Physiological and climate controls on foliar mercury uptake by ...

Biogeosciences 19 1335ndash1353 2022httpsdoiorg105194bg-19-1335-2022copy Author(s) 2022 This work is distributed underthe Creative Commons Attribution 40 License

Research

article

Physiological and climate controls on foliar mercuryuptake by European tree speciesLena Wohlgemuth1 Pasi Rautio2 Bernd Ahrends3 Alexander Russ4 Lars Vesterdal5 Peter Waldner6Volkmar Timmermann7 Nadine Eickenscheidt8 Alfred Fuumlrst9 Martin Greve10 Peter Roskams11 Anne Thimonier6Manuel Nicolas12 Anna Kowalska13 Morten Ingerslev5 Paumlivi Merilauml14 Sue Benham15 Carmen Iacoban16Guumlnter Hoch1 Christine Alewell1 and Martin Jiskra1

1Department of Environmental Sciences University of Basel Basel 4056 Switzerland2Natural Resources Institute Finland (Luke) Ounasjoentie 6 96200 Rovaniemi Finland3Department of Environmental Control Northwest German Forest Research Institute (NW-FVA)Graumltzelstr 2 37079 Goumlttingen Germany4Landesbetrieb Forst Brandenburg Alfred-Moumlller-Straszlige 1 16225 Eberswalde Germany5Department of Geosciences and Natural Resource Management University of CopenhagenRolighedsvej 23 1958 Frederiksberg C Denmark6Swiss Federal Institute for Forest Snow and Landscape Research (WSL)Zuumlrcherstrasse 111 8903 Birmensdorf Switzerland7Division of Biotechnology and Plant Health Norwegian Institute of Bioeconomy Research (NIBIO) 1431 Arings Norway8State Agency for Nature Environment and Consumer Protection of North Rhine-Westphalia (LANUV)Leibnizstr 10 45659 Recklinghausen Germany9Department of Forest Protection Austrian Federal Research Centre for Forests Vienna 1130 Austria10Research Institute for Forest Ecology and Forestry Rhineland-Palatinate (FAWF)Hauptstr 16 67705 Trippstadt Germany11Research Institute for Nature and Forest (INBO) Gaverstraat 4 9500 Geraardsbergen Belgium12 Deacutepartement Recherche-Deacuteveloppement-Innovation Office National des Forecircts (ONF) 77300 Fontainebleau France13Laboratory of Natural Environment Chemistry Forest Research Institute Sekocin StaryBraci Lesnej 3 05-090 Raszyn Poland14Natural Resources Institute Finland (Luke) Paavo Havaksentie 3 90570 Oulu Finland15Forest Research Alice Holt Lodge Farnham Surrey GU51 3QE United Kingdom16Department of Ecology ldquoMarin Draceardquo National Institute for Research and Development in ForestryCampulung Moldovenesc Station 73 bis Calea Bucovinei 725100 Campulung Moldovenesc Romania

Correspondence Lena Wohlgemuth (lenawohlgemuthunibasch) and Martin Jiskra (martinjiskraunibasch)

Received 9 September 2021 ndash Discussion started 4 October 2021Revised 27 December 2021 ndash Accepted 16 January 2022 ndash Published 4 March 2022

Abstract Despite the importance of vegetation uptake ofatmospheric gaseous elemental mercury (Hg(0)) within theglobal Hg cycle little knowledge exists on the physiologicalclimatic and geographic factors controlling stomatal uptakeof atmospheric Hg(0) by tree foliage We investigate controlson foliar stomatal Hg(0) uptake by combining Hg measure-ments of 3569 foliage samples across Europe with data ontree speciesrsquo traits and environmental conditions To account

for foliar Hg accumulation over time we normalized foliarHg concentration over the foliar life period from the simu-lated start of the growing season to sample harvest

The most relevant parameter impacting daily foliar stom-atal Hg uptake was tree functional group (deciduous ver-sus coniferous trees) On average we measured 32 timeshigher daily foliar stomatal Hg uptake rates in deciduousleaves than in coniferous needles of the same age Across

Published by Copernicus Publications on behalf of the European Geosciences Union

1336 L Wohlgemuth et al Physiological and climate controls on foliar mercury

tree species for foliage of beech and fir and at two out ofthree forest plots with more than 20 samples we found asignificant (p lt 0001) increase in foliar Hg values with re-spective leaf nitrogen concentrations We therefore suggestthat foliar stomatal Hg uptake is controlled by tree functionaltraits with uptake rates increasing from low to high nutrientcontent representing low to high physiological activity Forpine and spruce needles we detected a significant linear de-crease in daily foliar stomatal Hg uptake with the proportionof time during which water vapor pressure deficit (VPD) ex-ceeded the species-specific threshold values of 12 and 3 kParespectively The proportion of time within the growing sea-son during which surface soil water content (ERA5-Land)in the region of forest plots was low correlated negativelywith foliar Hg uptake rates of beech and pine These find-ings suggest that stomatal uptake of atmospheric Hg(0) isinhibited under high VPD conditions andor low soil watercontent due to the regulation of stomatal conductance to re-duce water loss under dry conditions Other parameters as-sociated with forest sampling sites (latitude and altitude)sampled trees (average age and diameter at breast height)or regional satellite-observation-based transpiration product(Global Land Evaporation Amsterdam Model GLEAM) didnot significantly correlate with daily foliar Hg uptake ratesWe conclude that tree physiological activity and stomatal re-sponse to VPD and soil water content should be implementedin a stomatal Hg model to assess future Hg cycling underdifferent anthropogenic emission scenarios and global warm-ing

1 Introduction

Mercury (Hg) is a toxic pollutant that is emitted by anthro-pogenic and geogenic activities into the atmosphere whereit can be transported over large distances and is eventuallytransferred to terrestrial and ocean surfaces by dry or wetdeposition (Bishop et al 2020) From a public health per-spective transfer rates of Hg to aquatic ecosystems are par-ticularly relevant within this cycle since Hg bioaccumulationin fish for consumption represents the most important Hg ex-posure pathway to many communities internationally (UNEnvironment 2019) In order to constrain future Hg levelsin edible fish and to assess how Hg exposure responds tocurbed anthropogenic Hg emissions under the policies imple-mented by the 2017 UN Minamata convention on mercury itis essential to understand and quantify all major net deposi-tion fluxes within the global Hg cycle Wet deposition occurswhen water-soluble oxidized Hg(II) is washed out from theatmosphere with rainwater (Driscoll et al 2013 Sprovieriet al 2017) or by cloud water interception (Weiss-Penziaset al 2012) In a dry deposition process gaseous elemen-tal Hg(0) and Hg(II) directly bind to surfaces (Bishop et al2020) or Hg(0) is taken up by plants (Zhou et al 2021) For

more than 2 decades vegetation has been recognized as animportant vector for Hg(0) dry deposition within the terres-trial Hg cycle (Rea et al 1996 2002 Grigal 2003) Basedon this seminal work researchers have since highlighted thatvegetation impacts Hg levels of all other environmental com-partments within the active Hg cycle (AMAP and UNEP2019 Bishop et al 2020 Zhou et al 2021) Vegetation up-take of Hg(0) governs the seasonality of atmospheric Hg(0)in the Northern Hemisphere with concentration minima insummer at the end of the growing season (Jiskra et al 2018)Thus vegetation has been suggested to operate like a globalHg pump (Obrist 2007 Jiskra et al 2018) AtmosphericHg(0) taken up by vegetation is oxidized to Hg(II) within theplant tissue (Manceau et al 2018) and transferred to soils vialitterfall (Iverfeldt 1991 Schwesig and Matzner 2000 Reaet al 2001 Graydon et al 2008 Risch et al 2012 2017Jiskra et al 2015 Wright et al 2016 Wang et al 2016)Moreover in forests Hg deposition to the ground may oc-cur by wash-off of Hg(0) from plant surfaces via through-fall and by Hg(0) uptake into woody tissues lichen mossesand soil litter (Wang et al 2020 Obrist et al 2021) Mer-cury sequestered by forest ecosystems accumulates in soiland may subsequently be transported from watersheds tostreams rivers and the ocean where it can bioaccumulatein fish (Drenner et al 2013 Jiskra et al 2017 Sonke et al2018)

Concerning the mechanism of Hg accumulation in foliagethere are multiple lines of evidence that leaf stomata con-trol the foliar Hg(0) uptake flux to terrestrial ecosystems (i)Hg concentrations were found to be higher in internal fo-liar tissues than on leaf surfaces (Laacouri et al 2013) (ii)experiments revealed that isotopic Hg tracers are transferredfrom the air to the leaf interior (Rutter et al 2011) (iii) fo-liar Hg concentrations are associated with leaf stomatal den-sity and net photosynthesis (Laacouri et al 2013 Teixeiraet al 2018) (iv) the isotopic composition of foliage is dis-criminated in heavy isotopes compared to atmospheric Hg(0)(Demers et al 2013 Enrico et al 2016 Yu et al 2016Jiskra et al 2019) and (v) temporal and vertical variationsin net foliar Hg(0) uptake fluxes in trees agree with the mech-anism of stomatal Hg(0) uptake (Wohlgemuth et al 2020)While there is increasing consensus that vegetation uptakeof atmospheric Hg(0) occurs via the stomatal pathway thereremain research gaps regarding parameters regulating thisstomatal Hg(0) uptake (Zhou et al 2021) Consequentlythe Hg(0) dry deposition flux to terrestrial surfaces in HgEarth system models is generally parametrized by static in-ferential or resistance-in-series approaches (Travnikov et al2017) Ecosystem processes including canopy gas exchangeare sensitive to climate conditions (Running and Coughlan1988) and vary between different plant species (Reich et al2003) Trees control leaf diffusive gas fluxes through theirstomata in order to optimize the diffusive influx of carbondioxide for photosynthesis while averting excessive loss ofwater vapor to the atmosphere (Koumlrner 2013) The regu-

Biogeosciences 19 1335ndash1353 2022 httpsdoiorg105194bg-19-1335-2022

L Wohlgemuth et al Physiological and climate controls on foliar mercury 1337

lation of stomata allows trees to dynamically adjust theirmetabolism to climatic conditions (temperature atmospherichumidity water vapor pressure deficit solar radiation) andsite-specific limitations (soil moisture nutrient availability)under the constraints of tree-specific prerequisites (leaf struc-ture leaf life span water use efficiency)

In this study we aim to improve the process understand-ing of the stomatal Hg(0) uptake with the long-term goal ofadvancing the parameterization of the foliar Hg(0) uptake inHg Earth system models The objectives of the study were(i) to investigate how foliar Hg(0) uptake depends on thephysiological traits of tree species and (ii) to study how thestomatal response of trees to climate conditions controls fo-liar Hg(0) uptake We address these objectives by analyzinga large dataset of foliar Hg uptake rates tree functional traitsand climate conditions across natural gradients in Europeanforests covering various tree species and climate conditions

2 Material and methods

21 Foliage sampling and dataset description

The final dataset for this study comprises Hg concentra-tions of 3569 foliage samples from 2015 and 2017 of which2129 samples were provided by 17 participating countriesof the United Nations Economic Commission for Europe(UNECE) International Cooperative Programme on Assess-ment and Monitoring of Air Pollution Effects on Forests (ICPForests) The samples include sun-exposed leaves and nee-dles from the upper third of the tree canopy of five trees(Austrian Bio-Indicator Grid two trees) of the main specieson the plot taken during full development in summer (de-ciduous species) or at the onset of dormant season in au-tumn (evergreen species) using harmonized national methodsaccording to the ICP Forests Manual (Rautio et al 2016)as described for example in Jonard et al (2015) Around10 of samples were taken during winter needle samplingcampaigns (December until March) Sample preparation pro-cedure typically includes separation of needle age classesdrying milling and chemical analyses for macronutrientsand further drying of a subsample at 105 C to constantweight for the determination of dry weight The participatingICP Forests countries harvested and carried out these pre-processing steps and collected the associated metadata Hgmeasurements of samples from ICP Forests Level II plotswere performed at the University of Basel Additional fo-liar Hg concentration data of 1440 samples from the Aus-trian Bio-Indicator Grid organized by the Austrian FederalResearch Center for Forests (German acronym BFW) (Aus-trian Bio-Indicator Grid 2016) were included in the analy-sis The combined dataset consists of 3569 foliage samplesencompassing 23 species of coniferous and deciduous trees(Table S1 in the Supplement) The most frequent (numberof samples gt 100) species within the dataset are Norway

spruce (Picea abies n= 2073) Scots pine (Pinus sylvestrisn= 413) European beech (Fagus sylvatica n= 372) sil-ver fir (Abies alba n= 162) sessile oak (Quercus petraean= 133) Austrian pine (Pinus nigra n= 125) and com-mon oak (Quercus robur n= 101) We pooled individualtree species into groups of tree species genera (eg beechoak pine spruce see Table S1) Coniferous samples consistof needles of different age classes most of the needle sam-ples (n= 1958) flushed in the sampling season (current sea-son y0) 600 samples are 1-year-old (y1) 121 samples are2-year-old (y2) 125 are 3-year-old (y3) 22 samples are 4-year-old (y4) 60 samples are 5-year-old (y5) and 3 samplesare 6-year-old (y6) needles All data analysis of this studyconcerning tree species foliage structure nutrient contentsand meteorological and site-specific parameters (Sects 31ndash36) is based on Hg values of current-season (y0) foliage Fo-liage samples originate from 995 European sites 232 sitesare ICP Forests Level II forest monitoring plots 737 loca-tions are sampling sites of the Austrian Bio-Indicator Gridand the remaining sites (26) are not classified within the ICPForests program See Fig 1 for a geographic overview of fo-liage sampling sites from the sampling year 2017

We assembled the foliar Hg concentration dataset includ-ing the following metadata sampling date needle age classleaf mass per area (LMA 19 of samples) drying temper-ature leaf nitrogen (N) and organic carbon (Corg) concen-tration Foliage concentrations of N and Corg were measuredin laboratories in respective ICP Forests countries followingstrict quality assurance (QA) procedures The tolerable qual-ity limit for N concentration measurements is plusmn15 (forN concentration gt 5 mggminus1) of the mean inter-laboratory Nconcentration in foliar reference material distributed for ICPForests laboratory comparison tests (Rautio et al 2016)

The measurements and observation from ICP ForestsLevel II forest plots additionally included the beginning ofthe growing season for the sampling years 2015 and 2017(where available) (Vilhar et al 2013) main tree species onthe plot mean age of trees on the plot (estimated during sys-tem installment) basal area and trees per hectare on the plot(Dobbertin and Neumann 2016) soil texture of the uppersoil layer (mineral soil between 0ndash5 cm or 0ndash10 cm from thesurvey years 2003ndash2019) (Fleck et al 2016 Cools and DeVos 2020) altitude and geographic coordinates At the treelevel metadata consist of tree species tree number and di-ameter at breast height (Dobbertin and Neumann 2016) Me-teorological in situ measurements of hourly temperature andrelative humidity (Raspe et al 2013) were available for 82forest Level II plots for both 2015 and 2017

Furthermore we amended the dataset with satellite-basedvalues of transpiration from the Global Land EvaporationAmsterdam Model (GLEAM) (Miralles et al 2011 Martenset al 2017) and of hourly soil water (layer 1 0ndash7 cm)and surface air temperature (2 m height) from ERA5-Land(Muntildeoz Sabater 2019) for the respective regions of everyforest plot GLEAM (v 33a) data were available at daily

httpsdoiorg105194bg-19-1335-2022 Biogeosciences 19 1335ndash1353 2022

Figure 1 Overview of forest plots at which Hg foliage samples were harvested from different tree species groups during the sampling year2017 At around 12 of plots in 2017 foliage from more than one tree species group was sampled Geographic distribution of sampling sitesin 2015 is similar except there were no samples from the ICP Forests partners Brandenburg (Germany) Baden-Wuumlrttemberg (Germany) andPoland and there were samples from five additional plots in North Rhine-Westphalia (Germany) see Fig S1 in the Supplement The enlargedmap view at the top right depicts sampling locations of the Austrian Bio-Indicator Grid in 2015 and 2017 Use of base map authorized underEuropean Commission reuse policy (EU 2011)

resolution and on a 025 latitudendashlongitude regular gridERA5-Land values were available at hourly resolution andon a 01 latitudendashlongitude regular grid For each forest plotwe calculated average daily GLEAM (v33a) transpirationwithin the life period of foliage samples from the beginningof the growing season to harvest Similarly from ERA5-Land values we calculated the average 2 m air temperaturewithin respective sample life periods We detected outliersof time-normalized foliar Hg concentrations (see Sect 23)within each tree species and needle age class by applying themodified Z score method according to Iglewicz and Hoaglin

(1993) using an absolute threshold value of 35 above whicha modified Z score value was considered an outlier As a re-sult 32 of values within the dataset were removed as out-liers

22 Correction of foliar Hg concentrations for dryingtemperature

Drying and grinding of foliage samples were carried out byICP Forests laboratories and BFW All foliar concentrationvalues (Hg N and Corg) within the dataset are normalized todry weight for a sample drying temperature of 105 C in or-

der to make values internally consistent The actual dryingtemperature differed between foliage samples (40ndash80 C) Inorder to adjust for actual drying temperature the laborato-ries determined the drying factor to correct for water con-tent of each sample by drying an aliquot of foliage sample atactual drying temperature and subsequently at 105 C Thedrying factor was available for 62 of samples within thedataset For the rest of the samples an average drying factorper tree species and needle age class was applied for dry-ing temperature correction The smallest average drying fac-tor was 103plusmn 0003 (mean plusmn sd) for 1-year-old (y1) Pinuspinaster needles and the biggest average drying factor was107plusmn 002 (mean plusmn sd) for Quercus robur leaves Previousstudies did not detect Hg losses with drying temperature infoliage (Wohlgemuth et al 2020 Pleijel et al 2021) wood(Yang et al 2017) or moss (Lodenius et al 2003)

23 Foliage Hg analysis

Total Hg concentrations in foliage samples from ICP ForestsLevel II plots were measured at the University of Basel usinga direct mercury analyzer (Milestone DMA-80 HeerbruggSwitzerland) Standard operation procedure involved mea-suring a pre-sequence of four blanks (three empty sampleholders and wheat flour) and three liquid primary referencestandards (50 mg of 100 nggminus1 NIST-3133 in 1 BrCl)If the three liquid primary reference standards were within90 ndash110 of expected value we corrected all measure-ment results of the respective sequence accordingly Other-wise we discarded the sequence and re-calibrated the instru-ment Standard reference materials (SRMs) (NIST-1515 ap-ple leaves and spruce needle sample B from the 19th ICPForests needlendashleaf interlaboratory comparison test ILC)were measured in each sequence (four SRMs in a sequenceof 40 samples) and the sequence was discarded if the mea-sured SRM value was outside the certified uncertainty range(NIST-1515) or outside plusmn10 of the expected concentra-tion (ICP Forests spruce B) Absolute Hg content in wheatblanks within the sequence had to be lt 03 ng We success-fully participated in the 21st (20182019) 22nd (20192020)and 23rd (20202021) ICP Forests needlendashleaf ILC test To-tal Hg concentrations in foliage samples from the AustrianBio-Indicator Grid were measured using a Hg analyzer (Al-tec AMA 254 HCS Prague Czech Republic) Standard op-eration procedure at BFW involved a pre-sequence of fiveblanks (empty nickel boats) and measurements of three sam-ples of reference material (BCR-62 olive leaves or spruceneedle samples from the 17th or 19th ICP Forests needlendashleaf ILC test) after every 40th sample within a sequence Ifthe measurement results of the three reference samples wereoutside of 93 ndash107 of expected value a drift correctionwas performed Final foliage Hg concentrations within theAustrian Bio-Indicator Grid represent average values of atleast two replicates

24 Determination of the beginning of the growingseason for calculating daily foliage Hg uptake rates

Mercury concentrations in leaves and needles have beendemonstrated to increase linearly over the course of thegrowing season (Rea et al 2002 Laacouri et al 2013Blackwell et al 2014 Wohlgemuth et al 2020) In thisstudy foliage samples within the dataset were harvested atvarious points in time making a direct comparison of mea-sured Hg concentrations unfeasible We therefore calculatedfoliar Hg uptake rates (in ngHggminus1

dw dminus1) of current-seasonsamples by normalizing foliar Hg concentrations to their re-spective life period in days from the beginning of the growingseason (emergence of new foliage) to date of harvest Theseresulting foliar Hg uptake rates are net Hg accumulation ratesper gram dry weight on a leaf basis and should not be con-fused with foliar Hg fluxes on a whole-tree basis Please alsonote that daily foliar Hg uptake rates in this study representaverage values over the growing season The actual daily fo-liar Hg uptake on a given day might differ from the averagevalue depending on the time period within the growing sea-son (eg early season versus peak season) (Laacouri et al2013) Needles 1 year of age or older were excluded fromcalculating daily foliage Hg uptake fluxes since Hg uptakemight slow down in physiologically less active older needles(Wohlgemuth et al 2020) and it is unclear to what extentHg uptake occurs in older needles in winter and in earlyspring before the emergence of new foliage While datesof harvest were available for all samples we determinedthe start of the growing season of current-season foliage bycombining available data sources with start-of-season mod-eling These data sources comprise in situ phenological ob-servations which were available for 15 of samples andobservations of the emergence of current-season needles ofconiferous tree species from the Pan European Phenologicaldatabase PEP725 (Templ et al 2018) We assigned observa-tions from PEP725 to the corresponding closest forest plotof the respective sampling year (2015 or 2017) by using thenearest neighbor function matchpt from the Biobase pack-age in R (Huber et al 2015) such that differences betweenPEP725 observation and forest plots did not exceed 3 in lat-itude or 30 m in altitude but matched longitude as closely aspossible For details on the matching procedure and resultssee Sect S31 in the Supplement To model the beginning ofthe growing season for deciduous trees we utilized the leafarea index (LAI) product of Copernicus Global Land Servicebased on PROBA-V satellite imagery at a resolution of 300 mand 10 d (Dierckx et al 2014 Fuster et al 2020) followinga recommendation by Boacuternez et al (2020) For informationon the model and quality assurance refer to Sect S32 in theSupplement

25 Evaluation of data on water vapor pressure deficit(VPD)

At 82 ICP Forests Level II plots (in total from both sam-pling years 2015 and 2017) in situ meteorological data at anhourly resolution were recorded in 2015 and 2017 for whichwe calculated hourly water vapor pressure deficit (VPD) val-ues for daytime (0600ndash1800 LT) The VPD represents thedifference between the water vapor pressure at saturation andthe actual water vapor pressure We calculated saturated wa-ter vapor pressure from average hourly air temperature us-ing the AugustndashRochendashMagnus formula (Yuan et al 2019)and actual water vapor pressure as the saturation water va-por pressure multiplied by the average hourly relative humid-ity These VPD values were calculated exclusively for day-time hours (0600ndash1800 LT) because both Hg(0) and pho-tosynthetic CO2 uptake by trees are at maximum during theday (Obrist et al 2021) From these daytime hourly VPDvalues at each forest plot we calculated the proportion ofhours within the daytime life period of the samples (from thebeginning of the growing season to sampling day) duringwhich VPD exceeded the four threshold values of 12 162 and 3 kPa respectively We chose these four VPD thresh-olds as test values because they were reported in the litera-ture to incrementally induce leaf stomatal closure of temper-ate forest trees ranging from initial stomatal closure (around08ndash1 kPa Koumlrner 2013) to maximum stomatal closure (ataround 3ndash32 kPa CLRTAP 2017) We calculated the aver-age proportion of daily daytime exceedance hours of VPD gt

respective threshold value by normalizing the total numberof respective daytime VPD exceedance hours with the totalnumber of daytime hours during the corresponding samplelife period

26 Evaluation of ERA5-Land volumetric soil watercontents

We calculated the time proportion within sample life peri-ods during which the volumetric soil water content in theregion of the respective forest plots fell below a soil-texture-dependent threshold value (PAWcrit) in which plants are ex-pected to close their stomata due to limited water availabil-ity To this we used the satellite-derived ERA5-Land data ofhourly soil water in soil layer 1 (vertical resolution 0ndash7 cmhorizontal resolution 01times01) (Muntildeoz Sabater 2019) anddata on soil texture of the respective forest plots where avail-able (Fleck et al 2016) Field data from literature suggestthat plant stomata start to close once the plant available wa-ter (PAW) in the soil falls below a critical value (PAWcrit)(Domec et al 2009 Gruumlnhage et al 2011 2012) The soilPAW represents the difference between soil water at field ca-pacity (SWFC) and soil water at the permanent wilting point(SWPWP) We calculated PAWcrit = 05timesPAW+SWPWP fol-lowing a recommendation by Buumlker et al (2012) and usedPAWcrit as the threshold value to calculate the proportion of

hours within the respective sample life periods during whichsoil water lt PAWcrit See Fig S11 in the Supplement for anexemplary time series of ERA5 soil water in the region ofa forest plot in France in 2015 Soil-texture-specific valuesfor SWFC and SWPWP (Table S4 in the Supplement) wereobtained from Saxton and Rawls (2006)

3 Results and discussion

31 Variation in foliar Hg concentrations with foliar lifeperiod

Average foliar Hg concentrations (mean plusmn sd) differedbetween tree species groups (see Table S1 for definitionof tree species groups) Ash leaves exhibited the high-est Hg concentrations (322plusmn 57 ngHggminus1

dw n= 10) fol-lowed by beech leaves (255plusmn 96 ngHggminus1

dw n= 372)current-season Douglas fir needles (229plusmn 67 ngHggminus1

dwn= 27) hornbeam leaves (322plusmn 57 ngHggminus1

dw n= 10)oak leaves (208plusmn 91 ngHggminus1

dw n= 287) larch needles(134plusmn34 ngHggminus1

dw n= 3) current-season spruce needles(118plusmn34 ngHggminus1

dw n= 1509) current-season fir needles(114plusmn28 ngHggminus1

dw n= 66) and current-season pine nee-dles (110plusmn 51 ngHggminus1

dw n= 344) For all tree speciessampled at more than 20 forest plots we found significant(p lt 005) positive trends of foliar Hg concentrations withrespective sampling date within the growing season (seeFig 2 for beech and oak and Fig S4 in the Supplement forpine and spruce)

Increasing foliar Hg concentrations with progressing sam-pling date are in line with previous observations demon-strating that at individual sites Hg concentrations increasedlinearly over the growing season (Rea et al 2002 Laa-couri et al 2013 Wohlgemuth et al 2020 Pleijel et al2021) To make Hg levels in foliage sampled at differenttimes comparable we calculated daily foliar Hg uptake ratesby normalizing foliar Hg concentrations with the life pe-riod of samples These daily foliar Hg uptake rates repre-sent average values over the life period The average lifeperiod (mean plusmn sd) of samples was 104plusmn 30 d for beech104plusmn 24 d for oak 159plusmn 12 d for pine and 148plusmn 14 d forspruce At 5 of spruce plots sampling took place in win-ter (December until March) Spruce and pine trees have beenfound to reduce their physiological activity (transpiration netphotosynthesis) at low soil temperatures (lt 8ndash10 C) poten-tially impacting stomatal Hg(0) uptake in winter (Schwarzet al 1997 Mellander et al 2004) The average daily Hguptake rates of current-season spruce needles sampled dur-ing peak season (0084 ngHggminus1

dw dminus1) and sampled dur-ing winter (0067 ngHggminus1

dw dminus1) were significantly differ-ent (Welch two-sample t test p = 0015 at 95 confidencelevel) If spruce trees continue to accumulate Hg through-out the winter Hg needle concentrations should be higher in

Figure 2 Average leaf Hg concentrations (ngHggminus1dw) in beech and

oak samples at multiple ICP Forests plots versus sampling dates(day of year DOY) of respective samples Sampling took place bothin 2015 and 2017 Two plots of holm oak (Quercus ilex) are locatedin Greece and were sampled in December 2015 (DOY= 348) andDecember 2017 (DOY= 346) Error bars denote plusmn 1 standard de-viation between multiple samples at each forest plot

winter samples than in samples harvested earlier during thegrowing season and Hg uptake rates per day should be com-parable between winter and growing season samples Thusthe difference of average daily Hg uptake between winter andgrowing season spruce needle samples indicates a decrease inHg accumulation in spruce needles during winter Howeverthe potential of needle Hg uptake in winter needles requiresfurther investigation for example by performing a full win-ter sampling at multiple forest plots For this study we short-ened the calculated life period of spruce needles from wintersampling plots to 15 November (Roumltzer and Chmielewski2001) to improve comparability of spruce needle Hg uptakerates within the dataset

32 Variation in foliar Hg uptake rates with tree speciesgroups

Median daily foliar Hg uptake rates (Fig 3) in de-creasing order are ash (026 ngHggminus1

dw dminus1) beech(025 ngHggminus1

dw dminus1) oak (022 ngHggminus1dw dminus1) horn-

beam (020 ngHg gminus1dw dminus1) larch (014 ngHggminus1

dw dminus1)current-season Douglas fir needles (013 ngHggminus1

dw dminus1)current-season spruce needles (007 ngHggminus1

dw dminus1) current-

season fir needles (007 ngHggminus1dw dminus1) and current-season

pine needles (005 ngHggminus1dw dminus1) The range of daily

foliar Hg uptake of beech (012ndash042 ngHggminus1dw dminus1)

is in agreement with the daily foliar Hg uptake rate of035plusmn 003 ngHggminus1 dminus1 which Bushey et al (2008) haddetermined in beech leaves growing in New York State in2005 There are distinct differences in median daily Hguptake rates between current-season foliage of tree speciesgroups (Fig 3) The median daily foliar Hg uptake rate ofdeciduous leaf samples is 023 ngHggminus1

dw dminus1 a factor of32 larger than the median daily foliar Hg uptake rate ofcurrent-season conifer needle values (007 ngHggminus1

dw dminus1)The difference between deciduous and coniferous leaves inthe European dataset is smaller than a previous observationfrom a mixed forest site in Switzerland in 2018 where Hguptake rates of coniferous species were reported to be 5 timeslower than those of deciduous trees (Wohlgemuth et al2020) Similarly Navraacutetil et al (2016) reported higher foliarHg concentrations in beech leaves (363 ngHggminus1) thanin current-season spruce needles (141 ngHggminus1) of twoadjacent forest plots sampled during peak season (August)Higher Hg concentrations in deciduous leaves (median28 ngHggminus1 from 341 remote sites) than in compositemulti-age coniferous needles (median 15 ngHggminus1 from535 remote sites) were also reported in a global literaturecompilation (Zhou et al 2021) Differences in daily foliarHg uptake between tree species within one genus (egQuercus petraea and Quercus robur) were negligible (seeFig S5 in the Supplement) We were not able to normalizedaily foliar Hg uptake rates with atmospheric Hg(0) con-centrations at each respective sampling site and sample lifeperiod as air Hg(0) measurements were unavailable for oursampling sites The relative standard deviation of averageair Hg(0) concentrations at six European measurement siteswithin the EMEP network (Toslashrseth et al 2012 EMEP2021) between May and September 2015 and 2017 (seeTable S2 in the Supplement for details) was 006 whichis lower than the relative standard deviation of the averagedaily Hg uptake rates between tree species and forest plotsof 064 (Fig 3) We therefore argue that the pronounceddifferences in median daily foliar Hg uptake rates betweentree species cannot exclusively be explained by differencesin atmospheric Hg(0) concentrations but we rather suggesta tree physiological cause However foliar Hg uptakerates should be normalized to ambient atmospheric Hg(0)concentrations in particular when comparing foliar Hgobservation between the Northern Hemisphere and SouthernHemisphere or over multi-decadal timescales

33 Foliar Hg uptake and sample-specific Nconcentration

Foliar N concentration serves as a surrogate for the maximumphotosynthetic capacity of foliage (Reich et al 1998) as thebulk amount of foliar N is contained in the photosynthetic

Figure 3 Median daily foliar Hg uptake (ngHggminus1dw dminus1) between different tree species groups (see Table S1 for definition of tree species

groups) arranged from highest to lowest value Error bars give the value range within each tree species group and n indicates the number ofsites at which the respective tree species were sampled in both the years 2015 and 2017 Foliar samples of evergreen coniferous tree species(Douglas fir spruce fir and pine) consist of needles of the current season

systems like chlorophylls thylakoid proteins and RuBisCO(Evans 1989 Koumlrner 2013 Loomis 1997) Furthermorefoliar N represents an indirect proxy for foliar maximumstomatal diffusive conductance for water vapor independentof tree species (Koumlrner et al 1979 Bolton and Brown 1980Schulze et al 1994 Reich et al 1999 Meziane and Ship-ley 2001 Koumlrner 2013 2018) Note that for this analysis wesolely compared N and Hg concentrations for foliage sam-ples harvested within a period of the growing season dur-ing which leaf N concentrations are relatively stable (JulyndashAugust for broadleaves Wilson et al 2000 Mediavilla andEscudero 2003 and SeptemberndashMarch for conifer needlesAdams et al 1987 Hatcher 1990) To assess the possibil-ity of physiological factors controlling the large variationin foliar Hg(0) uptake between different tree species groups(Fig 3) we compared average daily foliar Hg uptake ratesper tree species group with respective average foliar N con-centrations We found a positive linear correlation betweenfoliar N concentration and Hg uptake rates as tree speciesgroups with high average foliar N exhibited higher daily fo-liar Hg uptake rates (Table 1) This observation supports thenotion that the physiological activity of trees controls foliarHg(0) uptake thereby explaining the large variation amongtree species groups (Wohlgemuth et al 2020) We comparedfoliar Hg uptake rates and leaf N concentrations with valuesof median stomatal conductance for beech oak pine andspruce included in a global leaf-level gas exchange databasecompiled by Lin et al (2015) (see description of databasecalculation in Sect S7 in the Supplement) Albeit stomatalconductance measurements for tree species of interest within

the database (Lin et al 2015) originated from one or only afew sites (n= 1ndash5 Table 1) beech and oak exhibited highermedian stomatal conductance values than spruce and pinecorresponding to higher daily Hg uptake rates and foliar Nconcentrations in beech and oak compared to spruce andpine Thus we observed a strong control of plant functionaltraits on foliar Hg(0) uptake with tree species of high pho-tosynthetic activity (high N concentration) and stomatal con-ductance exhibiting the highest foliar Hg(0) uptake rates

Within tree species groups linear regression coefficientsof daily Hg uptake and foliar N concentration were signif-icant (p lt 0001) for beech (R2

= 015 n= 312) and fir(R2= 027 n= 66) Corresponding statistical significance

for hornbeam oak pine and spruce could not be evaluatedsince the respective data used for the linear regression washeteroscedastic Blackwell and Driscoll (2015) found a sig-nificant relationship between foliar Hg concentration and fo-liar N percentage for yellow birch sugar maple and Amer-ican beech but not for pine (red pine and white pine) redspruce or balsam fir We examined whether unaccountedsite-specific differences (eg soil N concentration) betweenforest plots could have caused the variability (low R2) indaily Hg uptake versus foliar N concentration within treespecies by individually analyzing foliar Hg concentrationversus foliar N concentration at two oak and one beech forestplot from which 20 or more foliage samples were availableLinear regression coefficients of foliar Hg concentrationsversus foliar N concentrations were significant (p lt 0001)at two (oak and beech) of the three plots but not at the thirdplot (p = 01 oak) (see Fig S7 in the Supplement) This

Table 1 Mean plusmn standard deviation of daily Hg uptake and foliar N concentration per tree species group from a subset of foliage samplesharvested during JulyndashAugust (broadleaf samples) or SeptemberndashMarch (coniferous needle samples) Values are ordered from highest tolowest mean daily Hg uptake All values from evergreen tree species groups (Douglas fir fir pine spruce) were evaluated in current-seasonneedles Median stomatal conductance values (minndashmax) were calculated from a global database of leaf-level gas exchange parameterscompiled by Lin et al (2015)

Tree species Daily Hg uptake Foliar N conc n samples Median stomatal conductance n sitesgroup

(ngHggminus1

dw dminus1) (mgNgminus1

(molmminus2 sminus1) (Lin et al 2015) (Lin et al 2015)

Beech 025plusmn 005 231plusmn 29 312 010 (003ndash031) 2Oak 020plusmn 005 251plusmn 28 252 015 (001ndash035) 1Hornbeam 019plusmn 003 194plusmn 21 10Douglas fir 013plusmn 002 170plusmn 35 26Spruce 008plusmn 002 129plusmn 17 1509 005 (001ndash016) 1Fir 007plusmn 002 130plusmn 17 66Pine 006plusmn 002 144plusmn 30 355 006 (000ndash033) 5

finding suggests that foliar N concentrations represent an in-dicator of foliar Hg concentrations at individual forest sitesas it does for foliar Hg uptake of different tree species (Ta-ble 1) However given the heterogeneity of nutrient avail-ability between sites (Vesterdal et al 2008) and the com-plexity of internal foliar allocation of N to different parts ofthe photosynthetic apparatus (Hikosaka 2004) a generallyvalid correlation of foliar Hg uptake versus foliar N may notexist

34 Foliar Hg uptake and leaf mass per area

Within the whole dataset leaf mass per area (LMAgdw mminus2

leaf) data were available in a subset of 349 foliage sam-ples from 48 sites (from both 2015 and 2017) LMA is animportant parameter in plant ecophysiology because carbongains of plants via photosynthetic activity and gas diffusionare optimized per unit of leaf area as plants adapt their LMAie their foliage thickness andor tissue density to the avail-ability of sunlight during growth (Ellsworth and Reich 1993Niinemets and Tenhunen 1997 Rosati et al 1999) ThisLMA adaptation of foliage to sunlight had been suggestedto be more effective for optimizing photosynthetic capacitythan within-leaf N partitioning of photosynthesizing biomass(Evans and Poorter 2001) Therefore we analyzed the con-nection of foliar Hg uptake to LMA across tree species Fig-ure 4 shows average LMA values (mean plusmn sd) of the sub-set of samples in which LMA was reported resolved by treespecies along with respective average daily Hg uptake ratesand associated foliar N concentrations (all values displayedin Fig 4 are listed in Table S3 in the Supplement see Fig S8in the Supplement for density plots of datasets from Table 1and Fig 4)

Current-season needle samples of coniferous tree speciesgroups (Douglas fir pine spruce) exhibited higher medianLMA values (308 gdw mminus2

leaf) lower median daily Hg uptakerates (010 ngHggminus1

dw dminus1) and lower median foliar N con-centrations (154 mgNgminus1

dw) compared to leaf samples of de-

ciduous tree species groups (beech oak hornbeam) (Fig 4)Wright et al (2004) illustrated that different evolutionary sur-vival strategies of plant species are positioned along a singleaxis of foliage characteristics ranging from plant species withhigh photosynthetic capacity and respiration high foliar Nconcentration low LMA and short leaf life spans to plantspecies with the respective opposite attributes Comparisonof average daily foliar Hg uptake LMA and foliar N concen-trations (Fig 4) across tree species in this study suggests thatfoliar Hg(0) uptake aligns along this plant species economicsspectrum with deciduous leaves with high leaf N concentra-tions and thus high physiological capacity (photosynthesisrespiration) taking up more Hg(0) per gram dry weight overthe same time span than coniferous needles with low leaf Nconcentrations and physiological capacity

35 Foliar Hg uptake and water vapor pressure deficit(VPD)

Trees regulate their transpiration rates in response to tempo-rary changes in water vapor pressure deficit (VPD) by con-trolling leaf stomatal aperture (Franks and Farquhar 1999McAdam and Brodribb 2015 Grossiord et al 2020) Whena critical VPD threshold is exceeded trees close their stom-ata to resist cavitation and excessive water loss in conditionsof high atmospheric evaporative forcing (ie high VPD)(Koumlrner 2013 Grossiord et al 2020) This decrease in leafstomatal conductivity in response to high VPD suppressesstomatal uptake fluxes of gaseous pollutants like ozone (Em-berson et al 2000 Koumlrner 2013) We investigated whetherVPD impacts the foliar uptake of gaseous Hg(0) by relatingspecies-specific average daily foliar Hg uptake rates to theproportion of daytime (0600ndash1800 LT) hours of an averageday within the respective sample life periods during whichhourly daytime VPD exceeded the threshold values of 1216 2 and 3 kPa at all forest plots with hourly meteorologi-cal data (n= 82 including both sampling years)

Figure 4 (a) Average daily Hg uptake rates (ngHggminus1dw dminus1) (b) average foliar nitrogen concentrations (mgNgminus1

dw) and (c) average

LMA (gdw mminus2leaf) determined in 349 foliage samples and resolved by tree species group and foliage type (leafneedle) Error bars denote

plusmn 1 standard deviation Number of samples (n) differs between tree species beech (n= 164) Douglas fir (n= 2) hornbeam (n= 9) oak(n= 106) pine (n= 35) and spruce (n= 33)

Figure 5 Average daily Hg uptake rates (ngHggminus1dw dminus1) of

current-season pine needles from multiple forest plots (n plots=19) versus the proportion of daytime hours (0600ndash1800 LT) withinan average day of the respective sample life periods during whichthe hourly daytime water vapor pressure deficit (VPD) exceeded athreshold value of 12 kPa Data points originate from both samplingyears 2015 and 2017 All forest plots are located in Central Europe(latitude 46ndash54) for which ambient air Hg(0) concentrations arerelatively constant (see Table S2 and Fig S6 in the Supplement)Error bars denote plusmn 1 standard deviation of daily needle Hg uptakerates between multiple samples at each forest plot

The linear regression coefficients of average daily Hg up-take versus daily proportion of daytime hours during whichVPD exceeded a threshold value (12 16 2 or 3 kPa) weresignificant (p lt 001) for pine at all VPD threshold val-ues (Figs 5 and S9 in the Supplement) and for spruce at a

VPD threshold value of 3 kPa (R2= 044 p = 001 n= 14)

(Fig S9) and they were not significant for beech and oak atany VPD threshold value (Fig S10 in the Supplement) Lin-ear regression coefficients were negative for all species andVPD threshold values ie there is a tendency that averagedaily foliar Hg uptake rates decreased with the average pro-portion of daytime hours during which VPD gt respectivethreshold value (12 16 2 or 3 kPa) We excluded Douglasfir fir hornbeam and larch from the regression analysis dueto a low number of forest plots (n= 1ndash5) Average daily nee-dle Hg uptake rates of spruce needles were clustered betweentwo groups of forest plots with high and low daytime propor-tions of VPD gt threshold (Fig S9) relative to each other Thet test revealed a significant (p = 0008) difference in averagedaily spruce needle Hg uptake rates between the two clus-ters for a VPD threshold value of 3 kPa and non-significant(p gt 005) differences for all other VPD threshold valuesThe timing and degree of stomatal closure during dry con-ditions is specific to tree species (Zweifel et al 2009 Tsujiet al 2020) Tree species like pine and spruce are isohydricie they tend to respond to drought stress under high evapo-rative demand by closing their stomata earlier than anisohy-dric species like beech and oak (Martiacutenez-Ferri et al 2000Zweifel et al 2007 Carnicer et al 2013 Coll et al 2013Caacutercer et al 2018) Among isohydric species pine has beendiscovered to reduce tree conductance and stomatal apertureduring the onset of dry conditions earlier and at a greaterrate than spruce (Lagergren and Lindroth 2002 Zweifel etal 2009) Spruce has been observed to keep stomata almostcompletely closed under drought stress ie high VPD andorsoil water deficit (Zweifel et al 2009) We hypothesize thatthe significantly decreasing average foliar stomatal Hg up-

Figure 6 Average daily Hg uptake rates (ngHggminus1dw dminus1) of beech

oak and current-season pine foliage from multiple forest plots(beech plots n= 38 oak plots n= 45 pine plots n= 19 latitude41ndash55) versus the proportion of hours within the respective samplelife periods during which the geographically associated hourly soilwater from the ERA5-Land dataset (Muntildeoz Sabater 2019) fell be-low a soil-texture-specific threshold value PAWcrit (see Sect 26)Data points originate from both sampling years 2015 and 2017 Er-ror bars denote plusmn 1 standard deviation of daily needle Hg uptakerates between multiple samples at each forest plot

take rates with daytime proportion of VPD gt 12 kPa for pine(Fig 5) and of VPD gt 3 kPa for spruce (Fig S9) possibly re-flect the early physiological response of pine and the high de-gree of stomatal closure under drought stress of spruce Oakexhibits later stomatal closure at the onset of dry conditionsand higher stomatal aperture under drought stress than forexample pine (Zweifel et al 2007 2009) which may be thereason why there was a tendency for a negative but not signif-icant correlation coefficient of average foliar Hg uptake witha daytime proportion of VPD gt any threshold value for oak(Fig S10)

36 Foliar Hg uptake and soil water content

Linear regression coefficients of average daily foliar Hg up-take rates at each forest plot versus proportion of hourswithin sample life periods during which ERA5-Land soil wa-ter fell below a soil-texture-specific threshold value (PAWcrit)(see Sect 26) were negative for all tree species groupsand significant for beech (p = 0036) and pine (p = 0031)(Fig 6) The linear regression coefficient was not significantfor oak (p = 0169) and not available for spruce due to a lownumber of data points

Linear regression results (Fig 6) indicate that foliar Hguptake rates decrease at forest plots where plant availablewater in the upper soil layer (0ndash7 cm) falls below specificthresholds (PAWcrit) for a relatively long time period over thegrowing season Studies on the atmospherendashplant transportof ozone have highlighted that plant stomatal ozone uptakedeclines with increasing soil water deficit because droughtprompts stomatal closure (Panek and Goldstein 2001 Simp-son et al 2003 Nunn et al 2005) We hypothesize thatstomatal uptake of Hg(0) is impacted by soil conditions oflow plant available water in a similar way to ozone In thefuture in situ soil matrix potential measurements should beused to better quantify the response rate of foliar Hg(0) up-take to soil water content in order to overcome the limitationsof the coarse satellite-derived soil water measurements usedhere We also suggest determining the possible influence ofadditional parameters like gravel content and density of soilsrooting depth of trees and atmospheric Hg(0) which couldvary within the range of latitude (41ndash55) of examined forestplots

37 Foliar Hg uptake and geographic and tree-specificparameters

We performed linear regressions of average daily foliar Hguptake rates per forest plot and tree species group (beechoak pine spruce) versus geographic and tree-specific pa-rameters These parameters include altitude latitude aver-age age of trees on plot average tree diameter at breastheight average daily GLEAM transpiration values and av-erage ERA5-Land 2 m air temperature over the course of therespective sample life periods (see Sect 21) None of theresulting 54 linear regression coefficients were significantgiven a Bonferroni adjusted p value= 0000925 The differ-ences between 2015 and 2017 species-specific averages ofdaily foliar Hg uptake rates from forest plots at which fo-liage sampling took place during both sampling years weresmall compared to the standard deviation of daily foliar Hguptake rates within each sampling year and species (see Ta-ble S5 in the Supplement for average and standard devia-tion values) From the sampling year 2015 to the samplingyear 2017 this difference was 004 ngHggminus1

dw dminus1 for beech2times10minus4 ngHggminus1

dw dminus1 for oak 8times10minus5 ngHggminus1dw dminus1 for

pine (current-season needles) andminus3times10minus3 ngHggminus1dw dminus1

for spruce (current-season needles) We therefore suggestthat differences in daily foliar Hg uptake rates between thesampling years 2015 and 2017 are negligible In agreementwith previous studies (Ollerova et al 2010 Hutnik et al2014 Navraacutetil et al 2019 Wohlgemuth et al 2020 Plei-jel et al 2021) we found a trend of Hg concentrations indifferently aged spruce needles with older needles exhibit-ing higher Hg concentrations (Fig 7) demonstrating that Hgaccumulation continues in older needles Annual Hg net ac-cumulation seems to slow down in older spruce needles ofage classes y3ndashy6 in contrast to needles of age classes y0ndash

Figure 7 Average Hg concentrations (ngHggminus1dw) in spruce nee-

dle samples of different ages Needle age class y0 corresponds tocurrent-season needles flushed in the year of sampling y1 corre-sponds to 1-year-old needles y2 corresponds to 2-year-old needlesetc Error bars denoteplusmn 1 standard deviation between multiple sam-ples n indicates number of samples

y2 (Fig 7) albeit ranges of average Hg concentrations plusmnstandard deviation overlap among older and younger spruceneedles which might be the result of relatively low samplenumbers of older needles compared to younger needles (eg3 samples for y6 versus 301 samples for y1) A decline infoliar Hg uptake by older needles could be caused by lowerphysiological activity cuticular wax degradation or an in-crease in Hg re-emission with needle age (Wohlgemuth etal 2020)

38 Implications for Hg cycle modeling

Our findings suggest that VPD impacts stomatal Hg(0) up-take by isohydric tree species due to stomatal closure duringconditions of high VPD (Fig 5) Similarly elongated timeperiods of low soil water content within the growing sea-son possibly result in a decrease in stomatal conductance toHg(0) and thus in less foliar Hg(0) uptake by tree speciessuch as beech and pine (Fig 6) Other meteorological param-eters such as temperature may also have an effect on stomatalclosure and consequently stomatal Hg(0) uptake (Sect 37)We therefore propose to refine existing stomatal uptake mod-els for the purpose of exploring the stomatal uptake flux ofHg(0) for common vegetation types across different globalregions over the course of the growing season For this thesensitivity of species-specific foliar Hg uptake normalizedto air Hg(0) concentrations has to be determined in labo-ratory experiments with regards to elevated VPD low soilwater content or temperature Eventually the effect of treespecies VPD soil water and point in time within the grow-ing season could be implemented in a stomatal Hg depo-

sition model We propose that the stomatal flux module ofthe DO3SE (Deposition of Ozone for Stomatal Exchange)model could serve as a prototype for a stomatal Hg deposi-tion model because DO3SE provides estimates of stomatalozone deposition based on plant phenological and meteoro-logical conditions (Emberson et al 2000 2018) Projectionsfrom stomatal Hg models are particularly relevant for theevaluation of future global environmental Hg cycling as thestomatal Hg(0) uptake flux exceeds direct Hg(II) wet deposi-tion (Wohlgemuth et al 2020 Obrist et al 2021) and quan-titatively represents the most relevant deposition pathways toland surfaces driving the seasonality of Hg(0) in the atmo-sphere (Obrist 2007 Jiskra et al 2018) VPD is projectedto increase with rising temperatures under global warming(Yuan et al 2019 Grossiord et al 2020) potentially causinga decrease in stomatal foliar Hg(0) uptake fluxes A dimin-ished global stomatal foliar Hg(0) uptake flux would resultin higher Hg(0) concentrations in the atmosphere and higherHg deposition fluxes to the ocean (Zhou et al 2021)

4 Conclusions

We created a large European forest dataset for investigatingthe control of tree physiology and climatic conditions on fo-liar stomatal Hg(0) uptake We observed that foliar Hg con-centrations were highly correlated with foliage sampling date(Fig 2) confirming the notion that foliage takes up Hg(0)over the entire growing season and over multiple growingseasons in the case of coniferous needles (Fig 7) Conse-quently it is necessary to calculate foliar Hg uptake rates bynormalizing foliar Hg concentrations by the time period ofHg(0) accumulation to make foliar Hg values from differentsites comparable For reasons of comparability foliar Hg up-take rates should ideally be normalized to ambient air Hg(0)concentrations when large variation in atmospheric Hg(0) isexpected (eg between Northern Hemisphere and SouthernHemisphere in polluted regions or over long timescales) Wefound notable differences in daily foliar Hg uptake rates be-tween tree functional groups (broadleaves versus coniferousneedles) ie Hg uptake rates of broadleaves were highercompared to coniferous needles of the same age by a fac-tor of 32 (Fig 3) Across tree species and within beechand fir the linear regression coefficients of daily foliar Hguptake rates versus foliar N concentration were significant(Sect 33) Tree species groups with foliage of lower LMAexhibited higher daily rates of Hg uptake per dry weight offoliage (Sect 34) We set these results within the contextof stomatal foliar uptake of atmospheric Hg(0) deciduoustree species like beech and oak which exhibit functionaltraits of high physiological activity (photosynthesis transpi-ration) over the time span of one growing season as repre-sented by high foliar N concentration and low LMA retaina higher stomatal conductance for diffusive gas exchangeThus beech and oak leaves accumulate more Hg per unit

dry weight over the same time span relative to needles ofconiferous tree species In addition to tree-species-specificmetabolism climatic conditions like current VPD or soil wa-ter content which impacts stomatal gas exchange can affectfoliar Hg uptake For current-season pine needles we founda significant negative linear regression coefficient of dailyHg uptake rates versus the average daily proportion of hourswithin sample life period during which atmospheric evapo-rative forcing was high (VPD gt 12 kPa) (Fig 5) suggestingthat a reduction in stomatal conductance during conditionsof high VPD suppresses foliar Hg(0) uptake In a similar lineof argument low surface soil water content lowers stomatalconductance and consequently foliar stomatal Hg(0) uptake(Fig 6) We therefore suggest that foliar Hg measurementsbear the potential to serve as a proxy for stomatal conduc-tance providing a time-integrated measure for stomatal aper-ture when taking into account the spatial and temporal varia-tion in atmospheric Hg(0) We call for the implementation ofa stomatal Hg(0) deposition model that takes tree physiologyand environmental conditions like VPD or soil water contentinto account in order to make projections about this impor-tant Hg deposition flux under climate change The diminu-tion of the vegetation mercury pump in response to droughtstress as a result of climate change could result in elevatedHg concentrations in the ocean and potentially in marine fishin future a potential risk which warrants further quantitativestudies

Data availability Foliar Hg concentrations foliar Hg uptakerates and Hg-related metadata are available for download athttpsdoiorg105281zenodo5495179 Please note that coordi-nates (latitude longitude) were rounded to minutes R scripts fordata analysis and plots of this paper can be found at httpsgithubcomwohleHg_Forests (Wohlgemuth 2022)

ICP Forests proprietary data (N concentrations and forest plotattributes) fall under the publication policy of ICP Forests (An-nex II of Seidling et al 2017) and can be accessed from theICP Forests database (httpicp-forestsnetpagedata-requests ICPForests 2022) upon request from the Programme Co-ordinatingCenter (PCC) in Eberswalde Germany Foliar Hg concentrationvalues from the Austrian Bio-Indicator Grid can be obtained fromBFW upon request (httpsbfwacatrzbfwcmswebdok=3687BFW 2022) ERA5-Land data (Muntildeoz Sabater 2019) were down-loaded from the Copernicus Climate Change Service (C3S) ClimateData Store (httpscdsclimatecopernicuseuhome Copernicus2022a) Data for the beginning of the growing season of coniferoustrees in 2015 and 2017 were provided by members of the PEP725project (httpwwwpep725eu PEP725 2022) PROBA-V leafarea index values of 300 m resolution and GLEAM transpirationvalues from 2015 and 2017 were obtained from the VITO ProductDistribution Portal (httpslandcopernicuseuglobalproductslaiCopernicus 2022b) and the GLEAM server (httpswwwgleameu GLEAM 2022) respectively

Supplement The supplement related to this article is available on-line at httpsdoiorg105194bg-19-1335-2022-supplement

Author contributions LW managed the project coordinated foliarHg measurements assembled the Hg dataset performed the dataanalysis and wrote the manuscript PR BA AR LV PW VT NEMG PR AT MN AK MI PM SB DZ and CI supplied foliagesamples and metadata and gave scientific input to the manuscriptAF contributed foliar Hg concentrations from the Austrian Bio-Indicator Grid and gave scientific input to the manuscript GH andCA provided valuable scientific support MJ designed and set up theSNSF project (174101) provided valuable scientific support andcontributed to manuscript writing

Competing interests The contact author has declared that neitherthey nor their co-authors have any competing interests

Disclaimer Publisherrsquos note Copernicus Publications remainsneutral with regard to jurisdictional claims in published maps andinstitutional affiliations

Acknowledgements We thank Fabienne Bracher and Ju-dith Kobler Waldis for assistance in foliage sample analy-sis The evaluation was based on data that were collectedby partners of the official UNECE ICP Forests network(httpicp-forestsnetcontributors last access 16 February2022) We are grateful to all ICP Forests participants who sup-ported the project through foliage sampling nutrient analysisand cooperation in the logistics of this project In this contextwe particularly thank Martin Maier Andrea Houmllscher and theirteam from the Department of Soil and Environment at FVABaden-Wuumlrttemberg Daniel Žlindra from the Slovenian ForestryInstitute Nils Koumlnig from Northwest German Forest ResearchInstitute (NW-FVA) Hans-Peter Dietrich and Stephan Raspefrom the Bavarian State Institute of Forestry (LWF Bayern)Michael Tatzber from the Austrian Research Centre for Forests(BFW) Arne Verstraeten and Luc De Geest from the BelgianResearch Institute for Nature and Forest (INBO) Seacutebastien Maceacutefrom the French National Forest Office (ONF) and PanagiotisMichopoulos from the Forest Research Institute of Athens (FRIA)We are grateful to Samantha Wittke and Christian Koumlrner for theirhelpful advice and support on leaf area indices and plant phenologySpecial thanks go to Till Kirchner and Anne-Katrin Prescher fromThuumlnen Institute for their assistance in accessing the ICP ForestsDatabase

Financial support This research was funded by the Swiss NationalScience Foundation (SNSF (grant no 174101)) The participatingcountries from the UNECE ICP Forests Network funded the sam-pling using national funding among funding agencies are the Nat-ural Resources Institute Finland (Luke) Swiss Federal Institute forForest Snow and Landscape Research (WSL) Norwegian Instituteof Bioeconomy Research (NIBIO) Norwegian Ministry of Agricul-ture and Food (LMD) Polish Forest Research Institute (IBL) Pol-

ish Ministry of Environment (grants no 650412-650415) North-west German Forest Research Institute (NW-FVA) French Na-tional Forest Office (ONF) French Ministry of Agriculture FrenchAgency for Environment and Energy (ADEME) Slovenian Min-istry of Agriculture Forestry and Food (Public Forest Service As-signment 13) and National Institute for Research and Develop-ment in Forestry ldquoMarin Draceardquo Romania (INCDS) Part of thedata were co-financed by the European Commission The AustrianBio-Indicator Grid is operated by BFW with funding from the Aus-trian Ministry of Agriculture Regions and Tourism

Review statement This paper was edited by Anja Rammig and re-viewed by Haringkan Pleijel Frank Wania and Charles T Driscoll

References

Adams M B Campbell R G Allen H L and DaveyC B Root and foliar nutrient concentrations in loblollypine effects of season site and fertilization Forest Sci33 984ndash996 httpsacademicoupcomforestsciencearticle3349844641975login=true (last access 16 February 2022)1987

AMAP and UNEP Technical background report to the globalmercury assessment 2018 Arctic Monitoring and Assess-ment Programme Oslo NorwayUN Environment Pro-gramme Chemicals and Health Branch Geneva Switzerlandhttpswwwamapnodocuments (last access 16 February2022) 2019

Austrian Bio-Indicator Grid httpsbfwacatrzbfwcms2webdok=3687 (last access 16 February 2022) 2016

BFW Official Homepage to Austrian Bio-Indicator Grid httpsbfwacatrzbfwcmswebdok=3687 last access 16 February2022

Bishop K Shanley J B Riscassi A de Wit H A Ek-loumlf K Meng B Mitchell C Osterwalder S SchusterP F Webster J and Zhu W Recent advances in un-derstanding and measurement of mercury in the environ-ment Terrestrial Hg cycling Sci Total Environ 721 137647httpsdoiorg101016jscitotenv2020137647 2020

Blackwell B D and Driscoll C T Deposition of mercury inforests along a montane elevation gradient Environ Sci Tech-nol 49 5363ndash5370 httpsdoiorg101021es505928w 2015

Blackwell B D Driscoll C T Maxwell J A and Holsen TM Changing climate alters inputs and pathways of mercury de-position to forested ecosystems Biogeochemistry 119 215ndash228httpsdoiorg101007s10533-014-9961-6 2014

Bolton J K and Brown R H Photosynthesis of grass speciesdiffering in carbon dioxide fixation pathways V Response ofPanicum maximum Panicum milioides and tall fescue (Festucaarundinacea) to nitrogen nutrition Plant Physiol 66 97ndash100httpsdoiorg101104pp66197 1980

Boacuternez K Descals A Verger A and Pentildeuelas J Land sur-face phenology from VEGETATION and PROBA-V data As-sessment over deciduous forests Int J Appl Earth Obs 84101974 httpsdoiorg101016jjag2019101974 2020

Buumlker P Morrissey T Briolat A Falk R Simpson D Tuovi-nen J-P Alonso R Barth S Baumgarten M Grulke N

Karlsson P E King J Lagergren F Matyssek R NunnA Ogaya R Pentildeuelas J Rhea L Schaub M UddlingJ Werner W and Emberson L D DO3SE modelling ofsoil moisture to determine ozone flux to forest trees AtmosChem Phys 12 5537ndash5562 httpsdoiorg105194acp-12-5537-2012 2012

Bushey J T Nallana A G Montesdeoca M Rand Driscoll C T Mercury dynamics of a north-ern hardwood canopy Atmos Environ 42 6905ndash6914httpsdoiorg101016jatmosenv200805043 2008

Caacutercer P S de Vitasse Y Pentildeuelas J Jassey V E J Buttler Aand Signarbieux C Vaporndashpressure deficit and extreme climaticvariables limit tree growth Glob Change Biol 24 1108ndash1122httpsdoiorg101111gcb13973 2018

Carnicer J Barbeta A Sperlich D Coll M and PenuelasJ Contrasting trait syndromes in angiosperms and conifersare associated with different responses of tree growth totemperature on a large scale Front Plant Sci 4 409httpsdoiorg103389fpls201300409 2013

CLRTAP Revised Chapter 3 of the Manual on Methodologies andCriteria for Modelling and Mapping Critical Loads and Lev-els and Air Pollution Effects Risks and Trends Mapping Crit-ical Levels for Vegetation httpswwwumweltbundesamtdeenmanual-for-modelling-mapping-critical-loads-levels (last ac-cess 16 February 2022) 2017

Coll M Pentildeuelas J Ninyerola M Pons X and CarnicerJ Multivariate effect gradients driving forest demographic re-sponses in the Iberian Peninsula Forest Ecol Manag 303 195ndash209 httpsdoiorg101016jforeco201304010 2013

Cools N and De Vos B Part X Sampling and analysis of soilin Manual on methods and criteria for harmonized sampling as-sessment monitoring and analysis of the effects of air pollutionon forests edited by UNECE ICP Forests Programme Coor-dinating Centre Thuumlnen Institute of Forest Ecosystems Eber-swalde Germany httpicp-forestsnetpageicp-forests-manual(last access 16 February 2022) 2020

Copernicus Official Homepage of the Copernicus Climate ChangeService Climate Data Store httpscdsclimatecopernicuseuhome last access 16 February 2022a

Copernicus Official Homepage of the Copernicus Global Land Ser-vice httpslandcopernicuseuglobalproductslai last access16 February 2022b

Demers J D Blum J D and Zak D R Mercury isotopes ina forested ecosystem Implications for air-surface exchange dy-namics and the global mercury cycle Global Biochem Cy 27222ndash238 httpsdoiorg101002gbc20021 2013

Dierckx W Sterckx S Benhadj I Livens S Duhoux G VanAchteren T Francois M Mellab K and Saint G PROBA-V mission for global vegetation monitoring standard prod-ucts and image quality Int J Remote Sens 35 2589ndash2614httpsdoiorg101080014311612014883097 2014

Dobbertin M and Neumann M Part V Tree Growth in Man-ual on methods and criteria for harmonized sampling assess-ment monitoring and analysis of the effects of air pollution onforests edited by UNECE ICP Forests Programme Coordinat-ing Centre Thuumlnen Institute of Forest Ecosystems EberswaldeGermany httpicp-forestsnetpageicp-forests-manual (last ac-cess 16 February 2022) 2016

Domec J-C Noormets A King J S Sun G Mcnulty S GGavazzi M J Boggs J L and Treasure E A Decoupling theinfluence of leaf and root hydraulic conductances on stomatalconductance and its sensitivity to vapour pressure deficit as soildries in a drained loblolly pine plantation Plant Cell Environ32 980ndash991 httpsdoiorg101111j1365-3040200901981x2009

Drenner R W Chumchal M M Jones C M Lehmann C MB Gay D A and Donato D I Effects of mercury deposi-tion and coniferous forests on the mercury contamination of fishin the South Central United States Environ Sci Technol 471274ndash1279 httpsdoiorg101021es303734n 2013

Driscoll C T Mason R P Chan H M Jacob D J andPirrone N Mercury as a global pollutant sources path-ways and effects Environ Sci Technol 47 4967ndash4983httpsdoiorg101021es305071v 2013

Ellsworth D S and Reich P B Canopy structure andvertical patterns of photosynthesis and related leaftraits in a deciduous forest Oecologia 96 169ndash178httpsdoiorg101007BF00317729 1993

Emberson L D Ashmore M R Cambridge H M and Simp-son D Modelling stomatal ozone flux across Europe EnvironPollut 11 403ndash413 2000

Emberson L D Pleijel H Ainsworth E A van den Berg MRen W Osborne S Mills G Pandey D Dentener F BuumlkerP Ewert F Koeble R and Van Dingenen R Ozone effectson crops and consideration in crop models Eur J Agron 10019ndash34 httpsdoiorg101016jeja201806002 2018

EMEP Present state of Hg(0) data httpebas-datanilunodefaultaspx (last access 16 February 2022) 2021

Enrico M Roux G L Marusczak N Heimbuumlrger L-E Claus-tres A Fu X Sun R and Sonke J E Atmospheric mer-cury transfer to peat bogs dominated by gaseous elemental mer-cury dry deposition Environ Sci Technol 50 2405ndash2412httpsdoiorg101021acsest5b06058 2016

EU Commission Decision of 12 December 2011 on the reuse ofCommission documents European Commission 2011833EU2011

Evans J R Photosynthesis and nitrogen relation-ships in leaves of C3 plants Oecologia 78 9ndash19httpsdoiorg101007BF00377192 1989

Evans J R and Poorter H Photosynthetic acclimation of plantsto growth irradiance the relative importance of specific leafarea and nitrogen partitioning in maximizing carbon gainPlant Cell Environ 24 755ndash767 httpsdoiorg101046j1365-3040200100724x 2001

Fleck S Cools N De Vos B Meesenburg H and Fischer RThe Level II aggregated forest soil condition database links soilphysicochemical and hydraulic properties with long-term obser-vations of forest condition in Europe Ann For Sci 73 945ndash957 httpsdoiorg101007s13595-016-0571-4 2016

Franks P J and Farquhar G D A relationship betweenhumidity response growth form and photosynthetic operat-ing point in C3 plants Plant Cell Environ 22 1337ndash1349httpsdoiorg101046j1365-3040199900494x 1999

Fuster B Saacutenchez-Zapero J Camacho F Garciacutea-Santos VVerger A Lacaze R Weiss M Baret F and Smets B Qual-ity assessment of PROBA-V LAI fAPAR and fCOVER collec-

tion 300 m products of Copernicus Global Land Service RemoteSens 12 1017 httpsdoiorg103390rs12061017 2020

GLEAM Official Homepage to the Global Land Evaporation Am-sterdam Model httpswwwgleameu last access 16 February2022

Graydon J A St Louis V L Hintelmann H LindbergS E Sandilands K A Rudd J W M Kelly C AHall B D and Mowat L D Long-term wet and dry de-position of total and methyl mercury in the remote borealecoregion of Canada Environ Sci Technol 42 8345ndash8351httpsdoiorg101021es801056j 2008

Grigal D F Mercury sequestration in forests and peatlands J Env-iron Qual 32 393 httpsdoiorg102134jeq20033930 2003

Grossiord C Buckley T N Cernusak L A Novick K A Poul-ter B Siegwolf R T W Sperry J S and McDowell N GPlant responses to rising vapor pressure deficit New Phytol226 1550ndash1566 httpsdoiorg101111nph16485 2020

Gruumlnhage L Braden H Bender J Burkart S Lehmann Y andSchroumlder M Evaluation of the ozone-related risk for winterwheat at local scale with the CRO3PS model Gefahrstoffendash Reinhaltung der Luft 71 9 httpswwwresearchgatenetpublication278672267_Evaluation_of_the_ozone-related_risk_for_winter_wheat_at_local_scale_with_the_CRO3PS_model(last access 16 February 2022) 2011

Gruumlnhage L Pleijel H Mills G Bender J Daniels-son H Lehmann Y Castell J-F and Bethenod OUpdated stomatal flux and flux-effect models for wheatfor quantifying effects of ozone on grain yield grainmass and protein yield Environ Pollut 165 147ndash157httpsdoiorg101016jenvpol201202026 2012

Hatcher P E Seasonal and age-related variation in the nee-dle quality of five conifer species Oecologia 85 200ndash212httpsdoiorg101007BF00319402 1990

Hikosaka K Interspecific difference in the photosynthesisndashnitrogen relationship patterns physiological causesand ecological importance J Plant Res 117 481ndash494httpsdoiorg101007s10265-004-0174-2 2004

Huber W Carey V J Gentleman R Anders S Carlson MCarvalho B S Bravo H C Davis S Gatto L Girke T Got-tardo R Hahne F Hansen K D Irizarry R A LawrenceM Love M I MacDonald J Obenchain V Oles A KPagegraves H Reyes A Shannon P Smyth G K Tenenbaum DWaldron L and Morgan M Orchestrating high-throughput ge-nomic analysis with Bioconductor Nat Methods 12 115ndash121httpsdoiorg101038nmeth3252 2015

Hutnik R J McClenahen J R Long R P and Davis D DMercury accumulation in Pinus nigra (Austrian Pine) NortheastNat 21 529ndash540 httpsdoiorg1016560450210402 2014

ICP Forests Official Homepage httpicp-forestsnetpagedata-requests last access 16 February 2022

Iglewicz B and Hoaglin D C How to detect and handle outliersASQC Quality Press Milwaukee Wis 87 pp ISBN 0-87389-247-X 1993

Iverfeldt Aring Mercury in forest canopy throughfall water and itsrelation to atmospheric deposition Water Air Soil Poll 56 553ndash564 httpsdoiorg101007BF00342299 1991

Jiskra M Wiederhold J G Skyllberg U Kronberg R-M Haj-das I and Kretzschmar R Mercury deposition and re-emission

pathways in boreal forest soils investigated with Hg isotope sig-natures Environ Sci Technol 49 7188ndash7196 2015

Jiskra M Wiederhold J G Skyllberg U Kronberg R-M and Kretzschmar R Source tracing of natural organicmatter bound mercury in boreal forest runoff with mer-cury stable isotopes Environ Sci-Proc Imp 19 1235ndash1248httpsdoiorg101039C7EM00245A 2017

Jiskra M Sonke J E Obrist D Bieser J Ebinghaus R MyhreC L Pfaffhuber K A Waumlngberg I Kylloumlnen K WorthyD Martin L G Labuschagne C Mkololo T Ramonet MMagand O and Dommergue A A vegetation control on sea-sonal variations in global atmospheric mercury concentrationsNat Geosci 11 1ndash7 httpsdoiorg101038s41561-018-0078-8 2018

Jiskra M Sonke J E Agnan Y Helmig D and Obrist D In-sights from mercury stable isotopes on terrestrialndashatmosphere ex-change of Hg(0) in the Arctic tundra Biogeosciences 16 4051ndash4064 httpsdoiorg105194bg-16-4051-2019 2019

Jonard M Fuumlrst A Verstraeten A Thimonier A Timmer-mann V Potocic N Waldner P Benham S Hansen KMerilauml P Ponette Q Cruz A C de la Roskams P Nico-las M Croiseacute L Ingerslev M Matteucci G DecintiB Bascietto M and Rautio P Tree mineral nutrition isdeteriorating in Europe Glob Change Biol 21 418ndash430httpsdoiorg101111gcb12657 2015

Koumlrner C PlantndashEnvironment Interactions in StrasburgerrsquosPlant Sciences Including Prokaryotes and Fungi edited byBresinsky A Koumlrner C Kadereit J W Neuhaus Gand Sonnewald U Springer Berlin Heidelberg 1065ndash1166httpsdoiorg101007978-3-642-15518-5_12 2013

Koumlrner C Concepts in empirical plant ecology Plant Ecol Divers11 405ndash428 httpsdoiorg10108017550874201815400212018

Koumlrner Ch Scheel J and Bauer H Maximum leaf diffusive con-ductance in vascular plants Photosynthetica 13 45ndash82 1979

Laacouri A Nater E A and Kolka R K Distribution and uptakedynamics of mercury in leaves of common deciduous tree speciesin Minnesota USA Environ Sci Technol 47 10462ndash10470httpsdoiorg101021es401357z 2013

Lagergren F and Lindroth A Transpiration response to soil mois-ture in pine and spruce trees in Sweden Agr Forest Meteo-rol 112 67ndash85 httpsdoiorg101016S0168-1923(02)00060-6 2002

Lin Y-S Medlyn B E Duursma R A Prentice I C WangH Baig S Eamus D de Dios V R Mitchell P EllsworthD S de Beeck M O Wallin G Uddling J TarvainenL Linderson M-L Cernusak L A Nippert J B Ochel-tree T W Tissue D T Martin-StPaul N K Rogers AWarren J M De Angelis P Hikosaka K Han Q OnodaY Gimeno T E Barton C V M Bennie J Bonal DBosc A Loumlw M Macinins-Ng C Rey A Rowland L Set-terfield S A Tausz-Posch S Zaragoza-Castells J Broad-meadow M S J Drake J E Freeman M Ghannoum OHutley L B Kelly J W Kikuzawa K Kolari P KoyamaK Limousin J-M Meir P Lola da Costa A C MikkelsenT N Salinas N Sun W and Wingate L Optimal stomatalbehaviour around the world Nat Clim Change 5 459ndash464httpsdoiorg101038nclimate2550 2015

Lodenius M Tulisalo E and Soltanpour-Gargari A Ex-change of mercury between atmosphere and vegetation undercontaminated conditions Sci Total Environ 304 169ndash174httpsdoiorg101016S0048-9697(02)00566-1 2003

Loomis R S On the utility of nitrogen in leavesP Natl Acad Sci USA 94 13378ndash13379httpsdoiorg101073pnas942513378 1997

Manceau A Wang J Rovezzi M Glatzel P and Feng X Bio-genesis of mercuryndashsulfur nanoparticles in plant leaves from at-mospheric gaseous mercury Environ Sci Technol 52 3935ndash3948 httpsdoiorg101021acsest7b05452 2018

Martens B Miralles D G Lievens H van der Schalie R deJeu R A M Fernaacutendez-Prieto D Beck H E Dorigo W Aand Verhoest N E C GLEAM v3 satellite-based land evapora-tion and root-zone soil moisture Geosci Model Dev 10 1903ndash1925 httpsdoiorg105194gmd-10-1903-2017 2017

Martiacutenez-Ferri E Balaguer L Valladares F Chico JM and Manrique E Energy dissipation in drought-avoiding and drought-tolerant tree species at midday dur-ing the Mediterranean summer Tree Physiol 20 131ndash138httpsdoiorg101093treephys202131 2000

McAdam S A M and Brodribb T J The evolutionof mechanisms driving the stomatal response to va-por pressure deficit Plant Physiol 167 833ndash843httpsdoiorg101104pp114252940 2015

Mediavilla S and Escudero A Relative growth rateof leaf biomass and leaf nitrogen content in severalmediterranean woody species Plant Ecol 168 321ndash332httpsdoiorg101023A1024496717918 2003

Mellander P-E Bishop K and Lundmark T The influence ofsoil temperature on transpiration a plot scale manipulation ina young Scots pine stand Forest Ecol Manag 195 15ndash28httpsdoiorg101016jforeco200402051 2004

Meziane D and Shipley B Direct and indirect relationships be-tween specific leaf area leaf nitrogen and leaf gas exchange Ef-fects of irradiance and nutrient supply Ann Bot 88 915ndash927httpsdoiorg101006anbo20011536 2001

Miralles D G Holmes T R H De Jeu R A M Gash J HMeesters A G C A and Dolman A J Global land-surfaceevaporation estimated from satellite-based observations HydrolEarth Syst Sci 15 453ndash469 httpsdoiorg105194hess-15-453-2011 2011

Muntildeoz Sabater J ERA5-Land hourly data from 1981 topresent Copernicus Climate Change Service (C3S) Climate DataStore (CDS) httpscdsclimatecopernicuseucdsappdatasetreanalysis-era5-landtab=overview (last access 16 February2022) 2019

Navraacutetil T Shanley J B Rohovec J Oulehle F Šimecek MHouška J and Cudliacuten P Soil mercury distribution in adjacentconiferous and deciduous stands highly impacted by acid rain inthe Ore Mountains Czech Republic Appl Geochem 75 63ndash75httpsdoiorg101016japgeochem201610005 2016

Navraacutetil T Novaacutekovaacute T Roll M Shanley J B Kopaacutecek JRohovec J Kana J and Cudliacuten P Decreasing litterfall mer-cury deposition in central European coniferous forests and ef-fects of bark beetle infestation Sci Total Environ 682 213ndash225 httpsdoiorg101016jscitotenv201905093 2019

Niinemets Uuml and Tenhunen J D A model separating leafstructural and physiological effects on carbon gain along

light gradients for the shade-tolerant species Acer saccharumPlant Cell Environ 20 845ndash866 httpsdoiorg101046j1365-30401997d01-133x 1997

Nunn A J Kozovits A R Reiter I M Heerdt C LeuchnerM Luumltz C Liu X Loumlw M Winkler J B Grams T EE Haumlberle K-H Werner H Fabian P Rennenberg H andMatyssek R Comparison of ozone uptake and sensitivity be-tween a phytotron study with young beech and a field experi-ment with adult beech (Fagus sylvatica) Environ Pollut 137494ndash506 httpsdoiorg101016jenvpol200501036 2005

Obrist D Atmospheric mercury pollution due to losses ofterrestrial carbon pools Biogeochemistry 85 119ndash123httpsdoiorg101007s10533-007-9108-0 2007

Obrist D Roy E M Harrison J L Kwong C F Munger J WMoosmuumlller H Romero C D Sun S Zhou J and CommaneR Previously unaccounted atmospheric mercury deposition ina midlatitude deciduous forest P Natl Acad Sci USA 118e210547711 httpsdoiorg101073pnas2105477118 2021

Ollerova H Maruskova A Kontrisova O and Pliestikova LMercury accumulation in Picea abies (L) Karst needles with re-gard to needle age Pol J Environ Stud 19 1401ndash1404 2010

Panek J A and Goldstein A H Response of stomatalconductance to drought in ponderosa pine implicationsfor carbon and ozone uptake Tree Physiol 21 337ndash344httpsdoiorg101093treephys215337 2001

PEP725 Official Homepage to the Pan European PhenologyProject httpwwwpep725eu last access 16 February 2022

Pleijel H Klingberg J Nerentorp M Broberg M C Nyiram-bangutse B Munthe J and Wallin G Mercury accumula-tion in leaves of different plant types ndash the significance of tis-sue age and specific leaf area Biogeosciences 18 6313ndash6328httpsdoiorg105194bg-18-6313-2021 2021

Raspe S Bastrup-Birk A Fleck S Weis W Mayer HMeesenburg H Wagner M Schindler D and Gartner KChapter 17 ndash Meteorology in Developments in EnvironmentalScience vol 12 edited by Ferretti M and Fischer R Elsevier319ndash336 httpsdoiorg101016B978-0-08-098222-900017-02013

Rautio P Fuumlrst A Stefan K Raitio H and Bartels U PartXII Sampling and analysis of needles and leaves in Manualon methods and criteria for harmonized sampling assessmentmonitoring and analysis of the effects of air pollution on forestsedited by UNECE ICP Forests Programme Coordinating Cen-tre Thuumlnen Institute of Forest Ecosystems Eberswalde Ger-many httpicp-forestsnetpageicp-forests-manual (last access16 February 2022) 2016

Rea A W Keeler G J and Scherbatskoy T The deposition ofmercury in throughfall and litterfall in the Lake Champlain Wa-tershed A short-term study Atmos Environ 30 3257ndash3263httpsdoiorg1010161352-2310(96)00087-8 1996

Rea A W Lindberg S E and Keeler G J Dry depositionand foliar leaching of mercury and selected trace elements indeciduous forest throughfall Atmos Environ 35 3453ndash3462httpsdoiorg101016S1352-2310(01)00133-9 2001

Rea A W Lindberg S E Scherbatskoy T and Keeler GJ Mercury accumulation in foliage over time in two northernmixed-hardwood forests Water Air Soil Poll 133 49ndash67 2002

Reich P B Walters M B Ellsworth D S Vose J M VolinJ C Gresham C and Bowman W D Relationships of leaf

dark respiration to leaf nitrogen specific leaf area and leaf life-span a test across biomes and functional groups Oecologia 114471ndash482 httpsdoiorg101007s004420050471 1998

Reich P B Ellsworth D S Walters M B Vose JM Gresham C Volin J C and Bowman W DGenerality of leaf trait relationships a test across sixbiomes Ecology 80 1955ndash1969 httpsdoiorg1018900012-9658(1999)080[1955GOLTRA]20CO2 1999

Reich P B Wright I J Cavender-Bares J Craine J MOleksyn J Westoby M and Walters M B The evolutionof plant functional variation traits spectra and strategies IntJ Plant Sci 164 S143ndashS164 httpsdoiorg1010863743682003

Risch M R DeWild J F Krabbenhoft D P KolkaR K and Zhang L Litterfall mercury dry deposi-tion in the eastern USA Environ Pollut 161 284ndash290httpsdoiorg101016jenvpol201106005 2012

Risch M R DeWild J F Gay D A Zhang L Boyer EW and Krabbenhoft D P Atmospheric mercury depositionto forests in the eastern USA Environ Pollut 228 8ndash18httpsdoiorg101016jenvpol201705004 2017

Rosati A Esparza G DeJong T M and Pearcy R W Influ-ence of canopy light environment and nitrogen availability onleaf photosynthetic characteristics and photosynthetic nitrogen-use efficiency of field-grown nectarine trees Tree Physiol 19173ndash180 httpsdoiorg101093treephys193173 1999

Roumltzer T and Chmielewski F Phenological maps of Eu-rope Clim Res 18 249ndash257 httpsdoiorg103354cr0182492001

Running S W and Coughlan J C A general model of for-est ecosystem processes for regional applications I Hydrologicbalance canopy gas exchange and primary production pro-cesses Ecol Model 42 125ndash154 httpsdoiorg1010160304-3800(88)90112-3 1988

Rutter A P Schauer J J Shafer M M Creswell J E Ol-son M R Robinson M Collins R M Parman A M Katz-man T L and Mallek J L Dry deposition of gaseous ele-mental mercury to plants and soils using mercury stable iso-topes in a controlled environment Atmos Environ 45 848ndash855 httpsdoiorg101016jatmosenv201011025 2011

Saxton K E and Rawls W J Soil water characteris-tic estimates by texture and organic matter for hydro-logic solutions Soil Sci Soc Am J 70 1569ndash1578httpsdoiorg102136sssaj20050117 2006

Schulze E-D Kelliher F M Koumlrner C Lloyd J and Le-uning R Relationships among maximum stomatal conduc-tance ecosystem surface conductance carbon assimilation rateand plant nitrogen nutrition A global ecology scaling exerciseAnnu Rev Ecol Sys 25 629ndash660 1994

Schwarz P A Fahey T J and Dawson T E Seasonalair and soil temperature effects on photosynthesis in redspruce (Picea rubens) saplings Tree Physiol 17 187ndash194httpsdoiorg101093treephys173187 1997

Schwesig D and Matzner E Pools and fluxes of mercury andmethylmercury in two forested catchments in Germany SciTotal Environ 260 213ndash223 httpsdoiorg101016S0048-9697(00)00565-9 2000

Seidling W Hansen K Strich S and Lorenz M Part I Ob-jectives Strategy and Implementation of ICP Forests in Man-

ual on methods and criteria for harmonized sampling assess-ment monitoring and analysis of the effects of air pollution onforests edited by UNECE ICP Forests Programme Coordinat-ing Centre Thuumlnen Institute of Forest Ecosystems EberswaldeGermany httpicp-forestsnetpageicp-forests-manual (last ac-cess 16 February 2022) 2017

Simpson D Tuovinen J-P Emberson L and AshmoreM R Characteristics of an ozone deposition module IISensitivity analysis Water Air Soil Poll 143 123ndash137httpsdoiorg101023A1022890603066 2003

Sonke J E Teisserenc R Heimbuumlrger-Boavida L-E PetrovaM V Marusczak N Dantec T L Chupakov A V LiC Thackray C P Sunderland E M Tananaev N andPokrovsky O S Eurasian river spring flood observations sup-port net Arctic Ocean mercury export to the atmosphere and At-lantic Ocean P Natl Acad Sci USA 115 E11586ndashE11594httpsdoiorg101073pnas1811957115 2018

Sprovieri F Pirrone N Bencardino M DrsquoAmore F AngotH Barbante C Brunke E-G Arcega-Cabrera F Cairns WComero S Dieacuteguez M D C Dommergue A Ebinghaus RFeng X B Fu X Garcia P E Gawlik B M HagestroumlmU Hansson K Horvat M Kotnik J Labuschagne C Ma-gand O Martin L Mashyanov N Mkololo T Munthe JObolkin V Ramirez Islas M Sena F Somerset V SpandowP Vardegrave M Walters C Waumlngberg I Weigelt A Yang Xand Zhang H Five-year records of mercury wet deposition fluxat GMOS sites in the Northern and Southern hemispheres At-mos Chem Phys 17 2689ndash2708 httpsdoiorg105194acp-17-2689-2017 2017

Teixeira D C Lacerda L D and Silva-Filho E V Foliar mer-cury content from tropical trees and its correlation with phys-iological parameters in situ Environ Pollut 242 1050ndash1057httpsdoiorg101016jenvpol201807120 2018

Templ B Koch E Bolmgren K Ungersboumlck M Paul AScheifinger H Rutishauser T Busto M Chmielewski F-MHaacutejkovaacute L Hodzic S Kaspar F Pietragalla B Romero-Fresneda R Tolvanen A Vucetic V Zimmermann K andZust A Pan European Phenological database (PEP725) a sin-gle point of access for European data Int J Biometeorol 621109ndash1113 httpsdoiorg101007s00484-018-1512-8 2018

Toslashrseth K Aas W Breivik K Fjaeligraa A M Fiebig MHjellbrekke A G Lund Myhre C Solberg S and YttriK E Introduction to the European Monitoring and Evalua-tion Programme (EMEP) and observed atmospheric compositionchange during 1972ndash2009 Atmos Chem Phys 12 5447ndash5481httpsdoiorg105194acp-12-5447-2012 2012

Travnikov O Angot H Artaxo P Bencardino M Bieser JDrsquoAmore F Dastoor A De Simone F Dieacuteguez M DC Dommergue A Ebinghaus R Feng X B GencarelliC N Hedgecock I M Magand O Martin L MatthiasV Mashyanov N Pirrone N Ramachandran R Read KA Ryjkov A Selin N E Sena F Song S Sprovieri FWip D Waumlngberg I and Yang X Multi-model study ofmercury dispersion in the atmosphere atmospheric processesand model evaluation Atmos Chem Phys 17 5271ndash5295httpsdoiorg105194acp-17-5271-2017 2017

Tsuji S Nakashizuka T Kuraji K Kume A and Hanba YT Sensitivity of stomatal conductance to vapor pressure deficitand its dependence on leaf water relations and wood anatomy

in nine canopy tree species in a Malaysian wet tropical rainfor-est Trees 34 1299ndash1311 httpsdoiorg101007s00468-020-01998-5 2020

UN Environment Global Mercury Assessment 2018 Chemicalsand Health Branch Geneva Switzerland ISBN 978-92-807-3744-8 2019

Vesterdal L Schmidt I K Callesen I Nilsson L O and Gun-dersen P Carbon and nitrogen in forest floor and mineral soilunder six common European tree species Forest Ecol Manag255 35ndash48 httpsdoiorg101016jforeco200708015 2008

Vilhar U Beuker E Mizunuma T Skudnik M LebourgeoisF Soudani K and Wilkinson M Chapter 9 ndash Tree Phe-nology in Developments in Environmental Science vol 12edited by Ferretti M and Fischer R Elsevier 169ndash182httpsdoiorg101016B978-0-08-098222-900009-1 2013

Wang X Bao Z Lin C-J Yuan W and FengX Assessment of global mercury deposition throughlitterfall Environ Sci Technol 50 8548ndash8557httpsdoiorg101021acsest5b06351 2016

Wang X Yuan W Lin C-J Luo J Wang F Feng X FuX and Liu C Underestimated sink of atmospheric mercury ina deglaciated forest chronosequence Environ Sci Technol 548083ndash8093 httpsdoiorg101021acsest0c01667 2020

Weiss-Penzias P S Ortiz C Acosta R P Heim WRyan J P Fernandez D Collett J L and Flegal AR Total and monomethyl mercury in fog water from thecentral California coast Geophys Res Lett 39 L03804httpsdoiorg1010292011GL050324 2012

Wilson K B Baldocchi D D and Hanson P J Spatial and sea-sonal variability of photosynthetic parameters and their relation-ship to leaf nitrogen in a deciduous forest Tree Physiol 20 565ndash578 httpsdoiorg101093treephys209565 2000

Wohlgemuth L GitHub repository to Publication Physiologi-cal and climate controls on foliar mercury uptake by Euro-pean tree species Zenodo httpszenodoorgrecord5495180YcnzjlnTU2w GitHub httpsgithubcomwohleHg_Forestslast access 16 February 2022

Wohlgemuth L Osterwalder S Joseph C Kahmen A HochG Alewell C and Jiskra M A bottom-up quantification offoliar mercury uptake fluxes across Europe Biogeosciences 176441ndash6456 httpsdoiorg105194bg-17-6441-2020 2020

Wright I J Reich P B Westoby M Ackerly D D BaruchZ Bongers F Cavender-Bares J Chapin T Cornelissen JH C Diemer M Flexas J Garnier E Groom P K Gu-lias J Hikosaka K Lamont B B Lee T Lee W Lusk CMidgley J J Navas M-L Niinemets Uuml Oleksyn J OsadaN Poorter H Poot P Prior L Pyankov V I Roumet CThomas S C Tjoelker M G Veneklaas E J and Villar RThe worldwide leaf economics spectrum Nature 428 821ndash827httpsdoiorg101038nature02403 2004

Wright L P Zhang L and Marsik F J Overview of mercurydry deposition litterfall and throughfall studies Atmos ChemPhys 16 13399ndash13416 httpsdoiorg105194acp-16-13399-2016 2016

Yang Y Yanai R D Montesdeoca M and DriscollC T Measuring mercury in wood challenging butimportant Int J Environ An Ch 97 456ndash467httpsdoiorg1010800306731920171324852 2017

Yu B Fu X Yin R Zhang H Wang X Lin C-J WuC Zhang Y He N Fu P Wang Z Shang L Som-mar J Sonke J E Maurice L Guinot B and FengX Isotopic composition of atmospheric mercury in chinanew evidence for sources and transformation processes in airand in vegetation Environ Sci Technol 50 9262ndash9269httpsdoiorg101021acsest6b01782 2016

Yuan W Zheng Y Piao S Ciais P Lombardozzi D WangY Ryu Y Chen G Dong W Hu Z Jain A K JiangC Kato E Li S Lienert S Liu S Nabel J E M SQin Z Quine T Sitch S Smith W K Wang F Wu CXiao Z and Yang S Increased atmospheric vapor pressuredeficit reduces global vegetation growth Sci Adv 5 eaax1396httpsdoiorg101126sciadvaax1396 2019

Zhou J Obrist D Dastoor A Jiskra M and Ryjkov A Vegeta-tion uptake of mercury and impacts on global cycling Nat RevEarth Environ 2 269ndash284 httpsdoiorg101038s43017-021-00146-y 2021

Zweifel R Steppe K and Sterck F J Stomatal regulation by mi-croclimate and tree water relations interpreting ecophysiologicalfield data with a hydraulic plant model J Exp Bot 58 2113ndash2131 httpsdoiorg101093jxberm050 2007

Zweifel R Rigling A and Dobbertin M Species-specific stom-atal response of trees to drought ndash a link to vegetation dynam-ics J Veg Sci 20 442ndash454 httpsdoiorg101111j1654-1103200905701x 2009

Abstract

Introduction

Material and methods

Foliage sampling and dataset description

Correction of foliar Hg concentrations for drying temperature

Foliage Hg analysis

Determination of the beginning of the growing season for calculating daily foliage Hg uptake rates

Evaluation of data on water vapor pressure deficit (VPD)

Evaluation of ERA5-Land volumetric soil water contents

Results and discussion

Variation in foliar Hg concentrations with foliar life period

Variation in foliar Hg uptake rates with tree species groups

Foliar Hg uptake and sample-specific N concentration

Foliar Hg uptake and leaf mass per area

Foliar Hg uptake and water vapor pressure deficit (VPD)

Foliar Hg uptake and soil water content

Foliar Hg uptake and geographic and tree-specific parameters

Implications for Hg cycle modeling

Conclusions

Data availability

Supplement

Author contributions

Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

tree species for foliage of beech and fir and at two out ofthree forest plots with more than 20 samples we found asignificant (p lt 0001) increase in foliar Hg values with re-spective leaf nitrogen concentrations We therefore suggestthat foliar stomatal Hg uptake is controlled by tree functionaltraits with uptake rates increasing from low to high nutrientcontent representing low to high physiological activity Forpine and spruce needles we detected a significant linear de-crease in daily foliar stomatal Hg uptake with the proportionof time during which water vapor pressure deficit (VPD) ex-ceeded the species-specific threshold values of 12 and 3 kParespectively The proportion of time within the growing sea-son during which surface soil water content (ERA5-Land)in the region of forest plots was low correlated negativelywith foliar Hg uptake rates of beech and pine These find-ings suggest that stomatal uptake of atmospheric Hg(0) isinhibited under high VPD conditions andor low soil watercontent due to the regulation of stomatal conductance to re-duce water loss under dry conditions Other parameters as-sociated with forest sampling sites (latitude and altitude)sampled trees (average age and diameter at breast height)or regional satellite-observation-based transpiration product(Global Land Evaporation Amsterdam Model GLEAM) didnot significantly correlate with daily foliar Hg uptake ratesWe conclude that tree physiological activity and stomatal re-sponse to VPD and soil water content should be implementedin a stomatal Hg model to assess future Hg cycling underdifferent anthropogenic emission scenarios and global warm-ing

1 Introduction

Mercury (Hg) is a toxic pollutant that is emitted by anthro-pogenic and geogenic activities into the atmosphere whereit can be transported over large distances and is eventuallytransferred to terrestrial and ocean surfaces by dry or wetdeposition (Bishop et al 2020) From a public health per-spective transfer rates of Hg to aquatic ecosystems are par-ticularly relevant within this cycle since Hg bioaccumulationin fish for consumption represents the most important Hg ex-posure pathway to many communities internationally (UNEnvironment 2019) In order to constrain future Hg levelsin edible fish and to assess how Hg exposure responds tocurbed anthropogenic Hg emissions under the policies imple-mented by the 2017 UN Minamata convention on mercury itis essential to understand and quantify all major net deposi-tion fluxes within the global Hg cycle Wet deposition occurswhen water-soluble oxidized Hg(II) is washed out from theatmosphere with rainwater (Driscoll et al 2013 Sprovieriet al 2017) or by cloud water interception (Weiss-Penziaset al 2012) In a dry deposition process gaseous elemen-tal Hg(0) and Hg(II) directly bind to surfaces (Bishop et al2020) or Hg(0) is taken up by plants (Zhou et al 2021) For

more than 2 decades vegetation has been recognized as animportant vector for Hg(0) dry deposition within the terres-trial Hg cycle (Rea et al 1996 2002 Grigal 2003) Basedon this seminal work researchers have since highlighted thatvegetation impacts Hg levels of all other environmental com-partments within the active Hg cycle (AMAP and UNEP2019 Bishop et al 2020 Zhou et al 2021) Vegetation up-take of Hg(0) governs the seasonality of atmospheric Hg(0)in the Northern Hemisphere with concentration minima insummer at the end of the growing season (Jiskra et al 2018)Thus vegetation has been suggested to operate like a globalHg pump (Obrist 2007 Jiskra et al 2018) AtmosphericHg(0) taken up by vegetation is oxidized to Hg(II) within theplant tissue (Manceau et al 2018) and transferred to soils vialitterfall (Iverfeldt 1991 Schwesig and Matzner 2000 Reaet al 2001 Graydon et al 2008 Risch et al 2012 2017Jiskra et al 2015 Wright et al 2016 Wang et al 2016)Moreover in forests Hg deposition to the ground may oc-cur by wash-off of Hg(0) from plant surfaces via through-fall and by Hg(0) uptake into woody tissues lichen mossesand soil litter (Wang et al 2020 Obrist et al 2021) Mer-cury sequestered by forest ecosystems accumulates in soiland may subsequently be transported from watersheds tostreams rivers and the ocean where it can bioaccumulatein fish (Drenner et al 2013 Jiskra et al 2017 Sonke et al2018)

Concerning the mechanism of Hg accumulation in foliagethere are multiple lines of evidence that leaf stomata con-trol the foliar Hg(0) uptake flux to terrestrial ecosystems (i)Hg concentrations were found to be higher in internal fo-liar tissues than on leaf surfaces (Laacouri et al 2013) (ii)experiments revealed that isotopic Hg tracers are transferredfrom the air to the leaf interior (Rutter et al 2011) (iii) fo-liar Hg concentrations are associated with leaf stomatal den-sity and net photosynthesis (Laacouri et al 2013 Teixeiraet al 2018) (iv) the isotopic composition of foliage is dis-criminated in heavy isotopes compared to atmospheric Hg(0)(Demers et al 2013 Enrico et al 2016 Yu et al 2016Jiskra et al 2019) and (v) temporal and vertical variationsin net foliar Hg(0) uptake fluxes in trees agree with the mech-anism of stomatal Hg(0) uptake (Wohlgemuth et al 2020)While there is increasing consensus that vegetation uptakeof atmospheric Hg(0) occurs via the stomatal pathway thereremain research gaps regarding parameters regulating thisstomatal Hg(0) uptake (Zhou et al 2021) Consequentlythe Hg(0) dry deposition flux to terrestrial surfaces in HgEarth system models is generally parametrized by static in-ferential or resistance-in-series approaches (Travnikov et al2017) Ecosystem processes including canopy gas exchangeare sensitive to climate conditions (Running and Coughlan1988) and vary between different plant species (Reich et al2003) Trees control leaf diffusive gas fluxes through theirstomata in order to optimize the diffusive influx of carbondioxide for photosynthesis while averting excessive loss ofwater vapor to the atmosphere (Koumlrner 2013) The regu-

(ngHggminus1

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

(ngHggminus1

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

(ngHggminus1

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

(ngHggminus1

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

(ngHggminus1

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

(ngHggminus1

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

(ngHggminus1

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

(ngHggminus1

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

dw) and (c) average

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

4 Conclusions

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

Abstract

Introduction



Foliage Hg analysis













Conclusions

Data availability

Supplement


Competing interests

Disclaimer

Acknowledgements

Financial support

Review statement

References

Physiological and climate controls on foliar mercury uptake by ...

Documents