Seasonal variation in throughfall and stemflowchemistry beneath a European beech (Fagussylvatica) tree in relation to canopy phenology

Jeroen Staelens, An De Schrijver, and Kris Verheyen

Abstract: The effect of canopy phenology on major ion fluxes beneath a mature European beech (Fagus sylvatica L.) treeis examined. Annual and seasonal ion fluxes to the forest floor were significantly higher than the incoming wet-only deposi-tion for all ions measured other than H+. The annual throughfall to wet deposition ratio generally ranged from 2.1 to 4.8.Stemflow contributed 9%–19% of the ion input to the forest floor, except for H+. Throughfall enrichment of K+, Ca2+, Mg2+,and NO3

– was significantly higher in the leafed than in the leafless season, in contrast to Na+, NH4+, and H+. The temporal

pattern of ion enrichment indicated canopy release of K+, Ca2+, and Mg2+ throughout the leafed season, of Na+, Cl–, andNH4

+ from emerging leaves, and of Cl– and SO42– from senescing leaves. The contribution of canopy leaching to annual net

throughfall and stemflow was estimated at 96% (K+), 54% (Ca2+), 40% (Mg2+), 12% (Cl–), and 7% (Na+, SO42–). Dry depo-

sition accounted for 58%–75% of the total deposition onto the canopy. The throughfall enrichment during the leaflessseason indicated high particulate and gaseous dry deposition onto the woody canopy as well as K+ release from Euro-pean beech branches.

Resume : Les effets de la phenologie de la canopee sur le flux des principaux ions sous un hetre (Fagus sylvatica L.) ma-ture sont examines. Pour tous les ions mesures autres que H+, les flux annuel et saisonnier d’ions vers le parterre forestieretaient significativement plus eleves que les seuls depots humides. Les valeurs du rapport entre la quantite de precipitationau sol et les depots humides variaient generalement de 2,1 a 4,8. L’ecoulement le long du tronc representait 9 a 19 % del’apport en ions au parterre forestier, sauf pour H+. L’enrichissement des precipitations au sol en K+, Ca2+, Mg2+ et NO3

–

au cours de la saison avec du feuillage etait significativement plus eleve qu’au cours de la saison sans feuillage, mais pasdans le cas de Na+, NH4

+ et H+. Le patron temporel d’enrichissement en ions indique que K+, Ca2+, Mg2+ sont liberes toutau long de la saison avec du feuillage, que Na+, Cl– et NH4

+ sont liberes lorsque les feuilles emergent et que Cl– et SO42–

sont liberes lorsque les feuilles sont senescentes. La contribution du lessivage de la canopee aux valeurs annuelles nettesde precipitation au sol et d’ecoulement le long du tronc a ete estimee a 96 % (K+), 54 % (Ca2+), 40 % (Mg2+), 12 % (Cl–)et 7 % (Na+, SO4

2–). La contribution des depots secs a l’ensemble des depots (humide et sec) sur la canopee variait de 58a 75 %. L’enrichissement des precipitations au sol pendant la saison sans feuillage indique qu’il y a d’importants depotssecs sous forme de gaz et de particules sur la canopee et que les branches du hetre liberent du K+.

[Traduit par la Redaction]

IntroductionThroughfall (TF) water can contribute large amounts of

nutrients to forest floors (Parker 1983) and may have somecontrol of the acid–base status of the soil by altering incom-ing precipitation (Belanger et al. 2004). Furthermore, TF iondeposition is often used for estimating total atmospheric in-put to forests, to assess potential effects of air pollution onthe diversity and function of forest ecosystems (de Vries etal. 2003). Interpreting TF measurements in nutrient-cyclingand atmospheric-deposition studies requires a distinction be-tween in-canopy sources and atmospheric input of chemicalcompounds, since both wash-off of dry deposition (DD) andion-exchange processes contribute to the modified ion com-position of the water flowing through a forest canopy(Lovett and Lindberg 1984). DD is the direct deposition ofparticles and gases from the atmosphere onto the vegetation

without the involvement of precipitation or fog droplets(Andersen and Hovmand 1999). Canopy exchange (CE)comprises passive ion diffusion and (or) exchange betweenthe water layer covering canopy tissues and the underlyingapoplast, as well as the uptake of gases through stomata(Draaijers et al. 1997).

The contribution of DD and CE to TF enrichment is ion-specific. CE of Na+ and Cl– is generally considered to benegligible (Draaijers et al. 1997) and Na+ and Cl– enrich-ment of TF compared with that of precipitation is attributedto the washing off of DD only. Likewise, S concentration isusually assumed to be conservative in the canopy, as stoma-tal uptake of SO2 is thought to be balanced by canopy leach-ing (CL) of SO4

2– to TF (Butler and Likens 1995; Kovacksand Horvath 2004). For N, in contrast, significant canopyuptake (CU) and assimilation of gaseous compounds and

Received 11 April 2006. Accepted 16 December 2006. Published on the NRC Research Press Web site at cjfr.nrc.ca on 30 August 2007.

J. Staelens,1 A. De Schrijver, and K. Verheyen. Laboratory of Forestry, Ghent University, Geraardsbergsesteenweg 267, B-9090Gontrode, Belgium.

1Corresponding author (e-mail: jeroen_staelens@yahoo.com).

1359

Can. J. For. Res. 37: 1359–1372 (2007) doi:10.1139/X07-003 # 2007 NRC Canada

dissolved NH4+ and NO3

– have been reported (Boyce et al.1996; Gessler et al. 2000; Harrison et al. 2000). Ca2+, Mg2+,and K+ are known to be leached from vegetative surfaces,and this leaching has been partly attributed to the retentionof H+ (Cronan and Reiners 1983) and NH4

+ (Roelofs et al.1985) in the canopy by ion-exchange reactions (Stachurskiand Zimka 2002).

Both DD and CE processes are influenced by a largenumber of variables and can vary significantly in spaceand time (Lovett et al. 1996; Morris et al. 2003). The DDof gases and particles from the atmosphere to a receptorsurface is governed by air concentration and by turbulenttransport processes in the aerodynamic boundary layer, aswell as by the chemical and physical nature of the deposit-ing compound and the capability of the surface to capturegases and particles (Erisman and Draaijers 2003). Conse-quently, the amount of DD at a given time and place de-pends on (i) meteorological factors such as wind speed andair humidity, (ii) the strength and proximity of emissionsources, and (iii) vegetation characteristics that affect sur-face roughness such as tree species, tree height, and leaf-area index (Lovett et al. 1996; Andersen and Hovmand1999). Precipitation characteristics influencing CE includethe composition and acidity of precipitation (Schaefer et al.1988; Hansen 1996; Lovett et al. 1996), and precipitationamount (Potter et al. 1991; Lovett et al. 1996), duration(Reiners and Olson 1984), and intensity (Hansen et al.1994). Furthermore, CE is affected by the quantity of freelyexchangeable ions in the canopy, which depends on treespecies, tree age, the nutrient status of the forest (Parker1983; Houle et al. 1999), canopy leaf area (Lovett et al.1989), and the physiological state of the vegetation (Morriset al. 2003).

Because several of these factors act at relatively largescales, the spatiotemporal variability of TF deposition at theplot scale can be expected to depend mostly on atmosphericconditions, precipitation characteristics, and local canopystructure. With respect to the spatial heterogeneity of TFwithin forest stands, previous research indicates that the in-put of water and nutrients to the forest floor may be relatedto the local canopy and branch structure (Beier et al. 1993;Hamburg and Lin 1998; Whelan et al. 1998). With respectto the temporal variability of TF, most attention has beenpaid to the relationship between TF chemistry and the tem-poral distribution and composition of precipitation events(Potter et al. 1991; Lovett et al. 1996).

In broad-leaved deciduous forest stands, considerabletemporal variation in canopy cover and physiological activ-ity occurs throughout the year, which likely affects both DDand CE. Nevertheless, TF and stemflow (SF) fluxes in de-ciduous forest types have mostly been measured during theleafed season. Even when ion fluxes are quantified duringthe leafless season, research often focuses on annual means,so there is relatively little information on the interaction ofincident precipitation with the woody canopy of trees duringthe dormant season (Houle et al. 1999; Levia and Frost2003; Pryor and Barthelmie 2005). Therefore, the aims ofthis study were (i) to quantify TF and SF ion depositionthroughout the year beneath a deciduous European beech(Fagus sylvatica L.) canopy, (ii) to examine the effect ofseasonal changes in canopy cover and phenology on TF and

SF chemistry, and (iii) to estimate ion exchange within thedeciduous canopy during each phenological canopy phase.

Materials and methods

Study siteThe Aelmoeseneie forest (50858’N, 3849’E, 16 m a.s.l) is

a mixed deciduous forest with a total area of 28 ha locatednear Ghent in the north of Belgium, approximately 60 kmfrom the North Sea. Mean annual precipitation (1980–2002)is 755 mm and distributed equally over the year, and meanannual temperature is 10.2 8C. The study tree is located in a85-year-old forest stand dominated by pedunculate oak(Quercus robur L.) and European beech. In 1997, stand den-sity was 345 trees�ha–1, mean canopy height was 27 m, andpedunculate oak and European beech represented 49% and27% of the total basal area of 26.6 m2�ha–1, respectively.The understory consists of rowan (Sorbus aucuparia L.), ha-zel (Corylus avellana L.), and sycamore maple (Acer pseu-doplatanus L.). Maximum leaf-area index during the 1996growing season was 5.5 (Mussche et al. 2001). The forestsoil has developed from a quaternary layer of sand loam(0.5–1 m) on a shallow impermeable clay and sand complexof tertiary origin (Gleyic Cambisol, United Nations Foodand Agriculture Organization classification). The foreststand contains a level-II observation plot of the Europeannetwork program for intensive monitoring of forest ecosys-tems (International Cooperative Programme on Forests), andTF and SF water have been measured since 1994. The re-gion is characterized by relatively high SO4

2–, NO3–, and

sea-salt deposition. In the level-II plot, the mean annual iondepositions to the forest floor between 1994 and 2002 were1.55 (NH4

+), 0.62 (NO3–), 1.76 (SO4

2–), 1.10 (Na+), and 1.25(Cl–) kmolc�ha –1�year–1 (G. Genouw, unpublished data).

In the pedunculate oak – European beech stand one dom-inant European beech tree with a diameter at 1.3 m of0.68 m and a height of 30 m was chosen. The studied treewas surrounded by several pedunculate oaks, a northern redoak (Quercus rubra L.), and a European beech tree (Fig. 1).Because of light competition, the vitality of the neighbour-ing pedunculate oak trees is declining. The understory layerwas removed from a 15 m � 12 m area to determine the ef-fect of the tree crown alone on TF. The European beech treewas in leaf from 25 April until 18 November 2003. These30 weeks will subsequently be called the leafed season,within which three phenological periods were distinguished(leaf emergence, full-leaf canopy phase, and leaf senes-cence). Mean canopy cover above the TF collectors was94% in August 2002 and 55% in March 2003. The spatialvariability of TF ion fluxes under the studied Europeanbeech tree has been previously discussed by Staelens et al.(2006a).

Data collection

Water fluxesBulk precipitation and TF were sampled weekly from 4

March 2003 to 4 March 2004. Wet-only precipitation wassampled weekly from 4 March 2003 to 9 December 2003.In the forest stand a 35 m high meteorological tower is lo-cated 50 m from the study tree. Wet-only precipitation was

1360 Can. J. For. Res. Vol. 37, 2007

# 2007 NRC Canada

measured on top of the tower by an automatic wet-onlysampler (G.K. Walter Eigenbrodt) with a funnel diameter of25.4 cm. An electrical resistance-driven rain sensor (RS 85,G.K. Walter Eigenbrodt) with a delay of 1 s activated theopening of the lid (Staelens et al. 2005). Bulk precipitationwas measured on top of the tower using two collectors witha funnel diameter of 24.2 cm. The funnels had a sharp-edged vertical rim and a slope of 458 and were set at aheight of 36.5 m above the ground. Precipitation drainedinto 2 L polyethylene bottles at 35 m height that werewrapped in aluminium foil to keep the samples cool. A nylon1 mm wire mesh was placed in the funnels to prevent con-tamination by large particles.

TF water was measured using 12 collectors that were setin a 3 m � 3 m grid under the European beech canopy(Fig. 1). TF collectors were of the same type as the bulkprecipitation samplers, and were set at a height of 1.5 mabove the ground. The collecting bottles were placed belowground level to avoid the growth of algae and keep the sam-ples cool. Funnels, wire meshes, and bottles were replacedweekly by equipment rinsed with demineralised water. Pre-cipitation and TF amounts were determined by weighing inthe laboratory. SF water was collected from the Europeanbeech tree using a spiral-type collector, and measured bymeans of a self-constructed 0.2 L tipping bucket. The out-flow was collected in a 200 L container and sampled at leastweekly from 4 March 2003 to 4 March 2004.

Chemical analysis and data qualityBulk and wet-only deposition (WD) and SF samples were

analysed weekly until December 2003 and biweekly after-

wards. Weekly samples of TF from each TF collector werepooled biweekly (on a volume-weighted (vw) basis) forchemical analysis. Water samples were transported andstored in darkness at 5 8C, and pH (ion-specific electrode)and conductivity were measured within 24 h after sampling.Within 1 week after sampling, the TF samples were filteredthrough a 0.45 mm mesh nylon membrane filter and NO3

–,SO4

2–, PO43–, and Cl– concentrations were determined with

ion chromatography (Dionex Corporation). The NH4+ con-

centration was determined using photometric determinationof a reaction product of NH4

+ at 660 nm (Dutch standardmethod NEN 6567) and K+, Ca2+, Mg2+, and Na+ concentra-tions were determined using flame atomic absorption spec-trophotometry. H+ concentrations were derived from the pHmeasurements.

The chemical analyses were validated by includingmethod blanks and by taking repeated measurements of in-ternal and certified standard reference samples (CRM 409,Quevauviller et al. 1993). Determination limits (mmolc�L–1)of the chemical analyses were 1.8 for Na+, 1.0 for K+, 4.4for Ca2+, 1.6 for Mg2+, 4.2 for NH4

+, 4.0 for NO3–, 5.2 for

SO42–, 5.6 for Cl–, and 6.3 for PO4

3–. The coefficient of var-iation of five repeated measurements of CRM 409 samplesthroughout the year was smaller than 5% and recovery washigher than 95% for all ions. Furthermore, the quality of theanalyses was evaluated by comparing the measured and cal-culated conductivities of water samples (r2 = 0.982 for pre-cipitation, r2 = 0.996 for TF, and r2 = 0.996 for SF) and bychecking the ion balance. When all major ions are analysed,the charge difference between cations and anions is ascribedto the presence of organic anions or to bicarbonate for sampleswith sufficiently high pH (Driscoll et al. 1989; Chiwa et al.2004). The bicarbonate concentration of water samples wasestimated from the pH using a dissociation constant (pKCO2

)of 7.8 and a partial CO2 pressure of 0.3 mbar (de Vries et al.2001). The PO4

3– concentration was below or around the de-termination limit, suggesting that the samples were not con-taminated by bird droppings (Erisman et al. 2003).

Data analysis

Ion fluxesIon deposition (mmolc�m–2) was calculated by multiplying

the amount of water collected by the ion concentration div-ided by the area of the collector openings. SF volume wastransformed to depth using the surface area of the horizontalcanopy projection, which was visually determined to be ap-proximately 180 m2. Mean ion concentrations of SF werecalculated as vw means. Snowfall occurred during 1 weekof the study period. Because of the undercatch of above-canopy snowfall measurements, the precipitation amountfor this sampling week was assumed to be equal to themaximum TF amount measured then. The bulk-precipitationmeasurements from 9 December 2003 to 4 March 2004were corrected for DD of particles and gases onto the funnelsby site-specific bulk:wet-only concentration factors derivedat the study site from 4 March 2003 to 9 December 2003(Staelens et al. 2005). WD was calculated by multiplyingwet-only rainfall ion concentrations by the rainfall amountscollected by the bulk collectors because the latter wereidentical to the TF collectors.

Fig. 1. Experimental setup in the Aelmoeseneie forest, Gontrode,Belgium. Tree labels indicate species (C, pedunculate oak; B, Eur-opean beech; R, northern red oak) and diameter (cm) at 1.3 mheight.

Staelens et al. 1361

# 2007 NRC Canada

Temporal variability of ion fluxes and correlation betweenions

The biweekly net throughfall (NTF) flux was calculatedas NTF = TF WD and the net throughfall and stemflow(NTS) flux as NTS = TF + SF – WD. In addition, the bi-weekly ratio of TF to WD (TF:WD) was calculated (Nearyand Gizyn 1994; Michopoulos et al. 2001). One-sample ttests (n = 12) were used to examine whether the annual andseasonal NTF and NTS depositions differed significantlyfrom zero. Positive net ion fluxes indicate ion enrichmentby the canopy due to DD and (or) internal CL, negative netion fluxes indicate a net uptake of ions by the canopy. Thecontribution of DD and CE to NTF can be estimated usingcanopy budget methods and will be described later.

The significance of the differences between the leafed (25April until 18 November 2003) and the leafless seasons wasdetermined. For each ion, WD, area-averaged TF, and SFwere compared between the leafed (13 sampling intervals)and the leafless (9 sampling intervals) seasons using Wilcox-on’s rank sum test for independent samples. Since TF iondeposition was determined for individual collectors (n = 12),TF, NTF, and TF:WD could also be compared statisticallybetween the leafed and leafless seasons using Wilcoxon’ssigned-ranks test for two related samples. As the leafed sea-son was longer than the leafless season, this analysis was per-formed on the mean biweekly ion deposition per collector forthe two periods, calculated by dividing the deposition in theleafed (30 weeks) and leafless (22 weeks) seasons by 15 and11, respectively. There were two biweekly sampling intervalswithout precipitation in both the leafed and leafless seasons.

Finally, relations between ions in NTF deposition wereexamined by visual inspection of plots of ion pairs and bycorrelation and regression analysis. Spearman’s coefficientof rank correlation (rS) were calculated between the NTFdeposition of 12 individual collectors for 22 sampling inter-vals (ntotal = 264) and by distinguishing four phenologicalphases: the leafed season was divided into (i) the period ofbud break and leaf unfolding (25 April to 6 May), herein-after termed leaf emergence, (ii) the full-leaf canopy period(6 May to 30 September), and (iii) the leaf senescence (1October to 18 November), in addition to (iv) the leaflessseason. All statistical analyses were performed with SPSS1

for Windows 12.0. The reported significance levels were ob-tained for two-tailed tests.

Estimated DD and CETo estimate the contribution of DD and CE to the NTS

deposition, we modified the canopy budget method reportedby Draaijers and Erisman (1995) that has been validated fora Douglas-fir stand in a region with high N and S depositionloads. In the original method, the NTS of Na+, Cl–, NO3

–,and SO4

2– is assumed to be due to DD only, and CE is as-sumed to be zero. Furthermore, the DD:WD ratio of Na+ isassumed to be valid for particles containing K+, Ca2+, andMg2+, implying that an equal deposition efficiency is as-sumed for particles containing Na+ as for Mg2+, Ca2+, andK+. The annual DD of the latter cations, X, is calculated bymultiplying WDX by ðDD:WDÞNaþ , and CL is estimated as(NTS – DD)X. CU of NH4

+ and H+ (CUNH þ4 þ Hþ) is assumed

to be equal to the CL of base cations K+, Ca2+, and Mg2+

(CLBC) that is not associated with weak acid release (CLwa):

½1� CUNH þ4 þ Hþ ¼ CLBC�CLwa

CLwa is calculated assuming that the DD of weak acidsequals their WD. NH4

+ and H+ are assumed to be taken upin a proportion equal to their average ratio in WD and TF +SF, accounting for more efficient uptake of H+ by a factor,xH, of 6 (Draaijers and Erisman 1995):

½2� CUHþ ¼ CLBC�CLwa

1 þ xH TFHþSFHTFNH4þSFNH4

þ WDH

WDNH4

� �=2

h i

In our study, we applied the canopy budget method toeach phenological canopy phase and modified it to accountfor CL of Na+, Cl–, and SO4

2– during leaf emergence and(or) senescence. The NTS:WD ratio of Na+ during emer-gence was set equal to its ratio during the full-leafed canopyphase. In this way, DD and leaching of Na+ from sproutingbuds and unfolding leaves were estimated. Then, DD of Cl–during leaf emergence and senescence was determined asthe product of DDNa

+ and the mean Cl–:Na+ ratio in NTSduring the rest of the year. This allowed the estimation ofCl– leaching from the emerging and senescent leaves. Next,DD and leaching of SO4

2– during leaf senescence was esti-mated by multiplying the WD of SO4

2– during senescenceby the NTS:WD ratio of SO4

2– during the rest of the year.Lastly, CU of H+ and NH4

+ were estimated from the leach-ing of Na+, K+, Ca2+, and Mg2+, corrected for the estimatedCL of the weak acids, Cl– and SO4

2–.

Results

Precipitation, TF and SF depositionThe annual precipitation amount of 613 mm (from 4

March 2003 to 2004) was partitioned into 67% TF, 8% SF,and 25% interception loss (Table 1). Annual and seasonal(leafed season, 30 weeks; leafless season, 22 weeks) iondeposition to the forest floor by TF and SF was significantlyhigher than the incoming WD for all measured ions exceptH+, which was significantly reduced after canopy passage(p < 0.001; Table 1). Consequently, the annual TF:WD ratioexceeded 1.0 for all ions other than H+, and generallyranged from 2.1 to 4.8 (Table 1). The annual TF:WD ratioof K+ (43) was much higher than that of the other ions. Themost important ions in NTS deposition were NH4

+, Na+,SO4

2–, and Cl–. The HCO3– fluxes calculated from the pH

hardly contributed to the annual fluxes of the sum of meas-ured anions in WD (3%), TF (1%), and SF (0.4%). The con-tribution of SF to the annual TF + SF ion deposition was38% for H+, 15%–19% for NH4

+, Na+, and Cl–, and approx-imately 10% for the other measured ions (Table 1).

According to the unrelated-samples rank test, the WD ofNa+, Cl–, and Mg2+ was significantly higher in the leaflessseason than in the leafed season (p < 0.05; Fig. 2, Table 2),while the fluxes of Na+, Cl–, and SO4

2– in TF were higher inthe leafless than in the leafed season (p < 0.05). However,for TF and its derived variables, NTF and TF:WD, themore powerful related-samples rank test demonstrated thatthe TF deposition of all ions differed significantly betweenthe leafed and leafless seasons (p < 0.01, p = 0.023 forNH4

+); Table 2, Fig. 2). The TF deposition of K+, Ca2+,NO3

–, and calculated weak acids was higher in the leafed

1362 Can. J. For. Res. Vol. 37, 2007

# 2007 NRC Canada

season, while TF of all other measured ions was higher inthe leafless season. In general, statistical results similar tothose for TF were obtained for NTF and TF:WD. For Mg2+,NTF did not differ significantly between the leafed and leaf-less seasons. TF:WD of Mg2+ was significantly higher in theleafed than in the leafless season (Table 2) because of thestrong Mg2+ increase in WD during the leafless season(130% compared with the leafed season) that was accompa-nied by only a moderate Mg2+ increase in TF (30%). For Cl–

and SO42–, TF:WD did not differ significantly between the

leafed and leafless seasons. The most pronounced temporalpattern was observed for K+, with strongly increasingTF:WD during leaf emergence and leaf senescence (Figs. 2and 3). Almost half (48%) of the annual NTF of K+ oc-curred during leaf emergence and senescence, although theseperiods accounted for only 15% of the study time. TF:WDalso increased for Na+, Mg2+, and Cl– during the short pe-riod of leaf emergence, and for Ca2+ and SO4

2– during leafsenescence (Fig. 3).

Ion deposition to the forest floor by SF was significantlyhigher in the leafless season than in the leafed season for allions other than NO3

– and NH4+ (p < 0.02; Table 2). The

higher SF deposition in the leafless season was due to sig-nificantly higher ion concentrations (p < 0.07) as well ashigher (although not statistically significant) water fluxes inSF. The mean vw ion concentrations of SF were similar inthe leafed and leafless seasons for NO3

–, NH4+, and K+, but

increased in the leafless season for the other ions by 1.7 forSO4

2–, 3.2 for Ca2+, and more than 5 times for Na+, Cl–, andMg2+. The increase in SF concentrations during the leaflessseason was not due to seasonal differences in rainfall con-centration of Ca2+ and SO4

2–, as the mean vw ion concentra-tions in WD were similar in the two periods. For Na+, Cl–,and Mg2+, the vw concentration in WD was 2–2.6 timeshigher in the leafless than in the leafed season, which isless than the fivefold increase in SF. For H+, the vw concen-tration in SF was one order of magnitude higher (13 times)in the leafless season than in the leafed season. In contrastto the negligible H+ input by TF (0.10 mmolc�m–2) and SF(0.02 mmolc�m–2) during the leafed season, higher H+ depo-sitions occurred during the leafless season (0.65 and

0.39 mmolc�m–2 by TF and SF, respectively). During theleafed season, the average vw pH increased from 5.2 inWD to 6.3 in TF (Fig. 4) and to 6.0 in SF. During the leaf-less season, the pH of WD (5.2) increased to only 5.6 in TFand decreased to 4.9 in SF.

Ion correlations in NTF depositionThe relationship between the biweekly NTF deposition of

the sum of measured cations (H+, NH4+, Na+, K+, Ca2+, and

Mg2+) and the sum of anions (measured Cl–, NO3–, SO4

2–,and calculated HCO3

–) clearly varied among the four canopyphases. The anion deficit in NTF, which represents the fluxof unmeasured anions of weak organic acids, was negligibleduring the leafless season, pronounced during leaf emer-gence, and moderate during the full-leaf canopy phase andleaf senescence (Fig. 5). During leaf emergence, the calcu-lated weak-acid flux in NTF was significantly positivelycorrelated with the NTF fluxes of Mg2+ (rs = 0.57, p = 0.05,n = 12) and NH4

+ (rs = 0.89, p < 0.001) (Fig. 6a). Duringthe full-leaf canopy phase, weak acid deposition in NTFwas significantly correlated with all cations other than H+

(p < 0.001, n = 108; rs = 0.81 for NH4+ (Fig. 6a), rs = 0.71

for K+ (Fig. 6b), rs = 0.50 for Ca2+ and Mg2+). The lesserNTF flux of weak acids during leaf senescence was signif-icantly correlated with Mg2+ (rs = 0.74), K+ (rs = 0.66),and Ca2+ (rs = 0.52) (p < 0.001, n = 36; Fig. 5).

The relationship between the concentrations of cationsand inorganic anions also varied during the year. Duringthe short period of leaf emergence, the NTF depositions ofboth Na+ and Cl– increased strongly compared with thoseof the full-leaf and leaf senescence canopy phases (Figs. 2and 7). During the last 2 weeks of leaf senescence, in con-trast, NTF deposition of Cl– increased while NTF deposi-tion of Na+ remained relatively low (Fig. 7). Hence, themean Cl– to Na+ ratio in NTF during this November sam-pling (2.7 molc�molc

–1) was significantly higher than duringthe rest of the year (1.3 molc�molc

–1). The biplots (notshown) suggested that the additional Cl– flux in NTF duringthe last 2 weeks of senescence, calculated from the Cl–:Na+

ratio of the rest of the year, was most closely correlatedwith K+ (rs = 0.76, p = 0.004, n = 12).

Table 1. Annual wet-only deposition (WD), bulk deposition (BD), throughfall (TF, mean ± standard deviation, n = 12),stemflow (SF), and net throughfall and stemflow (NTS) deposition (NTS = TF + SF – WD) of water and major ions, andTF:WD deposition ratio and SF to forest-floor deposition (SF : (TF + SF)).

Annual water (mm) and ion (mmolc�m–2) deposition Ratio

WD* BD TF SF NTS TF:WD SF : (TF + SF)

Water 612.6 612.6 410.4±31.9 51.6 –150.5 0.67 0.11H+ 4.0 1.9 0.6±0.1 0.4 –3.0 0.16 0.38Na+ 30.0 44.3 77.3±7.3 18.1 65.5 2.6 0.19K+ 1.2 2.5 49.6±6.9 5.1 53.4 42.8 0.09Ca2+ 10.9 19.2 47.4±6.6 5.3 41.8 4.3 0.10Mg2+ 6.8 10.2 25.7±3.7 3.3 22.2 3.8 0.11NH4

+ 33.5 36.9 89.5±16.7 15.2 71.2 2.7 0.15NO3

– 18.2 22.7 38.8±4.5 4.1 24.7 2.1 0.10SO4

2– 27.7 35.8 105.0±16.5 14.8 92.1 3.8 0.12Cl– 33.0 50.2 100.4±10.6 23.6 91.1 3.0 0.19Weak acids{ 7.5 6.4 46.0±10.3 4.7 43.2 4.8 0.09

*Calculated from wet-only rainfall concentrations and bulk precipitation amounts.{Calculated as the difference between sum of cations and sum of anions.

Staelens et al. 1363

# 2007 NRC Canada

A distinct seasonal difference was observed in the rela-tionship between the NTF depositions of Ca2+ and SO4

2–

(Fig. 8a). During the full-leaf canopy phase, similar meanNTF amounts of SO4

2– (22 mmolc�m–2) and Ca2+

(17 mmolc�m–2) were measured (regression slope = 1.1, rs =0.83, p < 0.001, n = 108), while the slope of the SO4

2– toCa2+ regression was between 2.2 and 2.9 during the otherthree canopy phases. This confirmed that the TF enrichmentof Ca2+ was highest during the full-leafed canopy phase, incontrast with that of SO4

2– (Fig. 3). An increase in theMg2+:SO4

2– ratio in NTF was observed during leaf emer-gence (rs = 0.78, p = 0.003, n = 12; Fig. 8b), in agreement

with the increased NTS:WD of Mg2+ at that time (Fig. 3).During leaf senescence, the NTF of SO4

2– was correlatedwith those of Mg2+ (rs = 0.84), Ca2+ (rs = 0.70), and K+

(rs = 0.59). During the leafless season, in contrast, theNTF of SO4

2– was closely correlated with that of NH4+

(rs = 0.88, p < 0.001). The NTF fluxes of SO42– and

NH4+ were also closely correlated during the full-leaf can-

opy phase (rs = 0.78, p < 0.001). The biweekly NTF depo-sition of NO3

– was higher during the full-leaf canopyphase than in the other phenological canopy phases(Fig. 2), and was correlated with Ca2+ (rs = 0.89), Mg2+,and NH4

+ (rs = 0.86) in the full-leaf canopy phase (plots

04/03/03 29/04/03 24/06/03 19/08/03 14/10/03 09/12/03 03/02/04

0

5

10

15

20

25Cl-

04/03/03 29/04/03 24/06/03 19/08/03 14/10/03 09/12/03 03/02/04

0

4

8

12

16SO4

2-

0

2

4

6NO3

-

0

4

8

12

16NH4

+

0

1

2

3

4Mg2+

0

2

4

6

8Ca2+

0

2

4

6

8

10K+

0

4

8

12

16Na+

0

0.2

0.4

0.6

0.8

1H+

0

20

40

60

80

100

Wet deposition Throughfall deposition Stemflow deposition

Water

Wate

r(m

m·

2w

eeks

)or

ion

(mm

ol

·m

·2

weeks

)flux

–1

c–2

–1

Fig. 2. Wet-only precipitation, throughfall (± standard deviation, n = 12), and stemflow water (mm�2 weeks–1) and ion fluxes(mmolc�m–2�2 weeks–1) from 4 March 2003 to 4 March 2004. The broken vertical lines separate the phenological canopy phases.

1364 Can. J. For. Res. Vol. 37, 2007

# 2007 NRC Canada

not shown). During leaf senescence, Ca2+ was the only cationthat was weakly correlated (rs = 0.38, p = 0.02) with NO3

–

in NTF.

Estimated DD and CEUsing the canopy budget method at the annual level after

Draaijers and Erisman (1995), CL was estimated to contribute95% for K+, 43% for Ca2+, and 33% for Mg2+ to the annualNTS deposition. The estimated DD of Ca2+ was 24 mmolcm–2�year–1. CE values of Na+, Cl–, SO4

2–, and NO3– were

assumed to be zero, so their NTS depositions were attrib-uted entirely to DD. To account for weak-acid leaching,the CU of NH4

+ was calculated as 29 mmolc m–2�year–1

and the DD of NH4+ as 100 mmolc m–2�year–1. The CU of

H+ was then estimated at 11 mmolc m–2�year–1, indicating aDD of 8.4 mmolc m–2�year–1. If weak-acid leaching from thecanopy is omitted, CUs of 55 mmolc NH4

+�m–2�year–1 and21 mmolc H+�m–2�year–1 were calculated.

Using the modified seasonal canopy budget method, theestimated contributions of CL to the annual NTS were 96%for K+, 54% for Ca2+, 40% for Mg2+, 12% for Cl–, and 7%for Na+ and SO4

2– (Table 3). During leaf burst, estimatedNa+ leaching was 78% of its NTS. CL of K+ contributedmore than 90% of the NTS, irrespective of the canopyphase. For Ca2+ and Mg2+, about 70% of the NTS was at-tributed to CL during the full-leaf canopy phase. Duringthe leafless season, 28% of Ca2+ and 5% of Mg2+ in theNTS was attributed to branch leaching. During leaf burst,estimated CL was low for Ca2+ (22% of NTS) but pro-nounced for Mg2+ (84%), while during leaf senescence,leaching was important for both Ca2+ (86%) and Mg2+

(67%). Similarly to Mg2+, the relative fraction of Cl– leach-ing in NTS was highest during leaf burst (78% of NTS) andimportant during leaf senescence (50%). The calculatedleaching of SO4

2– contributed 39% of the NTS during leaf

senescence. Net CU of NH4+ was estimated at 23 mmolc

m–2�year–1, with a small amount of CL of NH4+ calculated

during leaf burst.

Discussion

TF enrichment in the leafed and leafless seasonsThe TF ion deposition under the European beech tree was

significantly higher than the WD input during both theleafed and leafless seasons for all ions except H+. The an-nual TF and SF depositions of the principal ions (NH4

+,SO4

2–, Na+, and Cl–) deviated less than 12% between thesingle European beech tree and the adjacent level-II plotduring the study period (G. Genouw, unpublished data). Thefinding that leafed as well as leafless forest canopies interactwith atmospheric deposition is in line with other studies intemperate deciduous forest stands. In two Canadian hard-wood forests, TF deposition of all ions other than H+, NO3

–,and NH4

+ exceeded WD during the leafless season, with en-richment ratios varying from 1.04 to 4.7 (Neary and Gizyn1994) and from 1.3 to 7.5 (Houle et al. 1999). In an oak–hickory stand in Georgia — where only 24% of precipitationwas sampled in the 2 year study period — TF enrichment ratiosof ions other than H+ and NH4

+ ranged from 1.4 to 44 during theleafless season (Cappellato et al. 1993). In an oak–hickorystand in Kansas, TF deposition exceeded WD for five of eightelements during the leafless season, but the difference wasonly significant for PO4

3– (Hamburg and Lin 1998).While foliation decreased the net water flux to the forest

floor in the present study, owing to increased interceptionloss (Staelens et al. 2006b), varying seasonal patterns wereobserved for the major ions. The TF:WD ratios were signifi-cantly higher during the leafed season for K+, Ca2+, Mg2+,and NO3

–, but significantly higher during the leafless seasonfor Na+, NH4

+, and H+ (Table 2).

Table 2. Mean biweekly wet deposition (WD), throughfall (TF), stemflow (SF), and TF:WD ratios inthe leafed (L) and the leafless (NL) seasons.

Biweekly water (mm) and ion (mmolc�m–2) deposition Ratio

WD* TF{ SF* TF:WD{

L NL L NL L NL L NL

Water 22.40 25.14 12.88 19.75 1.53 2.61 0.57 0.79H+ 0.15 0.16 0.01 0.05 0.00 0.03 0.04 0.31Na+ 0.64 1.85 1.51 4.97 0.16 1.44 2.35 2.69K+ 0.04 0.05 2.49 1.11 0.14 0.26 58.77 23.39Ca2+ 0.42 0.42 2.09 1.46 0.07 0.38 4.95 3.52Mg2+ 0.17 0.39 0.87 1.15 0.03 0.26 5.25 2.93NH4

+ 1.30 1.27 3.26 3.69 0.50 0.70 2.51 2.90NO3

– 0.78 0.59 1.83 1.03 0.14 0.18 2.35 1.76SO4

2– 1.00 1.16 3.71 4.49 0.31 0.92 3.71 3.88Cl– 0.72 2.01 2.19 6.15 0.22 1.85 3.03 3.05Weak acids{ 0.23 0.37 2.51 0.75 0.23 0.11 11.14 2.01

Note: The leafed season includes leaf-emergence, full-leaf canopy phase, and leaf senescence. Values in boldfacetype indicate whether a deposition flux or ratio was significantly (p < 0.05) higher in the leafed or leafless season ofthe year.*Wilcoxon’s rank sum test between area-averaged deposition per sampling interval in the leafed (n = 15) and

leafless (n = 11) seasons.{Wilcoxon’s signed-rank test between time-averaged biweekly depositions from individual TF collectors (n = 12) in

the leafed and leafless seasons.{Calculated as the difference between sum of cations and sum of anions.

Staelens et al. 1365

# 2007 NRC Canada

Sodium and chlorideNa+ and Cl– are generally considered to be conservative

elements showing only minor CE (Parker 1983; Johnsonand Lindberg 1992; Draaijers et al. 1996). Our results indi-cate CL of Na+ and particularly of Cl–. The NTF depositionof both Na+ and Cl– strongly increased during the short leaf-emergence period compared with the rest of the full-leafcanopy phase (Figs. 3 and 7). Assuming that DD of theseions did not increase strongly in these 2 weeks, both Na+

and Cl– must have leached from the breaking buds and (or)the emerging leaves. During the last biweekly sampling pe-riod of leaf senescence, the high Cl– enrichment was not ac-companied by similar Na+ enrichment (Fig. 7), but wascorrelated with K+ concentrations, suggesting the leachedCl– occurred mainly as KCl. A similar increase in Cl– in TFduring autumn has previously been reported for other hard-wood species (Cronan and Reiners 1983; Neary and Gizyn1994; Houle et al. 1999). Experiments in which CaCl2 wasadded to the soil in sugar maple stands showed that Cl– istaken up, resulting in elevated Cl– levels in both foliage andTF (Berger et al. 2001). It therefore seemed reasonable tomodify the canopy budget method and include Na+ and Cl–

leaching during leaf emergence and senescence. The Cl–:Na+

ratio of 1.32 in NTF during the full-leaf canopy phase andthe leafless season (Fig. 7) was higher than in seawater(1.17; de Vries et al. 2003; UBA 2004). The additionalCl– may be due to DD of MgCl2, HCl, and (or) NH4Claerosols. TF deposition of Na+ was significantly higherthan WD of Na+ throughout the year, even though mean-ingful Na+ leaching most likely only occurred during leafemergence. Furthermore, the TF enrichment ratio of Na+

was higher in the leafless than in the leafed season, inline with previous results (Neary and Gizyn 1994; Houleet al. 1999; Devlaeminck et al. 2005). This demonstratesconsiderable DD of Na+ and Cl– to twigs and branches ofthe leafless tree, likely because of (i) increased atmos-pheric turbulence within the defoliated canopy and gener-ally higher wind speeds in winter (Beckett et al. 2000;

Fig. 3. Ratios of throughfall to wet ion deposition (TF:WD) during four phenological canopy phases.

Fig. 4. Relationship between pH of biweekly WD and TF (area-averaged, n = 12) in the leafed season (including leaf emergence andsenescence) and the leafless season. The lines indicate single linearregressions (y = ax + b) fitted to the seasonally divided data, forwhich the slope (a) is given (leafed season: y = 0.09x + 6.0 (r2 = 0.17,p = 0.18); leafless season: y = 0.80x + 1.5 (r2 = 0.87, p = 0.001)).

Fig. 5. Relationship between net throughfall (NTF) deposition of an-ions and cations for 12 individual TF collectors during four phenolo-gical canopy phases (ntotal = 264). For each period, the slope (a) ofthe linear regression (y = ax + b) is given. The diagonal line indicatesthe 1:1 ratio (leaf emergence: y = 0.56x – 1.6 (r2 = 0.81); full-leaf ca-nopy phase: y = 0.63x + 0.6 (r2 = 0.92); leaf senescence: y = 0.72x +0.8 (r2 = 0.92); leafless season: y = 0.95x – 0.1 (r2 = 0.99)). Allregressions are significant at p < 0.001.

1366 Can. J. For. Res. Vol. 37, 2007

# 2007 NRC Canada

Freer-Smith et al. 2004) and (ii) enhanced atmospheric seaspray in winter, as indicated by the WD pattern of Na+,Cl–, and Mg2+ (Fig. 2, Table 2).

Potassium, calcium, and magnesiumIn contrast to Na+, the TF:WD ratios of K+, Ca2+, and

Mg2+ were significantly higher during the leafed than theleafless season (Table 2), indicating CL of these cationsfrom physiologically active leaves. Similarly, enrichmentratios of K+ and Mg2+ beneath a deciduous Fagus moe-siaca Cz. canopy were significantly higher in the leafedseason (Michopoulos et al. 2001). CL was the main mech-anism of K+ enrichment in TF, in agreement with previousresearch (Parker 1983; Houle et al. 1999; Balestrini and

Tagliaferri 2001). K+ is more susceptible to CL than Ca2+

and Mg2+ are because it is not bound in structural tissuesor enzyme complexes (Marschner 1995). The estimated CLwas higher for Ca2+ than for Mg2+, but the contribution ofCa2+ leaching to NTS deposition was smaller during leafemergence than during senescence (Table 3). This is inline with the fact that Ca occurs in considerable quantitiesin cell walls as relatively insoluble pectates and is rarelyretranslocated from leaves to woody components in autumn(Marschner 1995).

While increased cation leaching during leaf senescence iscommonly reported (Cronan and Reiners 1983; Neary andGizyn 1994; Houle et al. 1999), the effect of bud break andleaf unfolding on TF chemistry has rarely been discussed.Our results suggest that Mg2+, NH4

+, and K+ were leachedduring leaf burst, accompanied by significant leaching of or-ganic acids. Leaching of Mg2+ and K+ from young oakleaves has also been found (Draaijers et al. 1992). The esti-mated leaching of weak organic acids during leaf senescencewas less than leaching during leaf emergence. This suggeststhat base cation leaching from senescent leaves was associ-ated not only with the release of organic anions (Cronan andReiners 1983; Balestrini and Tagliaferri 2001), but also withuptake of NH4

+ or H+ (Draaijers and Erisman 1995; Chiwaet al. 2004) and (or) the release of one or more of the meas-ured anions, such as Cl– and SO4

2–.Significant TF enrichment of base cations was also ob-

served during the leafless season (Table 2). We attributedthe TF enrichment of K+ and Ca2+ concentrations during theleafless season to the washing of DD from the leaves andleaching from the twigs and branches (Tukey 1970; Leviaand Herwitz 2002), while the enrichment of Mg2+ was at-tributed mainly to the wash-off of DD (Table 3) accordingto the modified seasonal canopy budget method.

Nitrogen, hydrogen, and sulphurTF enrichment ratios of NH4

+ and H+ were significantly

Fig. 6. Relationship between NTF deposition of weak organic acids (calculated as the anion deficit) and NH4+ (a) and K+ (b) for 12 indivi-

dual TF collectors during four phenological canopy phases (ntotal = 264). The lines indicate the 1:1 ratio.

Fig. 7. Relationship between NTF deposition of Cl– and Na+ for 12individual TF collectors during four phenological canopy phases(ntotal = 264). The diagonal line indicates the Cl–:Na+ ratio of sea-water (1.166; de Vries et al. 2003).

Staelens et al. 1367

# 2007 NRC Canada

lower during the leafed season than during the leafless sea-son (Table 2), in contrast to the TF enrichment ratios ofNO3

–. A decrease in TF enrichment of inorganic N duringthe leafed season is usually attributed to the retention of Nby canopy leaves and (or) their microbial communities(Neary and Gizyn 1994; Hamburg and Lin 1998). For decid-uous canopies, ion exchange of both NH4

+ (Roelofs et al.1985; Stachurski and Zimka 2002) and H+ (Puckett 1990;Cappellato et al. 1993; Lovett et al. 1996) with K+, Ca2+,and Mg2+ has been reported. During autumn, increased can-opy retention of NH4

+ (Cronan and Reiners 1983; Neary andGizyn 1994; Houle et al. 1999), NO3

–, and H+ (Cronan andReiners 1983; Houle et al. 1999) has been observed in hard-wood stands. Since this uptake occurs at leaf senescence, itis probably linked to N retranslocation by trees (Houle et al.

1999). Increased ion exchange between NH4+ and H+ and

base cations by senescent European beech leaves explainswhy the observed leaching of K+, Ca2+, and Mg2+ in autumndid not lead to an equivalent anion deficit in NTF (Fig. 5,Table 3). Considerable CU of N was also calculated duringthe leafless season. Although the main route of CU is con-sidered to be via the foliage, increasing evidence suggeststhat N uptake via twig, branch, and stem surfaces may beequally important (Bowden et al. 1989; Boyce et al. 1996;Macklon et al. 1996; Wilson and Tiley 1998). The estimatedCL of NH4

+ during leaf emergence (Table 3) was correlatedwith the release of weak acid anions (Fig. 6a). By includingthe estimated leaching of weak acid anions as well as Cl–and SO4

2–, the canopy absorption of NH4+ was calculated to

be 0.33 g N�m–2�year–1. This is in the lower range of values

Fig. 8. Relationship between NTF deposition of SO42– and Ca2+ (a) and Mg2+ (b) for 12 individual TF collectors during four phenological

canopy phases (ntotal = 264). The lines indicate the 1:1 ratio.

Table 3. Wet deposition (WD), estimated dry deposition (DD), and estimated canopy exchange (CE; CE > 0 indicates ca-nopy ion leaching, CE < 0 indicates canopy ion uptake) of ions during four phenological canopy phases and the total studyyear.

Ion flux (mmolc m–2�period–1)

Canopy phase Flux H+ Na+ K+ Ca2+ Mg2+ NH4+ NO3

– SO42– Cl– Weak acids*

Leaf emergence (2) WD 0.1 1.0 0.1 1.3 0.3 3.0 1.6 2.0 1.0 1.2DD –0.4 1.4 0.1 1.7 0.4 5.3 0.0 5.5 1.8 1.2CE 0.3 4.7 7.3 0.5 2.0 2.9 0.0 0.0 6.3 11.4

Full-leaf (21) WD 1.2 4.2 0.4 3.8 1.2 12.8 7.8 9.5 4.3 2.0DD 1.8 5.6 0.6 5.1 1.5 33.2 15.9 22.2 7.7 2.0CE –2.9 0.0 14.2 11.5 4.4 –10.6 0.0 0.0 0.0 16.7

Senescence (7) WD 1.0 4.4 0.1 1.2 1.1 3.6 2.2 3.5 5.6 0.2DD 2.1 3.6 0.1 1.0 0.9 9.8 1.9 10.8 4.8 0.2CE –3.1 0.0 16.7 6.2 1.8 –3.7 0.0 6.9 4.7 6.3

Leafless (22) WD 1.8 20.3 0.5 4.6 4.3 14.0 6.5 12.7 22.1 4.1DD 4.3 50.1 1.3 11.3 10.6 46.1 6.9 46.8 65.8 4.1CE –5.1 0.0 13.3 4.4 0.5 –11.8 0.0 0.0 0.0 1.3

Year (52) WD 4.0 30.0 1.2 10.9 6.8 33.5 18.2 27.7 33.0 7.5DD 7.8 60.7 2.1 19.1 13.4 94.4 24.7 85.2 80.1 7.5CE –10.8 4.7 51.4 22.7 8.8 –23.2 0.0 6.9 11.0 35.7

Note: The number of weeks in each canopy phase is given in parentheses.*Calculated as the difference between sum of cations and sum of anions.

1368 Can. J. For. Res. Vol. 37, 2007

# 2007 NRC Canada

reported in other studies (Brumme et al. 1992; Lovett andLindberg 1993). When leaching of K+, Ca2+, and Mg2+ wasattributed to CU of only NH4

+ and H+ (de Vries et al. 2003),however, the calculated retention of reduced N (NHx) wasmore than twice as high (0.77 g N�m–2�year–1).

NO3– enrichment of TF occurred in both the leafed and

leafless seasons but was significantly higher during theleafed season (Fig. 2, Table 2). CL or CU of NO3

– was as-sumed to be negligible, since there is no generally acceptedmethod for estimating NO3

– exchange within the canopy di-rectly from TF and precipitation measurements. NO3

– can betaken up by tree canopies, although NH4

+ is retained and as-similated more preferentially than NO3

– (Bowden et al.1989; Brumme et al. 1992; Boyce et al. 1996; Stachurskiand Zimka 2002). The TF enrichment of NO3

– in the leafedseason may be attributed mainly to the DD of NO2, HNO3,and particulate NO3

– (Hanson and Lindberg 1991; Duyzerand Fowler 1994; Gessler et al. 2000). The lower TF enrich-ment of NO3

– in the leafless than in the leafed season waslikely not found for NH4

+ (and SO42–) because NH3 is taken

up not only through stomata but also by moist surfaces(Wesely and Hicks 2000) and because of preferential NHxuptake by the leafed canopy (e.g., Stachurski and Zimka2002). Despite the uncertainty concerning N retention withinthe canopy, the study tree is clearly exposed to elevatedacidifying deposition and N deposition. The chronic increasein N and S deposition to forests in northern Belgium hascaused significant soil acidification (De Schrijver et al.2006) and NO3

– leaching below the rooting zone (Neiryncket al. 2002; De Schrijver et al. 2004).

Our results confirm that deciduous forest canopies exhibitsubstantial pH buffering during the leafed season and lesspH buffering during the leafless season (Houle et al. 1999;Michopoulos et al. 2001; Pryor and Barthelmie 2005). Basedon TF only, the canopy neutralized 96% of the free rainfallacidity during the leafed season and 69% during the leaflessseason. However, when the SF flux of H+ is included, only48% of the wet H+ input was neutralized in the leafless sea-son. The fraction of neutralized H+ is actually higher be-cause of the DD of H+ originating from H2SO4,(NH4)HSO4, HNO3, and HCl (Draaijers and Erisman 1995).The estimated DD and uptake of H+ (and NH4

+) (Table 3)depends on the CU equation (eq. 2) and exchange effi-ciency, the impact of which is much greater for H+ than forNH4

+ (J. Staelens, unpublished data). The reduction of rain-fall acidity after canopy passage is generally attributed tothe removal of free H+ by ion exchange by the canopy or toorganic anion leaching from the foliage (Cronan and Reiners1983), but it can also be due to proton consumption by dry-deposited NH3 gas (De Schrijver et al. 2004) and Ca2+ com-pounds (Lee and Longhurst 1992) or CU of HNO3 (Cappel-lato et al. 1993; Stachurski and Zimka 2002).

The TF:WD ratio of SO42– did not differ significantly be-

tween the leafed and leafless seasons (Table 2). When CUand CL of S are assumed to be negligible (Lindberg andGarten 1988; Butler and Likens 1995; Kovacks and Hor-vath 2004), considerable amounts of DD of S on boththe leafed (0.75 g�m–2) and leafless (0.72 g�m–2) canopies(1.47 g�m–2�year–1) are indicated. Some DD of S originatesfrom particulate SO4

2– but most originates from SO2, therate of which is strongly affected by stomatal resistance

(Wesely and Hicks 2000). Therefore, lower S enrichmentwas expected under leafless conditions. However, the dep-osition of SO2 onto leafless forests may be considerablyunderestimated based on stomatal uptake when ambientNH3 concentrations are sufficiently large because of the in-creased codeposition of SO2 and NH3 under humid condi-tions (Wesely and Hicks 2000). For example, Cape et al.(1998) found that fumigation of a Scots pine canopy withNH3 notably increased S deposition. The dissolution ofNH3 in a water film increases the pH, which promotes theDD of SO2 and results in the formation of (NH4)2SO4(Derome et al. 2004). The high TF enrichment of bothNH4

+ and SO42– (Table 2) and the close correlation be-

tween the NTFs of NH4+ and SO4

2– suggests increased co-deposition of NH3 and SO2. The molar N:S ratio in theestimated DD, accounting for CU of N and leaching of S(Table 3), was 2.71 in the leafed season and 1.97 in theleafless season.

Although the TF enrichment of SO42– beneath various

canopy types is generally attributed to the DD of S, bothcanopy absorption (Cappellato et al. 1993; Quilchano et al.2002) and leaching of S (Lovett and Lindberg 1984; Nearyand Gizyn 1994; Cappellato et al. 1998) have been reportedfor leafed deciduous canopies. SO4

2– enrichment increasedduring leaf senescence in the present study (Fig. 3). Assumingno increased DD of SOx during this period, 0.11 g S�m–2�year–1

was leached from the senescing leaves, which decreased theestimated DD slightly in the leafed season to 0.61 g S�m–2

(Table 3). This estimated CL contributed 33% of the total Sdeposition to the forest floor during senescence, but only6% of the S deposition over the 1 year study period. CLof S during senescence has also been reported for a beechforest in southern Sweden (Staaf 1982). Observations byMeiwes and Khanna (1981) suggest that 0.22 g S�m–2�year–1

could be leached from senescent European beech leaves.

StemflowSF contributed significantly to the annual fluxes of water

(11%) and ions (9%–19%), particularly H+ (38%), beneaththe single European beech tree canopy. In German beechforests, SF accounted for 15% (Chang and Matzner 2000a)and 37%–66% (Sah 1990) of the annual forest-floor deposi-tion of H+. The high contribution of H+ by SF in the presentstudy was entirely due to low pH values during the leaflessseason, when the SF was more acidic than the WD. As SFinfiltrates a small area around the trunk, the substantial SFfluxes of smooth-barked tree species such as Europeanbeech can contribute significantly to spatial patterns of waterpercolation under the rooting zone (Chang and Matzner2000a) and increase soil acidification (Chang and Matzner2000a; Matschonat and Falkengren-Grerup 2000), nitrifica-tion (Chang and Matzner 2000b), and mineral weathering(Rampazzo and Blum 1992) close to the European beechtrees.

SF concentrations of most ions were higher in the leaflessseason than in the leafed season, and could not be explainedby increased rainfall concentrations. According to Levia andFrost (2003), increased chemical enrichment of SF in wintermay be due to exposure of a larger portion of the woodytree crown to incident precipitation, which increases barkleaching, and to increased residence times of SF water on

Staelens et al. 1369

# 2007 NRC Canada

the bark surface by lower air temperatures and lower rainfallintensities. However, increased concentrations of K+ in SFduring the leafless season were not observed, which suggeststhat increased Na+, Cl–, and Mg2+ concentrations were dueto the efficient DD of aerosols derived from sea salt (Beckeret al. 2000; Freer-Smith et al. 2004) rather than to increasedion leaching.

ConclusionsThe chemical enrichment of TF and SF deposition be-

neath a deciduous European beech tree compared with thatof WD above the tree was clearly influenced by foliationand physiological canopy activity. TF:WD ratios were sig-nificantly higher during the leafed season for K+, Ca2+,Mg2+, and NO3

–, but significantly lower during the leaflessseason for Na+, NH4

+, and H+. The high enrichment ratiosin the leafless season indicated that K+ and Ca2+ wereleached from European beech branches and twigs by rain-fall. Furthermore, DD of sea salt derived aerosols, S, and re-duced N onto the leafless canopy was at least as high asonto the foliated canopy. The excretion of weak acid anionsby leaves, derived from the anion deficit in NTF, was high-est during leaf emergence and the full-leaf canopy phase,and should be taken into account when studying ion ex-change within forest canopies. The present research indicatesthat Na+, Cl–, and SO4

2– are not always conservative withrespect to the canopy, as is commonly assumed. Therefore,an existing canopy budget method was modified to includethe leaching of Na+, Cl–, and SO4

2– from emerging andsenescent European beech leaves.

AcknowledgementsWe thank Luc Willems and Greet De bruyn for assisting

in the laboratory work. Philip Van Avermaet of the FlemishEnvironment Agency is acknowledged for kindly granting uspermission to use the wet-only samplers. The first authorwas funded as a postdoctoral fellow of the Research Foun-dation - Flanders (FWO). We are grateful to two anonymousreviewers and the associate editor for valuable remarks onan earlier version of the manuscript.

ReferencesAndersen, H.V., and Hovmand, M.F. 1999. Review of dry deposi-

tion measurements of ammonia and nitric acid to forest. For.Ecol. Manage. 114: 5–18.

Balestrini, R., and Tagliaferri, A. 2001. Atmospheric depositionand canopy exchange processes in alpine forest ecosystems(northern Italy). Atmos. Environ. 35: 6421–6433. doi:10.1016/S1352-2310(01)00350-8.

Becker, J.S., Bellis, D., Staton, I., McLeod, C.W., and Dombovari,J. 2000. Determination of trace elements including platinum intree bark by ICP mass spectrometry. Fresenius’ J. Anal. Chem.368: 490–495. doi:10.1007/s002160000539.

Beckett, K.P., Freer-Smith, P.H., and Taylor, G. 2000. Particulatepollution capture by urban trees: effect of species and wind-speed. Global Change Biol. 6: 995–1003. doi:10.1046/j.1365-2486.2000.00376.x.

Beier, C., Hansen, K., and Gundersen, P. 1993. Spatial variabilityof throughfall fluxes in a spruce forest. Environ. Pollut. 81:257–267. doi:10.1016/0269-7491(93)90208-6. PMID:15091811.

Belanger, N., Pare, D., and Courchesne, F. 2004. Regression equa-

tions for estimating throughfall nutrient fluxes using wet deposi-tion data and their applicability for simulating the soil acid–basestatus using the dynamic forest soil–atmosphere model SAFE.Ecol. Modell. 175: 151–167. doi:10.1016/j.ecolmodel.2003.10.013.

Berger, T.W., Eager, C., Likens, G.E., and Stingeder, G. 2001. Ef-fects of calcium and aluminium chloride additions on foliar andthroughfall chemistry in sugar maples. For. Ecol. Manage. 149:75–90.

Bowden, R.D., Geballe, G.T., and Bowden, W.B. 1989. Foliar up-take of 15N from simulated cloud water by red spruce (Picea ru-bens) seedlings. Can. J. For. Res. 19: 382–386.

Boyce, R.L., Friedland, A.J., Chamberlain, C.P., and Poulson, S.R.1996. Direct canopy nitrogen uptake from N-15-labeled wet de-position by mature red spruce. Can. J. For. Res. 26: 1539–1547.

Brumme, R., Leimcke, U., and Matzner, E. 1992. Interception anduptake of NH4 and NO3 from wet deposition by abovegroundparts of young beech (Fagus sylvatica L.) trees. Plant Soil, 142:273–279. doi:10.1007/BF00010972.

Butler, T.J., and Likens, G.E. 1995. A direct comparison ofthroughfall plus stemflow to estimates of dry and total deposi-tion for sulfur and nitrogen. Atmos. Environ. 29: 1253–1265.doi:10.1016/1352-2310(94)00339-M.

Cape, J.N., Sheppard, L.J., Binnie, J., and Dickinson, A.L. 1998.Enhancement of the dry deposition of sulphur dioxide to a forestin the presence or ammonia. Atmos. Environ. 32: 519–524.doi:10.1016/S1352-2310(97)00202-1.

Cappellato, R., Peters, N.E., and Ragsdale, H.L. 1993. Acidic at-mospheric deposition and canopy interactions of adjacent decid-uous and coniferous forests in the Georgia Piedmont. Can. J.For. Res. 23: 1114–1125.

Cappellato, R., Peters, N.E., and Meyers, T.P. 1998. Above-ground sulfur cycling in adjacent coniferous and deciduous for-ests and watershed sulfur retention in the Georgia Piedmont,U.S.A. Water Air Soil Pollut. 103: 151–171. doi:10.1023/A:1004902816609.

Chang, S.-C., and Matzner, E. 2000a. The effect of beech stemflowon spatial patterns of soil solution chemistry and seepage fluxesin a mixed beech/oak stand. Hydrol. Processes, 14: 135–144.doi:10.1002/(SICI)1099-1085(200001)14:1<135::AID-HYP915>3.0.CO;2-R.

Chang, S.-C., and Matzner, E. 2000b. Soil nitrogen turnover inproximal and distal stem areas of European beech trees. PlantSoil, 218: 117–125. doi:10.1023/A:1014931625707.

Chiwa, M., Crossley, A., Sheppard, L.J., Sakugawa, H., and Cape,J.N. 2004. Throughfall chemistry and canopy interactions in aSitka spruce plantation sprayed with six different simulated pol-luted mist treatments. Environ. Pollut. 127: 57–64. doi:10.1016/S0269-7491(03)00259-8. PMID:14553995.

Cronan, C.S., and Reiners, W.A. 1983. Canopy processing of acidicprecipitation by coniferous and hardwood forests in New Eng-land. Oecologia, 59: 216–223. doi:10.1007/BF00378839.

Derome, J., Nieminen, T., and Saarsalmi, A. 2004. Sulphur dioxideadsorption in Scots pine canopies exposed to high ammoniaemissions near a Cu–Ni smelter in SW Finland. Environ. Pollut.129: 79–88. doi:10.1016/j.envpol.2003.09.021. PMID:14749072.

De Schrijver, A., Nachtergale, L., Staelens, J., Luyssaert, S., andDe Keersmaeker, L. 2004. Comparison of throughfall and soilsolution chemistry between a high-density Corsican pine standand a naturally regenerated silver birch stand. Environ. Pollut.131: 93–105. PMID:15210279.

De Schrijver, A., Mertens, J., Geudens, G., Staelens, J., Campforts,E., Luyssaert, S., De Temmerman, L., De Keersmaeker, L., DeNeve, S., and Verheyen, K. 2006. Acidification of forested pod-

1370 Can. J. For. Res. Vol. 37, 2007

# 2007 NRC Canada

zols in North Belgium during the period 1950–2000. Sci. TotalEnviron. 361: 189–195. PMID:16061272.

Devlaeminck, R., De Schrijver, A., and Hermy, M. 2005. Variationin throughfall deposition across a deciduous beech (Fagus sylva-tica L.) forest edge in Flanders. Sci. Total Environ. 337: 241–252.PMID:15626394.

de Vries, W., Reinds, G.J., van der Salm, C., Draaijers, G.P.J.,Bleeker, A., Erisman, J.W., Auee, J., Gundersen, P., Kristensen,H.L., van Dobben, H., de Zwart, D., Derome, J., Voogd, J.C.H.,and Vel, E.M. 2001. Intensive monitoring of forest ecosystemsin Europe. European Commission — United Nations EconomicCommission for Europe Tech. Rep. Brussels and Geneva.

de Vries, W., Reinds, G.J., and Vel, E. 2003. Intensive monitoringof forest ecosystems in Europe 2: atmospheric deposition and itsimpacts on soil solution chemistry. For. Ecol. Manage. 174: 97–115.

Draaijers, G.P.J., and Erisman, J.W. 1995. A canopy budget modelto assess atmospheric deposition from troughfall measurements.Water Air Soil Pollut. 85: 2253–2258. doi:10.1007/BF01186169.

Draaijers, G.P.J., van Ek, R., and Meijers, R. 1992. Research onthe impact of forest stand structure on atmopsheric deposition.Environ. Pollut. 75: 243–249. doi:10.1016/0269-7491(92)90046-D. PMID:15092040.

Draaijers, G.P.J., Erisman, J.W., Spranger, T., and Wyers, G.P.1996. The application of throughfall measurements for atmo-spheric deposition monitoring. Atmos. Environ. 30: 3349–3361.doi:10.1016/1352-2310(96)00030-1.

Draaijers, G.P.J., Erisman, J.W., Van Leeuwen, N.F.M., Romer,F.G., te Winkel, B.H., Veltkamp, A.C., Vermeulen, A.T., andWyers, G.P. 1997. The impact of canopy exchange on differ-ences observed between atmospheric deposition and throughfallfluxes. Atmos. Environ. 31: 387–397. doi:10.1016/S1352-2310(96)00164-1.

Driscoll, C.T., Fuller, R.D., and Schecher, W.D. 1989. The role oforganic acids in the acidification of surface waters in the easternUS. Water Air Soil Pollut. 43: 21–40. doi:10.1007/BF00175580.

Duyzer, J., and Fowler, D. 1994. Modelling land atmosphere ex-change of gaseous oxides of nitrogen in Europe. Tellus Ser. BChem. Phys. Meterol. 46: 353–372.

Erisman, J.W., and Draaijers, G. 2003. Deposition to forests in Eur-ope: most important factors influencing dry deposition and modelsused for generalisation. Environ. Pollut. 124: 379–388. doi:10.1016/S0269-7491(03)00049-6. PMID:12758019.

Erisman, J.W., Mols, H., Fonteijn, P., Geusebroek, M., Draaijers,G., Bleeker, A., and van der Veen, D. 2003. Field intercompari-son of precipitation measurements performed within the frame-work of the Pan European Intensive Monitering Program ofEU/ICP Forest. Environ. Pollut. 125: 139–155. doi:10.1016/S0269-7491(03)00082-4. PMID:12810309.

Freer-Smith, P.H., El-Khatib, A.A., and Taylor, G. 2004. Captureof particulate pollution by trees: a comparison of species typicalof semi-arid areas with European and North American species.Water Air Soil Pollut. 155: 173–187. doi:10.1023/B:WATE.0000026521.99552.fd.

Gessler, A., Rienks, M., and Rennenberg, H. 2000. NH3 and NO2

fluxes between beech trees and the atmosphere — correlationwith climatic and physiological parameters. New Phytol. 147:539–560. doi:10.1046/j.1469-8137.2000.00712.x.

Hamburg, S.P., and Lin, T.C. 1998. Throughfall chemistry of anecotonal forest on the edge of the Great Plains. Can. J. For.Res. 28: 1456–1463. doi:10.1139/cjfr-28-10-1456.

Hansen, K. 1996. In-canopy throughfall measurements of ion fluxesin Norway spruce. Atmos. Environ. 30: 4065–4076. doi:10.1016/1352-2310(95)00444-0.

Hansen, K., Draaijers, G.P.J., Ivens, W.P.M.F., Gundersen, P., andvan Leeuwen, N.F.M. 1994. Concentration variations in rain andcanopy throughfall collected sequentially during individual rainevents. Atmos. Environ. 28: 3195–3205. doi:10.1016/1352-2310(94)00176-L.

Hanson, P.J., and Lindberg, S.E. 1991. Dry deposition of reactivenitrogen compounds: a review of leaf, canopy and non-foliarmeasurements. Atmos. Environ. Ser. A, 25: 1615–1634.

Harrison, A.F., Schulze, E.D., Gebauer, G., and Bruckner, G. 2000.Canopy uptake and utilization of atmospheric pollutant nitrogen.In Carbon and nitrogen cycling in European forest ecosystems.Edited by E.-D. Schulze. Springer-Verlag, Berlin. pp. 172–188.

Houle, D., Ouimet, R., Paquin, R., and Laflamme, J.-G. 1999. In-teractions of atmospheric deposition with a mixed hardwoodand coniferous forest canopy at the Lake Clair watershed (Duch-esnay, Quebec). Can. J. For. Res. 29: 1944–1957. doi:10.1139/cjfr-29-12-1944.

Johnson, D.W., and Lindberg, S.E. (Editors). 1992. Atmosphericdeposition and forest nutrient cycling. Ecol. Stud. 91.

Kovacks, E.A., and Horvath, L. 2004. Determination of sulfur bal-ance between the atmosphere and a Norway spruce forest eco-system: comparison of gradient dry + wet and throughfalldeposition measurements. J. Atmos. Chem. 48: 235–240.

Lee, D.S., and Longhurst, A. 1992. A comparison between wet andbulk deposition at an urban site in the UK. Water Air Soil Pol-lut. 64: 635–648. doi:10.1007/BF00483372.

Levia, D.F., and Frost, E.E. 2003. A review and evaluation ofstemflow literature in the hydrologic and biogeochemical cyclesof forested and agricultural ecosystems. J. Hydrol. 274: 1–29.doi:10.1016/S0022-1694(02)00399-2.

Levia, D.F., and Herwitz, S.R. 2002. Winter chemical leachingfrom deciduous tree branches as a function of branch inclinationangle in central Massachusetts. Hydrol. Process. 16: 2867–2879.doi:10.1002/hyp.1077.

Lindberg, S.E., and Garten, C.T., Jr. 1988. Sources of sulphur inforest canopy throughfall. Nature (London), 336: 148–151.doi:10.1038/336148a0.

Lovett, G.M., and Lindberg, S.E. 1984. Dry deposition and canopyexchange in a mixed oak forest as determined by analysis ofthroughfall. J. Appl. Ecol. 21: 1013–1027.

Lovett, G.M., and Lindberg, S.E. 1993. Atmospheric depositionand canopy interactions of nitrogen in forests. Can. J. For. Res.23: 1603–1616.

Lovett, G.M., Reiners, W.A., and Olson, R.K. 1989. Factors con-trolling throughfall chemistry in a balsam fir canopy — a mod-elling approach. Biogeochemistry, 8: 239–264.

Lovett, G.M., Nolan, S.S., Driscoll, C.T., and Fahey, T.J. 1996.Factors regulating throughfall flux in a New Hampshire forestedlandscape. Can. J. For. Res. 26: 2134–2144.

Macklon, A.E.S., Sheppard, L.J., Sim, S., and Leith, I.D. 1996. Up-take of ammonium and nitrate ions from acid mist applied toSitka spruce (Picea sitchensis (Bong.) Carr.) grafts over thecourse of one growing season. Trees Struct. Funct. 10: 261–267.

Marschner, H. 1995. Mineral nutrition of higher plants. 2nd ed.Academic Press, San Diego.

Matschonat, G., and Falkengren-Grerup, U. 2000. Recovery of soilpH, cation exchange capacity and the saturation of exchangesites from stemflow induced soil acidification in three Swedishbeech (Fagus sylvatica L.) forests. Scand. J. For. Res. 15: 39–48.

Meiwes, K.J., and Khanna, P.K. 1981. Distribution and cycling ofsulphur in the vegetation of two forest ecosystems in an acidrain environment. Plant Soil, 60: 369–375. doi:10.1007/BF02149633.

Michopoulos, P., Baloutsos, G., Nakos, G., and Economou, A.

Staelens et al. 1371

# 2007 NRC Canada

2001. Effects of bulk precipitation pH and growth period on ca-tion enrichment in precipitation beneath the canopy of a beech(Fagus moesiaca) forest stand. Environ. Pollut. 281: 79–85.

Morris, D.M., Gordon, A.G., and Gordon, A.M. 2003. Patterns ofcanopy interception and throughfall along a topographic se-quence for black spruce dominated forest ecosystems in north-western Ontario. Can. J. For. Res. 33: 1046–1060. doi:10.1139/x03-027.

Mussche, S., Samson, R., Nachtergale, L., De Schrijver, A., Le-meur, R., and Lust, N. 2001. A comparison of optical and directmethods for monitoring the seasonal dynamics of leaf area indexin deciduous forests. Silva Fenn. 35: 373–384.

Neary, A.J., and Gizyn, W.I. 1994. Throughfall and stemflowchemistry under deciduous and coniferous forest canopies insouth-central Ontario. Can. J. For. Res. 24: 1089–1100.

Neirynck, J., Van Ranst, E., Roskams, P., and Lust, N. 2002. Im-pact of decreasing throughfall depositions on soil solution chem-istry at coniferous monitoring sites in northern Belgium. For.Ecol. Manage. 160: 127–142.

Parker, G.G. 1983. Throughfall and stemflow in the forest nutrientcycle. Adv. Ecol. Res. 13: 57–133.

Potter, C.S., Ragsdale, H.L., and Swank, W.T. 1991. Atmosphericdeposition and foliar leaching in a regenerating southern Appa-lachian forest canopy. J. Ecol. 79: 97–115.

Pryor, S.C., and Barthelmie, R.J. 2005. Liquid and chemical fluxesin precipitation, throughfall and stemflow: observations from adeciduous forest and a red pine plantation in the midwesternU.S.A. Water Air Soil Pollut. 163: 203–227. doi:10.1007/s11270-005-0278-x.

Puckett, L. 1990. Estimates of ion sources in deciduous and coni-ferous throughfall. Atmos. Environ. Ser. A, 24: 545–555.

Quevauviller, P., Van Renterghem, D., Griepink, B., Reijnders,H.F.R., and Van der Jagt, H. 1993. The certification of the con-tents (amount of substance contents) of ammonium, calcium,chloride, hydronium, magnesium, nitrate, potassium, sodiumand sulphate in simulated rainwater (low content CRM 408)(high content CRM 409). Commission of the European Commu-nities, Community Bureau of Reference. Final Rep. EUR 15024EN.

Quilchano, C., Haneklaus, S., Gallardo, J.F., Schnug, E., and Mor-eno, G. 2002. Sulphur balance in a broadleaf, non-polluted forestecosystem (central-western Spain). For. Ecol. Manage. 161:205–214.

Rampazzo, N., and Blum, W. 1992. Changes in chemistry andmineralogy of forest soils by acid rain. Water Air Soil Pollut.61: 209–220. doi:10.1007/BF00482605.

Reiners, W.A., and Olson, R.K. 1984. Effects of canopy compo-nents on throughfall chemistry: an experimental analysis. Oeco-logia, 63: 320–330. doi:10.1007/BF00390660.

Roelofs, J.G.M., Kempers, A.J., Houdijk, A.L.F.M., and Jansen, J.1985. The effect of air-borne ammonium sulphate on Pinus nigravar. maritima in the Netherlands. Plant Soil, 84: 45–56.doi:10.1007/BF02197866.

Sah, S.P. 1990. Vergleich des Staffhaushaltes zweier Buchenwaldo-kosysteme auf Kalkgestein und auf Buntsandstein. Ber. Forsch.Waldoekosyst. Waldsterben Univ. Goett.. Reihe A, 59: 1–140.

Schaefer, D.A., Reiners, W.A., and Olson, R.K. 1988. Factors con-trolling the chemical alteration of throughfall in a subalpine bal-sam fir canopy. Environ. Exp. Bot. 28: 175–189. doi:10.1016/0098-8472(88)90027-5.

Staaf, H. 1982. Plant nutrient changes in beech leaves during se-nescence as influenced by site characteristics. Acta Oecol. 3:161–170.

Stachurski, A., and Zimka, J.R. 2002. Atmospheric deposition andionic interactions within a beech canopy in the KarkonoszeMountains. Environ. Pollut. 118: 75–87. doi:10.1016/S0269-7491(01)00238-X. PMID:11996385.

Staelens, J., De Schrijver, A., Van Avermaet, P., Genouw, G., andVerhoest, N. 2005. A comparison of bulk and wet-only deposi-tion at two adjacent sites in Melle (Belgium). Atmos. Environ.39: 7–15. doi:10.1016/j.atmosenv.2004.09.055.

Staelens, J., De Schrijver, A., Verheyen, K., and Verhoest, N.E.C.2006a. Spatial variability and temporal stability of throughfalldeposition under beech (Fagus sylvatica L.) in relationship tocanopy structure. Environ. Pollut. 142: 254–263. doi:10.1016/j.envpol.2005.10.002. PMID:16338040.

Staelens, J., De Schrijver, A., Verheyen, K., and Verhoest, N.E.C.2006b. Spatial variability and temporal stability of throughfallwater under a dominant beech (Fagus sylvatica L.) tree in rela-tionship to canopy cover. J. Hydrol. 330: 651–662.

Tukey, H.B. 1970. The leaching of substances from plants. Annu.Rev. Plant Physiol. 21: 305–324. doi:10.1146/annurev.pp.21.060170.001513.

Umweltbundesamt (UBA). 2004. Mapping manual 2004. Manualon methodologies and criteria for modelling and mapping criti-cal loads and levels and air pollution effects, risks and trends.Available from www.icpmapping.org [accessed 25 May 2005].

Wesely, M.L., and Hicks, B.B. 2000. A review of the current statusof knowledge on dry deposition. Atmos. Environ. 34: 2261–2282.doi:10.1016/S1352-2310(99)00467-7.

Whelan, M.J., Sanger, L.J., Baker, M., and Anderson, J.M. 1998.Spatial patterns of throughfall and mineral ion deposition in alowland Norway spruce (Picea abies) plantation at the plotscale. Atmos. Environ. 32: 3493–3501. doi:10.1016/S1352-2310(98)00054-5.

Wilson, E.J., and Tiley, C. 1998. Foliar uptake of wet-depositednitrogen by Norway spruce: an experiment using 15N. Atmos.Environ. 32: 513–518. doi:10.1016/S1352-2310(97)00042-3.

1372 Can. J. For. Res. Vol. 37, 2007

# 2007 NRC Canada