europe andFLUXNET databases as well as various publications in to-tal leaf CN measurements were available for 31 sites Bycontrast data were much rarer for CN ratios for other com-partments of the forest ecosystem with data available at only15 sites for litter and at only five sites for roots stems andbranches

26 Modelling of C and N fluxes and pools by theBASFOR ecosystem model

The BASic FORest model (BASFOR) is a deterministic for-est ecosystem model that simulates the growth (from plantingor natural regeneration) and the biogeochemistry of temper-ate deciduous and coniferous even-aged stands (van Oijen etal 2005 Cameron et al 2013) A description of the modeland the fortran code are available in BASFOR (2016) Themodel was calibrated through a multiple site Bayesian cali-bration (BC) procedure applied to three groups of plant func-tional types (DBF EN-spruce EN-pine) based on CNH2Oflux and pool data from the CEIPNEU databases (Cameronet al 2018) Details on model implementation as part of thisstudy are provided in Flechard et al (2020)

Briefly the C N and water cycles are simulated at adaily time step in interaction with the soil and climate en-

vironments and constrained by management (pruning andthinning) Carbon and nitrogen pools are simulated in thedifferent ecosystem compartments (tree stems branchesleaves and roots litter layers and SOM with fast andslow turnover) which are interconnected by internal flowsand transformations (eg SOM mineralization nitrogen re-translocation) Carbon nitrogen and water enter the ecosys-tem from the atmosphere (photosynthesis Nr depositionrainfall) Inorganic nitrogen is taken up from the soil by treeroots C and N return to the litter and soil pools upon senes-cence of leaves branches and roots and also when trees arepruned or thinned Losses of C occur through autotrophic(root and shoot) respiration and microbial decompositioninto CO2 of litter and SOM (heterotrophic respiration) lossesof N occur through nitrate leaching below the root zone andsoil emissions to the atmosphere of NO and N2O The waterbalance is constrained by incoming rainfall soil water hold-ing capacity and evapotranspiration (ET) simulated by thePenman equation

3 Results

31 Nitrogen inputs and outputs

311 Nitrogen deposition

Total inorganic Nr deposition ranged from 01 to43 g N mminus2 yrminus1 across the CEIPNEU networks (Ta-ble 1) with the largest values observed in the Netherlandsnorthern Belgium and southern Germany and the lowestlevels observed at latitudes gt 60 N (Fennoscandia) Nitro-gen deposition was dominated by the dry fraction in forests(Fig 2) with an average contribution to total deposition of63 versus 39 for short semi-natural vegetation Thiscontribution was even larger (gt 23) for high-depositionsites (Ndep gt 2 g N mminus2 yrminus1) Total Ndep was more stronglycorrelated to dry deposition across all sites (R2

= 094) thanto wet deposition (R2

= 056) Important differences in theratio of dry to wet deposition are evident across climatic re-gions with the share of dry deposition being especially largeat Mediterranean sites (eg Sanz et al 2002) where annualrainfall is smaller However the share of dry deposition wasalso large for sites that are located near (large) anthropogenic(industrial vehicular agricultural) Nr emission sourcesTotal Nr deposition was around 25 smaller on average atshort semi-natural vegetation sites compared with forests(Fig S2) even though the mean total atmospheric Nrconcentrations (reduced and oxidized N-containing gas andaerosol compounds) were quite similar between the twodatasets (Flechard et al 2011) The difference was drivenby higher dry deposition rates over forests due to higheraerodynamic roughness and deposition velocities (Fig S3see also Schwede et al 2018) Reduced Nr (NH3 gas andNH+4 in aerosol and rain collectively NHx) contributed on



average 56 of total deposition oxidized Nr (HNO3+NO2gas and NOminus3 in aerosol and rain collectively NOy) wasdominant at only six forest sites of the network (EN7 EN10EN18 EB2 SN3 SN5 Fig 2)

For comparison dry deposition calculated here as the en-semble average of four inferential model estimates based onin situ Nr concentration measurements was on average morethan a factor of 2 larger than the ca 50 kmtimes 50 km gridsquare-averaged EMEP model estimate (taken from EMEP2013) (see Fig S2) However since each EMEP grid squarecontains variable proportions of different land uses with dif-ferent deposition velocities it is more meaningful to compareDELTA-based inferential estimates for each study site withecosystem-specific EMEP dry deposition rates in the relevantgrid squares In this case the EMEP dry deposition rates areon average 32 smaller than the inferential estimates

By contrast wet deposition was generally reasonably con-sistent between the different data sources for inorganic Nr(in situ bulk or wet-only measurement kriging of monitoringnetwork data EMEP model output) For the 18 sites whereall three sources of data were available the mean CV ofthe three estimates was 21 (range 2 ndash56 with 15 CVvalues out of 18 below 30 ) and the mean (plusmn95 confi-dence interval) wet deposition estimates across the 18 siteswere 063plusmn 014 064plusmn 015 and 068plusmn 016 g N mminus2 yrminus1

for the three methods respectively (Fig S2) showing no sys-tematic bias between methods Wet deposition of organic ni-trogen measured at 16 sites represented on average 11 (range 2 ndash36 ) of total inorganic+ organic wet deposition(Fig S2) but only 4 (range 1 ndash30 ) of total dry+wetNr deposition since total Ndep was dominated by dry depo-sition at most forest sites

312 Nitrogen losses

Total ecosystem losses of inorganic Nr were computed forthe forest sites as the sum of DIN leaching and NO and N2Oemissions (Fig 3andashd) We assumed that NH3 emissions bysoil and vegetation were negligible due to generally acidicforest soils as well as low values of the stomatal compen-sation point (the leaf NH3 emission potential) respectively(Flechard et al 2013) Inorganic Nr losses (Fig 3d) in-creased sharply with Nr deposition and were largely domi-nated by DIN leaching at Ndep levels above 2 g N mminus2 yrminus1

(Fig 3c) For these large Ndep levels the fraction of de-posited Nr lost as DIN NO or N2O was generally larger than50 (Fig 3f) The inorganic Nr balance (Nr deposition mi-nus NO N2O and DIN losses) was probably still positive formost sites (Fig 3e) although the confidence intervals of thebudget term (accounting for uncertainties in all terms includ-ing deposition) were very large for the elevated Nr deposi-tion sites Note that the DIN leaching estimate by BASFORshown for comparison on Fig 3c was not used in the calcu-lation of total inorganic N losses in Fig 3d this is becauseBASFOR does not simulate N2 loss by denitrification and

thus part of the soil N surplus that would in reality denitrify isassumed to drain resulting in an overestimation of the leach-ing term though not necessarily of the total N losses

Emissions of NO estimated from bioassay measurements(Schaufler et al 2010) and by BASFOR modelling were gen-erally of the same order in forests (average values acrossall forest sites of 022 and 021 g N mminus2 yrminus1 respectively)but validation by in situ chamber flux data was difficult ow-ing to the limited number of available measurements (onlyfive forest sites mean value 027 g N mminus2 yrminus1) Nonethe-less the largest NO emissions by the three methods wereall found at Ndep levels above 2 g N mminus2 yrminus1 (Fig 3a) Bycontrast N2O emissions did not show any marked depen-dence on Ndep and were on average smaller than NO emis-sions by a factor of 2 to 5 with mean values across all sitesof 012 008 and 004 g N mminus2 yrminus1 for bioassay BASFORand chamber fluxes respectively The mean N2O fluxes (av-eraged over the different methods) were larger than mean NOfluxes at only one-third of the forests sites by contrast at SNsites N2O emissions were larger than NO emissions at allbut one location The dominance of NO over N2O in forestscould in principle reflect the generally well aerated condi-tions of (especially coniferous) forest litter layers on well-drained topsoils which are more conducive to NO formationby nitrification than N2O by denitrification (Davidson et al2000 Pilegaard et al 2006) This would be perhaps espe-cially true for the four highest (gt 3 g N mminus2 yrminus1) Nr deposi-tion sites (EN2 EN8 EN15 and EN16 all coniferous forests)with the highest NO emissions (Fig 3) which all had sand-dominated (64 ndash96 ) soil textures (Table S4) On the otherhand given the acidity of many forest topsoils (Table S4) ni-trification could be inhibited but chemodenitrification couldproduce significant amounts of NO (Pilegaard 2013)

For a complete ecosystem net N budget additional mea-surements of dissolved organic nitrogen (DON) leaching aswell as dinitrogen (N2) fluxes (biological fixation and totaldenitrification) would be required (Fig 1) but they were notquantified in most cases A tentative ballpark estimate of thepotential magnitude of denitrification N2 emissions for theDB2 forest site may be calculated by considering the meanN2N2O ratio of 74 (plusmn085 st err) which was measuredin HendashO2 mixture soil incubation experiments performed onDB2 soil cores (unpublished data) This mean ratio mul-tiplied by the mean field-measured N2O emission flux of0074 g N2O-N mminus2 yrminus1 (Pilegaard et al 2006) yields anestimate of the order of 55 g N2-N mminus2 yrminus1 There is con-siderable uncertainty in this number since the mean N2N2Oratio was calculated from short-term investigations in the lab-oratory which may or may not be representative of the pre-vailing soil and weather conditions in the field This uncer-tainty is reinforced by the low sensitivity of the N2 detectorwhich was a factor of 20ndash80 lower than that of the N2O de-tector used in the experiment (Buchen et al 2019) Anotherestimate of forest soil denitrification loss obtained through asoil core incubation method was given by Butterbach-Bahl et



Figure 2 Total reactive nitrogen deposition (Ndep) and breakdown into inorganic wet and dry oxidized (NOy ) and reduced (NHx ) depositionestimates at the 31 forest sites (evergreen needleleaf EN1-7 (spruce) EN8-18 (pine) mixed ndash MF deciduous broadleaf ndash DB evergreenbroadleaf ndash EB) and at nine short semi-natural (SN) vegetation sites of the NitroEurope monitoring network Data are arithmetic meansover the years 2007ndash2010 of (i) inferential dry deposition estimates by four different models based on in situ atmospheric Nr measurementsand of (ii) different wet deposition estimates from precipitation monitoring datasets and from European-scale atmospheric chemistry andtransport modelling (EMEP) Error bars indicate standard deviations of the four dry deposition models (red bars) and standard deviationsof the different data sources for inorganic Nr wet deposition (blue bars) Wet deposition of water-soluble organic nitrogen (WSON) wasmeasured at a few selected sites and is shown here for comparison with total inorganic Nr deposition

al (2002) for the EN2 spruce site with an annual N2 emis-sion flux of 072 g N2-N mminus2 yrminus1 and a mean N2N2O ratioof 7 The N2 emissions thus estimated suggest that total deni-trification may be a very significant term in the total N budgetof forests possibly of the same order as atmospheric Nr de-position

Measurements of DON leaching were available at veryfew sites but proved to be significant At the pine forest site ofEN8 DON leaching was of the order of 03 g N mminus2 yrminus1 iea factor of 3 lower than DIN losses (Verstraeten et al 2014)At the beech forest site of DB2 DIN and DON leachingwere of the same order (007ndash008 g N mminus2 yrminus1) but bothwere very small in comparison to Ndep (215 g N mminus2 yrminus1)while at the pine forest site of EN10 the leachingrunoff Nrloss was actually dominated by DON (0012 g N mminus2 yrminus1)which was around an order of magnitude larger than DINleaching (Korhonen et al 2013) and a factor of 4 smallerthan Ndep

32 Net carbon and greenhouse gas balance

321 Spatial variability of the carbon sink in relationto climate and nitrogen deposition

The ultimate objective of the project was to quantify the re-sponse of C sequestration to atmospheric Nr deposition (ad-dressed in Flechard et al 2020) but this is not straightfor-ward We follow first in this paper a descriptive approach

whereby variations of C fluxes and other productivity indica-tors (eg leaf area index and N content) are examined graph-ically as a function of Ndep (Fig 4) However this is donewith the strong reservation that a simple empirical relation-ship does not necessarily prove causality as other confound-ing and co-varying factors eg climate soil and age mayexist Figures 4ndash5 show for example that the large inter-sitedifferences in MAT and MAP at the European scale also needto be considered beside the variability in Ndep Note that inassessing the variability of ecosystem carbon sink strengthwithin the network we use EC-derived NEP (the long-termNEE sum) as a proxy for the net ecosystem carbon balancebecause estimates of DICDOC leaching CH4 emissions andother C loss processes were not systematically measured atall sites

Inter-annual mean NEP ranged from a small net sourceofminus70 g C mminus2 yrminus1 (EN6 a waterlogged peat-based sprucestand in the southern Russian taiga) to a large net sink of+826 g C mminus2 yrminus1 (EN5 upland spruce forest in northernItaly) (Table 1 Fig 4c) GPP ranged from 377 g C mminus2 yrminus1

(SN3 a boreal peatland site with the lowest MAT=minus06 C)to 2256 g C mminus2 yrminus1 (EN14 a pine stand in Italy one of thewarmest sites with MAT of 149 C and non-limiting rain-fall with MAP= 920 mm) (Fig 4a) Ecosystem respirationpeaked at 1767 g C mminus2 yrminus1 at EN4 (upland spruce forestin eastern Germany) and was lowest at 345 g C mminus2 yrminus1 atSN3 (boreal peatland) the coldest site (Fig 4b) Reco wasstrongly and positively related to GPP (Fig 4f) (R2

= 062



Figure 3 Comparison of measured and estimated ecosystem inorganic Nr losses and their relationships to total atmospheric Nr deposition(x axis) at the forest sites NO fluxes (a) and N2O fluxes (b) were either (i) measured in situ using static or dynamic flux chambers (ii) scaledup from laboratory-bioassay-derived T ndashWFPS relationships or (iii) simulated using the BASFOR ecosystem model (see text for details)DIN leaching (c) was either measured (lysimeter or suction cups) or predicted from the Dise et al (2009) empirical algorithm The sum ofinorganic Nr losses (d) (DINleach+NO+ N2O) was computed as the mean of measured values and modelled estimates In panels (a)ndash(c)site names are indicated for sites where in situ measurements were available

slope= 064) The resulting carbon sequestration efficiencyvalues based on the ratio of observed NEP GPP (CSEobs)varied widely among observation sites ranging from minus9 to 61 with an average of 25

The data show a positive correlation between GPP andNdep in the range 0ndash25 g N mminus2 yrminus1 (R2

= 055 plt001)By contrast the five sites with Ndep gt 25 g N mminus2 yrminus1 tendto show visually an inverse relationship (Fig 4a) despitethe fact that they lie in comparatively favourable climatesSimilar patterns are observed for Reco and NEP (Fig 4bndashc)but with much larger scatter and lower R2 (024 plt001and 030 plt001 respectively for the Ndep range 0ndash25 g N mminus2 yrminus1) with the same apparent decline for higher-deposition sites However a closer inspection of Fig 4andashcreveals a potential cross-correlation with climate (see alsoFig S4) (i) the lower end of the Ndep range coincidingwith the lowest GPP Reco and NEP also coincides withthe lowest MAT and MAP (eg Finnish sites) and (ii) thesites in the intermediate Ndep range (15ndash25 g N mminus2 yrminus1)coinciding mostly with the largest observed GPP values(gt 1500 g C mminus2 yrminus1) were on average 18 C warmer (102vs 84 C) and 89 mm yrminus1 wetter (887 vs 798 mm) than thesites in the lower Ndep range (0-15 g N mminus2 yrminus1)

Other proxies of the ecosystem C and N cycles and pro-ductivity such as the LAI (defined as one-sided for broadleafor half of the total for needleleaf Table 1 and Fig 4d) and thefoliar N content (LeafN Fig 4e) also showed positive rela-tionships to Ndep (see below for differences between vegeta-tion types) The inter-annual mean value of the annual max-imum leaf area index (LAImax) increased from around 1 to7 m2 mminus2 for Ndep increasing from 01 to 45 g N mminus2 yrminus1with the lower half of the LAImax distribution (lt 45 m2 mminus2)mostly occurring at boreal Mediterranean and upland sitesand thus under temperature andor water limitations

Clearly therefore the continental-scale variability inecosystemndashatmosphere CO2 fluxes was to a large extent con-trolled by climate namely by limitations in temperature andwater availability Gross ecosystem productivity was limitedas expected by low temperatures at high latitudes (or highelevations) and by low rainfall andor high evaporative de-mand at Mediterranean boreal and continental sites The dis-tribution of the forest monitoring sites in the European cli-mate space with MAP and MAT on the x and y axes re-spectively (Fig 5andashb) shows that for sites with MAT gt 7 Cthere was a broad negative correlation between MAT andMAP (R =minus049 p = 001) ie the warmest sites in south-



Figure 4 Overview of inter-annual mean EC-derived C flux estimates (GPP Reco and NEP) ecosystem LAI and leaf N content in relationto total (dry+wet) atmospheric Nr deposition (andashe) and relationship of Reco to GPP (f) for forests (filled circles black labels) and shortsemi-natural vegetation (filled stars magenta labels) In all plots the colour scale indicates mean annual temperature (MAT) while thesymbol size is proportional to mean annual precipitation (MAP scale provided in panel a)



ern Europe tend to be the driest and therefore potentiallywater-limited Maximum GPP (and alsoReco not shown) oc-curred in the mid-climate range around 9ndash15 C MAT andaround 700ndash1000 mm MAP Similarly the largerNdep values(gt 2 g N mminus2 yrminus1) occurred almost exclusively at sites withMAT in the narrow range of 6ndash11 C and although theselarge Ndep values were found in a broad MAP range (550ndash1200 mm) they peaked sharply around 800ndash900 mm MAP(Figs 5a S4) Modelled Ndep values from the EMEP CTM(Fig 5cndashd) show that this is a generic pattern at the Europeanscale

Ecosystem DIC+DOC losses estimated by Kindler etal (2011) for four forest sites of this study (DB1 DB2EN4 EN15) were on average 13plusmn 7 g C mminus2 yrminus1 (range3ndash35 g C mminus2 yrminus1) with contributions by DIC to total(DIC+DOC) losses varying between 18 and 83 Bycontrast Gielen et al (2011) estimated DOC leaching lossesof 10plusmn 2 g C mminus2 yrminus1 for the EN8 pine stand on an acidicsandy soil in which DIC concentrations in soil water werenegligibly small Ilvesniemi et al (2009) found DOC lossesin runoff at EN10 of 08 g C mminus2 yrminus1 which was negligi-ble compared with NEP These leaching or runoff losses ofDOC and DIC were on average over all forest sites equiv-alent to a very small mean fraction of 06 of GPP (range01 ndash19 ) but a more significant fraction of NEP (mean6 range 03 ndash13 ) At the SN7 peatland site fluxesof total dissolved carbon (including CH4) through seepageinfiltration and drainage were relatively small by compari-son to NEP and to other peat bogs (17 g C mminus2 yrminus1 only5 of NEP) (Hendriks et al 2007) by contrast at the SN9peatland site net stream C export (including DIC DOC andPOC) was on average 291 g C mminus2 yrminus1 (81 of which be-ing DOC) equivalent to a mean leached fraction of 37 ofNEP (Dinsmore et al 2010)

322 Differences between plant functional types

Forests (F) and short semi-natural (SN) vegetation showedsimilar relationships with GPP as a function ofNr depositionincreasing with a broadly similar slope at low Ndep valuesand then levelling off beyond 2g N mminus2 yrminus1 except for thefact that GPP was lower by typically 200ndash500 g C mminus2 yrminus1

in SN compared with F sites for a given Ndep level (Fig 4)The behaviour was different for NEP where the slope againstNdep in the range 0ndash2 g N mminus2 yrminus1 was much steeper for Fthan for SN which occurred because Reco values are of thesame order for F and SN at a given Ndep level No system-atic difference was observed between the forest PFT basedon the available data in the apparent relationships of the Cfluxes vs Ndep However this may be a result of the smallnumber ndash and large diversity ndash of deciduous broadleaf (DB)and evergreen broadleaf (EB) forest sites in the dataset com-pared with evergreen needleleaf (EN) sites (Table 1)

The relationship of LeafN to Ndep (Fig 4e) showed threedistinct groups with the smallest values (08 ndash18 N in

dry weight DW) for evergreen needleleaf and broadleaf (ENEB) forests being positively correlated to Ndep in the range05ndash43 g N mminus2 yrminus1 (R2

= 071 plt001) Values for shortsemi-natural (SN) vegetation were found in an intermediaterange (1 ndash27 N DW) with a steep and significant rela-tionship toNdep (R2

= 051 plt005) The largest values oc-curred for deciduous broadleaf (DB) forests (mostly gt 2 NDW) but with little relationship to Ndep (R2

= 018 not sig-nificant) Seasonal variations in forest LeafN could reach afactor of 2 as did differences between tree species withinthe same forest which may account for some of the scatterobserved in Fig 4e

323 Carbon fluxes and pools derived from forestecosystem modelling

In the BASFOR base run (Fig 6) reasonable overall modelperformance was achieved for GPP ecosystem C pools HDBH LAI and LeafN while more scatter was present forReco NEP and ET In particular in apparent contrast to GPPReco stands out as a more challenging variable to model Pre-dictably because BASFOR was calibrated using a subset of22 sites from this dataset (Cameron et al 2018) the rangeand mean values of modelled Reco were close to mean obser-vations by EC across the study sites but differences betweensites were poorly reproduced with much scatter around the11 line and a low R2 The modelled carbon sequestrationefficiency (CSEmod) simulated over the same time period asthe flux measurements was much less variable (range 17 ndash31 mean 22 ) than observation-based values (CSEobs)(comparison made for the 22 sites used in model calibra-tion) One possible reason was that BASFOR assumed thatautotrophic respiration (Raut) is a constant fraction of GPPwhich may be an oversimplification (Collalti and Prentice2019) Also heterotrophic respiration (Rhet) appeared to bea much more variable fraction of Reco in reality (Table S7)than was predicted by the model leading to sizable diver-gence in the overall modelled Reco As the direct measure-ment NEP was the least uncertain term in EC-derived datacompared with GPP and daytime Reco which were inferredfrom measured (half-hourly) EC NEE by empirical partition-ing models By contrast in BASFOR NEP was calculated asthe residual between two large numbers (GPP and Reco) andthus compounds the uncertainties of both component termsThe modelled result for NEP appeared to be an overestima-tion of net C uptake at low-productivity sites and an under-estimation at high-productivity ones (slope lt 1) A broadlysimilar pattern emerged for ET

324 Net ecosystem greenhouse gas budgets

Carbon dioxide largely dominated the net GHG budget atall forest sites with only three sites where either N2O orCH4 GWP-equivalent fluxes were larger than 10 of NEPin absolute terms (Fig 7) Most of the forest soils (22 out



Figure 5 Distribution of observation-based nitrogen deposition (Ndep) (a) and gross primary productivity (GPP) (b) for the forest sitesof this study within the European climate space represented by mean annual temperature (MAT) and precipitation (MAP) In plot (a) thesymbol colour indicates Ndep while the symbol size is proportional to GPP in plot (b) the symbol colour indicates GPP while the symbolsize is proportional to Ndep Plot (c) shows modelled Ndep from the EMEP model over coniferous forests (year 2010) represented in climatespace (one data point for each grid square of the EMEP domain containing coniferous forests) also shown as a map (d) The MAT axiscan be seen as a proxy for latitude andor elevation while the MAP axis expresses to some extent longitude (distance to the ocean) andororographic precipitation enhancement

of 27 sites) investigated in the bioassay experiment behavedas small net sinks for CH4 with a mean (plusmnSE) net oxi-dation flux of minus014plusmn 003 g C mminus2 yrminus1 (range minus061 to+016 g C mminus2 yrminus1) The mean CH4 flux measured by soilchambers at the six forest sites where such measurementswere available (EN2 EN6 EN10 EN16 DB2 EB5) wasalso a net oxidation flux ofminus032plusmn015 g C mminus2 yrminus1 (rangeminus10 to minus00 g C mminus2 yrminus1) For these six sites there was asignificant correlation (R2

= 074 plt005) between annualsoil CH4 flux estimates derived from the bioassay experimentand from in situ flux measurements (Fig S5 in Supplement)with the largest net annual soil CH4 uptake flux being ob-

served by both methods at the EN10 pine forest site (Skibaet al 2009) By contrast at the elevated Ndep sites EN2 andEN16 the net soil CH4 flux was close to zero consistentwith previous research (eg Steudler et al 1989 Smith etal 2000) showing that the CH4 oxidation capacity of forestsoils in negatively affected by Nr addition or deposition Interms of C uptake soil CH4 oxidation was negligible com-pared to CO2 fluxes representing on average only 01 ofNEP (range 00 ndash04 ) In terms of GWP the CH4 flux waslarger being equivalent to 08 of NEP (range 0 ndash45 )but on average still a factor of 3 smaller than the warming byN2O emissions equal to 39 of NEP (range 0 ndash185 )



Figure 6 BASFOR baseline simulations for all forest sites model outputs and observation-based values were averaged over the yearsbetween the first and last available observations Note that model simulations include MF and EBF sites for which the model was notcalibrated in Cameron et al (2018) the two MF runs were made using the parameter table for DBF while the five EBF runs were madeusing the parameter table for ENF to allow continued growth throughout the year H mean tree height DBH mean diameter at breast heightCLBS CSOM CR CLITT carbon stocks in leaves branches and stems in soil organic matter in roots and in litter layers respectivelyR2 coefficient of determination MAE mean absolute error NRMSE root mean square error normalized to the mean



Figure 7 Net greenhouse gas (GHG) budgets calculated from a combination of inter-annual mean (around 2005ndash2010) net ecosystemproductivity (NEP) from eddy covariance and N2O and CH4 flux data measured in situ or estimated by extrapolated bioassay data andforest ecosystem BASFOR modelling Global warming potential values (100-year time horizon) of 265 and 28 were used for N2O and CH4respectively the sign convention is with respect to the atmosphere negative for a sink and positive for a source The data were grouped byecosystem type (evergreen needleleaf EN-spruce and EN-pine MF mixed forests DB deciduous broadleaf EB evergreen broadleaf SNshort semi-natural vegetation) within each group the data were sorted by increasing Nr deposition

In contrast to forests at semi-natural short vegetation sitesN2O or CH4 emissions had a larger impact on the net GHGbalance where most (seven out of nine) sites showed non-CO2 GHG contributions larger than 10 of NEP Three ofthese seven sites were unfertilized extensively grazed up-land (SN2 SN5 SN6) grasslands (small N2O sources) whilethree sites (SN3 SN7 SN8) were CH4-emitting peatlandsor wetlands (EC-CH4 and chamber flux data from Dreweret al 2010 Hendriks et al 2007 Juszczak and Augustin2013 Kowalska et al 2013) At SN3 and SN8 the smallto moderate NEP sinks were turned by large CH4 emissionsinto net GHG sources (net warming budgets of +127 and+242 g CO2-C Eq mminus2 yrminus1 respectively) though not intoactual net C sources (Fig 7) At SN8 CH4 emissions gen-erally ranged from 25 to 45 g CH4-C mminus2 yrminus1 but reached86 g CH4-C mminus2 yrminus1 during a particularly wet year whenthe whole area was flooded At the SN9 peatland site Dins-more et al (2010) calculated that stream GHG evasion ndash atthe scale of the 335 ha peat bog encompassing the flux towerfootprint ndash together with downstream export represented 50ndash60 g CO2-Eq mminus2 yrminus1 (13ndash16 g CO2-C Eq mminus2 yrminus1) 96 of which being degassed CO2 ie in the range 11 ndash23 ofthe GHG budget from the tower footprint

4 Discussion

Previous observations of simple empirical relationshipsfound between N deposition and forest productivity havebeen criticized for amongst other things their low numberof replications unreasonably high sensitivities of productiv-

ity to N additions and limitations of the data and simplisticunivariate statistical approaches used (Magnani et al 2007Houmlgberg 2007 de Vries et al 2008 Sutton et al 2008)Other attempts have subsequently been made to assess im-pacts of N deposition on forest growth and carbon seques-tration while accounting for other drivers at more than 350long-term monitoring plots in Europe (Solberg et al 2009Laubhann et al 2009 De Vries et al 2008) A special fea-ture of the present study is that it aims to assemble N deposi-tion rates and budgets together with variables of the carboncycle for a wide range of sites across the European continentin more depth and completeness than hitherto attempted inorder to seek more robust empirical evidence for the responseof the terrestrial carbon cycle to different regimes of atmo-spheric N inputs The quality of the individual datasets ishowever not uniformly high Some of the data were mea-sured in situ with known uncertainty while others were sim-ulated derived from laboratory experiments and adapted tothe field situation using measured time series of soil T andsoil moisture or taken from existing databases and literatureAlso data may not be fully comparable between sites (dif-ferent methods used) or even fully representative of each site(spatial heterogeneity) In the following sections we discusslimitations of the measured empirical and simulated data interms of the component C and N fluxes their budgets andinteractions and the challenges faced when attempting to es-tablish empiricalstatistical evidence for possible N effectson carbon sequestration in natural and semi-natural terres-trial ecosystems in Europe



41 Constraining the ecosystem nitrogen balancethrough combined measurements and modelling

The compilation of Nr flux data (Fig 3) based on severalindependent sources for each component term provides a re-alistic picture of inorganic Nr inputs and losses their bal-ance suggests that for forests subjected to large depositionloads (gt 2 g N mminus2 yrminus1) typically more than half of the in-coming Nr is lost to neighbouring environmental compart-ments such as groundwater and the atmosphere and is thusnot available to promote C storage in the forest ecosystemSince N losses increase ndash and N retention decreases ndash expo-nentially when Ndep exceeds a critical load of approximately2ndash25 g N mminus2 yrminus1 (Fig 3) it seems unlikely that the C sinkstrength of semi-natural ecosystems including forests in-creases linearly with Nr deposition especially not with wetN deposition only Based on a review of experimental Naddition studies (eg Houmlgberg et al 2006 Pregitzer et al2008) and monitoring-based field studies along N depositiongradients (eg Solberg et al 2009 Laubhann et al 2009Thomas et al 2010) De Vries et al (2014) suggested thatthe C response reaches a plateau near 15ndash20 g N mminus2 yrminus1

and then starts to decrease The linear relationship betweenC sequestration and wet Nr deposition as proposed for ex-ample by Magnani et al (2007) is also challenged by thelarge contribution of dry Nr deposition and therefore by thepoor correlation between total Ndep and wet deposition Weargue that our multiple-constraint approach for the nitrogenbalance (measurementndashmodel combination model ensembleaveraging alternative data sources) provides overall a morerobust basis for studying the impact of Ndep on the C cycleeven though uncertainties in individual terms remain signifi-cant

411 Reducing uncertainty in nitrogen deposition

The uncertainty in dry deposition based on measuredNr con-centrations and inferential modelling is likely not smallerthan 30 due to limitations in process understanding Thedifference between ecosystem-specific EMEP values and themean inferential estimates (Fig S2) reflects discrepanciesand uncertainties in the four dry deposition schemes used(Flechard et al 2011) the mean coefficient of variation(CV= σmicro) between the four inferential model estimateswas 36 ie larger than the difference between ecosystem-specific EMEP values and the mean inferential estimatesOther sources of discrepancy between the two methods in-clude the use of measured vs modelled meteorology to drivethe deposition models and site-specific vs generic values ofcanopy height and leaf area index as discussed in Flechardet al (2011)

The uncertainty in total Nr deposition is probably of thesame order since even wet deposition can be deceptively dif-ficult to measure (Daumlmmgen et al 2005) and organic N es-pecially water-soluble organic N may be significant but chal-

lenging to quantify (Cape et al 2012) and generally ignoredin the literature WSON appears to be a generally small frac-tion of total (wet+ dry) Ndep at most sites except at remotelocations in Fennoscandia (EN10 SN3) where WSON depo-sition could represent up to 20 ndash30 of total Ndep Alsopotential double-counting due to dry deposition to the bulkdeposition collectors (eg Thimonier et al 2018) was notconsidered in this study although on the basis of the com-parison to other data sources (Fig S2) bulk samplers did notappear to significantly overestimate wet deposition

Despite these uncertainties measuring gas-phase andaerosolNr concentrations locally should provide a better esti-mate of total ecosystem Nr inputs than the outputs of a large-scale chemical transport model In addition the partitioningof wet vs dry deposition reduced vs oxidized N and canopyabsorption vs soil deposition should also be improved allof which are useful in interpreting ecosystem N-cycling pro-cesses In particular for ammonia with its high spatial vari-ability on a local scale the inferential modelling approachbased on local measurements is likely to provide more re-alistic deposition estimates than a coarse-resolution chemi-cal transport model (Flechard et al 2013 Thimonier et al2018) In addition to low-cost methods forNr concentrationsmore actual micrometeorological Nr flux measurements areneeded to further process understanding and better constrainsurface exchange models over many ecosystems (Fowler etal 2009) For example ammonia flux measurements at DB2have revealed unexpected features such as net NH3 emissionsfrom the forest in summer and autumn in particular in re-sponse to leaf fall (Hansen et al 2013 2017) DB2 is likelynot a net NH3 source at the annual scale but short-term emis-sion pulses which are not represented in most dry depositionmodels (Flechard et al 2011) could significantly offset totalNr deposition

An improved knowledge of Nr exchange patterns overCO2 flux monitoring sites either through inferential mod-elling or direct flux measurements is also essential to quan-tify the fraction of deposited Nr that is absorbed by thecanopy reaching more or less directly the seat of photosyn-thesis in leaves thus favouring a higher nitrogen use effi-ciency (NUE) (Nair et al 2016 Wortman et al 2012 Gaigeet al 2007) Canopy nitrogen retention occurs via severalprocesses including gaseous uptake by stomatal diffusiona well-documented process (Monteith and Unsworth 1990)but also through cuticular diffusion and stomatal penetrationby aqueous solutions with surface-deposited and dissolvedgases and particles acting as direct leaf nutrients (Burkhardt2010 Burkhardt et al 2012) By contrast the Nr fractioninitially deposited to soil (as simulated by the majority offertilization tracer experiments eg Nadelhoffer et al 1999)is subject to various losses via nitrification denitrificationand microbial uptake before being eventually taken up byroots and moving upwards in xylem flow The more ad-vanced emerging multi-layer canopy exchange models foratmospheric pollutants (Nr species but also O3 SO2 etc) can



now partition dry deposition into stomatal non-stomatal andsoil pathways with increasing detail (Zhou et al 2017 Simp-son and Tuovinen 2014 Flechard et al 2013) thanks to im-proved understanding and parameterizations of surface andair column interactions and of photosynthesis-driven stom-atal conductance (Buumlker et al 2007 Grote et al 2014)However particular attention must be paid to measurementquality for an improved deposition accuracy because suchmodels are still very much dependent on local atmosphericconcentration data for all main Nr forms (gas and aerosolreduced and oxidized mineral and organic)

412 Uncertainty in ecosystem nitrogen losses and netbalance

The comparison of DIN leaching values by different meth-ods shows that the Dise et al (2009) algorithm performsreasonably well for low to moderate Nr deposition lev-els but underestimates DIN losses for some of the high-est (gt 4 g N mminus2 yrminus1) deposition sites This observation wasalso made by Dise et al (2009) themselves who argued thattheir simple relationships involving external forcings (Ndep)and internal factors (soil N status) are adequate ldquofor earlyto intermediate stages of nitrogen saturationrdquo but may failat sites where historical chronically enhanced Nr depositionhas so strongly impacted forest ecosystems that N leachinghas become dependent also on stochastic factors such as in-sect defoliation or a drought period followed by re-wettingof the soil As was the case for field-measured NO emis-sions (Fig 3a) the four highest DIN leaching fluxes (09ndash32 g N mminus2 yrminus1) occurred in the four highest Ndep forestsgrowing on well-drained acidic sandy soils In addition itis noteworthy that the two sites with the largest Ndep andDIN leaching rates (EN15 EN16) were dominated by pineor Douglas fir (Table S2) These species have been shownin a common garden experiment (Legout et al 2016) tocause larger nitrification NOminus3 leaching and acidificationrates (as well as larger losses of calcium magnesium andaluminium) compared with other tree species such as beechor oak This is consistent with deciduous trees being knownto take up and store more nitrogen per unit biomass in stemsand branches than coniferous trees (Jacobsen et al 2003)Typical stem N content values proposed for N uptake cal-culations in the Convention on Long-range TransboundaryAir Pollution (CLRTAP) manual for critical loads mappingare 1 and 15 g N kgminus1 dry matter for conifers and decidu-ous trees respectively for steady-state conditions (CLRTAP2017) Tree species traits may therefore in our study haveexacerbated an existing DIN leaching predisposition result-ing from edaphic factors and pollution climate At the lowerend of the Ndep range the dataset is consistent with previ-ous studies which have shown that DIN leaching is unlikelyto occur in forests where Ndep lt 1 g N mminus2 yrminus1 (de Vries etal 2009) although under these conditions there may still besignificant N losses as NO and N2O (Fig 3)

The best empirical fit for the relationship of the sumDIN+NO+N2O to Ndep was slightly non-linear (Fig 3d)and may indicate that at the upper end of the Ndep rangeabove 4 g N mminus2 yrminus1 the sum of inorganic Nr losses mightapproach or even exceed the estimated atmospheric deposi-tion which corresponds to one of the several existing defini-tions of ecosystem N saturation (see below) Whether theseecosystems turn into net N sources depends on the relativemagnitudes of the missing terms N2 fixation (likely smallin temperate compared with tropical forests Vitousek et al2002) N2 losses from denitrification (possibly the largest ofthe unknown terms at forest sites that are frequently water-logged) N2O losses from the litter layers of the forest floorDON leaching and also incoming organic nitrogen in pre-cipitation (WSON) as well as dry deposition of organic Nrspecies not quantified here (Fig 1) The presumably smalland unaccounted for N inputs via N2 fixation and organicNr deposition are at least partly compensated for by denitri-fication N2 losses and DON leaching losses Moreover DONleaching typically responds much less strongly than DINleaching to N inputs (Siemens and Kaupenjohann 2002)Under these assumptions the inorganic Nr budget calculatedfrom Fig 3 may provide a reasonable proxy for the overallecosystem N balance In this case N outputs by gaseous anddissolved losses represent on average across all forest sites43 of N inputs More important than the average N lossfor judging Nr deposition effects on C sequestration is thelarge range of losses from 6 to 85 with on average a27 loss (range 6 ndash54 ) for Ndep lt 1 g N mminus2 yrminus1 45 loss (12 ndash78 ) for intermediate Ndep levels and 65 loss(35 ndash85 ) for Ndep gt 3 g N mminus2 yrminus1 However if the veryfew available data or estimates for DON leaching and espe-cially denitrification N2 fluxes are correct and may be ex-trapolated to other sites they may often outweigh the inputsthrough organicNr deposition and biological N2 fixation andthus the inorganic Nr budget (Fig 3) may underestimate theoverall N losses

42 Drivers and uncertainties of the carbon and GHGbalance

421 Variability of carbon sequestration efficiency

The CSE ratio (=NEP GPP) calculated over theCEIPNEU project observation periods provides an in-dicator of the fraction of accumulated carbon in theecosystem relative to gross CO2 uptake by photosynthesisThis is a useful metric to compare carbon cycling in differentterrestrial ecosystems and it is directly related to climateeffects and other drivers such as site fertility (Vicca et al2012) and management (Campioli et al 2015) By contrastquantifying the accumulated carbon in terrestrial ecosystemsrequires much longer observations (one or several decades)to ensure statistical significance of a small change overa large C stock particularly when soils are considered



This is often impractical but also of limited use because Ndeposition rates are unlikely to be constant over such longperiods

Over the time frame of the CEIPNEU projectsobservation-based CSEobs values were much more variablethan their modelled counterparts Negative CSEobs values(EN6 EN11) imply a net carbon source and may be ex-plained by a number of factors including soil carbon losslateral DOC DIC water flow from adjacent ecosystems treemortality low fertility poor ecosystem health a recentlyplanted forest or other disturbances with long-lasting con-sequences on the C budget At EN6 the main reasons may bea large SOC concentration leading to large Reco values anda relatively old age of the forest responsible for a small GPP

However the large discrepancy between observation-based and modelled CSE estimates may not be entirelycaused by the modelrsquos inability to reproduce all fine patternsof GPP and especially Reco across all ecosystems (Fig 6)Some of the largest CSEobs values may be less ecologicallyplausible and might result from methodological biases andorincorrect interpretation of the EC measurements in terms oftheir representativeness for the ecosystem considered Multi-annual values of GPP and Reco derived from EC flux data arenot measurements sensu stricto they compound problems inEC measurements post-processing of high-frequency datagap-filling and partitioning Some partitioning algorithms(Barr et al 2004 Reichstein et al 2005) evaluate GPP asthe difference between measured daytime NEE and an esti-mate of daytime Reco that is based on an empirical modelof night-time Reco measurements In this case any problemwith night-time and thus with estimated daytime Reco woulddirectly impact GPP in the same way (Vickers et al 2009)GPP and Reco would both be underestimated or both overes-timated in absolute terms and by the same absolute magni-tude thereby impacting the annual or long-term NEP GPP(CSEobs) ratio

In this study however the use of the daytime-data-basedpartitioning method by Lasslop et al (2010) within theREddyProc algorithm embedded in the European FluxesDatabase Cluster was intended to ensure the independenceof GPP and Reco estimates since Reco was estimated fromthe intercept of the MichaelisndashMenten light response curvefitted to daytime-measured NEE This partitioning procedureshould avoid the propagation into the GPP estimate of poten-tial errors in night-time Reco data although it still assumessimilar dependencies of day- and night-time respiration onenvironmental factors which is debatable from a biologi-cal standpoint (eg Kok 1949 Wehr et al 2016 Wohlfahrtand Galvagno 2017) From a micrometeorological perspec-tive the night-time flux can be underestimated due to low-turbulence conditions and the transport of CO2 by horizontalandor vertical advection as well as the decoupling of soil-level and understorey fluxes from the turbulent fluxes mea-sured above the canopy (Feigenwinter et al 2008 Etzoldet al 2010 Montagnani et al 2010 Paul-Limoges et al

2017) Further in principle the ulowast threshold filtering (Gu etal 2005 Papale et al 2006) carried out to discard low-turbulence flux data at the start of the gap-filling and parti-tioning algorithm (REddyProc 2019) should alleviate the is-sue of night-time Reco underestimation which affects annualReco and CSEobs even if the error does not propagate intoGPP in the Lasslop et al (2010) method However the choiceof the value for the ulowast threshold can be critical if advection-affected flux values are to be discarded especially for sitesand datasets where the independence of the gap-filled annualNEP value from the ulowast threshold value cannot be demon-strated Advective flux contributions remain a largely unre-solved issue as Aubinet et al (2010) conclude that ldquodirectadvection measurements do not help to solve the night-timeCO2 closure problemrdquo Others (eg Kutsch and Kolari 2015)have commented on the need to assign appropriate uncertain-ties when dealing with CSE and C balances derived from ECflux towers which only measure turbulent fluxes and CO2storage change in the air column underneath the sensor butnot the other terms of the conservation equation of a scalar inthe atmospheric boundary layer (see Eq 1 in Aubinet et al2000)

Despite all these precautions at sloping or complex terrainsites where advection can be important it cannot be excludedthat the Lasslop et al (2010) daytime-data-based approachmay still underestimate Reco (and overestimate CSEobs) ifadvection is not accounted for explicitly This is because theReco estimate based on the intercept of the light responsecurve for the measured NEE (at PAR= 0) is strongly in-fluenced by measurements made around sunrise and sunsetwhen a clear impact of advection on the light response curveordinate has been observed as shown at the EN5 subalpinesite by Montagnani et al (2009) (see their Fig 13)

It is important to note that advection may also be a prob-lem at flat lowland sites if there is strong spatial land surfaceheterogeneity eg differences in albedo or in Bowen ratio agradient in tree species a nearby lake and a gradient in wa-ter availability Conversely there may also be sites where ECunderestimates CSEobs for similar reasons albeit in the op-posite direction for example additional CO2 being advectedinto the ecosystem and then released by turbulent diffusionto the atmosphere within the tower footprint Another pos-sibility is that basal Reco measured at dawn or dusk over adifferent (larger) footprint is lower than during the day Fluxpartitioning may again in this case underestimateReco duringthe warmer daytime hours and therefore also underestimateGPP resulting in overestimated NEP GPP (CSEobs) ratios

Given this uncertainty the fact that most of the foreststands with CSEobs values larger than 40 (EN1 EN5 DB6MF2) were located at elevations above 700 m asl (Table 1and Fig 8a) ie in hilly or mountainous areas with topo-graphically more complex terrain than typically encounteredat lowland sites may be coincidental or partly a conse-quence of advection or decoupling issues (Paul-Limoges etal 2017) In such conditions consistency cross-checks in-



Figure 8 Variability of observation-based and modelled carbon sequestration efficiency (CSE defined as the NEP GPP ratio) as a functionof (a) site elevation above mean sea level (m) and (b) MAP mean annual precipitation (mm) Site labels are provided for observations only

volving additional flux advection soil and biometric mea-surements even ecosystem modelling provide useful refer-ence points to assess the plausibility of EC-derived C bud-gets and to better constrain the problem At the EN5 site theannual total tree biomass C increment based on biometricmeasurements was on average 218 g C mminus2 yrminus1 over the pe-riod 2010ndash2017 (Leonardo Montagnani unpublished data)ie 26 of the reported mean EC-derived NEP value of826 g C mminus2 yrminus1 for the CEIPndashNEU period and it seemsunlikely that the increase in soil carbon and fine root stockscould account for the large difference By contrast the DB6site was a fertile and managed beech forest with a signif-icantly higher efficiency conversion of photosynthates intobiomass compared to less fertile and unmanaged sites (Viccaet al 2012 Campioli et al 2015) The long-term annualtotal NPP at the site was 780 g C mminus2 yrminus1 over the period1992ndash2007 with a significant part allocated below ground(Alberti et al 2015) while heterotrophic respiration esti-mated at the site using either bomb carbon (Harrison etal 2000) or mineralization rates (Persson et al 2000) wasaround 200 g C mminus2 yrminus1 resulting in similar NEP estimatesby EC flux measurements versus biometric data combinedwith process studies

At the MF2 site Etzold et al (2011) calculated inter-annual mean EC-derived NEP GPP and Reco values (for thesame 2005ndash2009 period used in this study) of 415 1830and 1383 g C mminus2 yrminus1 respectively using customized gap-filling and partitioning algorithms and thus providing al-ternative estimates to those from the REddyProc algorithmwithin the European Fluxes Database Cluster (Table 1) Val-ues of Reco and NEP were 82 larger and 40 lower re-spectively in Etzold et al (2011) compared with the defaultdatabase values that do not explicitly correct for advectionHowever the Etzold et al (2011) mean EC-derived NEP wasmuch closer to NEP values calculated from the net annual

increment in the woody and non-woody biomass and soil Cstorage using four different biometric and modelling meth-ods (range 307ndash514 g C mminus2 yrminus1 mean 421 g C mminus2 yrminus1)The CSEobs value derived from Etzold et al (2011) was23 and comparable to the value of 25 that can be cal-culated from the decoupling-corrected EC budget computedby Paul-Limoges et al (2017) for the same site for the years2014ndash2015 in which the decoupling correction to accountfor undetected below-canopy fluxes doubled Reco and re-duced NEP from 758 to 327 g C mminus2 yrminus1 These alterna-tive CSEobs estimates were thus much lower than the defaultCSEobs value of 48 but fully consistent with model predic-tions (Fig 8)

The four upland sites EN1 EN5 DB6 and MF2 were alsoamong the wettest with MAP gt 1000 mm (Fig 8b) in princi-ple promoting larger leaching and runoff The overall distri-bution of CSEobs as a function of MAP (Fig 8b) shows anapparent increase in CSEobs with precipitation though withlarge scatter which would be consistent with a reduction inEC-tower-based Reco through an increase in the dissolvedleached fraction At sites where significant leaching occursReco determined from the atmospheric flux is no longer areliable indicator of total C losses by respiration since thedissolved and then leached fraction of Rsoil is not capturedby the flux tower (Gielen et al 2011) which implies thatCSEobs is overestimated As observed in the case of GPPsuch apparent correlations of CSEobs to single factors like el-evation or MAP may not be (entirely) causal potentially con-cealing underlying cross-correlations (such as large but un-measured advection components occurring at the same siteswhere MAP is largest) The data by Kindler et al (2011) andGielen et al (2011) do suggest that the overestimation of Csequestration (as estimated by EC-derived NEP) caused bynot accounting for dissolved C leaching was likely smallerthan 10 for forests (7 of NEP on average) but all five



sites they investigated had MAP lt 1000 mm and only one(EN4) was an upland site (785 m)

To summarize a set of unresolved issues the largestCSEobs values (gt 45 ) are likely to result from a com-bination of ecological factors and methodological biasesbut they occurred at sites in the mid-range for Ndep (12ndash22 g N mminus2 yrminus1) and thus did not introduce confoundingtrends in the overall CndashN relationships we seek to establishacross the whole Ndep spectrum in this study

422 Forest net greenhouse gas balance dominated bycarbon

Based on the available data the net GHG balance of the31 forests investigated was generally not significantly af-fected by N2O or CH4 (Fig 7) with the caveat that thesefluxes were not actually measured in situ everywhere or withthe same intensity and duration as CO2 Thus the uncertaintyin non-CO2 GHG fluxes is much larger (possibly gt 100 )than for multi-annual EC-based CO2 datasets where a typi-cal uncertainty is of the order of 10 ndash30 (Loescher et al2006) Nonetheless the N2O and CH4 emissions observedby different methods in forest soils were typically 2 ordersof magnitude smaller than the CO2 sink (in GWP equiva-lents) which means that the quality of CO2 estimates domi-nates the overall uncertainty in our forest GHG budgets Notethat such results cannot be extended to waterlogged organicsoils of temperate and boreal zones where CH4 emissionscan be large (Morison et al 2012) nor can they be extendedto the tropics especially in degraded forests (Pearson et al2017) Also N2O fluxes can be highly episodic with emis-sion events linked to for example freezendashthaw cycles (Risket al 2013 Medinets et al 2017) and such episodes wouldhave been missed by the bioassay approach and in somecases by discontinuous (manual) chamber measurements

By contrast for the short semi-natural vegetation sites ofour study NEP was on average a factor of 27 smaller than inforests but only a factor of 15 smaller for GPP which impliesthat total C losses were much larger in proportion to grossassimilation especially non-respiratory non-CO2 losses (iea much lower CSE) Large wetland CH4 emissions and dis-solved DICDOC fluxes were much more likely to offset oreven determine the C and GHG balance (Fig 7 Kindler et al2011) In these systems studying the impact of Nr deposi-tion on C sequestration requires much more robust estimatesof the gaseous and dissolved budgets for all components andover the long term since the estimation of NECB requiresin addition to EC CO2 the knowledge of non-atmosphericnon-CO2 fluxes (Fig 1) Technological developments in thefield of (routine) EC measurements for N2O and CH4 (egNemitz et al 2018) are likely to reduce uncertainties in netGHG budgets in the foreseeable future but DICDOC lossesin wetlands probably represent a bigger challenge

It should however be remembered that such short-termGHG budgets based on a few years of flux data and GWP

multipliers for a 100-year time horizon do not actually re-flect the long-term climate impact of northern mires whichmay be thousands of years old and despite their CH4 emis-sions typically have an overall climate-cooling effect Asshown by Frolking et al (2006) pristine mires typically startcooling the climate some hundreds of years after their for-mation with the exact timing of course depending on themagnitude of the CH4 and CO2 fluxes thus the history ofthe site should be accounted for when dealing with ecosys-tem radiative forcing assessments For the SN3 site Dreweret al (2010) actually used a 500-year time horizon GWP(instead of the usual 100-year) for CH4 reducing the GHGsource strength of the site by a factor of 4 to 10 dependingon the year considered

43 Challenges in understanding the coupling ofcarbon and nitrogen budgets

431 Tangled effects of nitrogen deposition and climateon ecosystem productivity

The analysis of Ndep variability and spatial patterns at thescale of the monitoring network as well as the Europeanscale (Fig 5) showed that the impact of Nr depositionon ecosystem C sequestration cannot be considered inde-pendently of climate in the regional context of this studyNitrogen deposition patterns at the European scale resultfrom the continent-wide geographical distribution of popu-lation human industrial and agricultural activities and ofprecursor emissions combined with mesoscale patterns ofmeteorology-driven atmospheric circulation and chemistryThrough the interplay of these factors the elevated Ndeplevels in this study happened to co-occur geographicallywith temperate climatic zones of central and western Europe(Fig 5cndashd) that are the most conducive to vegetation growthat the continental scale This means adequate water supplyas precipitation reasonably low summertime evaporative de-mand mild winters and temperate summers and long grow-ing seasons In other words there are many gaps in the multi-dimensional variable space which is incompletely exploredby the available dataset Thus any regression analysis thatwould correlate NEP and other C fluxes with Ndep withoutsimultaneously accounting for climate would be flawed asSutton et al (2008) concluded from their reanalysis of thedata used by Magnani et al (2007) A dCdN slope calcu-lated directly from a (linear or non-linear) mono-factorial re-gression analysis of GPP or NEP vs Ndep would mislead-ingly attribute the whole C flux variability to Ndep while ig-noring climate effects (Fleischer et al 2013) In additiona range of other potential explanatory variables such as soiltype especially the water holding capacity (8FCminus8WP) soilfertility (Vicca et al 2012 Legout et al 2014) tree speciesand stand age (Besnard et al 2018) are potentially neededto explain the observed variability In order to account forand untangle the multiple inter-relationships we chose a



Figure 9 Relationships of leaf (a) and topsoil (b) CN ratios with atmospheric nitrogen deposition (Ndep) and to each other (c) in differentecosystem types (DBF deciduous broadleaf forests MF mixed forests ENF evergreen needleleaf forests EBF evergreen broadleaf forestsSN short semi-natural vegetation)

mechanistic model (BASFOR) based approach described inFlechard et al (2020) whereby most of the known interac-tions of plant soil climate age and species are encoded andparameterized to the best of our current knowledge Giventhe limited size and very large diversity of the dataset suchan approach appears to be preferable to regression-based sta-tistical analyses since a simple pattern to explain the cou-pling of carbon and nitrogen budgets with the available dataand knowledge is unlikely

432 Evidence of nitrogen saturation from variousindicators

Various definitions of nitrogen saturation have been proposed(Aber 1992 De Schrijver et al 2008 Binkley and Houmlg-berg 2016) including (i) the absence of a growth responsein the case of further N addition (dCdN= 0) (ii) the on-set of NOminus3 leaching andor gaseous emissions and (iii) theequivalence of N inputs and N losses The underlying con-cept of a dCdN response is that the C and N cycles areclosely coupled through stoichiometric ratios in the differ-ent parts of the ecosystem with very different CN ratiosin soil organic matter roots leaves tree branches and stems(de Vries et al 2009 Zechmeister-Boltenstern et al 2015)

A difference in dCdN response could for example be ex-pected between forests where carbon is stored in both woodyand root biomass (CN ratio 300ndash500) and below groundin SOM (CN ratio 30ndash40) versus short semi-natural veg-etation where most of the stock is in SOM and thus witha much lower overall ecosystem CN ratio This would beconsistent with the observations in Fig 4 where the appar-ent increase in NEP with increasing Ndep is smaller in shortsemi-natural vegetation than in forests But the theoreticalstoichiometric approach becomes more uncertain in the eventof N saturation as the C and N cycles have become much lesstightly coupled than in pristine N-limited environments andthus defining a dose-response relationship requires a precisequantification of all C and N inputs and losses not just pro-ductivity and Nr deposition

Another possible indicator of N saturation in the presentdataset may be provided by the comparison of the rela-tionships of CN ratios of foliage and topsoil (5 cm) to at-mospheric Nr deposition (Fig 9andashb) Since leaf N contentwas not only dependent on Nr deposition but also on theecosystem type (Fig 4e) CN ratios are shown separatelyfor the different vegetation classes in Fig 9 There was aclear negative correlation of leaf CN ratio to Ndep for conif-



erous forests (ENF spruce and pine pooled exponential fitR2= 086 plt001) and a similar but not significant trend

for SN (linear R2= 029) (Fig 9a) for the other ecosys-

tems (DBF MF EBF) there were not enough data to derivetrends In topsoils (Fig 9b) there was also a broad downwardtrend of CN ratios with increasing Ndep within the ENF andSN classes but only for Ndep up to 25 g N mminus2 yrminus1 Againas for GPP and NEP the relationship is highly non-linearas the four ENF sites above this Ndep threshold break thetrend observed in the lower Ndep sites and the overall bestfit is quadratic (R2

= 049 plt001) with an inflexion pointaround this threshold While the relationship of foliar CNratio to Ndep was almost linear for ENF (a consequence ofthe linear trend in ENF leaf N content Fig 4e) the non-linear behaviour of the topsoil CN ratio and its stabilizationor increase for Ndep gt 25 g N mminus2 yrminus1 indicate a possiblethreshold for saturation Atmospheric nitrogen was thereforeapparently efficiently taken up by vegetation when reachingthe leaves but after leaf fall and following litter decomposi-tion and incorporation into the topsoil there appeared to be alimit to the amount of nitrogen that can be stabilized into soilorganic matter of the ENF sites However forest soil organicN stocks are very large (in the range 200ndash700 g N mminus2 at thesites we investigated) and therefore changes in CN ratiosin response to atmospheric Nr deposition must be very slowThe soil CN ratio at a given time reflects centuries of landuse as well as a more recent history of multi-decadal changesin Nr deposition (Flechard et al 2020) This complicates theinterpretation of the downward trends observed from instan-taneous snapshots of soil and foliar CN ratios versus Ndepsince the ecosystems cannot be considered to be in steadystate neither for Ndep nor for growth or productivity Therewas a positive correlation across all vegetation types betweentopsoil and foliar CN ratios (Fig 9c R2

= 019 plt005)but this was mostly driven by differences between plant func-tional types (no significant correlation within each PFT)

Following definition (ii) of N saturation given abovethe sum of inorganic Nr losses heavily dominated byDIN leaching at the upper end of the Ndep range in ourdatasets (Fig 3) may indicate various stages of N satura-tion in all forests with Ndep gt 1ndash15 g N mminus2 yrminus1 A thresh-old for a more advanced saturation stage could be placedat 2ndash25 g N mminus2 yrminus1 where inorganic Nr losses are con-sistently larger than 50 of Ndep Such numbers are en-tirely consistent with the leaching risk classification of Eu-ropean forests proposed by Dise and Wright (1995) withlow leaching risk at Ndep lt 1 g N mminus2 yrminus1 intermediate riskat Ndep in the range 1ndash25 g N mminus2 yrminus1 and high risk atNdep gt 25 g N mminus2 yrminus1 The results are also in line with thereview by De Vries et al (2014) based on literature results ofdCdN responses derived from stoichiometric scaling meta-analysis of N addition experiments and field observationsof both growth changes and Nr deposition accounting forother drivers the data showed beneficial Nr deposition ef-fects up to 2ndash3 g N mminus2 yrminus1 and adverse effects at higher

levels A lower Ndep threshold of 1 g N mminus2 yrminus1 had alsobeen suggested by de Vries et al (2007) but this was usingthroughfall deposition which generally underestimates totaldeposition through canopy retention processes (Thimonier etal 2018) It must be stressed however that the definition ofan all-purpose generic Ndep threshold for N saturation maybe misleading or at least qualified with an uncertainty sincesome tree species (Douglas fir pine spruce) grown on thesame soil and under the same climate and Ndep regime mayresult in significantly higher NOminus3 leaching rates than others(Legout et al 2016) This also means that the NOminus3 leach-ing flux is not necessarily a good proxy of the severity of Nsaturation though this depends on which of the several defi-nitions of N saturation is considered

The upper threshold of 2ndash25 g N mminus2 yrminus1 happens to co-incide with the levelling off of GPP Reco and NEP as wellas the further reduction in C fluxes at higher Ndep levels(Fig 4a-c) Whether this should be interpreted as a nega-tive impact of advanced N saturation on soil processes andplant functioning and hence C sequestration potential isnot straightforward (Binkley and Houmlgberg 2016) If the par-allel effects of climate soil fertility other nutrient limita-tions tree species traits age and planting density are over-looked in a simplistic first-order interpretation the datasethints at an optimum Ndep level around 2 g N mminus2 yrminus1 be-yond which no further benefits (in carbon terms) could begained from further atmospheric Nr additions which wouldbe consistent with the 2ndash25 g N mminus2 yrminus1 Ndep threshold de-rived by Etzold et al (2014) for Swiss forests The high soilNr losses observed in these ecosystems growing under rel-atively favourable climates would then suggest that what-ever fertilization effect Nr deposition may have at low tomoderate deposition rates (lt 2 g N mminus2 yrminus1) is unlikely tobe sustained at high-deposition levels especially on acidicsandy soils However the very limited number of affectedsites with Ndep gt 3 g N mminus2 yrminus1 leaves too few degrees offreedom to make the argument statistically compelling Moreimportantly a knowledge of all other limitations to growth(climate soil fertility nutrients age structure) would be re-quired to confirm the hypothesis Additional measurement-and model-based investigations to untangle theNdep effect onC sequestration (the dCdN term) are presented in Flechardet al (2020) drawing from the results fluxes and budgetspresented here

5 Conclusions

We provided estimates of carbon nitrogen and greenhousegas budgets for 40 flux tower sites over European forests andsemi-natural vegetation compiled from a large variability ofstate-of-the-art methods that can be applied in such a networkapproach The CO2 budgets from well-established EC meth-ods were the least uncertain followed by GHG budgets offorests and then the CH4 and DICDOC fluxes of wetlands



uncertainty levels were likely highest in the net N budgetsespecially at the elevated Nr deposition sites where NOminus3leaching was almost of the same order as Ndep The uncer-tainty was still compounded by the lack of some data on bi-ological N2 fixation N2 loss by denitrification and organicNr in rainwater in dry deposition and in soil leaching butsome of these unknown terms would compensate mutuallyto some extent Nevertheless the low-cost network to moni-tor atmospheric gas-phase and aerosol Nr contributed to sub-stantially reducing the large uncertainty in total Ndep rates atindividual sites (compared with gridded outputs of a regionalchemical transport model) This was because dry depositionalmost systematically heavily dominates over wet depositionin forests except at very remote sites (away from sources ofatmospheric pollution) and directly measured Nr concentra-tions reduced the uncertainty in dry deposition fluxes

The greenhouse gas balances of the 31 forest sites includedin this study were almost entirely determined by the CO2budgets with small to negligible contributions by N2O andCH4 The GHG balance of nine extensively managed andupland grasslands moorlands and wetlands was much moredependent on CH4 and N2O fluxes Ecosystem productivity(GPP NEP) data across Europe showed an apparent increasewith atmospheric Ndep though only up to 25 g N mminus2 yrminus1while the larger Ndep rates also happen to coincide geo-graphically with regions of Europe where climate is opti-mal for tree growth (neither too cold nor too dry) The datathus underpinned a strong covariation of Nr deposition withvariables like elevation and climate and they indicated thatthe ecosystem response of carbon sequestration to nitrogendeposition cannot be calculated simply and directly fromthe observed apparent dNEP dNdep using bivariate statis-tics Other co-varying influences such as climate soil fer-tility nutrient availability forest age and ecophysiologicalprocesses should be analysed alongside so the nitrogen de-position effect can be isolated

The site-specific analysis of C and N fluxes and bud-gets across a large geographical and climatic gradient sup-port the concept of a non-linear response of C sequestra-tion to N deposition Large nitrogen losses (especially ni-trate) from forests suggest that up to one-third of the sitesinvestigated can be classified as in the early to advancedstages of N saturation At the sites with the largest Nr depo-sition rates (gt 25 g N mminus2 yrminus1) a stagnation or reduction inforest productivity compared to mid-range deposition siteswas observed Beyond the conclusion that the apparent C re-sponse to increased Nr deposition was non-linear we do nothave enough data to test the hypothesis that the reduction inproductivity and C sequestration is linked to N-saturation-induced ecological impacts on soil and ecosystem function-ing rather than just the confounding effects of variability inmeteorological and other drivers Further efforts are requiredto disentangle Ndep effects and climatic as well as pedologi-cal effects on C sequestration at the continental scale

Code and data availability The data used in this study are publiclyavailable from online databases and from the literature as describedin the ldquoMaterials and methodsrdquo section

The codes of models and other software used in this study arepublicly available online as described in the ldquoMaterials and meth-odsrdquo section

Supplement The supplement related to this article is available on-line at httpsdoiorg105194bg-17-1583-2020-supplement

Author contributions CRF MAS AI WdV MvO and UMS con-ceived the paper MAS EN UMS KBB and WdV conceived ordesigned the NEU study CRF performed the data analyses ranmodel simulations and wrote the text MvO and DRC wrote andprovided the BASFOR model code and performed the Bayesian cal-ibration YST NvD HU UD SV VD MM FS and YF performedDELTA and bulk deposition chemical analyses AF collected wetdeposition databases DS provided modelled EMEP Nr depositiondata BaKi and SZB conceived and performed the soil bioassay ex-periment JKS provided foliar nitrogen analyses AI UMS JFJKArL MJS MaA DL LM JN IAJ MP RK JA AV JO RJ MiABHC JD WE AF AG PG YH CH AH LH BaKr WLK RLdVALoh BL MVM GM VM JMO KP GP MTS MU TV CV andTW provided eddy covariance andor other field data or contributedto data collection from external databases and literature AI WdVMAS UMS MvO EN KBB SZB DRC NBD JFJK NB ArLDS MJS MaA DL LM JN IAJ MP RK JS AJF JA AV JO RJMiA WE BaKi BaKr RLdV AnL GM AN and MU contributedsubstantially to discussions and revisions

Competing interests The authors declare that they have no conflictof interest

Acknowledgements The authors gratefully acknowledge finan-cial support by the European Commission through the twoFP6 integrated projects CarboEurope Integrated Project (projectno GOCE-CT-2003-505572) and NitroEurope Integrated Project(project no 017841) the FP7 ECLAIRE project (grant agreementno 282910) and the ABBA COST Action ES0804 We are alsothankful for funding from the French GIP-ECOFOR consortiumunder the F-ORE-T forest observation and experimentation net-work as well as from the MDM-2017-0714 Spanish grant Weare grateful to Christian Bernhofer Robert Clement Han DolmanAxel Don Eric Dufrecircne Damiano Gianelle Ruediger Grote An-ders Lindroth John Moncrieff Dario Papale Corinna Rebmann andAlex Vermeulen for the data they provided as well as to Klau-dia Ziemblinska for her comments on the paper Computer timefor EMEP model runs was supported by the Research Council ofNorway through the NOTUR project EMEP (NN2890K) Finaliza-tion of the paper was supported by the UK Natural EnvironmentResearch Council award number NER0164291 as part of the UK-SCAPE programme delivering national capability We also wish tothank two anonymous referees for their constructive criticism of thepaper



Financial support This research has been supported by theEuropean Commissionrsquos Sixth Framework Programme (grantnos 017841 and GOCE-CT-2003-505572) and the EuropeanCommissionrsquos Seventh Framework Programme (ECLAIRE grantno 282910)

Review statement This paper was edited by Soumlnke Zaehle and re-viewed by A Agravevila and one anonymous referee

References

Aber J D Nitrogen cycling and nitrogen saturationin temperate forest ecosystems Tree 7 220ndash224httpsdoiorg1010160169-5347(92)90048-G 1992

Alberti G Vicca S Inglima I Belelli-Marchesini L Gene-sio L Miglietta F Marjanovic H Martinez C MatteucciG DrsquoAndrea E Peressotti A Petrella F Rodeghiero Mand Cotrufo M F Soil C N stoichiometry controls carbonsink partitioning between above-ground tree biomass and soilorganic matter in high fertility forests iForest 8 195ndash206httpsdoiorg103832ifor1196-008 2015

Aubinet M A Grelle A Ibrom A Rannik Uuml MoncrieffJ Foken T Kowalski T A S Martin PH Berbigier PBernhofer C Clement R Elbers J Granier A GruumlnwaldT Morgenstern K Pilegaard K Rebmann C Snijders WValentini R and Vesala T Estimates of the annual net carbonand water exchange of forests The EUROFLUX methodologyAdv Ecol Res 30 113ndash175 httpsdoiorg101016S0065-2504(08)60018-5 2000

Aubinet M Feigenwinter C Bernhofer C Canepa E HeineschB Lindroth A Montagnani L Rebmann C Sedlak P andvan Gorsel E Direct advection measurements do not help tosolve the nighttime CO2 closure problem ndash evidence from threeinherently different forests Agr Forest Meteorol 150 655ndash664httpsdoiorg101016jagrformet201001016 2010

Baccini A Walker W Carvalho L Farina M Sulla-MenasheD and Houghton R A Tropical forests are a net carbon sourcebased on aboveground measurements of gain and loss Science358 230ndash234 httpsdoiorg101126scienceaam5962 2017

Barr A G Black T A Hogg E H Kljun N Morgen-stern K and Nesic Z Inter-annual variability in the leafarea index of a boreal aspen-hazelnut forest in relation to netecosystem production Agr Forest Meteorol 126 237ndash255httpsdoiorg101016jagrformet200406011 2004

Barton L Wolf B Rowlings D Scheer C Kiese R GraceP Stefanova K and Butterbach-Bahl K Sampling frequencyaffects estimates of annual nitrous oxide fluxes Sci Rep-UK 515912 httpsdoiorg101038srep15912 2015

BASFOR BASic FORest ecosystem model httpsgithubcomMarcelVanOijenBASFOR (last access 22 August 2019) 2016

Bertolini T Flechard C R Fattore F Nicolini G Ste-fani P Materia S Valentini R Laurin G V andCastaldi S DRY and BULK atmospheric nitrogen deposi-tion to a West-African humid forest exposed to terrestrialand oceanic sources Agr Forest Meteorol 218 184ndash195httpsdoiorg101016jagrformet201512026 2016

Besnard S Carvalhais N Arain A Black A de Bruin SBuchmann N Cescatti A Chen J Clevers J G P W De-sai AR Gough C M Havrankova K Herold M Houmlrt-nagl L Jung M Knohl A Kruijt B Krupkova L LawB E Lindroth A Noormets A Roupsard O SteinbrecherR Varlagin A Vincke C and Reichstein M Quantifyingthe effect of forest age in annual net forest carbon balanceEnviron Res Lett 13 124018 httpsdoiorg1010881748-9326aaeaeb 2018

Binkley D and Houmlgberg P Tamm Review Re-visiting the influence of nitrogen deposition onSwedish forests Forest Ecol Manag 368 222ndash239httpsdoiorg101016jforeco201602035 2016

Boone D R Ecology of methanogenesis in Microbial Produc-tion and Consumption of Greenhouse Gases Methane Nitro-gen oxides and Halomethanes edited by by Rogers J E andWhitman W E American Society for Microbiology Washing-ton DC 57ndash70 httpsdoiorg1010160169-5347(93)90172-L1991

Bousquet P Ciais P Peylin P Ramonet M and Monfray PInverse modeling of annual atmospheric CO2 sources and sinks1 Method and control inversion J Geophys Res 104 26161ndash26178 httpsdoiorg1010291999JD900342 1999

Buchen C Roobroeck D Augustin J Behrendt U BoeckxP and Ulrich A High N2O consumption potential ofweakly disturbed fen mires with dissimilar denitrifiercommunity structure Soil Biol Biochem 130 63ndash72httpsdoiorg101016jsoilbio201812001 2019

Buumlker P Emberson L D Ashmore M R Cambridge HM Jacobs C M J Massman W J Muumlller J NikolovN Novak K Oksanen E Schaub M and de la TorreD Comparison of different stomatal conductance algorithmsfor ozone flux modelling Environ Pollut 146 726ndash735httpsdoiorg101016jenvpol200604007 2007

Burkhardt J Hygroscopic particles on leaves nutrients or desic-cants Ecol Monogr 80 369ndash399 httpsdoiorg10189009-19881 2010

Burkhardt J Basi S Pariyar S and Hunsche M Stomatal up-take of aqueous solutions ndash an update involving leaf surface parti-cles New Phytol 196 774ndash787 httpsdoiorg101111j1469-8137201204307x 2012

Butterbach-Bahl K Willibald G and Papen H Soil coremethod for direct simultaneous determination of N2 andN2O emissions from forest soils Plant Soil 240 105ndash116httpsdoiorg101023A1015870518723 2002

Butterbach-Bahl K and Gundersen P Nitrogen processes in ter-restrial ecosystems in The European Nitrogen Assessmentedited by Sutton M Howard C M Erisman J W BillenG Bleeker A Grennfelt P van Grinsven H and GrizzettiB Cambridge University Press Cambridge UK 99ndash125 avail-able at httpwwwnine-esforgfilesena_docENA_pdfsENA_c6pdf (last access 22 August 2019) 2011

Cameron D R Van Oijen M Werner C Butterbach-Bahl KGrote R Haas E Heuvelink G B M Kiese R KrosJ Kuhnert M Leip A Reinds G J Reuter H I Schel-haas M J De Vries W and Yeluripati J Environmentalchange impacts on the C- and N-cycle of European forestsa model comparison study Biogeosciences 10 1751ndash1773httpsdoiorg105194bg-10-1751-2013 2013



Cameron D Flechard C and van Oijen M Calibrat-ing a process-based forest model with a rich observationaldataset at 22 European forest sites Biogeosciences Discusshttpsdoiorg105194bg-2018-156 2018

Campioli M Vicca S Luyssaert S Bilcke J Ceschia EChapin F S Ciais P Fernaacutendez-Martiacutenez M Malhi YObersteiner M Olefeldt D Papale D Piao S L Pentildeue-las J Sullivan P F Wang X Zenone T and Janssens IA Biomass production efficiency controlled by managementin temperate and boreal ecosystems Nat Geosci 8 843ndash846httpsdoiorg101038NGEO2553 2015

Canadell J G Kirschbaum M U F Kurz W A SanzM-J Schlamadinger B and Yamagata Y Factoringout natural and indirect human effects on terrestrial car-bon sources and sinks Environ Sci Policy 10 370ndash384httpsdoiorg101016jenvsci200701009 2007

Cape J N Tang Y S Gonzalez-Benitez J M MitosinkovaM Makkonen U Jocher M and Stolk A Organic nitrogenin precipitation across Europe Biogeosciences 9 4401ndash4409httpsdoiorg105194bg-9-4401-2012 2012

CEIP (CarboEurope Integrated Project) available at httpwwwcarboeuropeorg (last access 22 August 2019) 2011

Chapin F S Woodwell G M Randerson J T Rastetter E BLovett G M Baldocchi D D Clark D A Harmon M ESchimel D S Valentini R Wirth C Aber J D Cole J JGoulden M L Harden J W Heimann M Howarth R WMatson P A McGuire A D Melillo J M Mooney H ANeff J C Houghton R A Pace M L Ryan M G Run-ning S W Sala O E Schlesinger W H and Schulze E-DReconciling carbon-cycle concepts terminology and methodsEcosystems 9 1041ndash1050 httpsdoiorg101007s10021-005-0105-7 2006

Ciais P Canadell J G Luyssaert S Chevallier F ShvidenkoA Poussi Z Jonas M Peylin P King A W Schulze E-D Piao S L Rodenbeck C Pters W and Breon F M Canwe reconcile atmospheric estimates of Northern terrestrial car-bon sink with land-based accounting Curr Opin Env Sust 2225ndash230 httpsdoiorg101016jcosust201006008 2010

Ciais P Sabine C Bala G Bopp L Brovkin V CanadellJ Chhabra A DeFries R Galloway J Heimann MJones C Le Queacutereacute C Myneni R B Piao S and Thorn-ton P Carbon and Other Biogeochemical Cycles in Cli-mate Change 2013 The Physical Science Basis Contribu-tion of Working Group I to the Fifth Assessment Report ofthe Intergovernmental Panel on Climate Change edited byStocker T F Qin D Plattner G-K Tignor M Allen SK Boschung J Nauels A Xia Y Bex V and Midgley PM Cambridge University Press Cambridge UK and New YorkNY USA 465ndash570 available at httpswwwipccchreportar5wg1carbon-and-other-biogeochemical-cycles (last access22 August 2019) 2013

CLRTAP Mapping critical loads for ecosystems Chapter V ofManual on methodologies and criteria for modelling and map-ping critical loads and levels and air pollution effects risks andtrends UNECE Convention on Long-range Transboundary AirPollution available at httpwwwicpmappingorg (last access24 May 2019) 2017

Collalti A and Prentice I C Is NPP proportional to GPP War-ingrsquos hypothesis 20 years on Tree Physiol 39 1473ndash1483httpsdoiorg101093treephystpz034 2019

Daumlmmgen U Atmospheric nitrogen dynamics in Hesse GermanyCreating the data base 1 Bulk deposition of acidifying and eu-trophying species Landbauforschung-Ger 56 117ndash138 2006

Daumlmmgen U Erisman J W Cape J N Gruumlnhage L andFowler D Practical considerations for addressing uncertaintiesin monitoring bulk deposition Environ Pollut 134 535ndash548httpsdoiorg101016jenvpol200408013 2005

Davidson E A Keller M Erickson H E VerchotL V and Veldkamp E Testing a conceptual modelof soil emissions of nitrous and nitric oxides Bio-science 50 667ndash680 httpsdoiorg1016410006-3568(2000)050[0667tacmos]20co2 2000

De Schrijver A Verheyen K Mertens J Staelens J Wuyts Kand Muys B Nitrogen saturation and net ecosystem productionNature 451 E1 httpsdoiorg101038nature06578 2008

De Vries W van der Salm C Reinds G J and Erisman J WElement fluxes through European forest ecosystems and theirrelationships with stand and site characteristics Environ Pollut148 501ndash513 httpsdoiorg101016jenvpol2006120012007

De Vries W Solberg S Dobbertin M Sterba H Laubhann DReinds G J Nabuurs G J Gundersen P and Sutton M AEcologically implausible carbon response Nature 451 E1ndashE3httpsdoiorg101038nature06579 2008

De Vries W Solberg S Dobbertin M Sterba H LaubhannD van Oijen M Evans C Gundersen P Kros J WamelinkG W W Reinds G J and Sutton M A The impactof nitrogen deposition on carbon sequestration by Europeanforests and heathlands Forest Ecol Manag 258 1814ndash1823httpsdoiorg101016jforeco200902034 2009

De Vries W Du E and Butterbach-Bahl K Short and long-term impacts of nitrogen deposition on carbon sequestrationby forest ecosystems Curr Opin Env Sust 910 90ndash104httpsdoiorg101016jcosust201409001 2014

De Vries W Posch M Simpson D and Reinds GJ Mod-elling long-term impacts of changes in climate nitrogen de-position and ozone exposure on carbon sequestration of Eu-ropean forest ecosystems Sci Total Environ 605606 1097ndash1116 httpsdoiorg101016jscitotenv201706132 2017

Dezi S Medlyn B E Tonon G and Magnani F The ef-fect of nitrogen deposition on forest carbon sequestration amodel-based analysis Glob Change Biol 16 1470ndash1486httpsdoiorg101111j1365-2486200902102x 2010

Dinsmore K J Billet M F Skiba U M Rees R M DrewerJ and Helfter C Role of the aquatic pathway in the car-bon and greenhouse gas budgets of a peatland catchment GlobChange Biol 16 2750ndash2762 httpsdoiorg101111j1365-2486200902119x 2010

Dise N B and Wright R F Nitrogen leaching from Europeanforests in relation to nitrogen deposition Forest Ecol Manag71 153ndash161 httpsdoiorg1010160378-1127(94)06092-W1995

Dise N B Rothwell J J Gauci V van der Salm Cand de Vries W Predicting dissolved inorganic nitro-gen leaching in European forests using two indepen-



dent databases Sci Total Environ 407 1798ndash1808httpsdoiorg101016jscitotenv200811003 2009

Drewer J Lohila A Aurela M Laurila T Minkkinen K Pent-tilauml T Dinsmore K M McKenzie R Helfter C Flechard CSutton M A and Skiba U M Comparison of greenhouse gasfluxes and nitrogen budgets from an ombotrophic bog in Scot-land and a minerotrophic sedge fen in Finland Eur J Soil Sci61 640ndash650 httpsdoiorg101111j1365-2389201001267x2010

Du E and de Vries W Nitrogen-induced new net primary produc-tion and carbon sequestration in global forests Environ Pollut242 1476ndash1487 httpsdoiorg101016jenvpol2018080412018

Eckelmann W Sponagel H Grottenthaler W Hartmann K-JHartwich R Janetzko P Joisten H Kuumlhn D Sabel K-Jand Traidl R Ad-hoc-Arbeitsgruppe Boden BodenkundlicheKartieranleitung (Manual of Soil Mapping in German) 5th EdnE Schweizerbart Hannover 438 pp 2005

EMEP (European Monitoring and Evaluation Programme) EMEPMSC-W modelled data available at httpwwwemepintmscwmscw_datahtml (last access 22 August 2019) 2013

Erisman J W Mennen M G Fowler D Flechard C RSpindler G Gruumlner A Duyzer J H Ruigrok W and WyersG P Towards development of a deposition monitoring networkfor air pollution in Europe Report no 722108015 RIVM theNetherlands available at httprivmopenrepositorycomrivmbitstream10029104321722108015pdf (last access 22 Au-gust 2019) 1996

Etzold S Buchmann N and Eugster W Contribution of advec-tion to the carbon budget measured by eddy covariance at a steepmountain slope forest in Switzerland Biogeosciences 7 2461ndash2475 httpsdoiorg105194bg-7-2461-2010 2010

Etzold S Ruehr N K Zweifel R Dobbertin M Zingg APluess P Haumlsler R Eugster W and Buchmann N The Car-bon Balance of Two Contrasting Mountain Forest Ecosystems inSwitzerland Similar Annual Trends but Seasonal DifferencesEcosystems 14 1289ndash1309 httpsdoiorg101007s10021-011-9481-3 2011

Etzold S Waldner P Thimonier A Schmitt M and Dob-bertin M Tree growth in Swiss forests between 1995and 2010 in relation to climate and stand conditions re-cent disturbances matter Forest Ecol Manag 311 41ndash55httpsdoiorg101016jforeco201305040 2014

European Fluxes Database Cluster available at httpwwweurope-fluxdataeu (last access 22 August 2019) 2012

Falge E Baldocchi D Olson R Anthoni P Aubinet MBernhofer C Burba G Ceulemans R Clement R Dol-man H Granier A Gross P Grunwald T Hollinger DJensen N O Katul G Keronen P Kowalski A Lai CT Law B E MeyersT Moncrieff H Moors E MungerJ W Pilegaard K RannikU Rebmann C Suyker ATenhunen J Tu K Verma S Vesala T Wilson K andWofsy S Gap filling strategies for defensible annual sums ofnet ecosystem exchange Agr Forest Meteorol 107 43ndash69httpsdoiorg101016S0168-1923(00)00225-2 2001

Fernaacutendez-Martiacutenez M Vicca S Janssens I A Ciais P Ober-steiner M Bartrons M Sardans J Verger A Canadell JG Chevallier F Wang X Bernhofer C Curtis P S Gi-anelle D Gruwald T Heinesch B Ibrom A Knohl A Lau-

rila T Law B E Limousin J M Longdoz B Loustau DMammarella I Matteucci G Monson R K Montagnani LMoors E J Munger J W Papale D Piao S L and Penue-las J Atmospheric deposition CO2 and change in the land car-bon sink Sci Rep-UK 7 9632 httpsdoiorg101038s41598-017-08755-8 2017

Feigenwinter C Bernhofer C Eichelmann U Heinesch BHertel M Janous D Kolle O Lagergren F Lindroth AMinerbi S Moderow U Moumllder M Montagnani L QueckR Rebmann C Vestin P Yernaux M Zeri M Ziegler Wand Aubinet M Comparison of horizontal and vertical advec-tive CO2 fluxes at three forest sites Agr Forest Meteorol 14812ndash24 httpsdoiorg101016jagrformet200708013 2008

Flechard C R Nemitz E Smith R I Fowler D Vermeulen AT Bleeker A Erisman J W Simpson D Zhang L Tang YS and Sutton M A Dry deposition of reactive nitrogen to Eu-ropean ecosystems a comparison of inferential models acrossthe NitroEurope network Atmos Chem Phys 11 2703ndash2728httpsdoiorg105194acp-11-2703-2011 2011

Flechard C R Massad R-S Loubet B Personne E Simp-son D Bash J O Cooter E J Nemitz E and Sutton MA Advances in understanding models and parameterizations ofbiosphere-atmosphere ammonia exchange Biogeosciences 105183ndash5225 httpsdoiorg105194bg-10-5183-2013 2013

Flechard C R van Oijen M Cameron D R de Vries W IbromA Buchmann N Dise N B Janssens I A Neirynck JMontagnani L Varlagin A Loustau D Legout A Ziem-blinska K Aubinet M Aurela M Chojnicki B H DrewerJ Eugster W Francez A-J Juszczak R Kitzler B KutschW L Lohila A Longdoz B Matteucci G Moreaux V Nef-tel A Olejnik J Sanz M J Siemens J Vesala T VinckeC Nemitz E Zechmeister-Boltenstern S Butterbach-BahlK Skiba U M and Sutton M A Carbonndashnitrogen interac-tions in European forests and semi-natural vegetation ndash Part 2Untangling climatic edaphic management and nitrogen deposi-tion effects on carbon sequestration potentials Biogeosciences17 1621ndash1654 httpsdoiorg105194bg-17-1621-2020 2020

Fleischer K Rebel K T Van Der Molen M K Erisman JW Wassen M J van Loon E E Montagnani L GoughC M Herbst M Janssens I A Gianelle D and DolmanA J The contribution of nitrogen deposition to the photosyn-thetic capacity of forests Global Biogeochem Cy 27 187ndash199httpsdoiorg101002gbc20026 2013

Foken T Goumlckede M Mauder M Mahrt L Amiro B Dand Munger J W Post-field data quality control in Handbookof Micrometeorology A guide for Surface Flux Measurementsedited by Lee X Massman W J and Law B E Kluwer Aca-demic Publishers Dordrecht 181ndash208 2004

F-ORE-T (Fonctionnement des Ecosystegravemes Forestiers) avail-able at httpwwwgip-ecofororgf-ore-treseauphp (last ac-cess 22 August 2019) 2012 (in French)

Fowler D Pitcairn C E R Sutton M A Flechard C Lou-bet B Coyle M and Munro R C The mass budget of atmo-spheric ammonia in woodland within 1 km of livestock buildingsEnviron Pollut 102 343ndash348 httpsdoiorg101016S0269-7491(98)80053-5 1998

Fowler D Pilegaard K Sutton M A Ambus P Raivonen MDuyzer J Simpson D Fagerli H Fuzzi S Schjoerring JK Granier C Neftel A Isaksen I S A Laj P Maione M



Monks P S Burkhardt J Daemmgen U Neirynck J Per-sonne E Wichink-Kruit R Butterbach-Bahl K Flechard CTuovinen J-P Coyle M Gerosa G Loubet B Altimir NGruenhage L Ammann C Cieslik S Paoletti E MikkelsenTN Ro-Poulsen H Cellier P Cape J N Horvaacuteth L LoretoF Niinemets Uuml Palmer P I Rinne J Misztal P NemitzE Nilsson D Pryor S Gallagher M W Vesala T SkibaU Bruumlggemann N Zechmeister-Boltenstern S Williams JOrsquoDowd C Facchini M C de Leeuw G Flossman AChaumerliac N and Erisman J W Atmospheric compositionchange EcosystemsndashAtmosphere interactions Atmos Environ43 5193ndash5267 httpsdoiorg101016jatmosenv2009070682009

Francez A J Pinay G Josselin N and Williams BL Denitrification triggered by nitrogen addition in Sphag-num magellanicum peat Biogeochemistry 106 435ndash441httpsdoiorg101007s10533-010-9523-5 2011

Fratini G Ibrom A Arriga N Burba G and Pa-pale D Relative humidity effects on water vapour fluxesmeasured with closed-path eddy-covariance systems withshort sampling lines Agr Forest Meteorol 165 53ndash63httpsdoiorg101016jagrformet201205018 2012

Frolking S Roulet N and Fuglestvedt J How northern peat-lands influence the Earthrsquos radiative budget Sustained methaneemission versus sustained carbon sequestration J Geophys Res111 G01008 httpsdoiorg1010292005JG000091 2006

Gaige E Dail D B Hollinger D Y Davidson E A Fernan-dez I J Sievering H White A and Halteman W Changesin canopy processes following whole-forest canopy nitrogen fer-tilization of a mature Spruce-Hemlock forest Ecosystems 101133ndash1147 httpsdoiorg101007s10021-007-9081-4 2007

GHG-Europe available at httpwwweurope-fluxdataeughg-europe (last access 22 August 2019) 2012

Gielen B Neirynck J Luyssaert S and Janssens I A The im-portance of dissolved organic carbon fluxes for the carbon bal-ance of a temperate Scots pine forest Agr Forest Meteorol151 270ndash278 httpsdoiorg101016jagrformet2010100122011

Goumlckede M Foken T Aubinet M Aurela M Banza J Bern-hofer C Bonnefond J M Brunet Y Carrara A ClementR Dellwik E Elbers J Eugster W Fuhrer J Granier AGruumlnwald T Heinesch B Janssens I A Knohl A KoebleR Laurila T Longdoz B Manca G Marek M Markka-nen T Mateus J Matteucci G Mauder M Migliavacca MMinerbi S Moncrieff J Montagnani L Moors E OurcivalJ-M Papale D Pereira J Pilegaard K Pita G Rambal SRebmann C Rodrigues A Rotenberg E Sanz M J Sed-lak P Seufert G Siebicke L Soussana J F Valentini RVesala T Verbeeck H and Yakir D Quality control of Car-boEurope flux data ndash part 1 coupling footprint analyses withflux data quality assessment to evaluate sites in forest ecosys-tems Biogeosciences 5 433ndash450 httpsdoiorg105194bg-5-433-2008 2008

Goodale C L Apps M J Birdsey R A Field C BHeath L S Houghton R A Jenkins J C KohlmaierG H Kurz W Liu S Nabuurs G J Nilsson Sand Shvidenko A Z Forest carbon sink in North Hemi-sphere Ecol Appl 12 891ndash899 httpsdoiorg1018901051-0761(2002)012[0891FCSITN]20CO2 2002

Grote R Morfopoulos C Niinemets U Sun Z KeenanT F Pacifico F and Butler T A fully integrated iso-prenoid emissions model coupling emissions to photosyn-thetic characteristics Plant Cell Environ 37 1965ndash1980httpsdoiorg101111pce12326 2014

Gu L Falge E Boden T Baldocchi D D Black TA Saleska S R Suni T Vesala T Wofsy S andXu L Observing threshold determination for night-timeeddy flux filtering Agr Forest Meteorol 128 179ndash197httpsdoiorg101016jagrformet200411006 2005

Gundale M J From F Back-Holmen L and Nordin ANitrogen deposition in boreal forests has a minor impact onthe global carbon cycle Glob Change Biol 20 276ndash286httpsdoiorg101111gcb12422 2014

Hansen K Soslashrensen L L Hertel O Geels C Skjoslashth CA Jensen B and Boegh E Ammonia emissions from de-ciduous forest after leaf fall Biogeosciences 10 4577ndash4589httpsdoiorg105194bg-10-4577-2013 2013

Hansen K Personne E Skjoslashth C A Loubet B Ibrom AJensen R Soslashrensen L L and Boegh E Investigating sourcesof measured forest-atmosphere ammonia fluxes using two-layerbi-directional modelling Agr Forest Meteorol 237238 80ndash94httpsdoiorg101016jagrformet201702008 2017

Harrison A F Harkness D D Rowland A P Garnett J S andBacon P J Annual carbon and nitrogen fluxes along the Eu-ropean forest transect determined using 14C-bomb in Carbonand nitrogen cycling in European forest ecosystems edited bySchulze E D Ecol Stud 142 Springer Berlin HeidelbergNew York 237ndash256 httpsdoiorg101007978-3-642-57219-7 2000

Hendriks D M D van Huissteden J Dolman A J andvan der Molen MK The full greenhouse gas balance ofan abandoned peat meadow Biogeosciences 4 411ndash424httpsdoiorg105194bg-4-411-2007 2007

Houmlgberg P Fan H Quist M Binkley D and Tamm C O Treegrowth and soil acidification in response to 30 years of exper-imental nitrogen loading on boreal forest Glob Change Biol12 489ndash499 httpsdoiorg101111j1365-2486200601102x2006

Houmlgberg P Nitrogen impacts on forest carbon Nature 447 781ndash782 httpsdoiorg101038447781a 2007

Ibrom A Dellwik E Larsen S E and Pilegaard K Strong low-pass filtering effects on water vapour flux measurements withclosed-path eddy correlation systems Agr Forest Meteorol147 140ndash156 httpsdoiorg101016jagrformet2007070072007

ICP International Co-operative Programme on Assessment andMonitoring of Air Pollution Effects on Forests available athttpicp-forestsnet (last access 22 August 2019) 2019

Ilvesniemi H Levula J Ojansuu R Kolari P Kulmala LPumpanen J Launiainen S Vesala T and Nikinmaa ELong-term measurements of the carbon balance of a borealScots pine dominated forest ecosystem Boreal Environ Res 14731ndash753 available at httpwwwborenvnetBERpdfsber14ber14-731pdf (last access 22 August 2019) 2009

IPCC Climate Change 2013 The Physical Science Basis Con-tribution of Working Group I to the Fifth Assessment Reportof the Intergovernmental Panel on Climate Change edited byStocker T F Qin D Plattner G-K Tignor M Allen S K



Boschung J Nauels A Xia Y Bex V and Midgley P MCambridge University Press Cambridge UK and New York NYUSA 1535 pp available at httpswwwipccchreportar5wg1(last access 22 August 2019) 2013

Jacobsen C Rademacher P Meesenburg H and Meiwes K JGehalte chemischer Elemente in Baumkompartimenten Nieder-saumlchsische Forstliche Versuchs-anstalt Goumlttingen im Auftragdes Bundesministeriums fuumlr Verbraucher-schutz Ernaumlhrung undLandwirtschaft (BMVEL) (in German) Bonn 80 pp availableat httpswwwnw-fvadefileadminuser_uploadVerwaltungPublikationen2003Jacobsen_et_al_2003_Elemengehalte_Biomassepdf (last access 22 August 2019) 2003

Juszczak R and Augustin J Exchange of the greenhousegases methane and nitrous oxide at a temperate pris-tine fen mire in Central Europe Wetlands 33 895ndash907httpsdoiorg101007s13157-013-0448-3 2013

Kanakidou M Myriokefalitakis S Daskalakis N FanourgakisGS Nenes A Baker ARTsigaridis K and MihalopoulosN Past present and future atmospheric nitrogen deposition JAtmos Sci 73 2039ndash2047 httpsdoiorg101175JAS-D-15-02781 2016

Kindler R Siemens J Kaiser K Walmsley D C BernhoferC Buchmann N Cellier P Eugster W Gleixner G Grun-wald T Heim A Ibrom A Jones S K Jones M KlumppK Kutsch W Larsen KS Lehuger S Loubet B McKen-zie R Moors E Osborne B Pilegaard K Rebmann CSaunders M Schmidt M W I Schrumpf M SeyfferthJ Skiba U Soussana J-F Sutton M A Tefs C Vow-inckel B Zeeman M J and Kaupenjohann M Dissolvedcarbon leaching from soil is a crucial component of the netecosystem carbon balance Glob Change Biol 17 1167ndash1185httpsdoiorg101111j1365-2486201002282x 2011

Kok B On the interrelation of respiration and photosynthe-sis in green plants Biochim Biophys Acta 3 625ndash631httpsdoiorg1010160006-3002(49)90136-5 1949

Korhonen J F J Pihlatie M Pumpanen J Aaltonen H HariP Levula J Kieloaho A-J Nikinmaa E Vesala T and Il-vesniemi H Nitrogen balance of a boreal Scots pine forestBiogeosciences 10 1083ndash1095 httpsdoiorg105194bg-10-1083-2013 2013

Kowalska N Chojnicki B H Rinne J Haapanala S SiedleckiP Urbaniak M Juszczak R and Olejnik J Measure-ments of methane emission from a temperate wetland bythe eddy covariance method Int Agrophys 27 283ndash290httpsdoiorg102478v10247-012-0096-5 2013

Kutsch W L and Kolari P Data quality and the role of nutrientsin forest carbon-use efficiency Nat Clim Change 5 959ndash960httpsdoiorg101038nclimate2793 2015

Lasslop G Reichstein M Papale D Richardson A DArneth A Barr A Stoy P and Wohlfahrt G Separa-tion of net ecosystem exchange into assimilation and res-piration using a light response curve approach critical is-sues and global evaluation Glob Change Biol 16 187ndash208httpsdoiorg101111j1365-2486200902041x 2010

Laubhann D Sterba H Reinds G J and de Vries WThe impact of atmospheric deposition and climate on for-est growth in European monitoring plots An empiricaltree growth model Forest Ecol Manag 258 1751ndash1761httpsdoiorg101016jforeco200809050 2009

Lee X Massman W and Law B (Eds) Handbook of microm-eteorology A guide for surface flux measurement and analysisAtmos Ocean Sci Lib 29 ISBN 1-4020-2264-6 Kluwer Aca-demic Publishers Dordrecht 250 pp 2004

Legout A Hansson K Van der Heijden G Laclau J-P Au-gusto L and Ranger J Fertiliteacute chimique des sols forestiersconcepts de base Revue forestiegravere franccedilaise 4 413ndash424httpsdoiorg104267204256556 2014 (in French)

Legout A van der Heijden G Jaffrain J Boudot J-P and Ranger J Tree species effects on solution chem-istry and major element fluxes A case study in the Mor-van (Breuil France) Forest Ecol Manag 378 244ndash258httpsdoiorg101016jforeco201607003 2016

Le Queacutereacute C Andrew R M Friedlingstein P Sitch S HauckJ Pongratz J Pickers P A Korsbakken J I Peters G PCanadell J G Arneth A Arora V K Barbero L BastosA Bopp L Chevallier F Chini L P Ciais P Doney S CGkritzalis T Goll D S Harris I Haverd V Hoffman F MHoppema M Houghton R A Hurtt G Ilyina T Jain AK Johannessen T Jones C D Kato E Keeling R F Gold-ewijk K K Landschuumltzer P Lefegravevre N Lienert S Liu ZLombardozzi D Metzl N Munro D R Nabel J E M SNakaoka S-I Neill C Olsen A Ono T Patra P PeregonA Peters W Peylin P Pfeil B Pierrot D Poulter B Re-hder G Resplandy L Robertson E Rocher M RoumldenbeckC Schuster U Schwinger J Seacutefeacuterian R Skjelvan I Stein-hoff T Sutton A Tans P P Tian H Tilbrook B TubielloF N van der Laan-Luijkx I T van der Werf G R Viovy NWalker A P Wiltshire A J Wright R Zaehle S and ZhengB Global Carbon Budget 2018 Earth Syst Sci Data 10 2141ndash2194 httpsdoiorg105194essd-10-2141-2018 2018

Limpens J Granath G Gunnarsson U Aerts R Bayley SBragazza L Bubier J Buttler A van den Berg L J LFrancez A-J Gerdol R Grosvernier P Heijmans M MP D Hoosbeek M R Hotes S Ilomets M Leith IMitchell E A D Moore T Nilsson M B Nordbakken J-F Rochefort L Rydin H Sheppard L J Thormann MWiedermann M M Williams B L and Xu B Climaticmodifiers of the response to nitrogen deposition in peat-formingSphagnum mosses a meta-analysis New Phytol 191 496ndash507httpsdoiorg101111j1469-8137201103680x 2011

Liu L and Greaver T L A review of nitrogen enrichment effectson three biogenic GHGs the CO2 sink may be largely offset bystimulated N2O and CH4 emission Ecol Lett 12 1103ndash1117httpsdoiorg101111j1461-0248200901351x 2009

Loescher H W Law B E Mahrt L Hollinger D YCampbell J and Wofsy S C Uncertainties in andinterpretation of carbon flux estimates using the eddycovariance technique J Geophys Res 111 D21S90httpsdoiorg1010292005JD006932 2006

Loubet B Asman W A Theobald M R Hertel O TangY S Robin P Hassouna M Daemmgen U GenermontS Cellier P and Sutton M A Ammonia Deposition nearHot Spots Processes Models and Monitoring Methods in At-mospheric Ammonia Detecting Emissions Changes and En-vironmental Impacts edited by Sutton M A Reis S andBaker S M Springer Dordrecht the Netherlands 205ndash267httpsdoiorg101007978-1-4020-9121-6_15 2009



Luo G J Bruumlggemann N Wolf B Gasche R Grote Rand Butterbach-Bahl K Decadal variability of soil CO2 NON2O and CH4 fluxes at the Houmlglwald Forest Germany Bio-geosciences 9 1741ndash1763 httpsdoiorg105194bg-9-1741-2012 2012

Luyssaert S Inglima I Jung M Richardson A D Reich-stein M Papale D Piao S L Schulzes E D Wingate LMatteucci G Aragao L Aubinet M Beers C BernhoferC Black K G Bonal D Bonnefond J M Chambers JCiais P Cook B Davis K J Dolman A J Gielen BGoulden M Grace J Granier A Grelle A Griffis T Gruumln-wald T Guidolotti G Hanson P J Harding R HollingerD Y Hutyra L R Kolar P Kruijt B Kutsch W Lager-gren F Laurila T Law B E Le Maire G Lindroth ALoustau D Malhi Y Mateus J Migliavacca M MissonL Montagnani L Moncrieff J Moors E Munger J WNikinmaa E Ollinger S V Pita G Rebmann C Roup-sard O Saigusa N Sanz M J Seufert G Sierra C SmithM L Tang J Valentini R Vesala T and Janssens I ACO2 balance of boreal temperate and tropical forests derivedfrom a global database Glob Change Biol 13 2509ndash2537httpsdoiorg101111j1365-2486200701439x 2007

Magnani F Mencuccini M Borghetti M Berbigier PBerninger F Delzon S Grelle A Hari P Jarvis P G Ko-lari P Kowalski A S Lankreijer H Law B E Lindroth ALoustau D Manca G Moncrieff J B Rayment M TedeschiV Valentini R and Grace J The human footprint in the car-bon cycle of temperate and boreal forests Nature 447 848ndash850httpsdoiorg101038nature05847 2007

Magnani F Mencuccini M Borghetti M Berninger F DelzonS Grelle A Hari P Jarvis P G Kolari P Kowalski A SLankreijer H Law B E Lindroth A Loustau D Manca GMoncrieff J B Tedeschi V Valentini R and Grace J Replyto A De Schrijver et al (2008) and W de Vries et al (2008)Nature 451 E3ndashE4 httpsdoiorg101038nature06580 2008

Mammarella I Peltola O Nordbo A Jaumlrvi L and Rannik Uuml Quantifying the uncertainty of eddy covariance fluxes due to theuse of different software packages and combinations of process-ing steps in two contrasting ecosystems Atmos Meas Tech 94915ndash4933 httpsdoiorg105194amt-9-4915-2016 2016

Mauder M Foken T Bernhofer C Clement R Elbers JEugster W Gruumlnwald T Heusinkveld B and Kolle OQuality control of CarboEurope flux data ndash Part 2 Inter-comparison of eddy-covariance software Biogeosciences 5451ndash462 httpsdoiorg105194bg-5-451-2008 2008

Medinets S Gasche R Skiba U Schindlbacher A Kiese Rand Butterbach-Bahl K Cold season soil NO fluxes from atemperate forest Drivers and contribution to annual budgetsEnviron Res Lett 11 114012 httpsdoiorg1010881748-93261111114012 2016

Montagnani L Manca G Canepa E Georgieva E Acosta MFeigenwinter C Janous D Kerschbaumer G Lindroth AMinach L Minerbi S Moumllder M Pavelka M Seufert GZeri M and Ziegler W A new mass conservation approach tothe study of CO2 advection in an alpine forest J Geophys Res-Atmos 114 D07306 httpsdoiorg1010292008JD0106502009

Montagnani L Manca G Canepa E and Georgieva EAssessing the method-specific differences in quantifica-

tion of CO2 advection at three forest sites during theADVEX campaign Agr Forest Meteorol 150 702ndash711httpsdoiorg101016jagrformet201001013 2010

Monteith J L and Unsworth M H Principles of EnvironmentalPhysics 2nd Edn Edward Arnold London 291 pp 1990

Morison J Matthews R Miller G Perks M Randle TVanguelova E White M and Yamulki S Understanding thecarbon and greenhouse gas balance of forests in Britain ForestryCommission Research Report Forestry Commission EdinburghUK 149 pp 2012

Nadelhoffer K J Emmett B A Gundersen P Kjoslashnaas OJ Koopmansk C J Schleppi P Tietemak A and WrightR F Nitrogen deposition makes a minor contribution to car-bon sequestration in temperate forests Nature 398 145ndash148httpsdoiorg10103818205 1999

Nair R K F Perks M P Weatherall A Baggs E M andMencuccini M Does canopy nitrogen uptake enhance car-bon sequestration by trees Glob Change Biol 22 875ndash888httpsdoiorg101111gcb13096 2016

Neirynck J Kowalski A S Carrara A Genouw G Bergh-mans P and Ceulemans R Fluxes of oxidised and reducednitrogen above a mixed coniferous forest exposed to vari-ous nitrogen emission sources Environ Pollut 149 31ndash43httpsdoiorg101016jenvpol200612029 2007

Nemitz E Mammarella I Ibrom A Aurela M Burba G GDengel S Gielen B Grelle A Heinesch B Herbst M Houmlrt-nagel L Klemedtsson L Lindroth A Lohila A McDermittD K Meier P Merbold L Nelson D Nicolini G NilssonM B Peltola O Rinne J and Zahniser M Standardisation ofeddy-covariance flux measurements of methane and nitrous ox-ide Int Agrophys 32 517ndash549 httpsdoiorg101515intag-2017-0042 2018

NEU (NitroEurope Integrated Project) available at httpwwwnitroeuropeeu (last access 22 August 2019) 2013

Nykaumlnen H Vasander H Huttunen J T and Martikainen P JEffect of experimental nitrogen load on methane and nitrous ox-ide fluxes on ombrotrophic boreal peatland Plant Soil 242 147ndash155 httpsdoiorg101023A1019658428402 2002

Pan Y Birdsey R A Fang J Houghton R Kauppi P E KurzW A Phillips O L Shvidenko A Lewis S L CanadellJ G Ciais P Jackson R B Pacala S W McGuire A DPiao S Rautiainen A Sitch S and Hayes D A large andpersistent carbon sink in the worldrsquos forests Science 333 988ndash993 httpsdoiorg101126science1201609 2011

Papale D Reichstein M Aubinet M Canfora E Bernhofer CKutsch W Longdoz B Rambal S Valentini R Vesala Tand Yakir D Towards a standardized processing of Net Ecosys-tem Exchange measured with eddy covariance technique algo-rithms and uncertainty estimation Biogeosciences 3 571ndash583httpsdoiorg105194bg-3-571-2006 2006

Parkin T B Effect of sampling frequency on estimates of cumu-lative nitrous oxide emissions J Environ Qual 37 1390ndash1395httpsdoiorg102134jeq20070333 2008

Paul-Limoges E Wolf S Eugster W Houmlrtnagl Land Buchmann N Below-canopy contributions toecosystem CO2 fluxes in a temperate mixed forestin Switzerland Agr Forest Meteorol 247 582ndash596httpsdoiorg101016jagrformet201708011 2017



Pearson T R H Brown S Murray L and Sidman G Green-house gas emissions from tropical forest degradation an un-derestimated source Carbon Balance Management 12 1ndash11httpsdoiorg101186s13021-017-0072-2 2017

Persson T Karlsson P S Seyferth U Sjoumlberg R M and Rude-beck A Carbon mineralization in European forest soils in Car-bon and nitrogen cycling in European forest ecosystems editedby Schulze E D Ecol Stud 142 Springer Berlin HeidelbergNew York 257ndash275 httpsdoiorg101007978-3-642-57219-7 2000

Pilegaard K Processes regulating nitric oxide emis-sions from soils Philos T R Soc B 368 1ndash8httpsdoiorg101098rstb20130126 2013

Pilegaard K Skiba U Ambus P Beier C Bruumlggemann NButterbach-Bahl K Dick J Dorsey J Duyzer J GallagherM Gasche R Horvath L Kitzler B Leip A Pihlatie MK Rosenkranz P Seufert G Vesala T Westrate H andZechmeister-Boltenstern S Factors controlling regional differ-ences in forest soil emission of nitrogen oxides (NO and N2O)Biogeosciences 3 651ndash661 httpsdoiorg105194bg-3-651-2006 2006

Pilegaard K Ibrom A Courtney M S Hummelshoslashj P andJensen N O Increasing net CO2 uptake by a Danish beech for-est during the period from 1996 to 2009 Agr Forest Meteorol151 934ndash946 httpsdoiorg101016jagrformet2011020132011

Pregitzer K S Burton A J Zak D R and Talhelm A F Sim-ulated chronic nitrogen deposition increases carbon storage inNorthern Temperate forests Glob Change Biol 14 142ndash153httpsdoiorg101111j1365-2486200701465x 2008

Reay D S Dentener F Smith P Grace J and Feely R AGlobal nitrogen deposition and carbon sinks Nat Geosci 1430ndash437 httpsdoiorg101038ngeo230 2008

REddyProc R package for Post Processing of (Half-)HourlyEddy-Covariance Measurements available at httpscranr-projectorgwebpackagesREddyProcREddyProcpdf (last ac-cess 22 August 2019) and httpswwwbgc-jenampgdebgiindexphpServicesREddyProcWeb (last access 22 Au-gust 2019) 2019

Reichstein M Falge E Baldocchi D Papale D AubinetM Berbigier P Bernhofer C Buchmann N Gilmanov TGranier A Gruumlnwald T Havraacutenkovaacute K Ilvesniemi HJanous D Knohl A Laurila T Lohila A Loustau D Mat-teucci G Meyers T Miglietta F Ourcival J-M PumpanenJ Rambal S Rotenberg E Sanz M Tenhunen J SeufertG Vaccari F Vesala T Yakir D and Valentini R Onthe separation of net ecosystem exchange into assimilation andecosystem respiration review and improved algorithm GlobChange Biol 11 1424ndash1439 httpsdoiorg101111j1365-24862005001002x 2005

Risk N Snider D and Wagner-Riddle C Mechanismsleading to enhanced soil nitrous oxide fluxes inducedby freeze-thaw cycles Can J Soil Sci 93 401ndash414httpsdoiorg104141CJSS2012-071 2013

Sanz M J Carratalaacute A Gimeno C and Millaacuten M M At-mospheric nitrogen deposition on the east coast of Spain rel-evance of dry deposition in semi-arid Mediterranean regionsEnviron Pollut 118 259ndash272 httpsdoiorg101016S0269-7491(01)00318-9 2002

Schaufler G Kitzler B Schindlbacher A Skiba U SuttonM A and Zechmeister-Boltenstern S Greenhouse gas emis-sions from European soils under different land use effects ofsoil moisture and temperature Eur J Soil Sci 61 683ndash696httpsdoiorg101111j1365-2389201001277x 2010

Schulte-Uebbing L and de Vries W Global-scale impacts of ni-trogen deposition on tree carbon sequestration in tropical tem-perate and boreal forests A meta-analysis Glob Change Biol24 416ndash431 httpsdoiorg101111gcb13862 2018

Schulze E-D Ciais P Luyssaert S Schrumpf M JanssensI A Thiruchittampalam B Theloke J Saurat M BringezuS Lelieveld J Lohila A Rebmann C Jung M BastvikenD Abril G Grassi G Leip A Freibauer A Kutsch WDon A Nieschulze J Boumlrner A Gash J H and Dolman AJ The European carbon balance Part 4 integration of carbonand other trace-gas uxes Glob Change Biol 16 1451ndash1469httpsdoiorg101111j1365-2486201002215x 2010

Schwede D B Simpson D Tan J Fu J S Dentener F Du Eand de Vries W Spatial variation of modelled total dry and wetnitrogen deposition to forests at global scale Environ Pollut243 1287ndash1301 httpsdoiorg101016jenvpol2018090842018

Shvaleva A Lobo-do-Vale R Cruz C Castaldi S Rosa AP Chaves M M and Pereira J S Soil-atmosphere green-house gases (CO2 CH4 and N2O) exchange in evergreen oakwoodland in southern Portugal Plant Soil Environ 57 471ndash477httpsdoiorg10172212232011-PSE 2011

Siemens J and Kaupenjohann M Contribution of dis-solved organic nitrogen to N leaching from four Germanagricultural soils J Plant Nutr Soil Sc 165 675ndash681httpsdoiorg101002jpln200290002 2002

Simpson D and Tuovinen J P ECLAIRE Ecosystem SurfaceExchange model (ESX) in Transboundary particulate matterphoto-oxidants acidifying and eutrophying components EMEPStatus Report 12014 Norwegian Meteorological InstituteNorway 147ndash154 available at httpsemepintpublreports2014EMEP_Status_Report_1_2014pdf (last access 22 Au-gust 2019) 2014

Simpson D Butterbach-Bahl K Fagerli H Kesik M SkibaU and Tang S Deposition and Emissions of Reactive Nitro-gen over European Forests A Modelling Study Atmos Environ40 5712ndash5726 httpsdoiorg101016jatmosenv2006040632006a

Simpson D Fagerli H Hellsten S Knulst J and Westling OComparison of modelled and monitored deposition fluxes of sul-phur and nitrogen to ICP-forest sites in Europe Biogeosciences3 337ndash355 httpsdoiorg105194bg-3-337-2006 2006b

Simpson D Benedictow A Berge H Bergstroumlm R Em-berson L D Fagerli H Flechard C R Hayman G DGauss M Jonson J E Jenkin M E Nyiacuteri A RichterC Semeena V S Tsyro S Tuovinen J-P Valdebenito Aacuteand Wind P The EMEP MSC-W chemical transport modelndash technical description Atmos Chem Phys 12 7825ndash7865httpsdoiorg105194acp-12-7825-2012 2012

Simpson D Andersson C Christensen J H Engardt MGeels C Nyiri A Posch M Soares J Sofiev M WindP and Langner J Impacts of climate and emission changeson nitrogen deposition in Europe a multi-model study At-



mos Chem Phys 14 6995ndash7017 httpsdoiorg105194acp-14-6995-2014 2014

Skiba U Drewer J Tang Y S van Dijk N Helfter C Ne-mitz E Famulari D Cape J N Jones S K Twigg MPihlatie M Vesala T Larsen K S Carter M S AmbusP Ibrom A Beier C Hensen A Frumau A Erisman JW Bruggemann N Gasche R Butterbach-Bahl K Nef-tel A Spirig C Horvath L Freibauer A Cellier P Lav-ille P Loubet B Magliulo E Bertolini T Seufert V An-dersson M Manca G Laurila T Aurela M Lohila AZechmeister-Boltenstern S Kitzler B Schaufler G SiemensJ Kindler R Flechard C R and Sutton M A Biosphere-atmosphere exchange of reactive nitrogen and greenhouse gasesat the NitroEurope core flux measurement sites Measurementstrategy and first data sets Agr Ecosyst Environ 133 139ndash149httpsdoiorg101016jagee200905018 2009

Smith K A Dobbie K E Ball B C Bakken L R SitaulaB K Hansen S Brumme R Borken W Kristensen SPrieme A Fowler D Macdonald J A Skiba U Klemedts-son L Kasimir-Klemedtsson A Degorska A and Orlan-ski P Oxidation of atmospheric methane in Northern Euro-pean soils comparison with other ecosystems and uncertaintiesin the global terrestrial sink Glob Change Biol 6 791ndash803httpsdoiorg101046j1365-2486200000356x 2000

Solberg S Dobbertin M Reinds G J Andreassen K LangeH Garcia Fernandez P Hildingsson A and de Vries WAnalyses of the impact of changes in atmospheric depositionand climate on forest growth in European monitoring plots Astand growth approach Forest Ecol Manag 258 1735ndash1750httpsdoiorg101016jforeco200809057 2009

Steudler P A Bowden R D Melillo J M and AberJ D Influence of nitrogen fertilization on methane up-take in temperate forest soils Nature 341 314ndash316httpsdoiorg101038341314a0 1989

Stephens B B Gurney K R Tans P P Sweeney C Pe-ters W Bruhwiler L Ciais P Ramonet M Bousquet PNakazawa T Aoki S Machida T Inoue G VinnichenkoN Lloyd J Jordan A Heimann M Shibistova O Lan-genfelds R L Steele L P Francey R J Denning A SWeak northern and strong tropical land carbon uptake from ver-tical profiles of atmospheric CO2 Science 316 1732ndash1735httpsdoiorg101126science1137004 2007

Subke J-A Inglima I and Cotrufo M F Trends andmethodological impacts in soil CO2 efflux partitioning Ameta-analytical review Glob Change Biol 12 921ndash943httpsdoiorg101111j1365-2486200601117x 2006

Sutton M A and Reis S (Eds) The nitrogen cycle and its influ-ence on the European greenhouse gas balance NitroEurope finalproject report Center for Ecology and Hydrology UK 44 pp2011

Sutton M A Tang Y S Miners B and Fowler D A new diffu-sion denuder system for long-term regional monitoring of atmo-spheric ammonia and ammonium Water Air Soil Poll Focus 1145ndash156 httpsdoiorg101023A1013138601753 2001

Sutton M A Simpson D Levy P E Smith R I ReisS van Oijen M and de Vries W Uncertainties in the re-lationship between atmospheric nitrogen deposition and for-est carbon sequestration Glob Change Biol 14 2057ndash2063httpsdoiorg101111j1365-2486200801636x 2008

Swinbank W C The measurement of vertical transfer of heat andwater vapor by eddies in the lower atmosphere J Meteorol 8135ndash145 1951

Tang Y S Simmons I van Dijk N Di Marco C NemitzE Daumlmmgen U Gilke K Djuricic V Vidic S Gliha ZBorovecki D Mitosinkova M Hanssen J E Uggerud T HSanz M J Sanz P Chorda J V Flechard C R Fauvel YFerm M Perrino C and Sutton M A European scale applica-tion of atmospheric reactive nitrogen measurements in a low-costapproach to infer dry deposition fluxes Agr Ecosyst Environ133 183ndash195 httpsdoiorg101016jagee200904027 2009

Thimonier A Kosonen Z Braun S Rihm B SchleppiP Schmitt M Seitler E Waldner P and Thoumlni LTotal deposition of nitrogen in Swiss forests Compar-ison of assessment methods and evaluation of changesover two decades Atmos Environ 198 335ndash350httpsdoiorg101016jatmosenv201810051 2018

Thomas R Q Canham C D Weathers K C andGoodale C L Increased tree carbon storage in responseto nitrogen deposition in the US Nat Geosci 3 13ndash17httpsdoiorg101038ngeo721 2010

Treseder K K Nitrogen additions and microbial biomass ameta-analysis of ecosystem studies Ecol Lett 11 1111ndash1120httpsdoiorg101111j1461-0248200801230x 2008

van Genuchten M T A closed-form equation forpredicting the hydraulic conductivity of unsatu-rated soils Soil Sci Soc Am J 44 892ndash898httpsdoiorg102136sssaj198003615995004400050002x1980

van Oijen M Rougier J and Smith R Bayesian cali-bration of process-based forest models bridging the gapbetween models and data Tree Physiol 25 915ndash927httpsdoiorg101093treephys257915 2005

Verstraeten A De Vos B Neirynck J Roskams P andHens M Impact of air-borne or canopy-derived dis-solved organic carbon (DOC) on forest soil solution DOCin Flanders Belgium Atmos Environ 83 155ndash165httpsdoiorg101016jatmosenv201310058 2014

Vicca S Luyssaert S Penuelas J Campioli M Chapin F SCiais P Heinemeyer A Houmlgberg P Kutsch W L Law BE Malhi Y Papale D Piao S L Reichstein M Schulze ED and Janssens I A Fertile forests produce biomass more effi-ciently Ecol Lett 15 520ndash526 httpsdoiorg101111j1461-0248201201775x 2012

Vickers D Thomas C K Martin J G and LawB Self-correlation between assimilation and respira-tion resulting from flux partitioning of eddy-covarianceCO2 fluxes Agr Forest Meteorol 149 1552ndash1555httpsdoiorg101016jagrformet200903009 2009

Vileacuten T Cienciala E Schelhaas M J Verkerk P J Lindner Mand Peltola H Increasing carbon sinks in European forests ef-fects of afforestation and changes in mean growing stock volumeForestry 89 82ndash90 httpsdoiorg101093forestrycpv0342016

Vitousek P M Cassman K Cleveland C Crews T Field CB Grimm N B Horwarth R W Marino R Martinelli LRastetter E B and Sprent J Towards an ecological under-standing of biological nitrogen fixation Biogeochemistry 57 1ndash45 httpsdoiorg101023A1015798428743 2002



Waldrop M P and Zak D R Response of Oxidative En-zyme Activities to Nitrogen Deposition Affects Soil Concen-trations of Dissolved Organic Carbon Ecosystems 9 921ndash933httpsdoiorg101007s10021-004-0149-0 2006

Wang L Ibrom A Korhonen J F J Frumau K F A WuJ Pihlatie M and Schjoslashrring J K Interactions between leafnitrogen status and longevity in relation to N cycling in threecontrasting European forest canopies Biogeosciences 10 999ndash1011 httpsdoiorg105194bg-10-999-2013 2013

Webb E K Pearman G I and Leuning R Correction offlux measurements for density effects due to heat and wa-ter vapour transfer Q J Roy Meteor Soc 106 85ndash100httpsdoiorg101002qj49710644707 1980

Wehr R Munger J W McManus J B Nelson D D ZahniserM S Davidson E A Wofsy S C and Saleska S R Sea-sonality of temperate forest photosynthesis and daytime respira-tion Nature 534 680ndash683 httpsdoiorg101038nature179662016

Wohlfahrt G and Galvagno M Revisiting the choice ofthe driving temperature for eddy covariance CO2 fluxpartitioning Agr Forest Meteorol 237238 135ndash142httpsdoiorg101016jagrformet201702012 2017

Wortman E Tomaszewski T Waldner P Schleppi PThimonier A Eugster W Buchmann N and Siever-ing H Atmospheric nitrogen deposition and canopy reten-tion influences on photosynthetic performance at two highnitrogen deposition Swiss forests Tellus B 64 17216httpsdoiorg103402tellusbv64i017216 2012

Wu J Jansson P E vd Linden L Pilegaard K Beier Cand Ibrom A Modelling the decadal trend of ecosystem car-bon fluxes demonstrates the important role of biotic changesin a temperate deciduous forest Ecol Model 260 50ndash61httpsdoiorg101016jecolmodel201303015 2013

Zechmeister-Boltenstern S Keiblinger K M Mooshammer MPenuelas J Rchter A Sardans J and Wanek W The ap-plication of ecological stoichiometry to plantndashmicrobialndashsoilorganic matter transformations Ecol Monogr 85 133ndash155httpsdoiorg10189014-07771 2015

Zhang L Vet R OrsquoBrien J M Mihele C Liang Z andWiebe A Dry deposition of individual nitrogen species ateight Canadian rural sites J Geophys Res 114 D02301httpsdoiorg1010292008JD010640 2009

Zhou P Ganzeveld L Rannik Uuml Zhou L Gierens RTaipale D Mammarella I and Boy M Simulating ozonedry deposition at a boreal forest with a multi-layer canopydeposition model Atmos Chem Phys 17 1361ndash1379httpsdoiorg105194acp-17-1361-2017 2017


Abstract

Introduction

Materials and methods

Monitoring sites

Nitrogen fluxes

Atmospheric deposition

Soil gaseous and leaching losses

Carbon fluxes

Ecosystemndashatmosphere CO2 exchange

Soil CO2 and CH4 fluxes

Dissolved carbon losses

Ecosystem greenhouse gas balance

Ancillary soil plant and ecosystem measurements

Modelling of C and N fluxes and pools by the BASFOR ecosystem model

Results

Nitrogen inputs and outputs

Nitrogen deposition

Nitrogen losses

Net carbon and greenhouse gas balance

Spatial variability of the carbon sink in relation to climate and nitrogen deposition

Differences between plant functional types

Carbon fluxes and pools derived from forest ecosystem modelling

Net ecosystem greenhouse gas budgets

Discussion

Constraining the ecosystem nitrogen balance through combined measurements and modelling

Reducing uncertainty in nitrogen deposition

Uncertainty in ecosystem nitrogen losses and net balance

Drivers and uncertainties of the carbon and GHG balance

Variability of carbon sequestration efficiency

Forest net greenhouse gas balance dominated by carbon

Challenges in understanding the coupling of carbon and nitrogen budgets

Tangled effects of nitrogen deposition and climate on ecosystem productivity

Evidence of nitrogen saturation from various indicators

Conclusions

Code and data availability

Supplement

Author contributions

Competing interests

Acknowledgements

Financial support

Review statement

References


18Karlsruhe Institute of Technology (KIT) Institute of Meteorology and Climate Research AtmosphericEnvironmental Research (IMK-IFU) Kreuzeckbahnstr 19 82467 Garmisch-Partenkirchen Germany19Institute of Soil Science and Soil Conservation iFZ Research Centre for Biosystems Land Use and NutritionJustus Liebig University Giessen Heinrich-Buff-Ring 26ndash32 35392 Giessen Germany20University of Rennes CNRS UMR6553 ECOBIO Campus de Beaulieu263 avenue du Geacuteneacuteral Leclerc 35042 Rennes France21Leibniz Centre for Agricultural Landscape Research (ZALF) Eberswalder Straszlige 8415374 Muumlncheberg Germany22AN Severtsov Institute of Ecology and Evolution Russian Academy of Sciences119071 Leninsky pr33 Moscow Russia23Department of Meteorology Poznan University of Life Sciences Piatkowska 94 60-649 Poznan Poland24Department of Matter and Energy Fluxes Global Change Research Centre AS CRvvi Belidla 9864a 603 00 Brno Czech Republic25Department of Ecology and Environmental Protection Laboratory of BioclimatologyPoznan University of Life Sciences Piatkowska 94 60-649 Poznan Poland26Finnish Meteorological Institute Climate System Research PL 503 00101 Helsinki Finland27Ecologie Systeacutematique Evolution Univ Paris-Sud CNRS AgroParisTechUniversiteacute Paris-Saclay 91400 Orsay France28Weststrasse 5 38162 Weddel Germany29Air Quality Department Meteorological and Hydrological Service Gric 3 10000 Zagreb Croatia30TNO Environmental Modelling Sensing and Analysis Petten the Netherlands31Institut National de la Recherche en Agriculture Alimentation et Environnement (INRAE) UMR1434 Silva Site de NancyRue drsquoAmance 54280 Champenoux France32Greengrass ndash Atmospheric Environment Expert Ltd fellowship Korneacutelia utca 14a 2030 Eacuterd Hungary33Federal Research and Training Centre for Forests Natural Hazards and LandscapeSeckendorff-Gudent-Weg 8 1131 Vienna Austria34Wageningen University and Research PO Box 47 6700AA Wageningen the Netherlands35Integrated Carbon Observation System (ICOS ERIC) Head Office Erik Palmeacutenin aukio 1 00560 Helsinki Finland36Centro de Estudos Florestais Instituto Superior de Agronomia Universidade de LisboaTapada da Ajuda 1349-017 Lisbon Portugal37Institute for Atmospheric and Earth System ResearchPhysics Faculty of SciencePO Box 68 00014 University of Helsinki Helsinki Finland38Gembloux Agro-Bio Tech Axe Echanges Ecosystegravemes Atmosphegravere 8 Avenue de la Faculteacute 5030 Gembloux Belgium39Global Change Research Institute Academy of Sciences Belidla 4a 603 00 Brno Czech Republic40National Research Council of Italy Institute for Agriculture and Forestry Systems in the Mediterranean(CNR-ISAFOM) Via Patacca 85 80056 Ercolano (NA) Italy41Department of Air Quality Slovak Hydrometeorological Institute Jeseniova 17 83315 Bratislava Slovakia42Institute for Geosciences and Environmental research (IGE) UMR 5001 Universiteacute Grenoble Alpes CNRS IRDGrenoble Institute of Technology 38000 Grenoble France43NRE Oberwohlenstrasse 27 3033 Wohlen bei Bern Switzerland44CEFE CNRS Univ Montpellier Univ Paul Valeacutery Montpellier 3 EPHE IRD Montpellier France45Mechanical Engineering Department Instituto Superior Teacutecnico (Technical University of Lisbon)Ave Rovisco Pais IST 1049-001 Lisbon Portugal46Fundacioacuten CEAM CCharles R Darwin 46980 Paterna (Valencia) Spain47Department of Plant and Environmental Sciences Faculty of Science University of CopenhagenThorvaldsensvej 40 1871 Frederiksberg C Denmark48Laboratory of Functional Ecology and Global Change (ECOFUN) Forest Science and Technology Centreof Catalonia (CTFC) Carretera de Sant Llorenccedil de Morunys 25280 Solsona Spain49Group GAMES amp Department of Horticulture Botany and Landscaping School of Agrifood and Forestry Scienceand Engineering University of Lleida Av Rovira Roure 191 25198 Lleida Spain50Norsk institutt for luftforskning Postboks 100 2027 Kjeller Norway51Earth and Life Institute (Environmental sciences) Universiteacute catholique de Louvain Louvain-la-Neuve Belgium52Department of Meteorology Eoumltvoumls Loraacutend University 1117 Budapest Paacutezmaacuteny Peacuteter s 1A Hungary53Department of Forest and Soil Sciences Institute of Soil Research University of Natural Resourcesand Life Sciences Vienna Peter Jordan Str 82 1190 Vienna Austria








1 Introduction




















21 Monitoring sites




22 Nitrogen fluxes







Tabl

e1

Ove

rvie

wof

ecos

yste

man

dcl

imat

icch

arac

teri

stic

san

din

ter-

annu

alm

ean

ecos

yste

mndasha

tmos

pher

eex

chan

geflu

xes

forf

ores

tand

sem

i-na

tura

lsho

rtve

geta

tion

site

s

Site

Site

nam

ePF

Ta

Dom

inan

tFo

rest

age

Hm

axb

LA

I max

cL

at

Lon

gE

leva

tiond

MA

Te

MA

PfN

depg

GPP

hR

ecoi

NE

Pj

acro

nym

coun

try

shor

tnam

eve

geta

tion

(201

0)(m

)(m

2mminus

2 )(

N)

(E

)(m

as

l)(

C)

(mm

)(g

Nmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )

DE

-Hai

Hai

nich

Ger

man

yD

B1

Fagu

ssy

lvat

ica

142

234

051

079

104

5243

08

477

52

315

5310

7447

9D

K-S

orSo

roslashD

enm

ark

DB

2Fa

gus

sylv

atic

a91

314

655

487

116

4640

89

730

22

1883

1581

301

FR-F

onFo

ntai

nebl

eau-

Bar

beau

Fra

nce

DB

3Q

uerc

uspe

trae

a11

128

51

484

762

780

9211

069

01

718

5011

8566

5FR

-Fgs

Foug

egraveres

Fra

nce

DB

4Fa

gus

sylv

atic

a41

206

048

383

minus1

185

140

103

900

24

1725

1316

409

FR-H

esH

esse

Fra

nce

DB

5Fa

gus

sylv

atic

a45

166

748

674

706

630

010

297

51

716

3411

8744

6IT

-Col

Col

lelo

ngo

Ital

yD

B6

Fagu

ssy

lvat

ica

120

225

741

849

135

8815

607

211

401

214

2577

665

0C

Z-B

K1

Bily

Kri

zC

zech

Rep

E

N1

Pic

eaab

ies

3313

98

495

0318

538

908

78

1200

21

1548

767

781

DE

-Hoe

Houmlg

lwal

dG

erm

any

EN

2P

icea

abie

s10

435

63

483

0011

100

540

89

870

32

1856

1229

627

DE

-Tha

Tha

rand

tG

erm

any

EN

3P

icea

abie

s12

027

67

509

6413

567

380

88

820

23

1997

1396

601

DE

-Wet

Wet

zste

inG

erm

any

EN

4P

icea

abie

s56

227

150

453

114

5878

56

695

02

218

0917

6743

IT-R

enR

enon

Ita

lyE

N5

Pic

eaab

ies

111

295

146

588

114

3517

304

610

101

313

5352

882

6R

U-F

yoFy

odor

ovsk

oye

Rus

sia

EN

6P

icea

abie

s19

021

28

564

6232

922

265

53

711

10

1488

1559

minus70

UK

-Gri

Gri

ffin

UK

EN

7P

icea

sitc

hens

is29

126

556

617

minus3

800

340

77

1200

07

989

677

311

BE

-Bra

Bra

ssch

aat

Bel

gium

EN

8P

inus

sylv

estr

is82

211

951

309

452

116

108

850

41

1272

1149

123

ES-

ES1

ElS

aler

Spa

inE

N9

Pin

usha

lepe

nsis

111

102

639

346

minus0

319

517

655

12

115

5296

059

3FI

-Hyy

Hyy

tiaumllauml

Fin

land

EN

10P

inus

sylv

estr

is48

183

461

848

242

9518

13

870

90

511

1484

526

8FI

-Sod

Soda

nkyl

aumlFi

nlan

dE

N11

Pin

ussy

lves

tris

100

131

267

362

266

3818

0minus

04

527

03

551

598

minus47

FR-B

ilB

ilos

Fran

ceE

N12

Pin

uspi

nast

er9

40

544

522

minus0

896

5012

493

00

811

7898

918

9FR

-LB

rL

eB

ray

Fran

ceE

N13

Pin

uspi

nast

er41

221

944

717

minus0

769

6112

997

21

619

0614

7942

7IT

-SR

oSa

nR

osso

reI

taly

EN

14P

inus

pina

ster

6118

40

437

2810

284

414

992

01

622

5617

0255

4N

L-L

ooL

oobo

sN

ethe

rlan

dsE

N15

Pin

ussy

lves

tris

101

181

552

168

574

425

100

786

42

1617

1141

476

NL

-Spe

Speu

lder

bos

Net

herl

ands

EN

16P

seud

otsu

gam

enzi

esii

5132

75

522

525

691

5210

083

44

314

1610

1540

1SE

-Nor

Nor

unda

Sw

eden

EN

17P

inus

sylv

estr

is11

228

46

600

8317

467

456

852

70

614

1413

5658

SE-S

k2Sk

ytto

rpS

wed

enE

N18

Pin

ussy

lves

tris

3916

32

601

2917

840

557

452

70

512

3595

328

2E

S-L

Ma

Las

Maj

adas

Spa

inE

B1

Que

rcus

ilex

111

80

639

941

minus5

773

258

161

528

09

1091

958

133

FR-P

uePu

eacutecha

bon

Fran

ceE

B2

Que

rcus

ilex

696

29

437

413

596

270

137

872

11

1309

1030

279

IT-R

o2R

occa

resp

ampa

niI

taly

EB

3Q

uerc

usce

rris

2116

38

423

9011

921

224

157

876

18

1707

886

821

PT-E

spE

spir

raP

ortu

gal

EB

4E

ucal

yptu

sgl

obul

us25

202

738

639

minus8

602

9516

170

91

214

7311

6331

1PT

-Mi1

Mitr

aPo

rtug

alE

B5

Que

rcus

ilex

Que

rcus

sube

r91

83

438

541

minus8

000

264

145

665

09

870

817

53B

E-V

ieV

iels

alm

Bel

gium

MF1

Fagu

ssy

lvat

ica

Pse

udot

suga

men

zies

ii86

305

150

305

599

745

08

110

001

717

9212

4754

5C

H-L

aeL

aumlger

enS

witz

erla

ndM

F2Fa

gus

sylv

atic

aP

icea

abie

s11

130

36

474

788

365

689

77

1100

22

1448

757

692

DE

-Meh

Meh

rste

dtG

erm

any

SN1

Aff

ores

ted

gras

slan

dn

a0

52

951

276

106

5729

39

154

71

511

7111

75minus

4E

S-V

DA

Val

ldrsquoA

linya

Spa

inSN

2U

plan

dgr

assl

and

na

01

14

421

521

448

1765

64

1064

12

669

528

140

FI-L

omL

ompo

lojauml

nkkauml

Fin

land

SN3

Peat

land

na

04

10

679

9824

209

269

minus1

052

10

137

734

532

HU

-Bug

Bug

acH

unga

rySN

4Se

mi-

arid

gras

slan

dn

a0

54

746

692

196

0211

110

750

01

410

4491

812

6IT

-Am

pA

mpl

ero

Ital

ySN

5U

plan

dgr

assl

and

na

04

25

419

0413

605

884

98

1365

09

1241

1028

213

IT-M

Bo

Mon

teB

ondo

neI

taly

SN6

Upl

and

gras

slan

dn

a0

32

546

029

110

8315

505

111

891

714

3513

4789

NL

-Hor

Hor

stem

eer

Net

herl

ands

SN7

Peat

land

na

25

69

520

295

068

minus2

108

800

31

1584

1224

361

PL-w

etPO

LWE

TR

zeci

nPo

land

SN8

Wet

land

(ree

dss

edge

sm

osse

s)n

a2

14

952

762

163

0954

85

550

14

937

642

295

UK

-AM

oA

uche

ncor

thM

oss

UK

SN9

Peat

land

na

06

21

557

92minus

323

927

07

611

650

878

670

581

aPF

T(p

lant

func

tiona

ltyp

es)

DB

ndashde

cidu

ous

broa

dlea

ffor

est

EN

ndashev

ergr

een

need

lele

afco

nife

rous

fore

stE

Bndash

ever

gree

nbr

oadl

eafM

edite

rran

ean

fore

stM

Fndash

mix

edde

cidu

ousndash

coni

fero

usfo

rest

SN

ndashsh

orts

emi-

natu

ral

incl

udin

gm

oorl

and

peat

land

shr

ubla

ndan

dun

impr

oved

upl

and

gras

slan

db

Max

imum

cano

pyhe

ight

cM

axim

umle

afar

eain

dex

defin

edas

one-

side

dor

half

ofth

eto

tal

dA

bove

mea

nse

ale

vel

eM

ean

annu

alte

mpe

ratu

ref

Mea

nan

nual

prec

ipita

tion

gN

itrog

ende

posi

tion

hG

ross

prim

ary

prod

uctiv

ityi

Eco

syst

emre

spir

atio

nj

Net

ecos

yste

mpr

oduc

tivity

na

not

appl

icab

le








Water budget terms































23 Carbon fluxes





























3 Results












312 Nitrogen losses

















= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References







1 Introduction




















21 Monitoring sites




22 Nitrogen fluxes







Tabl

e1

Ove

rvie

wof

ecos

yste

man

dcl

imat

icch

arac

teri

stic

san

din

ter-

annu

alm

ean

ecos

yste

mndasha

tmos

pher

eex

chan

geflu

xes

forf

ores

tand

sem

i-na

tura

lsho

rtve

geta

tion

site

s

Site

Site

nam

ePF

Ta

Dom

inan

tFo

rest

age

Hm

axb

LA

I max

cL

at

Lon

gE

leva

tiond

MA

Te

MA

PfN

depg

GPP

hR

ecoi

NE

Pj

acro

nym

coun

try

shor

tnam

eve

geta

tion

(201

0)(m

)(m

2mminus

2 )(

N)

(E

)(m

as

l)(

C)

(mm

)(g

Nmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )

DE

-Hai

Hai

nich

Ger

man

yD

B1

Fagu

ssy

lvat

ica

142

234

051

079

104

5243

08

477

52

315

5310

7447

9D

K-S

orSo

roslashD

enm

ark

DB

2Fa

gus

sylv

atic

a91

314

655

487

116

4640

89

730

22

1883

1581

301

FR-F

onFo

ntai

nebl

eau-

Bar

beau

Fra

nce

DB

3Q

uerc

uspe

trae

a11

128

51

484

762

780

9211

069

01

718

5011

8566

5FR

-Fgs

Foug

egraveres

Fra

nce

DB

4Fa

gus

sylv

atic

a41

206

048

383

minus1

185

140

103

900

24

1725

1316

409

FR-H

esH

esse

Fra

nce

DB

5Fa

gus

sylv

atic

a45

166

748

674

706

630

010

297

51

716

3411

8744

6IT

-Col

Col

lelo

ngo

Ital

yD

B6

Fagu

ssy

lvat

ica

120

225

741

849

135

8815

607

211

401

214

2577

665

0C

Z-B

K1

Bily

Kri

zC

zech

Rep

E

N1

Pic

eaab

ies

3313

98

495

0318

538

908

78

1200

21

1548

767

781

DE

-Hoe

Houmlg

lwal

dG

erm

any

EN

2P

icea

abie

s10

435

63

483

0011

100

540

89

870

32

1856

1229

627

DE

-Tha

Tha

rand

tG

erm

any

EN

3P

icea

abie

s12

027

67

509

6413

567

380

88

820

23

1997

1396

601

DE

-Wet

Wet

zste

inG

erm

any

EN

4P

icea

abie

s56

227

150

453

114

5878

56

695

02

218

0917

6743

IT-R

enR

enon

Ita

lyE

N5

Pic

eaab

ies

111

295

146

588

114

3517

304

610

101

313

5352

882

6R

U-F

yoFy

odor

ovsk

oye

Rus

sia

EN

6P

icea

abie

s19

021

28

564

6232

922

265

53

711

10

1488

1559

minus70

UK

-Gri

Gri

ffin

UK

EN

7P

icea

sitc

hens

is29

126

556

617

minus3

800

340

77

1200

07

989

677

311

BE

-Bra

Bra

ssch

aat

Bel

gium

EN

8P

inus

sylv

estr

is82

211

951

309

452

116

108

850

41

1272

1149

123

ES-

ES1

ElS

aler

Spa

inE

N9

Pin

usha

lepe

nsis

111

102

639

346

minus0

319

517

655

12

115

5296

059

3FI

-Hyy

Hyy

tiaumllauml

Fin

land

EN

10P

inus

sylv

estr

is48

183

461

848

242

9518

13

870

90

511

1484

526

8FI

-Sod

Soda

nkyl

aumlFi

nlan

dE

N11

Pin

ussy

lves

tris

100

131

267

362

266

3818

0minus

04

527

03

551

598

minus47

FR-B

ilB

ilos

Fran

ceE

N12

Pin

uspi

nast

er9

40

544

522

minus0

896

5012

493

00

811

7898

918

9FR

-LB

rL

eB

ray

Fran

ceE

N13

Pin

uspi

nast

er41

221

944

717

minus0

769

6112

997

21

619

0614

7942

7IT

-SR

oSa

nR

osso

reI

taly

EN

14P

inus

pina

ster

6118

40

437

2810

284

414

992

01

622

5617

0255

4N

L-L

ooL

oobo

sN

ethe

rlan

dsE

N15

Pin

ussy

lves

tris

101

181

552

168

574

425

100

786

42

1617

1141

476

NL

-Spe

Speu

lder

bos

Net

herl

ands

EN

16P

seud

otsu

gam

enzi

esii

5132

75

522

525

691

5210

083

44

314

1610

1540

1SE

-Nor

Nor

unda

Sw

eden

EN

17P

inus

sylv

estr

is11

228

46

600

8317

467

456

852

70

614

1413

5658

SE-S

k2Sk

ytto

rpS

wed

enE

N18

Pin

ussy

lves

tris

3916

32

601

2917

840

557

452

70

512

3595

328

2E

S-L

Ma

Las

Maj

adas

Spa

inE

B1

Que

rcus

ilex

111

80

639

941

minus5

773

258

161

528

09

1091

958

133

FR-P

uePu

eacutecha

bon

Fran

ceE

B2

Que

rcus

ilex

696

29

437

413

596

270

137

872

11

1309

1030

279

IT-R

o2R

occa

resp

ampa

niI

taly

EB

3Q

uerc

usce

rris

2116

38

423

9011

921

224

157

876

18

1707

886

821

PT-E

spE

spir

raP

ortu

gal

EB

4E

ucal

yptu

sgl

obul

us25

202

738

639

minus8

602

9516

170

91

214

7311

6331

1PT

-Mi1

Mitr

aPo

rtug

alE

B5

Que

rcus

ilex

Que

rcus

sube

r91

83

438

541

minus8

000

264

145

665

09

870

817

53B

E-V

ieV

iels

alm

Bel

gium

MF1

Fagu

ssy

lvat

ica

Pse

udot

suga

men

zies

ii86

305

150

305

599

745

08

110

001

717

9212

4754

5C

H-L

aeL

aumlger

enS

witz

erla

ndM

F2Fa

gus

sylv

atic

aP

icea

abie

s11

130

36

474

788

365

689

77

1100

22

1448

757

692

DE

-Meh

Meh

rste

dtG

erm

any

SN1

Aff

ores

ted

gras

slan

dn

a0

52

951

276

106

5729

39

154

71

511

7111

75minus

4E

S-V

DA

Val

ldrsquoA

linya

Spa

inSN

2U

plan

dgr

assl

and

na

01

14

421

521

448

1765

64

1064

12

669

528

140

FI-L

omL

ompo

lojauml

nkkauml

Fin

land

SN3

Peat

land

na

04

10

679

9824

209

269

minus1

052

10

137

734

532

HU

-Bug

Bug

acH

unga

rySN

4Se

mi-

arid

gras

slan

dn

a0

54

746

692

196

0211

110

750

01

410

4491

812

6IT

-Am

pA

mpl

ero

Ital

ySN

5U

plan

dgr

assl

and

na

04

25

419

0413

605

884

98

1365

09

1241

1028

213

IT-M

Bo

Mon

teB

ondo

neI

taly

SN6

Upl

and

gras

slan

dn

a0

32

546

029

110

8315

505

111

891

714

3513

4789

NL

-Hor

Hor

stem

eer

Net

herl

ands

SN7

Peat

land

na

25

69

520

295

068

minus2

108

800

31

1584

1224

361

PL-w

etPO

LWE

TR

zeci

nPo

land

SN8

Wet

land

(ree

dss

edge

sm

osse

s)n

a2

14

952

762

163

0954

85

550

14

937

642

295

UK

-AM

oA

uche

ncor

thM

oss

UK

SN9

Peat

land

na

06

21

557

92minus

323

927

07

611

650

878

670

581

aPF

T(p

lant

func

tiona

ltyp

es)

DB

ndashde

cidu

ous

broa

dlea

ffor

est

EN

ndashev

ergr

een

need

lele

afco

nife

rous

fore

stE

Bndash

ever

gree

nbr

oadl

eafM

edite

rran

ean

fore

stM

Fndash

mix

edde

cidu

ousndash

coni

fero

usfo

rest

SN

ndashsh

orts

emi-

natu

ral

incl

udin

gm

oorl

and

peat

land

shr

ubla

ndan

dun

impr

oved

upl

and

gras

slan

db

Max

imum

cano

pyhe

ight

cM

axim

umle

afar

eain

dex

defin

edas

one-

side

dor

half

ofth

eto

tal

dA

bove

mea

nse

ale

vel

eM

ean

annu

alte

mpe

ratu

ref

Mea

nan

nual

prec

ipita

tion

gN

itrog

ende

posi

tion

hG

ross

prim

ary

prod

uctiv

ityi

Eco

syst

emre

spir

atio

nj

Net

ecos

yste

mpr

oduc

tivity

na

not

appl

icab

le








Water budget terms































23 Carbon fluxes





























3 Results












312 Nitrogen losses

















= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References

















21 Monitoring sites




22 Nitrogen fluxes







Tabl

e1

Ove

rvie

wof

ecos

yste

man

dcl

imat

icch

arac

teri

stic

san

din

ter-

annu

alm

ean

ecos

yste

mndasha

tmos

pher

eex

chan

geflu

xes

forf

ores

tand

sem

i-na

tura

lsho

rtve

geta

tion

site

s

Site

Site

nam

ePF

Ta

Dom

inan

tFo

rest

age

Hm

axb

LA

I max

cL

at

Lon

gE

leva

tiond

MA

Te

MA

PfN

depg

GPP

hR

ecoi

NE

Pj

acro

nym

coun

try

shor

tnam

eve

geta

tion

(201

0)(m

)(m

2mminus

2 )(

N)

(E

)(m

as

l)(

C)

(mm

)(g

Nmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )

DE

-Hai

Hai

nich

Ger

man

yD

B1

Fagu

ssy

lvat

ica

142

234

051

079

104

5243

08

477

52

315

5310

7447

9D

K-S

orSo

roslashD

enm

ark

DB

2Fa

gus

sylv

atic

a91

314

655

487

116

4640

89

730

22

1883

1581

301

FR-F

onFo

ntai

nebl

eau-

Bar

beau

Fra

nce

DB

3Q

uerc

uspe

trae

a11

128

51

484

762

780

9211

069

01

718

5011

8566

5FR

-Fgs

Foug

egraveres

Fra

nce

DB

4Fa

gus

sylv

atic

a41

206

048

383

minus1

185

140

103

900

24

1725

1316

409

FR-H

esH

esse

Fra

nce

DB

5Fa

gus

sylv

atic

a45

166

748

674

706

630

010

297

51

716

3411

8744

6IT

-Col

Col

lelo

ngo

Ital

yD

B6

Fagu

ssy

lvat

ica

120

225

741

849

135

8815

607

211

401

214

2577

665

0C

Z-B

K1

Bily

Kri

zC

zech

Rep

E

N1

Pic

eaab

ies

3313

98

495

0318

538

908

78

1200

21

1548

767

781

DE

-Hoe

Houmlg

lwal

dG

erm

any

EN

2P

icea

abie

s10

435

63

483

0011

100

540

89

870

32

1856

1229

627

DE

-Tha

Tha

rand

tG

erm

any

EN

3P

icea

abie

s12

027

67

509

6413

567

380

88

820

23

1997

1396

601

DE

-Wet

Wet

zste

inG

erm

any

EN

4P

icea

abie

s56

227

150

453

114

5878

56

695

02

218

0917

6743

IT-R

enR

enon

Ita

lyE

N5

Pic

eaab

ies

111

295

146

588

114

3517

304

610

101

313

5352

882

6R

U-F

yoFy

odor

ovsk

oye

Rus

sia

EN

6P

icea

abie

s19

021

28

564

6232

922

265

53

711

10

1488

1559

minus70

UK

-Gri

Gri

ffin

UK

EN

7P

icea

sitc

hens

is29

126

556

617

minus3

800

340

77

1200

07

989

677

311

BE

-Bra

Bra

ssch

aat

Bel

gium

EN

8P

inus

sylv

estr

is82

211

951

309

452

116

108

850

41

1272

1149

123

ES-

ES1

ElS

aler

Spa

inE

N9

Pin

usha

lepe

nsis

111

102

639

346

minus0

319

517

655

12

115

5296

059

3FI

-Hyy

Hyy

tiaumllauml

Fin

land

EN

10P

inus

sylv

estr

is48

183

461

848

242

9518

13

870

90

511

1484

526

8FI

-Sod

Soda

nkyl

aumlFi

nlan

dE

N11

Pin

ussy

lves

tris

100

131

267

362

266

3818

0minus

04

527

03

551

598

minus47

FR-B

ilB

ilos

Fran

ceE

N12

Pin

uspi

nast

er9

40

544

522

minus0

896

5012

493

00

811

7898

918

9FR

-LB

rL

eB

ray

Fran

ceE

N13

Pin

uspi

nast

er41

221

944

717

minus0

769

6112

997

21

619

0614

7942

7IT

-SR

oSa

nR

osso

reI

taly

EN

14P

inus

pina

ster

6118

40

437

2810

284

414

992

01

622

5617

0255

4N

L-L

ooL

oobo

sN

ethe

rlan

dsE

N15

Pin

ussy

lves

tris

101

181

552

168

574

425

100

786

42

1617

1141

476

NL

-Spe

Speu

lder

bos

Net

herl

ands

EN

16P

seud

otsu

gam

enzi

esii

5132

75

522

525

691

5210

083

44

314

1610

1540

1SE

-Nor

Nor

unda

Sw

eden

EN

17P

inus

sylv

estr

is11

228

46

600

8317

467

456

852

70

614

1413

5658

SE-S

k2Sk

ytto

rpS

wed

enE

N18

Pin

ussy

lves

tris

3916

32

601

2917

840

557

452

70

512

3595

328

2E

S-L

Ma

Las

Maj

adas

Spa

inE

B1

Que

rcus

ilex

111

80

639

941

minus5

773

258

161

528

09

1091

958

133

FR-P

uePu

eacutecha

bon

Fran

ceE

B2

Que

rcus

ilex

696

29

437

413

596

270

137

872

11

1309

1030

279

IT-R

o2R

occa

resp

ampa

niI

taly

EB

3Q

uerc

usce

rris

2116

38

423

9011

921

224

157

876

18

1707

886

821

PT-E

spE

spir

raP

ortu

gal

EB

4E

ucal

yptu

sgl

obul

us25

202

738

639

minus8

602

9516

170

91

214

7311

6331

1PT

-Mi1

Mitr

aPo

rtug

alE

B5

Que

rcus

ilex

Que

rcus

sube

r91

83

438

541

minus8

000

264

145

665

09

870

817

53B

E-V

ieV

iels

alm

Bel

gium

MF1

Fagu

ssy

lvat

ica

Pse

udot

suga

men

zies

ii86

305

150

305

599

745

08

110

001

717

9212

4754

5C

H-L

aeL

aumlger

enS

witz

erla

ndM

F2Fa

gus

sylv

atic

aP

icea

abie

s11

130

36

474

788

365

689

77

1100

22

1448

757

692

DE

-Meh

Meh

rste

dtG

erm

any

SN1

Aff

ores

ted

gras

slan

dn

a0

52

951

276

106

5729

39

154

71

511

7111

75minus

4E

S-V

DA

Val

ldrsquoA

linya

Spa

inSN

2U

plan

dgr

assl

and

na

01

14

421

521

448

1765

64

1064

12

669

528

140

FI-L

omL

ompo

lojauml

nkkauml

Fin

land

SN3

Peat

land

na

04

10

679

9824

209

269

minus1

052

10

137

734

532

HU

-Bug

Bug

acH

unga

rySN

4Se

mi-

arid

gras

slan

dn

a0

54

746

692

196

0211

110

750

01

410

4491

812

6IT

-Am

pA

mpl

ero

Ital

ySN

5U

plan

dgr

assl

and

na

04

25

419

0413

605

884

98

1365

09

1241

1028

213

IT-M

Bo

Mon

teB

ondo

neI

taly

SN6

Upl

and

gras

slan

dn

a0

32

546

029

110

8315

505

111

891

714

3513

4789

NL

-Hor

Hor

stem

eer

Net

herl

ands

SN7

Peat

land

na

25

69

520

295

068

minus2

108

800

31

1584

1224

361

PL-w

etPO

LWE

TR

zeci

nPo

land

SN8

Wet

land

(ree

dss

edge

sm

osse

s)n

a2

14

952

762

163

0954

85

550

14

937

642

295

UK

-AM

oA

uche

ncor

thM

oss

UK

SN9

Peat

land

na

06

21

557

92minus

323

927

07

611

650

878

670

581

aPF

T(p

lant

func

tiona

ltyp

es)

DB

ndashde

cidu

ous

broa

dlea

ffor

est

EN

ndashev

ergr

een

need

lele

afco

nife

rous

fore

stE

Bndash

ever

gree

nbr

oadl

eafM

edite

rran

ean

fore

stM

Fndash

mix

edde

cidu

ousndash

coni

fero

usfo

rest

SN

ndashsh

orts

emi-

natu

ral

incl

udin

gm

oorl

and

peat

land

shr

ubla

ndan

dun

impr

oved

upl

and

gras

slan

db

Max

imum

cano

pyhe

ight

cM

axim

umle

afar

eain

dex

defin

edas

one-

side

dor

half

ofth

eto

tal

dA

bove

mea

nse

ale

vel

eM

ean

annu

alte

mpe

ratu

ref

Mea

nan

nual

prec

ipita

tion

gN

itrog

ende

posi

tion

hG

ross

prim

ary

prod

uctiv

ityi

Eco

syst

emre

spir

atio

nj

Net

ecos

yste

mpr

oduc

tivity

na

not

appl

icab

le








Water budget terms































23 Carbon fluxes





























3 Results












312 Nitrogen losses

















= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References


Tabl

e1

Ove

rvie

wof

ecos

yste

man

dcl

imat

icch

arac

teri

stic

san

din

ter-

annu

alm

ean

ecos

yste

mndasha

tmos

pher

eex

chan

geflu

xes

forf

ores

tand

sem

i-na

tura

lsho

rtve

geta

tion

site

s

Site

Site

nam

ePF

Ta

Dom

inan

tFo

rest

age

Hm

axb

LA

I max

cL

at

Lon

gE

leva

tiond

MA

Te

MA

PfN

depg

GPP

hR

ecoi

NE

Pj

acro

nym

coun

try

shor

tnam

eve

geta

tion

(201

0)(m

)(m

2mminus

2 )(

N)

(E

)(m

as

l)(

C)

(mm

)(g

Nmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )(g

Cmminus

2yrminus

1 )

DE

-Hai

Hai

nich

Ger

man

yD

B1

Fagu

ssy

lvat

ica

142

234

051

079

104

5243

08

477

52

315

5310

7447

9D

K-S

orSo

roslashD

enm

ark

DB

2Fa

gus

sylv

atic

a91

314

655

487

116

4640

89

730

22

1883

1581

301

FR-F

onFo

ntai

nebl

eau-

Bar

beau

Fra

nce

DB

3Q

uerc

uspe

trae

a11

128

51

484

762

780

9211

069

01

718

5011

8566

5FR

-Fgs

Foug

egraveres

Fra

nce

DB

4Fa

gus

sylv

atic

a41

206

048

383

minus1

185

140

103

900

24

1725

1316

409

FR-H

esH

esse

Fra

nce

DB

5Fa

gus

sylv

atic

a45

166

748

674

706

630

010

297

51

716

3411

8744

6IT

-Col

Col

lelo

ngo

Ital

yD

B6

Fagu

ssy

lvat

ica

120

225

741

849

135

8815

607

211

401

214

2577

665

0C

Z-B

K1

Bily

Kri

zC

zech

Rep

E

N1

Pic

eaab

ies

3313

98

495

0318

538

908

78

1200

21

1548

767

781

DE

-Hoe

Houmlg

lwal

dG

erm

any

EN

2P

icea

abie

s10

435

63

483

0011

100

540

89

870

32

1856

1229

627

DE

-Tha

Tha

rand

tG

erm

any

EN

3P

icea

abie

s12

027

67

509

6413

567

380

88

820

23

1997

1396

601

DE

-Wet

Wet

zste

inG

erm

any

EN

4P

icea

abie

s56

227

150

453

114

5878

56

695

02

218

0917

6743

IT-R

enR

enon

Ita

lyE

N5

Pic

eaab

ies

111

295

146

588

114

3517

304

610

101

313

5352

882

6R

U-F

yoFy

odor

ovsk

oye

Rus

sia

EN

6P

icea

abie

s19

021

28

564

6232

922

265

53

711

10

1488

1559

minus70

UK

-Gri

Gri

ffin

UK

EN

7P

icea

sitc

hens

is29

126

556

617

minus3

800

340

77

1200

07

989

677

311

BE

-Bra

Bra

ssch

aat

Bel

gium

EN

8P

inus

sylv

estr

is82

211

951

309

452

116

108

850

41

1272

1149

123

ES-

ES1

ElS

aler

Spa

inE

N9

Pin

usha

lepe

nsis

111

102

639

346

minus0

319

517

655

12

115

5296

059

3FI

-Hyy

Hyy

tiaumllauml

Fin

land

EN

10P

inus

sylv

estr

is48

183

461

848

242

9518

13

870

90

511

1484

526

8FI

-Sod

Soda

nkyl

aumlFi

nlan

dE

N11

Pin

ussy

lves

tris

100

131

267

362

266

3818

0minus

04

527

03

551

598

minus47

FR-B

ilB

ilos

Fran

ceE

N12

Pin

uspi

nast

er9

40

544

522

minus0

896

5012

493

00

811

7898

918

9FR

-LB

rL

eB

ray

Fran

ceE

N13

Pin

uspi

nast

er41

221

944

717

minus0

769

6112

997

21

619

0614

7942

7IT

-SR

oSa

nR

osso

reI

taly

EN

14P

inus

pina

ster

6118

40

437

2810

284

414

992

01

622

5617

0255

4N

L-L

ooL

oobo

sN

ethe

rlan

dsE

N15

Pin

ussy

lves

tris

101

181

552

168

574

425

100

786

42

1617

1141

476

NL

-Spe

Speu

lder

bos

Net

herl

ands

EN

16P

seud

otsu

gam

enzi

esii

5132

75

522

525

691

5210

083

44

314

1610

1540

1SE

-Nor

Nor

unda

Sw

eden

EN

17P

inus

sylv

estr

is11

228

46

600

8317

467

456

852

70

614

1413

5658

SE-S

k2Sk

ytto

rpS

wed

enE

N18

Pin

ussy

lves

tris

3916

32

601

2917

840

557

452

70

512

3595

328

2E

S-L

Ma

Las

Maj

adas

Spa

inE

B1

Que

rcus

ilex

111

80

639

941

minus5

773

258

161

528

09

1091

958

133

FR-P

uePu

eacutecha

bon

Fran

ceE

B2

Que

rcus

ilex

696

29

437

413

596

270

137

872

11

1309

1030

279

IT-R

o2R

occa

resp

ampa

niI

taly

EB

3Q

uerc

usce

rris

2116

38

423

9011

921

224

157

876

18

1707

886

821

PT-E

spE

spir

raP

ortu

gal

EB

4E

ucal

yptu

sgl

obul

us25

202

738

639

minus8

602

9516

170

91

214

7311

6331

1PT

-Mi1

Mitr

aPo

rtug

alE

B5

Que

rcus

ilex

Que

rcus

sube

r91

83

438

541

minus8

000

264

145

665

09

870

817

53B

E-V

ieV

iels

alm

Bel

gium

MF1

Fagu

ssy

lvat

ica

Pse

udot

suga

men

zies

ii86

305

150

305

599

745

08

110

001

717

9212

4754

5C

H-L

aeL

aumlger

enS

witz

erla

ndM

F2Fa

gus

sylv

atic

aP

icea

abie

s11

130

36

474

788

365

689

77

1100

22

1448

757

692

DE

-Meh

Meh

rste

dtG

erm

any

SN1

Aff

ores

ted

gras

slan

dn

a0

52

951

276

106

5729

39

154

71

511

7111

75minus

4E

S-V

DA

Val

ldrsquoA

linya

Spa

inSN

2U

plan

dgr

assl

and

na

01

14

421

521

448

1765

64

1064

12

669

528

140

FI-L

omL

ompo

lojauml

nkkauml

Fin

land

SN3

Peat

land

na

04

10

679

9824

209

269

minus1

052

10

137

734

532

HU

-Bug

Bug

acH

unga

rySN

4Se

mi-

arid

gras

slan

dn

a0

54

746

692

196

0211

110

750

01

410

4491

812

6IT

-Am

pA

mpl

ero

Ital

ySN

5U

plan

dgr

assl

and

na

04

25

419

0413

605

884

98

1365

09

1241

1028

213

IT-M

Bo

Mon

teB

ondo

neI

taly

SN6

Upl

and

gras

slan

dn

a0

32

546

029

110

8315

505

111

891

714

3513

4789

NL

-Hor

Hor

stem

eer

Net

herl

ands

SN7

Peat

land

na

25

69

520

295

068

minus2

108

800

31

1584

1224

361

PL-w

etPO

LWE

TR

zeci

nPo

land

SN8

Wet

land

(ree

dss

edge

sm

osse

s)n

a2

14

952

762

163

0954

85

550

14

937

642

295

UK

-AM

oA

uche

ncor

thM

oss

UK

SN9

Peat

land

na

06

21

557

92minus

323

927

07

611

650

878

670

581

aPF

T(p

lant

func

tiona

ltyp

es)

DB

ndashde

cidu

ous

broa

dlea

ffor

est

EN

ndashev

ergr

een

need

lele

afco

nife

rous

fore

stE

Bndash

ever

gree

nbr

oadl

eafM

edite

rran

ean

fore

stM

Fndash

mix

edde

cidu

ousndash

coni

fero

usfo

rest

SN

ndashsh

orts

emi-

natu

ral

incl

udin

gm

oorl

and

peat

land

shr

ubla

ndan

dun

impr

oved

upl

and

gras

slan

db

Max

imum

cano

pyhe

ight

cM

axim

umle

afar

eain

dex

defin

edas

one-

side

dor

half

ofth

eto

tal

dA

bove

mea

nse

ale

vel

eM

ean

annu

alte

mpe

ratu

ref

Mea

nan

nual

prec

ipita

tion

gN

itrog

ende

posi

tion

hG

ross

prim

ary

prod

uctiv

ityi

Eco

syst

emre

spir

atio

nj

Net

ecos

yste

mpr

oduc

tivity

na

not

appl

icab

le








Water budget terms































23 Carbon fluxes





























3 Results












312 Nitrogen losses

















= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References







Water budget terms































23 Carbon fluxes





























3 Results












312 Nitrogen losses

















= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References

























23 Carbon fluxes





























3 Results












312 Nitrogen losses

















= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References






















3 Results












312 Nitrogen losses

















= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References






312 Nitrogen losses

















= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References











= 062









































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References








































4 Discussion






































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References



































































5 Conclusions

















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References















References








































































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References



















































































































































































Abstract

Introduction


Monitoring sites

Nitrogen fluxes



Carbon fluxes







Results


Nitrogen deposition

Nitrogen losses






Discussion










Conclusions


Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References

carbon–nitrogen interactions in european forests...

Documents