Science of the Total Environment 506–507 (2015) 95–101

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Downwash of atmospherically deposited trace metals in peat and theinfluence of rainfall intensity: An experimental test

Sophia V. Hansson ⁎, Julie Tolu, Richard BindlerDepartment of Ecology and Environmental Science, Umeå University, SE-901 87 Umeå, Sweden

H I G H L I G H T S

• 4 rainfall treatments, including pollutant metals, were added to 13 peat cores.• The depth at which added metals was detected increased in the order Pb b Cu b Zn ≤ Ni.• Higher precipitation leads to deeper downwashing of metals.• Downwashing exerts a strong influence on the vertical distribution of metals.

⁎ Corresponding author. Tel.: +46 90 786 7101; fax: +E-mail address: sophia.v.hansson@gmail.com (S.V. Ha

http://dx.doi.org/10.1016/j.scitotenv.2014.10.0830048-9697/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 8 August 2014Received in revised form 23 October 2014Accepted 23 October 2014Available online 15 November 2014

Editor: Charlotte Poschenrieder

Keywords:DepositionDownwashElemental mobilityGeochemistryPeat

Accumulation records of pollutant metals in peat have been frequently used to reconstruct past atmosphericdeposition rates. While there is good support for peat as a record of relative changes in metal deposition overtime, questions remain whether peat archives represent a quantitative or a qualitative record. Several processescan potentially influence the quantitative record ofwhich downwashing is particularly pertinent as itwould havea direct influence on how and where atmospherically deposited metals are accumulated in peat. The aim ofour study was two-fold: first, to compare and contrast the retention of dissolved Pb, Cu, Zn and Ni in peatcores; and second, to test the influence of different precipitation intensities on the potential downwashingof metals. We applied four ‘rainfall’ treatments to 13 peat cores over a 3-week period, including both daily(2 or 5.3 mm day−1) and event-based additions (37 mm day−1, added over 1 h or over a 10 h rain event).Two main trends were apparent: 1) there was a difference in retention of the added dissolved metals in thesurface layer (0–2 cm): 21–85% for Pb, 18–63% for Cu, 10–25% for Zn and 10–20% for Ni. 2) For all metals andboth peat types (sphagnum lawn and fen), the addition treatments resulted in different downwashing depths,i.e., as the precipitation-addition increased so did the depth at which added metals could be detected. Althoughthe largest fraction of Pb and Cu was retained in the surface layer and the remainder effectively immobilized inthe upper peat (≤10 cm), there was a smearing effect on the overall retention, where precipitation intensity ex-erts an influence on the vertical distribution of added tracemetals. These results indicate that the relative positionof a deposition signal in peat records would be preserved, but it would be quantitatively attenuated.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Accumulation records of trace metals in peat cores have beenused in numerous studies to reconstruct past atmospheric metal de-position rates, not only for lead (Pb) – the most studied pollutantmetal – but also for a suite of other pollutant metals, e.g., copper(Cu), nickel (Ni), antimony (Sb) and zinc (Zn) (Shotyk, 1996; Cloyet al., 2005; Pontevedra-Pombal et al., 2013). Based on the generaltemporal coherence between peat and other natural archives, especiallyfor the record of Pb accumulation (Renberg et al., 2001), there is good

46 90 786 6705.nsson).

evidence that peat cores are reliable archives for studying relativechanges in atmospheric metal deposition over centennial to millennialtimescales (Shotyk, 1998; Martinez-Cortizas et al., 1999). Comparisonsof the stableisotopic composition of Pb accumulated in peat coresagainst herbaria samples with known dates of collection (spanningthe past 100–150 years) also demonstrate that peat tracks these gen-eral changes in deposition over shorter (decadal) timescales (Weisset al., 1999; Farmer et al., 2002). For other pollutant metals the utilityof peat has been justified largely on the co-variation with Pb overtime (e.g., Krachler et al., 2003; Cloy et al., 2005). While such studiesprovide strong support for peat as a record of relative changes inmetal deposition over time, questions remain regarding the temporalresolution possible with peat and the extent to which the accumulationrecords of different metals in peat represent absolute, quantitative

Table 1Recipe per liter stock artificial rain water as described by Wiederet al. (2010). In total 11 L stock synthetic rain was prepared andthe pH of the rainwater was adjusted to 4.7 using NaOH and H2SO4

to approximate the precipitation chemistry of the Umeå-region.

Compound Quantity (1 L)

(NH4)2SO4 0.5965 gNH4HCO3 0.5453 gCa(NO3)2 0.6973 gMg(NO3)2 ∗ 6H2O 0.4105 gNaNO3 0.0764 gKNO3 0.0818 gNH4NO3 0.1520 gNH4Cl 0.0588 gNaOH 0.8586 gH2SO4 221 μL

96 S.V. Hansson et al. / Science of the Total Environment 506–507 (2015) 95–101

records rather than only a qualitative record of past deposition (Bindleret al., 2004; Biester et al., 2007).

Some of the key processes that have been discussed as potentiallyinfluencing the quantitative record of metal accumulation rates includenutrient cycling (Malmer, 1988), decomposition (Biester et al., 2012)and post-depositional remobilization (Urban et al., 1990), spatialvariability of deposition (Bindler et al., 2004), and downwashing(Damman, 1978; Urban et al., 1990; Oldfield et al., 1995; Biester et al.,2007; Hansson et al., 2014b). Downwashing, our main interest here, isparticularly pertinent because it would have a direct influence on howand where atmospherically deposited metals are accumulated in thepeat record. We define downwashing here as the rapid percolation ofelements, deposited on the mire surface via rainfall, through the upperlayers of the peat.

Based on a recent study of the natural, short-lived radioisotopeberyllium-7 (7Be; t 1/2 = 53.4 day) in 5 peat cores from two sites inSweden (Hansson et al., 2014a), we found that 25–80% of the total in-ventory was located in the upper 2-cm layer, but that the atmospheri-cally supplied 7Be could be detected from 4 cm to as much as 20 cmdepth. Given the short-half life for 7Be, this required that the 7Be wastransported recently to the sub-surface peat. Our objective in thisstudy was to further test the importance of downwashing for tracemetals in peat, beyond that of 7Be, based on a controlled experimentalsetup. We used a set of 13 peat cores collected from Rödmossamyran,a site we have studied both for radiometric tracers (7Be, 210Pb, 137Csand 241Am) (Hansson et al., 2014a,b) and Hg (Rydberg et al., 2010).

Our first aim was to compare and contrast the retention or down-washing of dissolved Pb, Cu, Zn and Ni in peat cores. As noted above,the immobility or mobility of trace metals in peat records is oftenargued on the basis of similarities or dissimilarities to the distributionof Pb. Previous experimental studies of, e.g., Pb and Be suggest differentpotentials for downwashing, where 50% of the added Be (Wieder et al.,2010) and 30% of added Pb (Vile et al., 1999) were retained in the top2-cm layer, with the remainder of the additions distributed downwardstowards (Be) or just below (Pb) the water table at 15 cm. However,because these experimental studies were conducted at different timesusing cores from different sites, it is possible that differences indownwashing between those elements could be related to differencesin more basic peat properties (e.g., bulk density, decomposition andhydraulic connectivity). Using a similar experimental design andestablished approach, we directly compare the retention behavior offour commonly studied pollutant trace metals, Pb, Cu, Ni and Zn withinthe same cores simultaneously, and use peat cores collected from twocontrasting environments on a nutrient-poor mire (Sphagnum lawnand dominant fen section). The four different metals were also chosento represent a range in known retention from the relatively immobilePb and Cu to the generally considered more mobile metals Ni and Zn(Bunzl et al., 1976; Kerndorff and Schnitzer, 1980; Berg and Steinnes,1997; Ceburnis andValiulis, 1999; Steinnes et al., 2003).We also includ-ed the conservative element Ti in our analyses to serve as an immobilecheckpoint between cores.

Our second, novel aimwas to test the influence of different simulat-ed precipitation intensities on the potential downwashing of metalsin peat cores. Rainfall intensities are quite variable over time, and wehypothesized that the amount and intensity of rainfall would influencethe extent of downwashing of atmospherically deposited trace metals.We would predict that a low, steady addition would facilitate retentionat shallower depths whereas more rapid flushing of rainwater througha peat column during more intense events would reduce the contacttime possible between metal cations and the peat substrate, thusincreasing the depth at which the depositedmetals can be immobilizedand retained. Over seasonal and especially annual timescales both pre-cipitation extremes – slow, steady events and extreme rainfall – willoccur. In our study a total of 4 ‘rainfall’ treatments were applied to 13peat cores over a 3-week period, including both daily and event-basedapplications.

2. Material & method

2.1. Site description and rainfall history

The peat coreswere collected from Rödmossamyran, a small (~7 ha)oligotrophic mire located in Umeå, northern Sweden (63° 47′N, 20° 20′E; 40m a.s.l.). Field and laboratorymeasured pH values of the peat porewater have been in the range 3.5–4.0. For amore detailed description ofthe mire, including peat stratigraphy, see Rydberg et al. (2010). Basedon climate data from the Swedish Meteorological and HydrologicalInstitute (SMHI.se) themean (defined as the period 1961–1990) annualtemperature for Umeå is 3 °C and the mean annual precipitation is600 mm, but annual rainfall has varied from 400 to 1020 mm year−1.The average 24-hr rainfall during 2012 in Umeå was 2.2 mm day−1,but with 187 days of recorded precipitation, the 24-hr average perrainfall day would be 4.2 mm day−1. 10% of the 24-hr rainfall days in2012 exceeded 10.3 mm day−1, and the highest values were 25and 28 mm day−1 (5 and 2 October 2012, respectively); however,only within the past several years 24-hr rainfall as high as 67 and119mmday−1 were recorded (7 July 2009 and 27 August 2007, respec-tively). Rainfall intensities also vary within events; for example, atSMHI's Holmön station 26 km from Umeå, where hourly data are avail-able, the maximum hourly rainfall was 12.7 mm h−1 within a 24-hrrainfall event with 63 mm of rain in total (27 August 2007, 12 p.m. to28 August 2007, 12 p.m.).

2.2. Sampling and experimental setup

We collected thirteen surface peat cores from the mire in late June2013, of which 8 cores were retrieved on the open sphagnum lawn(dominated by Sphagnum centrale and Sphagnum subsecundum) and5 cores on the dominant fen area. The main fen area consists of a mix-ture of Sphagnum spp. (S. centrale, S. subsecundum and Sphagnumpalustre), shrubs (e.g., Calluna vulgaris and Ledum palustre L.). Small,scattered pine (Pinus sylvestris) also occurs on the fen area but not inclose proximity to our coring location. The two sampling sites werelocated approximately 5 m apart. Using a handsaw sharpened to aknife blade we removed 10 × 10 × 30 cm deep monoliths, to fit thedimensions of our plexiglas core boxes. Each core was placed into itsown core box directly in the field, and care was taken tominimize com-paction of the peat during coring and transferring to the core boxes. Thecores were then transported back to the laboratory, after which the pre-conditioning commenced immediately, and all cores were stored in thelaboratory overnight to ensure that no leakage occurred from the boxes.The cores were then placed outside in a sheltered courtyard and cov-ered by a tent built of see-through plastic with the short ends open toreduce any greenhouse effect. The cores were placed outside to ensurenatural conditions of daylight and temperature, and the tent preventedany addition to the core surface other than our additions.

Table 2Description of additions (presented as mm and mL), occurrence and total amount ofmetals (μg) added to the Stead-low, Steady, Event and Flush treatments.

Treatment: Event Steady Steady-low Flush

Precipitation: 37 mm 10 h−1/370 mL

5.3 mm day−1/53 mL

2 mm d−1/20 mL

37 mm b 1 h/370 mL

Occurrence: Once per week Daily Daily OnceElement

Cu μg 2523 2523 955 841Ni μg 5045 5045 1909 1682Zn μg 5045 5045 1909 1682Pb μg 5045 5045 1909 1682

97S.V. Hansson et al. / Science of the Total Environment 506–507 (2015) 95–101

We followed an experimental setup similar to those described inVile et al. (1999) and Wieder et al. (2010). In order to control thewater level throughout the treatment period, drain holes were drilledin each box 15 cm below the top surface of the Sphagnum. Regardlessof treatment, the water level never rose above the surface of theSphagnum. Prior to starting the addition treatments, all cores werepreconditioned by applying 50 mL Milli-Q water every second day fortwo weeks in order to ensure that the height of the water table wouldremain stable (i.e. that our drain holes would not clog and that therewas no leakage from the core boxes other than from the pre-drilleddrain hole). This was also repeated as a post-condition treatment afterthe third addition cycle was finished to remove metals not effectivelybound to the peat. The addition treatments took place over a three-week period and the cores were watered with specified amounts ofartificial rainwater (Table 1; recipe Wieder et al., 2010) containingseveral metals (EPA Method 200.7 Spiking standard 3 and 5). Here wefocus on 4 trace metals associated with atmospheric pollution: Cu, Ni,Pb and Zn. We adjusted the pH of the rainwater to 4.7 using NaOHand H2SO4, to approximate the precipitation chemistry of the Umeå-region (average pH at the nearest station, Rickleå, was 4.8 ± 0.1 for1992–2012; IVL.se). All cores received the additions from the samebatch of artificial rainwater, including the EPA spiking standards, withthe exception of the two controls (one sphagnum- and one fen-core),which received only the artificial rain (Table 1). All additions wereapplied using a small plastic container with pre-drilled holes at thebottom tomimic the flow of raindrops, and the container wasmanuallycirculated over the core surfaces during additions to ensure that theentire peat-surface received added water.

We applied 4 addition treatments: i) Steady-low (n= 2), ii) Steady(n = 4), iii) Event (n = 4) and iv) Flush (n = 1), along with a Control(n = 2). The treatments and the amount of metals added are summa-rized in Table 2. The Steady-low treatment corresponds to the meanannual 24-hr precipitation of the region, 2 mm day−1, which was

Fig. 1. Distribution of hourly additions (mm h−1 left axis and mL h−1 right axis) for theEvent treatment. (Based on data recorded at the weather station maintained by theDepartment of Applied Physics and Electronics, Umeå University. www.tfe.umu.se).

applied gradually over a 1-hr period each day. In the Steady treatment5.3 mm day−1 was applied gradually over a 1-hr period on a dailybasis. The weekly total of 37 mm was equal to that in our thirdtreatment, the Event treatment, which consisted of one 37 mm day−1

addition one time per week. This was applied over a 10-hr periodwith hourly additions ranging from 0.5 mm h−1 to a maximum of12mm h−1 in the 8th hour. This Event-treatment replicated an intense,yet natural, rainfall event recorded at theweather stationmaintained bythe Department of Applied Physics and Electronics, Umeå University(www.tfe.umu.se, Fig. 1). For these first 3 treatments the addition waslightly sprinkled over the core surface. Further to this, one corewas sub-jected to our fourth treatment, a one-time treatment – Flush – where37 mm was rapidly added over 1 h. The Flush treatment was onlyapplied to one core on only one occasion as an extreme test; however,it represents a test of the potential extent of downwashing from a singleintense rainfall event. Although this was an intense addition, as noted24-hr rainfalls of 67 and 119 mm have been recorded in Umeå overthe past several years.

The two control cores (one sphagnum and one fen) received onlyartificial rainwater applied at the same rate as the Steady treatment,i.e., 5.3 mm day−1. After a 2-week post-conditioning treatment,all cores were placed in a walk-in freezer at −20 °C before furtherprocessing.

2.3. Sample preparation and analytical methods

In the walk-in freezer, the cores were removed from the core boxesand cut into 2-cm-thick slices using a bandsaw with a stainless steelblade. 2-cm-increments were first pre-marked on the frozen surfaceto ensure that the total length remained intact and tomaintain constantsample thickness. All samples were dried to constant weight at 60 °C,after which their dry mass was recorded and bulk density calculated.Finally, the samples were ground in a Retsch ball mill before furtheranalysis.

The concentrations of Cu, Pb, Ni, Zn and titanium (Ti) were mea-sured on 0.5 g of dried and homogenized powder using a Bruker S8Tiger WD-XRF analyzer equipped with a Rh anticathode X-ray tube. Aspecific calibration modified from Rydberg (2014) was developed inorder to optimize the WD-XRF for the matrices of the powderedsamples (see supplementary information, Table S1-2 and Figure S1 fordetails). Method detection limits for trace elements are in the range of1–9 μg g−1, with an accuracy within 8% for all elements except lead(14%). Reproducibility of the analyses was ≥94%.

3. Results and discussion

In our results we observe twomain patterns, which relate in turn tothe two main aims of our study. First, there is a clear difference in thedepth to which elevated concentrations for each of the four metalscan be detected (all added in dissolved form) with the order fromshallowest to deepest being Pb b Cu b Zn≤Ni. This pattern is consistentfor all treatments and also with results from experimental work on ad-sorption rates using peat (Bunzl et al., 1976; Kerndorff and Schnitzer,1980) and results from field studies of mosses for biomonitoring(Ross, 1990; Berg and Steinnes, 1997; Ceburnis and Valiulis, 1999;Steinnes et al., 2003). Second, for all four metals and in both peattypes the individual application treatments resulted in different down-washing depths, i.e., as the artificial precipitation additions increased sodid the depth at which elevated concentrations of the metals could bedetected (Steady-low b Steady ≤ Event b Flush).

We present and discuss these two patterns in turn below. For brevitydetailed information on bulk density (g cm−3) and downcore distribu-tion of Ti (μg g−1), and the addedmetals, i.e., Pb, Cu, Zn, andNi (μg g−1),for all cores (including controls and replicates) and for both peat types ispresented in Fig. 2. We also present the results as inventories, i.e., thedistribution of the calculated excess inventory in % for each 2-cm slice

98 S.V. Hansson et al. / Science of the Total Environment 506–507 (2015) 95–101

99S.V. Hansson et al. / Science of the Total Environment 506–507 (2015) 95–101

based on the total addition (100%) and taking into account the contribu-tion of ambient metals by subtracting the background concentrationsmeasured in the control cores for each slice.

3.1. Depth of downwashing and the relationship to metal adsorption rates

The bulk density (Fig. 2a) varies between the sphagnum lawn andthe dominant fen section of the mire, but is consistent between coresfrom the same area (Rydberg et al., 2010; Hansson et al., 2014a). Inthose studies the lower level of decomposition in the sphagnum(0.05–0.06 g cm−3) versus the fen (0.07 to 0.10 g cm−3) peat is alsosupported by analyses of C/N ratios and light transmission. The spatialheterogeneity between the two mire settings is also reflected in thedistribution of the conservative, immobile element Ti (Fig. 2 b), whichwas not added in the treatments, where the total recovery varies from≤5.1mgg−1 in the sphagnumcores to≥7.5mg g−1 in the fen cores. Al-though small differences occur between coreswithin each peat type, thebulk density and general distribution of Ti are consistent and indicate nosignificant differences in the properties of the different cores withineach peat type.

The increasing depths at which elevated concentrations of eachmetal can be detected (Pb b Cu b Zn ≤ Ni) suggest differences in theadsorption efficiency of the organic matter for each metal, where Pb isadsorbed most efficiently and Ni least. For Pb our calculated (%) inven-tory retained in the top 2 cm for each treatment was 21–85%, whileadded Cu was in general flushed somewhat deeper into the peat thanPb (Fig. 2d) with calculated excess inventories of 18–63% retainedin the top 2 cm. Zn and Ni did not show the same level of near-surfaceretention, where only 10–25% of the calculated excess inventory forZn and 10–26% for Ni was retained in the top 2-cm for each treatment.This difference in the adsorption efficiency of peat for each of the dis-solved metals has previously been shown in batch experiments usingpeat as well as different organic compounds. For example, batch exper-iments using peat have shown a decrease in metal adsorption in theorder Pb N Cu N Zn (Bunzl et al., 1976), Pb N Cu N Ni N Zn (Kalmykovaet al., 2008) and Cu N Zn (Ringqvist and Oborn, 2002) in similar pH con-ditions to our study (pH = 3.5–4.0), with single metal solutions atlower to similar concentrations (3 μg L−1 to 5 mg L−1) to our additions(Table 2). The addition amounts are much lower than the sorptioncapacities experimentally determined by, e.g., Kalmykova et al. (2008)and Ho and McKay (2000). Despite the higher concentrations and mul-tiple metals added in our study as compared to Vile et al. (1999), whoadded soluble Pb at concentrations 5–14 times lower than our addi-tions, the inventories of added Pb retained in the 0–2 cm slice in ourcores are much higher (42–85% depending on the rainfall intensity,excluding the flush treatment) than in their study (15–64%). The lowestdepths at which added Pb could be detected in our cores (≤6 cm) arealso less than the depths (≥15 cm) reported by Vile et al. (1999). Thiswould suggest that the higher concentrations in our experiment donot explain the downcore distributions in our cores, i.e., sorption capac-ity is not a limiting factor. Other factors, such as the sorption rate of thepeat or the rate of addition are more important for the observed distri-bution of metals.

Batch experiment studies with fulvic and humic acids have shown asimilar order to that which occurs in our cores: Pb N Cu N Zn N Ni(Rashid, 1974; Kerndorff and Schnitzer, 1980; Logan et al., 1997), indi-cating that organicmatter is a predominant factor controlling the down-wash or retention of thesemetals in peat. For instance, Qin et al. (2006)had shown that Pb and Cu bind to the peat through carboxyl groups(and in some cases hydroxylic groups), thereby having very similarbinding mechanisms and binding environments in peat. Zn and Nihave not been studied with the same level of detail, however, Zn and

Fig. 2. Calculated bulk density g cm−3 and concentrations (μg g−1) of Ti, Pb, Cu, Zn and Ni forthe sphagnum and fen cores. Top axes in all graphs shows concentrations (filled circles) of mehorizontal bar graphs (average of replicate treatments). Controls (n = 2, black crosses) and re

Ni do have similar affinities for both humic acids and the more solublefulvic acids as compared to Cu, which has a significantly higher affinityfor humic acids than fulvic acids (Kalmykova et al., 2008). Our resultsconcur with this because both Pb and Cu are rather similarly affectedby the downwashing compared to Ni, Zn (maximal depth b12 cm vs.N12 cm). Overall, our results suggest that due to different adsorptionefficiency of organic matter, Pb and Cu are relatively well retained inthe upper aerated peat section near the peat surface, while Ni and Znare subject to a greater amount of downwash. Thus, the attenuation ofan addition signal of dissolved metals in the uppermost slice (0–2 cm)is less for Pb and Cu and greater for Zn and Ni.

3.2. Depth of downwashing and the influence of precipitation intensities

Previous experimental studies examining the retention of Pb (Vileet al., 1999) and Be (Wieder et al., 2010) in peat cores considered the in-fluence of different water table levels, where a high water table (3 cm)inhibits downward movement in contrast to low (15 cm) or variable(3–15 cm) water tables. Both studies applied only one rate of addition,with 50 mL added rapidly over 1 min every other day, correspondingto an average daily rainfall of 3.2mmday−1. Aswe noted in our summa-ry of local climate data (Section 2.1), rainfall does not occur at regularintervals and instead intensities vary substantially. In the Umeå region24-hr rainfalls greater than 20 mm occur on average 3.5 times year−1,and so-called “extreme” events (≥40 mm day−1) occur on average0.6 times year−1 for the last decade (SMHI.se). Rainfall also varies sea-sonally. For example, in a spatial analysis of extreme precipitation inSweden during 1961–2000, Hellstrom and Malmgren (2004) observedthat for the Umeå region ~80% of these events occurred mostly duringJune–August, when the height of thewater table in the peat is generallyat its lowest. These two factors combined – lower water table andhigher intensity precipitation –would facilitate downwashing to deeperdepths in the peat and a greater attenuation of themetal deposition sig-nal throughout the peat profile.

As expected the lowest addition intensity, Steady-low, showedthe least downwash with depth ranging from 2–4 cm (Pb) to ≥14–16 cm (Ni). Although the addition of 2 mm day−1 corresponds tothe annual daily average, this unrealistically even and slow rate ofaddition should be interpreted as a test of the minimum potentialinfluence of downwashing. The Steady treatment, with an additionof 5.3 mm day−1 designed to assess a higher degree of precipitationwhile still maintaining a steady rate of addition, increased the depthsto which downwashing was recorded in the range of 4–6 cm (Pb) to≥14–16 cm (Zn and Ni). Even though both of the steady-rate treat-ments (i.e., Steady-low and Steady) were applied on a daily basis,the increase from 2 mm day−1 to 5.3 mm day−1 resulted generally inonly a 2 cm increase in the depth of downwashing regardless of metaland peat type.

The greater application intensity of the Event treatment(37 mm 10 h−1), which equaled the amount in the Steady treatmenton a weekly basis, did not increase the depth of downwashing forPb (4–6 cm) in both peat types or Cu in the fen cores (8–10 cm),but did further increase the depth of downwashing by ≥2 cm forCu (10–12 cm) in the sphagnum cores and Zn and Ni in both peattypes (≥14–16 cm). Such rainfall intensities occur less frequently, but24-hr rainfall values N30 mm day−1 have occurred on average onceper year over the past 10 years in Umeå. By comparison to other regionswhere peat studies have taken place such as in northern Minnesota,USA (e.g., Urban et al., 1990; Benoit et al., 1998; Vile et al., 1999) dailyrainfall N 30 mm day−1 occurs on average 3.6 times per year (during2000–2013, Grand Rapids, MN, USA; NOAA National Weather Service:www.weather.gov).

the Steady-low (n = 2), Steady (n = 4), Event (n = 4) and Flush (n = 1) treatment intals in (μg g−1) and bottom axes shows % recovery of total addition, which are shown asplicates (open circles) are included in each graph for each parameter.

100 S.V. Hansson et al. / Science of the Total Environment 506–507 (2015) 95–101

As expected the one time Flush treatment (37 mm over 1 h) gener-ated the deepest depths of downwashing, with depths of 10–12 cm forPb, 12–14 cm for Cu, and≥14–16 cm for Zn and Ni. It was also the onlytreatment that showed a clear downward transport to the water tablefor all 4 metals, whereas for all other treatments the additions ofPb and Cu were completely immobilized within the upper aeratedlayers (≤12 cm). Although on an hourly basis the Flush additionexceeds the highest measured hourly rainfall values for the Umeå area(12–13 mm h−1), this total amount is still only half to one-third of thetwo highest 24-hr rainfall values recorded on 27 August 2007(118.5mmday−1) and 7 July 2009 (66.8mmday−1). Again for compar-ison, a similar maximum 24-hr rainfall value was recently reported forGrand Rapids, MN, 119.7 mm day−1 (19 June 2012), but maximumvalues as high as 197 and 298 mm day−1 have been recorded since1981 in coastal areas of Maine, USA (www.weather.gov) in proximityto other peatland studies (e.g., Norton et al., 1997; Roos-Barracloughet al., 2006). Thus, the Flush addition, though extreme as a rain event,would have its natural analog over decadal timescales.

The deepest depths we observe for downwashing of Pb (≤6 cm) andCu (≤12 cm), excluding for Flush, are less than the depths previouslyshown for Pb and Be (Vile et al., 1999; Wieder et al., 2010), whereadded Pb (both as dissolved and particulate additions) could be detectedat or even below the 15 cm water table and Be just above (12–14 cm).We suggest that the greater degree of downwashing indicated by thosestudies is related to the rapid addition of 6.4 mm applied over 1 minwith a squirt bottle every second day. This small, but rapid additionwould have a similar effect to our Flush (37 mm over 1 h) by speedingthe delivery of rainwater through the peat.

3.3. Implications for downwashing on quantitative and qualitativereconstructions of past metal deposition

The rapid exponential declinewith depth in added Pb and Cuwithinthe upper 4–10 cm in our cores is similar to, but not quite as deep as, theresults fromprevious experimentalwork (Vile et al., 1999;Wieder et al.,2010), as well as the overall exponential decline in 7Be we measured in5 natural cores (Hansson et al., 2014a). Although we do not necessarilyexpect that particulate inputs, such as that which occurs in proximity tosmelters (e.g. Nieminen et al., 2002), would be subjected to the samedegree of downwashing, Vile et al. (1999) did observe that Pb addedwith coal fly ash exhibited the same degree of downwashing as solublelead. These experimental approaches indicate that added pollutantmetals applied to the peat surface are not fully retained at the surfacelevel, but that a significant component is immobilized below the surfacelayer (~20–80% of the added Pb and ~20–40% of the added Cu). Theseresults suggest that a deposition signal – the peak input – would quali-tatively remain in place, but that the deposition signal might then be at-tenuated over a few to several cm and whatever number of years thiswould represent in peat accumulation. This has specific implicationsfor directly translating metal accumulation rates in peat to absoluteatmospheric deposition rates. From these results we can reasonablyinfer, for example, that Pb deposited in 2012 would not be locatedonly in peat accumulated during 2012, but that some fraction of thisdeposited Pb would be smeared over several years of accumulatedpeat. Nonetheless, for Pb and Cu, but less so for Zn and Ni, the attenua-tion of the addition signal and the depths to which they are transportedare constrained.

While the efficacy of peat as an archive of pollutionmetal depositionover longer timescales is well established (Weiss et al., 1999; Renberget al., 2001), the smearing in metal accumulation caused by down-washing at the peat surface does have implications for quantitativereconstructions of recent deposition and connecting peat metal accu-mulation rates to direct measurements of deposition from monitoring.The overall immobility of Pb in dated peat records has been argued pri-marily based on the fact that the accumulation record of Pb in peat is ingood accordance with the general pattern for the known history of Pb

usage over the industrial period and that peat, lake sediment and glacialice records are in good agreement over centennial and millennial time-scales (Renberg et al., 2001). Long-term records only support the factthat the peat records preserve a relative or qualitative history of metalinputs over longer timescales, and they cannot themselves be used toassess downwashing over shorter timescales. The attenuation of a Pband Cu deposition signal through downwashing would not be apparentover centennial to millennial timescales, where 1 cm of decomposedand compacted peat in the catotelm may represent 10–20 years. Withthat temporal resolution it is not possible to address the occurrence orinfluence of downwashing (or small-scale remobilization, which wedo not study here) within the upper acrotelm.

4. Conclusions

From our results we conclude that downwashing, particularly inconnection with precipitation events with higher intensities, exerts aninfluence on the vertical distribution of recently deposited trace metalsin the upper aerated section of peat cores. The experimental data indi-cate that the peak in added Pb and Cu is retained at the surface, butthis smearing effect of the addition signal in the upper, aerated peathas clear implications for calculating accumulation rates both directlybecause the pollutant metal signal itself is attenuated, as well as thefact that downwashing should have a similar effect on radioisotopesused for age-depth modeling (Hansson et al., 2014b). The effect ofdownwashing on Zn and Ni is greater. If the extent of metal down-washing in a core can be estimated, however, then it should be possibleto take this process into account when calculating accumulation rates.Thiswould enable a better quantitative calculation of recently depositedelements in surface peat cores and thus facilitate a more precise linkbetween peat accumulation records and direct monitoring data. Anext step would be to apply these types of artificial metal additionsunder natural conditions, i.e., directly onto controlled field plots usingisotopically labeled metals (e.g., Kaste et al., 2003) with addition levelscomparable to ambient levels of metal atmospheric deposition rates asmeasured bymonitoring data. Isotopic labelingwould facilitate separat-ing small additions of metals from those already present in the peat,which would provide a more natural assessment of downwashing innatural field conditions.

Acknowledgments

The authors wish to thank Carsten Meyer-Jacob for the help in thefield as well as in the lab. This research has been supported by grantsfrom the Swedish Research Council and the Kempe Foundations.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.scitotenv.2014.10.083.

References

Benoit JM, Fitzgerald WF, Damman AWH. The biogeochemistry of an ombrotrophic bog:evaluation of use as an archive of atmospheric mercury deposition. Environ Res1998;78(2):118–33.

Berg T, Steinnes E. Use of mosses (Hylocomium splendens and Pleurozium schreberi) asbiomonitors of heavy metal deposition: from relative to absolute deposition values.Environ Pollut 1997;98(1):61–71.

Biester H, Bindler R, Martinez-Cortizas A, Engstrom DR. Modeling the past atmosphericdeposition of mercury using natural archives. Environ Sci Technol 2007;41(14):4851–60.

Biester H, Hermanns YM, Cortizas AM. The influence of organic matter decay on the dis-tribution of major and trace elements in ombrotrophic mires— a case study from theHarz Mountains. Geochim Cosmochim Acta 2012;84:126–36.

Bindler R, Klarqvist M, Klaminder J, Forster J. Does within-bog spatial variability of mercuryand lead constrain reconstructions of absolute deposition rates from single peatrecords? The example of Store Mosse, Sweden. Glob Biogeochem Cycles 2004;18(3).

101S.V. Hansson et al. / Science of the Total Environment 506–507 (2015) 95–101

Bunzl K, Schmidt W, Sansoni B. Kinetics of ion-exchange in soil organic matter. IV.Adsorption and desorption of Pb2+, Cu2+, Cd2+, Zn2+ and Ca2+ by peat. J Soil Sci1976;27(1):32–41.

Ceburnis D, Valiulis D. Investigation of absolute metal uptake efficiency fromprecipitationin moss. Sci Total Environ 1999;226(2–3):247–53.

Cloy JM, Farmer JG, Graham MC, MacKenzie AB, Cook GT. A comparison of antimonyand lead profiles over the past 2500 years in Flanders Moss ombrotrophic peat bog,Scotland. J Environ Monit 2005;7(12):1137–47.

Damman AWH. Distribution and movement of elements in ombrotrophic peat bogs.Oikos 1978;30(3):480–95.

Farmer JG, Eades LJ, Atkins H, Chamberlain DF. Historical trends in the lead isotopic com-position of archival Sphagnum mosses from Scotland (1838–2000). Environ SciTechnol 2002;36(2):152–7.

Hansson SV, Kaste JM, Chen K, Bindler R. Beryllium-7 as a natural tracer for short-termdownwash in peat. Biogeochemistry 2014a;119:329–39.

Hansson SV, Kaste JM, Olid C, Bindler R. Incorporation of radiometric tracers in peat andimplications for estimating accumulation rates. Sci Total Environ 2014b;493:170–7.

Hellstrom C, Malmgren BA. Spatial analysis of extreme precipitation in Sweden1961–2000. Ambio 2004;33(4–5):187–92.

Ho YS, McKay G. The kinetics of sorption of divalent metal ions onto sphagnum mosspeat. Water Res 2000;34:735–42.

Kalmykova Y, Stromvall AM, Steenari BM. Adsorption of Cd, Cu, Ni, Pb and Zn onSphagnum peat from solutions with low metal concentrations. J Hazard Mater2008;152(2):885–91.

Kaste JM, Friedland AJ, Sturup S. Using stable and radioactive isotopes to trace atmospher-ically deposited Pb in montane forest soils. Environ Sci Technol 2003;37(16):3560–7.

Kerndorff H, Schnitzer M. Sorption of metals on humic-acid. Geochim Cosmochim Acta1980;44(11):1701–8.

KrachlerM, Mohl C, Emons H, ShotykW. Atmospheric deposition of V, Cr, and Ni since thelate glacial: effects of climatic cycles, human impacts, and comparison with crustalabundances. Environ Sci Technol 2003;37(12):2658–67.

Logan EM, Pulford ID, Cook GT, MacKenzie AB. Complexation of Cu2+ and Pb2+ by peatand humic acid. Eur J Soil Sci 1997;48(4):685–96.

Malmer N. Patterns in the growth and the accumulation of inorganic constituents in thesphagnum cover on ombrotrophic bogs in Scandinavia. Oikos 1988;53(1):105–20.

Martinez-Cortizas A, Pontevedra-Pombal X, Garcia-Rodeja E, Novoa-Munoz JC, Shotyk W.Mercury in a Spanish peat bog: archive of climate change and atmospheric metaldeposition. Science 1999;284:939–42.

Nieminen TM, Ukonmaanaho L, Shotyk W. Enrichment of Cu, Ni, Zn, Pb and As in anombrotrophic peat bog near a Cu–Ni smelter in southwest Finland. Sci Total Environ2002;292(1–2):81–9.

Norton SA, Evans GC, Kahl JS. Comparison of Hg and Pb fluxes to hummocks and hollowsof ombrotrophic big heath bog and to nearby Sargent Mt. Pond, Maine, USA. WaterAir Soil Pollut 1997;100(3–4):271–86.

Oldfield F, Richardson N, Appleby PG. Radiometric dating (Pb-210, Cs-137, Am-241) ofrecent ombrotrophic peat accumulation and evidence for changes in mass-balance.The Holocene 1995;5(2):141–8.

Pontevedra-Pombal X, Mighall TM, Novoa-Munoz JC, Peiteado-Varela E, Rodriguez-Racedo J, Garcia-Rodeja E, et al. Five thousand years of atmospheric Ni, Zn, As, andCd deposition recorded in bogs from NW Iberia: prehistoric and historic anthropo-genic contributions. J Archaeol Sci 2013;40(1):764–77.

Qin F, Wen B, Shan XQ, Xie YN, Liu T, Zhang SZ, et al. Mechanisms of competitive adsorp-tion of Pb, Cu, and Cd on peat. Environ Pollut 2006;144(2):669–80.

Rashid MA. Absorption of metals on sedimentary and peat humic acids. Chem Geol 1974;13(2):115–23.

Renberg I, Bindler R, Brannvall ML. Using the historical atmospheric lead-depositionrecord as a chronological marker in sediment deposits in Europe. The Holocene2001;11(5):511–6.

Ringqvist L, Oborn I. Copper and zinc adsorption onto poorly humified Sphagnum andCarex peat. Water Res 2002;36(9):2233–42.

Roos-Barraclough F, Givelet N, Cheburkin AK, Shotyk W, Norton SA. Use of Br and Sein peat to reconstruct the natural and anthropogenic fluxes of atmospheric Hg: a10000-year record from Caribou Bog, Maine. Environ Sci Technol 2006;40(10):3188–94.

Ross HB. On the use of mosses (Hylocomium-Splendens) and (Pleurozium-Schrebebri)for estimating atmospheric trace-metal deposition. Water Air Soil Pollut 1990;50(1–2):63–76.

Rydberg J, Karlsson J, Nyman R,Wanhatalo I, Nathe K, Bindler R. Importance of vegetationtype for mercury sequestration in the northern Swedish mire, Rodmossamyran.Geochim Cosmochim Acta 2010;74(24):7116–26.

Rydberg J. Wavelength dispersive X-ray fluorescence spectroscopy as a fast, non-destruc-tive and cost-effective analytical method for determining the geochemical composi-tion of small loose-powder sediment samples. Journal of Paleolimnology 2014;52(3):265–76.

Shotyk W. Natural and anthropogenic enrichments of As, Cu, Pb, Sb, and Zn inombrotrophic versus minerotrophic peat bog profiles, Jura Mountains, Switzerland.Water Air Soil Pollut 1996;90(3–4):375–405.

ShotykW. History of atmospheric lead deposition since 12,370 14C yr BP from a Peat Bog,Jura Mountains, Switzerland. Science 1998;281:1635–40.

Steinnes E, Berg T, Sjobakk TE. Temporal and spatial trends in Hg depositionmonitored bymoss analysis. Sci Total Environ 2003;304(1–3):215–9.

Urban NR, Eisenreich SJ, Grigal DF, Schurr KT. Mobility and diagenesis of Pb and 210Pb inpeat. Geochim Cosmochim Acta 1990;54(12):3329–46.

VileMA,Wieder RK, NovakM.Mobility of Pb in Sphagnum-derived peat. Biogeochemistry1999;45(1):35–52.

Weiss D, Shotyk W, Kramers JD, Gloor M. Sphagnum mosses as archives of recentand past atmospheric lead deposition in Switzerland. Atmos Environ 1999;33(23):3751–63.

Wieder RK, Vile MA, Scott KD. Cosmogenic Be-10 as a potential dating tool in peat.Biogeochemistry 2010;101(1–3):177–82.