author's personal copy - uni-muenster.de · organic carbon (doc) can alter substantially due...

(This is a sample cover image for this issue. The actual cover is not yet available at this time.)

This article appeared in a journal published by Elsevier. The attached

copy is furnished to the author for internal non-commercial research

and education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling or

licensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of the

article (e.g. in Word or Tex form) to their personal website or

institutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies are

encouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Selective adsorption of dissolved organic matter to mineral soils

D.N. Kothawala a,⁎, C. Roehm b,1, C. Blodau c,2, T.R. Moore a,3

a Department of Geography and Global Environmental & Climate Change Centre, McGill University, 805 Sherbrooke Street West, Montréal, Québec, Canada H3A 2K6b Department of Geography and Planning and Great Lakes Center, SUNY College at Buffalo, Buffalo, NY 14222, USAc School of Environmental Sciences, University of Guelph, 50 Stone Road East, Guelph, Ontario, Canada M1G 2W1

a b s t r a c ta r t i c l e i n f o

Article history:

Received 22 July 2011Received in revised form 27 April 2012Accepted 2 July 2012Available online xxxx

Keywords:

SpectroscopyFluorescenceAbsorbanceSize exclusionFourier transform infrared spectroscopyDissolved organic carbon

As soil solutions pass through forested mineral soils, the chemical and structural compositions of dissolvedorganic carbon (DOC) can alter substantially due to interactions with soil particle surfaces. Typically, adsorp-tion processes dominate in mineral soils and the resulting concentration of DOC is reduced substantially. Westudied changes in the molecular and structural compositions of DOC during equilibration with mineral soilscollected across Canada (n=43) and found that the overall aromatic content of DOC decreased with equili-bration in almost all cases from using specific absorbance (SUVA) and the fluorescence index. The fluores-cence index revealed that podzolic B horizons, with typically large adsorption capacity (Qmax), had thegreatest reduction in aromaticity, which was partially explained by the much lower aromatic content ofDOC desorbed from soils surfaces. In contrast, a decrease in DOC aromaticity for volcanic B horizons, alsowith high Qmax, was primarily due to adsorption. An unexpected finding was the release of extremelyhigh (2.6×106 Da) and low (420 Da) molecular weight (MW) organic materials during equilibration usinghigh performance size exclusion chromatography (HPLC), for luvisols and podzols, respectively. In general,the average number–average MW (Mn) of DOC decreased for all soil types, but the greatest decrease in Mn

was observed for mineral soils with large Qmax, including the podzolic and volcanic B horizons. Analysis ofchanges in FTIR spectra revealed that the most prominent change to DOC functional groups was a reductionin carboxyl groups, which was even greater than the removal of aromatic DOC. The findings of this study em-phasize that while DOC concentrations may decrease substantially during passage through mineral soils, it isvaluable to consider the contribution of DOC from desorption of pre-existing soil C. Essential to the findings ofthis study was the inclusion of multiple analytical techniques to track changes to DOC character, and the in-clusion of a wide range of mineral soils.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Dissolved organic carbon (DOC) found in forest soil–solutions consistsof a spectrum of molecules ranging in molecular size, solubility and func-tional group composition. The chemical and physical structures of DOChave been linked to ecological and environmentally relevant functions(Cabaniss et al., 2000; Chin et al., 1998). For instance, the transport andfate of heavy metals and organic contaminants are strongly influencedby chemical associations with DOC within soils (Tipping and Hurley,1992; Tipping et al., 2003). In addition, the quality of DOC leached fromforest soils can influence pH buffering capacity and nutrient availabilityin surface waters (Hruska et al., 1999; Qualls and Haines, 1991a).

An important abiotic process influencing the chemical compositionof DOC in soil solutions is adsorption and desorption on soil particle sur-faces (Kaiser et al., 2002; Qualls et al., 2002). Soil solutions leachingthrough mineral soil horizons often contain high concentrations ofDOC generated from decomposing litter within upper organic rich hori-zons (McDowell and Likens, 1988; Michalzik et al., 2001). DOC containsa broad mixture of organic polyelectrolytes, and likewise, soil particlesurfaces have positively and negatively charged sites. Consequently, amultitude of interactions can occur betweenDOC and soil surfaces.Min-eral soil horizonswith a high capacity to adsorbDOChave been found tocontain high amounts of amorphous oxides and oxyhydroxides of Feand Al, which are particularly abundant in the B horizons of acidic pod-zols and soils derived from volcanic parent materials (Kaiser et al.,1996; Kothawala et al., 2009; Moore et al., 1992). As a soil–solution in-filtrates through mineral horizons, the concentration and chemicalcomposition of DOC can vary substantially resulting from dominant in-teractions including ligand exchange, cation bridging, anion exchangeand weaker van der Waals forces (Sposito, 2004; Tan, 2003). Changesto DOC compositionmay thus be highly variable depending on physicaland mineralogical characteristics of the soil. Determining which soil

Geoderma 189–190 (2012) 334–342

⁎ Corresponding author at: Limnology Department, Uppsala University, Norbyvägen18D, Uppsala, SE-75 236, Sweden. Tel.: +46 524 7371.

E-mail addresses: [email protected] (D.N. Kothawala),[email protected] (C. Roehm), [email protected] (C. Blodau),[email protected] (T.R. Moore).

1 Tel.: +1 716 878 4508; fax: +1 716 878 6644.2 Tel.: +1 519 824 4120x56203.3 Tel.: +1 514 398 4961; fax: +1 514 398 7437.

0016-7061/$ – see front matter © 2012 Elsevier B.V. All rights reserved.doi:10.1016/j.geoderma.2012.07.001

Contents lists available at SciVerse ScienceDirect

Geoderma

j ourna l homepage: www.e lsev ie r .com/ locate /geoderma


properties influence the chemical partitioning of DOC during passagethrough mineral soils can further our understanding of how soil pro-cesses influence the quality and function of DOC entering aquaticsystems.

Several studies have examined changes to the chemical characteris-tics of DOC before and after sorption to mineral soils under controlledconditions, providing useful insights intomechanisms of soil–DOC inter-actions (Ghosh et al., 2009; Gu et al., 1994; Kaiser, 2003). But, there is agap in our understanding of how variable DOM partitioning may beover a broad range of mineral soils, as studies using soils collected fromanarrow range of sites have a limited ability to compare among soil char-acteristics (Kaiser et al., 2002; van Hees et al., 2005). Similarly, studiesusing well-identified pure soil minerals such as goethite (FeO(OH)) areunable to assess the potential contribution of soil C desorption, whichcould be an important source of DOC.

Numerous analytical techniques have been developed and adaptedover the past several decades to assess various measures of DOC quality(for example see Thacker et al., 2005), yet no single technique has provento be fully effective as ameans of characterizing DOC. Here, we apply fourtechniques to examine changes to soil–solution DOC before and afterequilibration with mineral soils. Two optical techniques are the fluores-cence index and SUVA, which provide surrogate measures of aromaticityand have been effective due to their simplicity to measure and interpret(McKnight et al., 2001; Weishaar et al., 2003). In addition, we includehigh performance size exclusion chromatography (HPSEC), allowing forthe identification of shifts in the molecular weight distribution, and Fou-rier transform infra red (FTIR) spectrophotometry, to examine changesto the prominent functional groups of DOC. Our objectives are to identifychanges to the chemical and structural character of DOC that result fromexchange reactions with mineral soil surfaces. We use a wide range ofmineral soils representative of Canadian temperate and boreal forests,that have beenwell-characterized in terms of adsorption capacity and de-sorption potential (Kothawala et al., 2009).

2. Methods

2.1. DOC source

DOC was extracted from the organic L, F, and H horizons of a podzol(Canada Soil Survey Committee, 1998) collected from Mt. Saint Hilaire,Québec (45°33′N, 73°08′W). The intact LFH horizon was soaked in de-ionizedwater overnight. The soil solutionwas decanted from the soil, fil-tered with 0.45 μmglass fiber filters (Rose Scientific), and stored at 4 °C,to produce an initial DOM solution with a concentration of 61 mg·L−1

for all batch sorption experiments, and had a pH of 4.8 and conductivityof 100 μS·cm−1.

2.2. Study sites

Mineral soil horizons included in the study were collected from 17soil profiles across Canada and represent five soils orders including pod-zol, brunisol, luvisol, gleysol and an organic soil (Canada Soil SurveyCommittee, 1998). Soil horizons were grouped into four horizon groups,based on previously determined sorption characteristics (Kothawalaet al., 2009). A horizons and B horizons are defined as not being enrichedwith Fe, while podzol B and volcanic B horizons are enriched with Fe(Table 1). Soils included in this study represent a sub-set (n=43) of aprevious study (n=52) (Kothawala et al., 2009) and horizons not in-cluded in the current study include twoAhorizons (A/Ae), three volcanicB horizons (Bfj), four podzol B horizons (Bf/Bcc/Bfh) and two C horizons.Thus, soil properties (Table 1) and sorption characteristics (Fig. 1) onlyreflect the sub-set of soil horizons included in this study as identified inTable 2. We collected a sandy orthic dystric brunisol (Bf, Bfj) developedon volcanic ash from a Douglas fir (Pseudotsuga menziessi) stand on Van-couver Island (49°35′N, 124°56′W). Clay rich orthic humic gleysol (Bg)and orthic dystric brunisol (Ah, Bm), as well as one sandy gray luvisol

(Ae, Bt), were collected from Waskesiu Lake, Saskatchewan (53°55′N,106°04′W), under black spruce (Picea mariana), jack pine (Pinusbanksiana) and aspen (Populus tremuloides), respectively. A heavily mot-tled clayey orthic gleysol, humic luvic gleysol and an organic soil (Ah, Ae,Bt, Bg, Bf, Bfj), were collected from Groundhog River, Northern Ontario(48°58′N, 82°19′W), under a mixed forest stand of aspen, black spruce,and white spruce (Picea glauca). Well-drained sandy lacustrine dystric

Table 1

Properties and sorption characteristics for four groups of mineral soil horizons. Valuesare averages±standard deviation (range).

Properties A horizons B horizons Podzol B Volcanic B

Soil horizons Ah/Ae Bm/Bg Bfj/Bf/Bfh/Bcc Bfj/Bfn 11 17 7 10Soil C (mg g−1) 18.4±20.0

(4.5 to 72.6)4.8±4.8(0.5 to 17.8)

46.0±25.4(7.2 to 84.8)

15.3±7.1(8.2 to 27.5)

Soil pHCaCl2 4.8+1.3(3.4 to 7.5)

5.7±0.9(4.5 to 7.2)

4.3±0.4(3.7 to 5.3)

5.2±0.3(4.7 to 5.5)

Feam (mg·g−1) 0.6±0.4(0.1 to 1.3)

0.8±0.4(0.1 to 1.8)

6.1±4.1(2.5 to 14.5)

5.4±2.6(1.40 to 10.5)

Alam (mg·g−1) 0.4±0.2(0.1 to 0.6)

0.7±0.9(0.1 to 3.8)

5.2±3.2(1.0 to 8.9)

8.7±7.7(0.9 to 25.9)

% sand 66±29(22 to 97)

55±28(22 to 98)

72±23(37 to 92)

70±12(48 to 88)

% clay 10±7(5 to 29)

16±11(5 to38)

10±3(6 to 17)

12±8(6 to 26)

Qmax (mg kg−1) 259±235(60 to 855)

463±466(182 to 2200)

2022±1891(572 to 2006)

1041±487(630 to 4913)

b (mg·kg−1) 70±43(14 to 146)

50±28(4 to 96)

117±64(52 to 256)

46±26(15 to 83)

Qmax

(mg

kg

-1)

Deso

rpti

on

po

ten

tial

(mg

kg

-1)

a

b

a

a,b

bb

a

bb

a,b

Fig. 1. Distribution of (a) adsorption capacity (Qmax), and (b) the desorption potentialfor soil horizon groups, A horizons (n=10), B horizons (n=17), volcanic B (n=10)and podzol B (n=10). Only groups not sharing the same lower case letter are signifi-cantly different from each other based on the Tukey–Kramer honestly significant testfor comparison of means with uneven sample sizes (pb0.05). The line in the middleof boxes represents the median, with lower and upper parts of the box representing25% and 75% of the distribution, respectively.

335D.N. Kothawala et al. / Geoderma 189–190 (2012) 334–342


brunisols (Ah, Bm), were collected from Turkey Point, Southern Ontario(42°71′N, 80°35′W), under white pine (Pinus strobus). We included fourhumo-ferric podzols (Ae, Bf, Bm, C) developed on the glacial till of thePrecambrian shield at Plastic Lake, south-central Ontario (45°10′N,78°50′W)under amixedmaple-beech deciduous forest (Acer saccharum,Acer rubrum, Fagus grandifolia), and three sites in Quebec, includingHermine (45°54′N, 74°00′W), a sugar maple forest, Lac Laflamme(47°17′N, 71°14′W), a balsam fir forest, and Lac Tirasse (49°12′N,73°29′W), a black spruce forest.

2.3. Soil properties

Mineral soils were sieved to b2 mm, air-dried, and stored at 4 °C.Soil pH (pHCaCl2) was measured in 0.01 M CaCl2 using a soil to solutionratio of 1:2. Total soil carbon (soil C) was determined by an ElementalAnalyzer Carlo Erba™ (model NC2500). Particle size analysis for sand,silt and clay was performed by the sieve pipette method, after the re-moval of organic carbon by hydrogen peroxide, with dispersion usingsodium metaphosphate and sodium carbonate (Sheldrick and Wang,1993). Two extractions (Ross and Wang, 1993) were used to identifyFe and Al components in the soil samples, including ammoniumoxalate(Feo, Alo) and sodium pyrophosphate (Fep, Alp). The amount of poorlycrystalline Fe and Al (Feam and Alam) was established by takingthe difference between oxalate and pyrophosphate extractions (Feo–Fep and Alo–Alp). Analysis of Fe and Al was performed by flame atomicabsorption spectroscopy (Perkin Elmer AAnalyst 100). DOC concentra-tionwas determined by a TOC analyzer (Shimadzu 5050) after acidifica-tion to pH 3 with 0.1 M HCl.

2.4. Batch sorption experiments

To create the original sorption isotherms, thirty mL of initial DOCsolution with concentrations ranging from 0 to ~120 mg·L−1 wasadded to 3 g of soil, as presented elsewhere (Kothawala et al., 2009).However, here we only focus on equilibrium solutions using a startingconcentration of 61 mg·L−1 and compare the chemical characteristicsof DOC in the initial and final equilibrium solutions. The ionic strengthand pH of initial solutions were adjusted to a total ionic strengthof 0.001 mol L−1, using a 0 mg DOC L−1 solution consisting of10 mg·L−1 NaCl, 20 mg·L−1 CaCl2·H2O and 24 mg·L−1 K2SO4, witha pH of 4.8. Soil and DOC solutionswere combined in 50 mL centrifugevials, shaken by hand, and laid flat on a horizontal shaker for 24 h at4 °C to minimize microbial decomposition. Vials were placed uprightto settle for ~30 min prior to filtration with 0.45 μm glass fiber filters.The initial and final equilibrium solutions were analyzed for DOC con-centration, fluorescence index, SUVA, HPSEC and FTIR. A total of 43mineral soil horizons were included in the fluorescence index andSUVA analyses, while 22 horizonswere included in the HPSEC analysisand 10 were analyzed for FTIR (Table 2).

2.5. DOC sorption characteristics for mineral soils

Detailed characterisation of soil properties and adsorption charac-teristics were described elsewhere (Kothawala et al., 2009). Briefly,the maximum adsorption capacity (Qmax) of DOC to mineral soils was

derived using the Langmuir isotherm (Eq. (1)). The Langmuir equationexpresses a relationship between the amount of solute adsorbed ordesorbed (RE), in mg C·kg−1 soil, and the final equilibrium solutionconcentration (Cf) in mg·kg−1 soil, the binding affinity (k), the maxi-mum adsorption capacity (Qmax), in mg·kg−1 soil, and the desorption(b) term in mg·kg−1 soil.

RE ¼k⋅Qmax⋅Cf

! "

1þ k⋅Cf

! " −b: ð1Þ

The null point (np) was determined as the final concentrationwhen adsorption and desorption are equal (RE=0 in mg C·kg−1 soil)(Kothawala et al., 2009), yet are not included here since npwas not signif-icant to the findings of this study. The desorption potential (b) is definedas the amount of soil C released when the initial solution concentration(Ci) is 0 mg·L−1, and was determined from a modified Langmuir iso-therm using Ci rather than Cf (Kothawala et al., 2008).

2.6. Fluorescence spectrophotometry

A fluorescence spectrophotometer (Shimadzu RF 5301) was used toestablish the fluorescence intensity at a fixed excitation wavelength of370 nm, and emission wavelengths ranging from 400 to 700 nm, with aresolution of 2 nm. We ensured that the emissions at wavelengths 450and 500 nm were on either side of the maximum peak intensity andused to establish the fluorescence index (Ex 370 nm, Em 450/500 nm)as described by McKnight et al. (2001). Fluorescence was measured onthe initial DOC solution, as well as 43 solutions after equilibration withmineral soils. Distilled de-ionized water was used as a blank, and theemission spectrum for the blank was subtracted from the spectra of allsamples. Solutions were acidified to pH 3.5±0.2 to minimize quenchingof the fluorescence intensity due to organic–metal complexes. The fluo-rescence index of DOC solutions at ambient pH was very similar to thatacidified to pH 3.5 (FIambient=0.97×FIacid+0.05; R2=0.75, pb0.0001).We chose to use the fluorescence index for acidified solutions for pH con-sistency across all samples.We also established that the variability withinthe observed DOC concentration (43 to 79 mg·L−1) did not significantlyinfluence the fluorescence index.

The fluorescence index was determined for DOC desorbed from pod-zolic B horizons. The equilibrium solutions of batch incubations with aninitial solution concentration of 0 mg DOC·L−1 (adjusted to an ionicstrength of 0.001 mol L−1 and pH of 4.8) were analyze for fluorescencespectra as described above. The fluorescence index for DOC desorbedfrom volcanic B horizons was problematic since emission spectra at exci-tation 370 nm did not produce a sizeable peak, and fluorescence intensi-ties for several samples were too close to the baseline to be meaningful.

2.7. SUVA

Specific absorbance, SUVA, was measured using a UV–visible spectro-photometer (GENESIS 10, Thermo Electron) at 254 nm and 280 nmwitha path length of 1 cm for the DOC solution prior to and after equilibrationwith mineral soils. While thermodynamically stable aromatic rings areknown to absorb UV at 280 nm (Chin et al., 1994; McKnight et al.,1997), we found a stronger regression at 254 nm (R2=0.99, pb0.0001)and present SUVA at exclusively at 254 nm. SUVA was calculated by di-viding the absorbance with DOC concentration and is reported in unitsof liter per milligram carbon per meter (Weishaar et al., 2003). SUVAwas measured at ambient solution pH, and after acidification to pH 3.We found that SUVA (254 nm) of acidified solutions were generallylower than that at the ambient pH (3.4 to 7.5). While the discrepancy inSUVA readings for acidified and non-acidified solutions is likely due tothe protonation of carboxyl groups and disassociation of organo-metalcomplexes at lower pH, they were strongly correlated (SUVAacid=0.86×SUVAambient+0.42; R2=0.89, pb0.001). To avoid problems of

Table 2

Number of samples included in each soil horizon grouping for specific absorbance(SUVA), fluorescence index (FI), high performance size exclusion chromatography(HPSEC) and Fourier transform infrared spectroscopy (FTIR).

Horizon group Horizons SUVA and FI HPSEC FTIR

A horizons Ae/Aeh/Ah 10 5 3B horizons Bm/Bg 17 9 3Podzolic B Bfj/Bf/Bfh/Bcc 6 4 2Volanic B Bfj/Bf 10 4 2Total 43 22 10

336 D.N. Kothawala et al. / Geoderma 189–190 (2012) 334–342


DOC–iron aggregation at higher pHvalues,weused SUVAmeasured at pH3.5. Thus, acidified samples were interpreted for both SUVA and the fluo-rescence index.

2.8. High performance size exclusion chromatography (HPSEC)

The molecular weight distribution (MWD) was determined for theDOC solution prior to adsorption and after equilibration to 22 mineralsoils. HPSEC was performed using aWaters 600 controller (Waters Asso-ciates, Milford, MA) with a flow rate of 0.5 mL·min−1, a waters photodi-ode array 996 detector set at 254 nm, and a Rheodyne rotary injectorvalve equipped with a 200 μL sample loop, and an injection volume of100 μL. Chromatograms were produced with a Waters YMC 300 silicadiol size exclusion columnwith a gel bead pore size of 300 Å (Waters As-sociates, Milford, MA). The column had a broad calibration range from2.62×106 Da to 210 Da determined from using polystyrene sulfonatestandards with the following molecular weights: 2.62×106, 2.6×105,34,000, 7900, 4000, 1800, 1400 Da (American Polymer Standards, Men-tor, OH) and 210 Da (Fluka, Canada). The void volume (Vo) was deter-mined using a polystyrene sulfonate (PSS) standard with a molecularweight of 2.62 million Da, and the total permeation volume (Vt) was de-terminedwith acetone (56 Da). Themobile phase consisted of 0.1 MNaCl(Fisher Scientific, Nepean, ON)+0.002 M KH2PO4 (Fisher Scientific, Ne-pean, ON)+0.002 M K2HPO4 (BDH, Mississauga, ON), as previouslyused by Chin and Gschwend (1991).

Maximum peak height, total peak area, peak start, peak end, and re-tention volume (Rv) for the sample peak(s) were determined by the Mil-lennium software (Waters Associates, Milford, MA). Peaks were assessedby setting the high and low MW cutoffs at 1% and 2% of the peak height,respectively, or 50 Da, whichever was the higher value, as suggested byZhou et al. (2000). The MWD of DOC for samples was expressed interms of the number averagemolecularweight (Mn) (Eq. (2)), weight av-eragemolecularweight (Mw) (Eq. (3)), andpolydispersity (p) (Eq. (4)), asdescribed by Yau et al. (1979).Mn andMw are twomethods of calculatingan average value of a distribution,withMn being themean of the distribu-tion andMw being a weighted mean. WhenMn andMw values are equal,the polydispersity, p, is equal to 1.0, which is representative of a pure sub-stance. Signal intensities were acquired at 1.0 s time increments and ana-lyzed for Mn, Mw and p using a spreadsheet designed to divide thechromatogram into sections (i) that were 0.05 min wide (3 readings).The height of each section (hi) represents the UV detector response at254 nm, andMi is the MW of time section i.

Mn ¼

P

n

i¼1hi

P

n

i¼1

hiMi

ð2Þ

Mw ¼

P

n

i¼1hiMi

P

n

i¼1hi

ð3Þ

p ¼Mw

Mn

: ð4Þ

2.9. Fourier transform infrared spectrophotometry (FTIR)

Solid state FTIR spectroscopy was used to analyze air-dried DOC sam-ples on a Bruker Vector 22 with ATR MIRacle (Brucker Optics, Ettligen,Germany). The sample was prepared for FTIR analysis by pressing ontoa quartz plate. FTIR was run from 650 to 4500 cm−1, with a spectral res-olution of 0.2 cm−1 and spectrawere baseline correctedwith some noisereduction to smooth peaks. While peaks at several wavelengths wereidentified (Table 3), we did not find meaningful changes to all peaks,

and focused our analysis to that between carboxylic and aromatic groups.Due to the low sample sizewedid not attempt to interpret differences be-tween soil groups for FTIR results, but rather examine trends across all10 soil horizons. Specifically, peak height at wavelengths 1620 and1720 cm−1 were used to assess changes in the carboxylic and aromaticfunctional group content (Kalbitz et al., 1999) of the DOC solution beforeand after equilibrium with 10 mineral soils.

3. Results

A wide range of soil properties including soil pHCaCl2, texture, soil C,soil N, Soil C:N, Feam and Alam were included among the 43 mineral soilsamples (Table 1). Podzolic B horizons had the highest soil C and lowestpH (Table 1). Both podzolic B and volcanic B horizons had high levels ofFeam and Alam and correspondingly highQmax (Fig. 1a). In contrast the Aand B horizons (without Fe enrichment) had lower levels of Feam andAlam and lower Qmax. Adsorption capacity was positively related to thesum of Feam and Alam (Qmax=0.11 (Feam+Alam)+0.22, R2=0.52,pb0.0001, n=43). Desorption capacity in order from greatest to low-est was podzolic B>A horizons>B horizons and volcanic B horizons(Fig. 1b). Desorption capacity was positively related to the soil C content(b=1.53 Soil C+42.2, R2=0.46, pb0.0001, n=43).

3.1. SUVA

The SUVA of the original DOC solution (acidified) was 3.86±0.02 L mg C−1 m−1, whichwas lowered significantly after equilibrationwithmineral soils to a SUVA of 3.24 to 3.85 L mg C−1 m−1, and an aver-age final SUVA of 3.60±0.16 L mg C−1 m−1 (Fig. 2a). The averagechange in SUVA corresponds to a decrease in percent aromaticity from29 to 27%, based on a previously established linear relationship(Weishaar et al., 2003). Mean SUVA for the equilibrium solutions wereonly statistically different for podzol B horizons (3.48±0.14) and B hori-zons (3.70±0.11) (Tukey HSD, pb0.0001). The final SUVA was not sig-nificantly correlated to the Qmax (R2=0.05, p=0.18), but was weaklyrelated to b (R2=0.15, p=0.02).

3.2. Fluorescence index

The fluorescence index of the original (pH 3.5) DOC solutionwas 1.21(Fig. 2b), suggesting an aromaticity >32% based on the study byMcKnight et al. (2001). After equilibration to mineral soils the fluores-cence index rose for all but three soils, and ranged from 1.19 to 1.35,which corresponds to a final aromaticity ranging from >32 to 28%(McKnight et al., 2001). The DOC equilibration solutions of podzol B hori-zons experienced a significantly greater shift in fluorescence index rela-tive to other horizon groupings, with an average of 1.32±0.03,corresponding to an aromaticity change from >32 to 27% (Fig. 2),while the other horizon groups were not statistically different fromeach other (Fig. 2). The shift in fluorescence index is not large but is sig-nificant (pb0.0001) and clearly demonstrates the removal of aromaticcompounds in the solution phase DOC across all horizons.

Table 3

Characteristics of major peaks identified in FTIR spectra of DOC prior to and after equil-ibration with mineral soils.

Peak wave number(cm−1)

Assignment

1086 \C\O stretching of polysaccharides1405 \CH deformation of \CH3 and \CH bending of \CH2

1620 \Aromatic C_C stretch and/or asymmetric \COO− stretch1730 \C_O stretch of \COOH2960 to 3230 \OH groups and aliphatic C\H stretching3430 \OH stretch of phenolic OH



Podzol B horizons and volcanic B horizons have a highQmaxhoweverthe podzol B horizons are unique as they also have an especially high ca-pacity to desorb (b term) soil C. Thus, we attempted to determine thefluorescence index of DOC material desorbed from the indigenous soilsurfaces of podzolic B horizons, using an initial solution of 0 mgDOC·L−1. The DOC desorbed from podzol B horizons had a fluorescenceindex ranging from 1.63 to 1.41, which suggests an aromatic content ofapproximately 15 to 21%. Thus the aromatic content of desorbed DOCwas much lower than the original (61 mg·L−1) DOC solution (>32%),or thefinal equilibrium solutions (>32 to 27%). In contrast, other soil ho-rizons did not desorb a sufficient amount of DOC into solution phase toaccurately measure the fluorescence index. The variability in fluores-cence index explained by Qmax and b parameters alone was 16 and27%, with p-values of 0.005 and 0.0002, respectively. When combined,Qmax and b could explain 35% of the variability in FI (pb0.0001). Therewas a significant negative relationship between SUVA and the fluores-cence index (SUVA=−2.3FI+6.5, R2=0.27, pb0.0013), as expected.

3.3. Molecular weight distribution

We established changes to the molecular weight distribution (MWD)of DOC in the solution phase after equilibration, using HPSEC on 22 min-eral soils. An illustration of four representative soils in Fig. 3 shows thatthe original DOC solution had peak start and end times of 18.5 to 24.0min, corresponding to a molecular weight range of 16,890 to 470 Da

(Fig. 3). TheMn,Mw and p of the original DOC solution were closely com-parable when diluted with de-ionized water (Fig. 3a). This ensures that ashift in MWD observed for equilibrium DOC solutions reflects changes tothe molecular composition of DOC (shown in Fig. 3a), rather than beingan artifact of dilution.

The number–average molecular weight (Mn), weight average–molecular weight (Mw) and polydispersity (p) for the original DOCsolution were 1791 Da, 2468 Da and 1.38 respectively (Fig. 4). Thenumber–average molecular weight (Mn), of equilibrium DOC solu-tions was significantly reduced from 1791 Da in the original DOC so-lution to an average±standard deviation of 927±298 Da, rangingfrom 507 to 1445 Da (Fig. 4a). The average Mn for A horizon was1102±230 Da (n=5), B horizon was 1146±194 Da (n=9), volca-nic B horizon was 546±53 (n=4), and podzol B horizon was635±67 (n=4). The weight–average molecular weight (Mw) forthe original DOC solution was 2468 Da, and changed to an averageof 5463±10,645 Da after equilibration to mineral soils, rangingfrom 692 to 44,188 Da. The Mw for A horizons was 2220±1019 Da,B horizons was 12,295±15,509 Da, volcanic B horizons was 795±70 Da, and podzol B horizons was 955±181 Da (Fig. 4b). The poly-dispersity (p) of the original DOC solution was 1.38 and changed toan average of 4.80±8.54 after equilibration to mineral soils, rangingfrom 1.15 to 35.42. The p for A horizons was 2.02±0.80, B horizonswas 10.05±12.64, volcanic B was 1.46±0.05, podzol B was 1.50±0.23 (Fig. 4c). There was no significant relationship between p andQmax or b (p=0.00 Qmax+6.1, p>0.10, p=0.15 b−3.6, p>0.05).

The equilibrium DOC solutions observed for some A and B horizonsresulted in a wider MWD (larger Mw and p) compared to the originalDOC, which was due to the appearance of a high molecular weightpeak near the void volume, corresponding to a MW of 2.62×106 Da(example, Fig. 3b). In contrast, theMWD of DOC equilibrated with pod-zol B and volcanic B horizons (lowerMw and p closer to 1) did not havethe appearance of the highMWpeak (Fig. 3c and d). In the case of somepodzol B horizons, a low MW peak at a retention time of 24.2 min wasapparent, corresponding to a MW of 420 Da, and resulting in a higherpolydispersity (p). DOC solutions in equilibrium with volcanic B hori-zons did not have the appearance of any additional peaks (Fig. 3d).DOC remaining after equilibration eluted at a retention time between22.0 and 24.0 min, corresponding to 1520 to 470 Da. When the Mn forequilibrium solutions was related to the Qmax, we found a negative lin-ear relationship, whereby the greatest decline in Mn was observed forsoils with large Qmax (Fig. 5).

3.4. Fourier transform infrared spectroscopy (FTIR)

The FTIR spectrum for the original DOC samples contained seven de-fined peaks common to spectra of humic and fulvic acids and which canbe related to chemical characteristics (Table 3) (Kaiser and Zech, 1997;Kalbitz et al., 1999; Niemeyer et al., 1992). An illustration of the spectraof the original DOC solution and those after equilibration with mineralsoils are presented in Fig. 6, for five soils. After adsorption, the peak at1730 cm−1 was reduced to a shoulder relative to the peak at1620 cm−1. Peak abundance at 1730 and 1620 cm−1 has been used toestablish the relative amount of carboxyl relative to aromatic structures(Kalbitz et al., 1999), and 1730:1620 was reduced from 0.81 in the orig-inal solution to an average of 0.56±0.12 after equilibration, rangingfrom 0.42 to 0.81 across 10 mineral horizons. This reduction in1730:1620 indicates that DOC molecules with carboxyl groups wereremoved from solution more readily than aromatic DOC. The change in1730:1620 was positively related to Qmax (log FTIR=0.08(LogQmax)+1.08, R2=0.45, pb0.05, n=10). Changes to other ex-plored ratios based on peaks in Table 3 were not significant. Forexample, peaks at 1050:1620 indicated the relative amount of polysac-charides to aromatics and was 0.84 for the original DOC solution, andranged from 0.56 to 0.92 for solutions after equilibration with mineral

SU

VA

(L

mg

C-1

m-1

)

3.2

3.3

3.4

3.5

3.6

3.7

3.8

3.9

Flu

ore

scen

ce In

dex

1.15

1.2

1.25

1.3

1.35

1.4

A Horizons B Horizons Volcanic B Podzol B

a

a,b

a,b

b

Original

bb

b

a

Original

a

b

Fig. 2. Changes to the spectral indexes (a) SUVA at 254 nm (acidified) and (b) fluores-cence index (FI) (acidified), for all DOC solutions after equilibrationwithmineral soils cat-egorized by soil horizons, A horizons (n=10), B horizons (n=17), volcanic B (n=10),and podzolic B horizons (n=6). The original DOC solution had a SUVA of 3.86 and is indi-catedwith dashed line. Only groups not sharing the same lower case letter are significant-ly different from each other based on the Tukey–Kramer honestly significant test forcomparison of means with uneven sample sizes (pb0.05). The line in the middle ofboxes represents the median, with lower and upper parts of the box representing 25%and 75% of the distribution, respectively.



soils, and final values were not significantly related with the Qmax or bof mineral soils.

4. Discussion

After a gentle water extraction from a podzol LFH horizon, we createda solution with initial DOC concentration of 61 mg·L−1, which is withinthe range typical (20 to 90 mg·L−1) of the forest floor (Michalzik et al.,2001). The same source of DOC was used to generate sorption character-istics for themineral soils using the Langmuir isotherm, including the ad-sorption maxima (Qmax), and the desorption capacity (b) (Kothawalaet al., 2009). While our results are based on exchange reactions at onestarting concentration, we expect that differences observed between ho-rizon groupsmay becomemore exaggeratedwith greater adsorption, anddepend on the starting point along the isotherm curve. Here we used awide rangeofmineral soilswith variable physical and chemical propertiesthat have been well-characterized in terms of adsorption and desorptioncharacteristics. We now distinguish how sorption characteristics influ-ence observed changes to the molecular composition of DOC afterequilibration.

4.1. SUVA and fluorescence index

Two optical spectroscopic techniques, SUVA and the fluorescenceindex, indicated a reduction in the aromatic content of DOC upon equili-brationwith awide range ofmineral soils (Fig. 2a and b).While the selec-tive adsorption of aromatic constituents onto pure goethite (Zhou et al.,2001) has been previously reported, here we present the variability ob-served between horizon groups collected from natural environments.The observed extent of change in aromaticity is not substantialwith either

index, yet a consistent reduction was observed across all mineral soils.The full range of thefluorescence index including terrestrial andmicrobialend-members spans from ~10 to 30% aromaticity, respectively. The aro-maticity of DOC shifted only slightly from >32% to b32% after equilibra-tion, yet it should be noted that the specificity of the fluorescence indexin the region near 32% aromaticity is weak due to a non-linear relation-ship (McKnight et al., 2001). The decrease in aromaticity was greatestfor podzol Bhorizonswhichhadhighest adsorption capacities,which sug-gest that DOC shifted from terrestrial-like tomicrobial-like DOC. Thus, theconcentration of DOC exported from podzols would not only be reducedsubstantially, but the chemical composition would be altered relative toDOC produced in the organic horizons. The variability in fluorescenceindex was best explained by Qmax and b combined, which suggests thatthe chemical composition of DOC leached through podzol mineral soilswas influenced by adsorption andmaterial released from the soil surface.This was confirmed by measuring the fluorescence index of desorbedDOC from podzol B horizons, which had much lower aromatic content(15 to 21%,moremicrobial) than found in the original or even the equilib-rium solutions of all horizons. This observation supports the possibilitythat ligand exchange occurred whereby solution phase aromatic struc-tures (strong ligands) replaced less aromatic (weaker ligands) soil C. Inthe case of volcanic soils, lower levels of pre-existing soil C resulted inlower levels of desorption (Kothawala et al., 2009), and thus the observedchange in aromaticitywasmore a result of selective adsorptionof aromat-ic rich DOC, and less so due to desorption.

The relationships between SUVA and fluorescence index were nega-tively related and thus provided comparable results regarding reducedaromaticity, however the fluorescence index resulted in greater differen-tiation of the podzol B horizons from other groups. This discrepancy ishard to attribute to spectral interferences since all solutionswere acidified

50

100

150

200

250

300

10 12 14 16 18 20 22 24 26 28 30

Retention Time (mins)

UV

Ab

so

rba

nc

e (

AU

at

25

4 n

m)

0

50

100

150

200

250

300

10 12 14 16 18 20 22 24 26 28 30


UV

Ab

so

rban

ce (

AU

at

254 n

m)

50

100

150

200

250

300

10 12 14 16 18 20 22 24 26 28 30


UV

ab

so

rban

ce (

AU

at

254 n

m)

0

50

100

150

200

250

300

10 12 14 16 18 20 22 24 26 28 30


UV

ab

so

rban

ce (

AU

at

254 n

m)

420 Da

a) Gleysol Ah

c) Podzol Bf

b) Luvisol Bt

d) Volcanic Bfj

16 890 Da470 Da

Original

Grey Luvisol Bt (SK: Old Aspen)

Original

Podzol Bf (ON: Plastic Lake)

Original

Original 10 X dilution

Ah Luvic Gleysol (ON: Groundhog)

Original

Volcanic Bfj (BC: Douglas Fir)

420 Da

2.6 x 106 Da 16 890 Da

470 Da

Fig. 3. HPSEC for DOC solutions after equilibration with four examples representative of soil horizons, (a) Ah horizons of a Luvic gleysol, (b) Bt horizon of a Gray luvisol, (c) Bf ho-rizons of a podzol, and (d) Bfj horizons of a Volanic soil. Gray lines represent the chromatogram of the original solution and the dark line represents the equilibrium solution. Thethin gray line in figure (A) is a 10 times dilution of the original solution, which shows no artificial shift in retention time.



to reduce Fe interferences (Weishaar et al., 2003) or possible quenching(Lakowicz, 2006). NO3

− was not considered to pose any interferences inSUVA as concentrations were ~1.8 mg·L−1, well below that expected tocause interferences (>5 mg·L−1) (Weishaar et al., 2003). Ultimately, in-formation generated from both indexes provided greater informationthan use of only one index.

4.2. Molecular weight distribution

The fluorescence index and SUVA are useful measures of aromaticity,however both are limited in that they provide a single average valuerepresenting the change in aromaticity over a broad distribution of mol-ecules. The HPSEC technique holds the advantage of determining the fullMWdistribution, andwhen used in combinationwith SUVA and fluores-cence index can be powerful in providing amore complete assessment ofchanges in DOC. HPSEC chromatograms effectively showed the disap-pearance of high MW material from the original DOC solution (~5000to 1000 Da), which is in agreement with studies using pure minerals(Chorover and Amistadi, 2001; Meier et al., 1999). While the decreaseinMn represents the change in average MW, the Mw reflects the changeinMWdistribution. Consequently, the largeMw for B horizons reflect theunexpected appearance of extremely high MW peaks representing DOCnot found in the original solution. This new peak was observed in equi-librium solutions at the void volume (2.62×106 Da) for B horizons,whichmay represent a small quantity of humifiedmaterial with high ar-omatic content. In contrast, low MW peaks (420 Da) appeared for thepodzol B horizons, which is likely desorbed DOC. This desorbed lowMW material likely also contributed to the lower aromaticity observedfrom absorbance and fluorescence indexes. The podzolic B horizonshave statistically greater desorption potential than other soil groups(Kothawala et al., 2009). The fluorescence index results showed signifi-cantly lower aromaticity for the podzol B horizons which supports theHPSEC findings, since low MW material is generally less aromatic thanhigh MWmaterial (Cabaniss et al., 2000). The release of low MWmate-rial likely contributed to the observation of significantly lower aromaticcontent in the equilibrium solutions of podzol B horizons. In contrast,the appearance of high MW material in B horizons was likely highly ar-omatic, but may have been balanced out by the adsorption of aromaticstructures in the original DOC solution.

The relationship between soil Qmax and the change in Mn of DOC(Fig. 6) suggests that increasing DOC adsorption capacity results in aproportional reduction in the average MW of DOC remaining in solu-tion. Meier et al. (1999) found that a shift to lower Mw was dependenton the position along the sorption isotherm for goethite and kaolinite.Our results focus on one point on the isotherm, however include awide range of sorption characteristics.We found that soils with greatestQmax resulted in the greatest lowering ofMn. The findings for Mwwereconfounded due to the appearance of extremely high MW material in

a

b

c

Original

Original

Fig. 4. The molecular weight distribution of dissolved organic carbon after adsorptionto mineral soil horizons, A horizons (n=5), B horizons (n=9), volcanic B (n=4),and podzolic B (n=4), based on (a) the number–average molecular weight (Mn) inDaltons (Da), (b) the weight–average molecular weight (Log Mw), and (c) polydisper-sity (Log p). Dashed lines represent the original DOC solution added prior to equilibrat-ing with soil horizons.

R2= 0.38

p < 0.002

Fig. 5. Relationship between the equilibrium number–average molecular weight (Mn)of dissolved organic carbon in solution phase (in Daltons) and the maximum adsorp-tion capacity (Qmax) for soil horizon groups, A horizons (open diamonds ◊, n=5), Bhorizons (open circles ○, n=9), podzol B (filled in circles ●, n=4), and volcanic B(stars *, n=4), based on the Langmuir adsorption isotherm.

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

700

900

1100

1300

1500

1700

1900

2100

2300

2500

2700

2900

3100

3300

3500

3700

3900

Wavenumber (cm-1)

17

30 1

63

5

14

05

10

86

34

30

32

30

29

60

Original

14

10

59

12

1

Fig. 6. FTIR spectra of the original DOC solution and after equilibration with five miner-al soils that are representative of the ten samples included in analysis. Numbered spec-tra correspond to mineral horizons; 1 volcanic Bfj, 12 brunisol Bm, 59 podzol Bf, 10gleysol Bg, and 14 luvisol Ae. Baseline of spectra has been off-shifted to improve visualclarity.



several B horizons.While thesefindings are not entirely unexpected,weclearly demonstrate the selective removal of highmolecularweightma-terial (>1000 Da) over low molecular weight material (b1000 Da),with an increasing shift towards lower Mn for soils with high Qmax.The observed lowering of Mn and reduction in aromaticity suggeststhat DOM retained to soil surfaces was the large molecular weightDOM with aromatic structures.

4.3. Fourier transform infrared spectroscopy

The observed decrease in the FTIR 1730:1620 ratio for DOC afterequilibrationwithmineral soils shows that the proportion of protonatedcarboxyl groups declined in comparison to aromatic structures. Sincewewere restricted to analysis of a smaller number of samples in FTIR analy-sis (n=10), we are unable to statistically differentiate the relative differ-ence between horizon groups, yet the observed decrease in carboxylgroups to aromatics was observed across all samples. Carboxyl groupshave been previously identified as important functional groups on DOCfor adsorption to mineral soils, especially by ligand exchange (Gu et al.,1995; Parfitt et al., 1977). Alternatively, another study suggested theprinciple mechanism of adsorption to iron oxides was from physical ad-sorption of hydrophobics due to favorable entropy changes (Jardineet al., 1989). A study examining the chemical composition of DOC perco-lating through a forest soil found that the percent of neutrals increasedwith depth, while hydrophobic and hydrophilic acids decreased withdepth (Qualls and Haines, 1991b). Since both hydrophobic and hydro-philic acids contain carboxyls groups, the findings of Qualls and Haines(1991b) may be similar to ours. When results suggesting a reduction incarboxyl groups relative to aromatic groups are combined with resultsfrom optical analysis (SUVA and fluorescence index), both clearly identi-fying a reduction in aromatic content, we can conclude that the reduc-tion in carboxyl groups was greater than the reduction in aromatics,whichwas greater than the reduction in other functional groups. Conclu-sions regarding changes in aromatic content alone would not have beenpossible from the FTIR results alone. Since FTIR peak intensities are notdirectly related to DOC concentration, the relative intensity of twopeaks, in this case 1730 and 1620, was necessary to establish changesin the functional group composition of DOC (Niemeyer et al., 1992).Other relationships such as the polysaccharides to aromatics ratio didnot change systematically which was somewhat surprising since poly-saccharides might be expected to be hydrophilic and weakly adsorbingrelative to aromatics. However it is challenging to mechanistically attri-bute if polysaccharides in equilibrium solutions would have resultedfrom absorption, desorption or dual exchange between soil and solutionphase based on result of this study.

4.4. Ecological significance of changes in DOC functional group

composition

When results from all four analyses are combined, the overall loss ofDOC functional groups from greatest to least was in the order: carboxylgroups>aromatic groups≫other functional groups, and an overallshift from higher to lower MW. It was not the definitive goal of thisstudy to define exact mechanisms for the selective loss of DOC struc-tures, since soils with widely differing chemical and physical propertieswere included. Yet, we suspect that a combination of several mecha-nisms were responsible for observed changes in DOC structure includ-ing ligand exchange, physical adsorption and cation bridging. Resultsof this study emphasize that unless a soil is enrichedwith Fe, the overallorganic carbon retention in soils remains low,with a proportionally lowchange in DOC structure. In fact, we found no significant difference be-tweenA and B horizons (without Fe enrichment) in terms of adsorptionor desorption capacity, and changes to the chemical composition afterequilibration were often comparable. Thus, catchment soils without Feenrichment may result in the export of greater amounts of DOC toaquatic ecosystems, which may be chemically similar to that produced

within the organic horizons. In contrast, podzolic and volcanic B hori-zons (with Fe enrichment) can be expected to retain much higheramounts of organic carbon,with greater alteration to the chemical com-position of DOC. The functional group composition of DOC is linked toits ecological function, such that smaller MW, less aromatic DOC tendto be more biologically available than larger MWmaterial with greateraromatic content (Cabaniss et al., 2000). Other studies have found thatDON is more soluble than DOC, resulting in lower DOC:DON ratios dur-ing passage through mineral soils (Kaiser and Zech, 2000; Kothawalaand Moore, 2009), and in organic matter of greater biological qualityto aquatic ecosystems. While the quality of organic matter exportedfrom soils with high Fe enrichment may be of greater quality from a bi-ological perspective, the amount of DOC would be far lower than soilswithout Fe enrichment. Thus, regardless of the ecological question athand, it may be of primary relevance to identify if catchment soils areenriched in Fe to predict the magnitude of changes to the quantityand quality of DOC during passage through mineral horizons.

5. Conclusions

By utilizingmultiple analytical techniques to characterize changes tothemolecular and chemical structure of DOC, we were able to generatea robust account of the selective adsorption of DOC to a range ofmineralsoil horizons.While results of this study are restricted to comparisons ofbatch incubations which are not directly transferable to the field, weidentified relative changes to the structural composition of DOC be-tween mineral soil horizons, using one DOC source. SUVA and fluores-cence index revealed that soil solution equilibration with mineral soilsconsistently reduced the aromatic content of DOC. The fluorescenceindex distinguished podzol B horizons to have greater reduction in aro-matic content than other horizons included in the study. In particular,we found that DOC desorbed from mineral soil surfaces had a substan-tially lower fluorescence index than equilibrium solution DOC. Thisfinding emphasizes the need to be aware that while overall DOC con-centration may reduce substantially (e.g. from 61 to b30 mg L in pod-zolic B horizons), the resulting composition of DOC in solution phasecan reflect not only just adsorption, but also desorption processes. Fu-ture studies do need to place more emphasis on characterisation ofDOC desorbed from soils using techniques in addition to fluorescenceas done here. In contrast, the adsorption capacity of volcanic soil hori-zons is large, and often comparable to that of podzolic B horizons, how-ever there was less release of DOC during equilibration. The resultingsoil solution chemistry of the volcanic B horizons was thus largely dueto adsorption. The release of extremely high (2.6×106 Da) and low(420 Da) molecular weight material was clearly illustrated from sizeexclusion chromatography, for luvisols and podzols, respectively. Byviewing the full size distribution of DOC, we are able to visualize thecontribution of newmaterial released from the soil surface. The analysisof functional group composition using FTIR revealed that while the aro-matic content may decrease due to equilibration, the carboxyl contentdecreases more so. Lastly, we found that soils with the greatest capacityto adsorb DOC had the greatest decrease in molecular weight distribu-tion and aromatic content (based on the fluorescence index). From anecological perspective, the Fe status of mineral horizons may be of pri-mary important to consider when attempting to predict the quantityand quality of DOC exported from soils into aquatic ecosystems.

Acknowledgments

This study was funded by the Natural Sciences and Engineering Re-search Council of Canada and BIOCAP Canada, and funding was providedto DNK by the Global Environmental and Climate Change Centre. Wethank Mike Dalva, Isabelle Gagnon, Ed Hudson, Doug Evans and PeterDillon for the technical support and scientific discussions.



References

Cabaniss, S.E., Zhou, Q.H., Maurice, P.A., Chin, Y.P., Aiken, G.R., 2000. A Log-normal dis-tribution model for the molecular weight of aquatic fulvic acids. EnvironmentalScience and Technology 34 (6), 1103–1109.

Canada Soil Survey Committee, S.o.S.C. (Ed.), 1998. The Canadian System of Soil Classi-fication. Can. Dep. Agric. Publ. 1646. Supply and Services Canada, Ottawa, Ontario.

Chin, Y.P., Gschwend, P.M., 1991. The abundance, distribution and configuration ofporewater organic colloids in recent sediments. Geochimica et CosmochimicaActa 55 (5), 1309–1317.

Chin, Y.-P., Aiken, G., O'Loughlin, E., 1994. Molecular weight, polydispersity, and spec-troscopic properties of aquatic humic substances. Environmental Science andTechnology 28, 1853–1858.

Chin, Y.P., Traina, S.J., Swank, C.R., Backhus, D., 1998. Abundance and properties ofdissolved organic matter in pore waters of a freshwater wetland. Limnology andOceanography 43 (6), 1287–1296.

Chorover, J., Amistadi, M.K., 2001. Reaction of forest floor organic matter at goethite,birnessite and smectite surfaces. Geochimica et Cosmochimica Acta 65 (1),95–109.

Ghosh, S., Wang, Z.Y., Kang, S., Bhowmik, P.C., Xing, B.S., 2009. Sorption and fraction-ation of a peat derived humic acid by kaolinite, montmorillonite, and goethite.Pedosphere 19 (1), 21–30.

Gu, B.J.Z., Schmitt, Z., Liang, L., McCarthy, J.F., 1994. Adsorption and desorption of nat-ural organic matter on iron oxides: mechanims and models. Environmental Sci-ence and Technology 28, 39–46.

Gu, B.H., Schmitt, J., Chen, Z., Liang, L.Y., Mccarthy, J.F., 1995. Adsorption and desorptionof different organic–matter fractions on iron-oxide. Geochimica et CosmochimicaActa 59 (2), 219–229.

Hruska, J., Köhler, S.J., Bishop, K., 1999. Buffering processes in a boreal dissolved organiccarbon-rich stream during experimental acidification. Environmental Pollution106, 55–65.

Jardine, P.M., Weber, N.L., Mccarthy, J.F., 1989. Mechanisms of dissolved organic-carbon adsorption on soil. Soil Science Society of America Journal 53 (5),1378–1385.

Kaiser, K., 2003. Sorption of natural organic matter fractions to goethite (alpha-FeOOH): effect of chemical composition as revealed by liquid-state C-13 NNRand wet-chemical analysis. Organic Geochemistry 34 (11), 1569–1579.

Kaiser, K., Zech, W., 1997. Competitive sorption of dissolved organic matter fractions tosoils and related mineral phases. Soil Science Society of America Journal 61 (1),64–69.

Kaiser, K., Zech, W., 2000. Sorption of dissolved organic nitrogen by acid subsoil hori-zons and individual mineral phases. European Journal of Soil Science 51 (3),403–411.

Kaiser, K., Guggenberger, G., Zech, W., 1996. Sorption of DOM and DOM fractions to for-est soils. Geoderma 74, 281–303.

Kaiser, K., Guggenberger, G., Haumaier, L., Zech, W., 2002. The composition of dissolvedorganic matter in forest soil solutions: changes induced by seasons and passagethrough the mineral soil. Organic Geochemistry 33, 307–318.

Kalbitz, K., Geyer, W., Geyer, S., 1999. Spectroscopic properties of dissolved humic sub-stances—a reflection of land use history in a fen area. Biogeochemistry 47,219–238.

Kothawala, D.N., Moore, T.R., 2009. Adsorption of dissolved nitrogen by forest mineralsoils. Canadian Journal of Forest Research (Revue Canadienne de RechercheForestiere) 39 (12), 2381–2390.

Kothawala, D.N., Moore, T.R., Hendershot, W.H., 2008. Adsorption of dissolved organiccarbon to mineral soils: a comparison of four isotherm approaches. Geoderma 148(1), 43–50.

Kothawala, D.N., Moore, T.R., Hendershot, W.H., 2009. Soil properties controlling theadsorption of dissolved organic carbon to mineral soils. Soil Science Society ofAmerica Journal 73, 1831–1842.

Lakowicz, J.R., 2006. Principles of Fluorescence Spectroscopy, 3rd ed. Springer-Verlag.McDowell, W.H., Likens, G.E., 1988. Origin, composition and flux of dissolved organic

carbon in the hubbard brook valley. Ecological Monographs 58, 177–195.

McKnight, D.M., Harnish, R., Wershaw, R.L., Baron, J.S., Schiff, S., 1997. Chemical charac-teristics of particulate, colloidal, and dissolved organic material in Loch Vale water-shed, Rocky Mountain National Park. Biogeochemistry 36, 99–124.

McKnight, D.M., Boyer, E.W., Westerhoff, P.K., Doran, P.T., Kulbe, T., Andersen, D.T.,2001. Spectrofluorometric characterization of dissolved organic matter for indica-tion of precursor organic material and aromaticity. Limnology and Oceanography46 (1), 38–48.

Meier, M., Namjesnik-Dejanovic, K., Maurice, P.A., Chin, Y.P., Aiken, G.R., 1999. Fraction-ation of aquatic natural organic matter upon sorption to goethite and kaolinite.Chemical Geology 157 (3–4), 275–284.

Michalzik, B., Kalbitz, K., Park, J.H., Solinger, S., Matzner, E., 2001. Fluxes and concentra-tions of dissolved organic carbon and nitrogen—a synthesis for temperate forests.Biogeochemistry 52 (2), 173–205.

Moore, T.R., DeSouza, W., Koprivnjak, J.F., 1992. Controls on the sorption of dissolvedorganic carbon by soils. Soil Science 154, 120–129.

Niemeyer, J., Chen, Y., Bollag, J.M., 1992. Characterization of humic acids, composts, andpeat by diffuse reflectance fourier-transform infrared-spectroscopy. Soil ScienceSociety of America Journal 56 (1), 135–140.

Parfitt, R.L., Fraser, A.R., Farmer, V.C., 1977. Adsorption on hydrous oxides .3. Fulvic-acidand humic-acid on goethite, gibbsite and imogolite. Journal of Soil Science 28 (2),289–296.

Qualls, R.G., Haines, B.L., 1991a. Fluxes of dissolved organic nutrients and humic sub-stances in a deciduous forest. Ecology 72, 254–266.

Qualls, R.G., Haines, B.L., 1991b. Geochemistry of dissolved organic nutrients in waterpercolating though a forest ecosystem. Soil Science Society of America Journal1112–1123.

Qualls, R.G., Haines, B.L., Swank, W.T., Tyler, S.W., 2002. Retention of soluble organicnutrients by a forested ecosystem. Biogeochemistry 61 (2), 135–171.

Ross, G.I., Wang, C., 1993. Extractable Al, Fe, Mn, and Si. In: Carter, M.R. (Ed.), Soil Sam-pling and Methods of Analysis. Lewis Publishers, Boca Raton, FL, pp. 239–246.

Sheldrick, B.H., Wang, C., 1993. Particle size distribution. In: Carter, M.R. (Ed.), SoilSampling and Methods of Analysis. Lewis Publishers, Boca Raton, FL, pp. 499–511.

Sposito, G. (Ed.), 2004. The Surface Chemistry of Natural Particles. Oxford UniversityPress, New York.

Tan, K.H. (Ed.), 2003. Humic Matter in Soil and the Environment: Principles and Con-troversies. Marcel Dekker, Inc., New York.

Thacker, S.A., Tipping, E., Baker, A., Gondar, D., 2005. Development and application offunctional assays for freshwater dissolved organic matter. Water Research 39(18), 4559–4573.

Tipping, E., Hurley, M.A., 1992. A unifying model of cation binding by humic sub-stances. Geochimica et Cosmochimica Acta 56, 3627–3641.

Tipping, E., Rieuwerts, J., Pan, G., Ashmore, M.R., Lofts, S., Hill, M.T.R., Farago, M.E.,Thornton, I., 2003. The solid-solution partitioning of heavy metals (Cu, Zn, Cd,Pb) in upland soils of England and Wales. Environmental Pollution 125 (2),213–225.

van Hees, P.A.W., Jones, D.L., Nyberg, L., Holmstrom, S.J.M., Godbold, D.L., Lundstrom,U.S., 2005. Modelling low molecular weight organic acid dynamics in forest soils.Soil Biology and Biochemistry 37 (3), 517–531.

Weishaar, J.L., Aiken, G.R., Bergamaschi, B.A., Fram, M.S., Fujii, R., Mopper, K., 2003.Evaluation of specific ultraviolet absorbance as an indicator of the chemical com-position and reactivity of dissolved organic carbon. Environmental Science andTechnology 37 (20), 4702–4708.

Yau, W.W., Kirkland, J.J., Bly, D.D. (Eds.), 1979. Modern Size-Exclusion Liquid Chroma-tography: Practice of Gel Permeation and Gel Filtration Chromatography. Wiley,New York, NY.

Zhou, Q.H., Cabaniss, S.E., Maurice, P.A., 2000. Considerations in the use of high-pressure size exclusion chromatography (Hpsec) for determining molecularweights of aquatic humic substances. Water Research 34 (14), 3505–3514.

Zhou, Q., Maurice, P.A., Cabaniss, S.E., 2001. Size fractionation upon adsorption of fulvicacid on geothite: equilibrium and kinetic studies. Geochimica et CosmochimicaActa 65, 803–812.


author's personal copy - uni-muenster.de · organic carbon (doc) can alter substantially due...

Documents