Geochemistry of Dissolved Organic Nutrients in Water Percolating through a Forest Ecosystem Robert G. Quails* and Bruce L. Haines ABSTRACT Dissolved organic matter (DOM) is a major vehicle for the trans- location and loss of N and P from forest ecosystems. The chemical properties of DOM and its interactions with soil surfaces are crucial in determining the mobility of these organic nutrients. We fraction- ated DOM from throughfall; all soil horizons (Ultisols and Incep- tisols), and stream water from an Appalachian mountain forest ecosystem into hydrophobic or hydrophilic acids, neutrals, and bas- es. \Ve analyzed each fraction for dissolved organic C (DOC), N (DON), and P (DOP). Most of the DOC was in the acid fractions, with the humic fractions (hydrophobic acids and phenols) comprising 35 to 57% of the DOC in all samples except summer throughfall. Concentrations of all fractions declined with depth in the soil. As a percent of total DOC, the humics declined with depth, whereas the hydrophilic neutrals increased. Bases, which we expected to contain cationic amino groups, were < 2.5% of the DOC. Instead, most DON was in the humic. hydrophilic acid, and hydrophilic neutral fractions. Most DOP occurred in the hydrophilic acid, humic, and hydrophilic neutral fractions. The functional groups in which N and P occur had little influence on the behavior of most of the DOM as a whole since: (i) cationic DOM was such a minor component, and (ii) P was simply too rare to influence the anionic behavior of many molecules. Never- theless, for those molecules in which P did occur, P may have in- fluenced their behavior since a large percentage of the DOP was in the hydrophilic acid (i.e., anionic) fraction. The carboxylic and phe- nolic functional groups, or in some cases the neutrality, of the DOM molecules appeared to be much more important than N-containing groups in influencing the behavior of the N carried passively by the DOM. D ISSOLVED ORGANIC MATTER plays an important role in interrestrial and stream ecosystems be- cause it: (i) is a major mode of export of N and P in many ecosystems that are not experiencing severe ero- sion (Sollins and McCorison, 1981); (ii) plays a major role in determining the balance of soil N and P over the time of soil development; (iii) affects soil structure, e.g., resulting in a deeper redistribution of soil organic matter and the coating of clay particles with organic matter, (iv) is the principal vehicle for movement of Al and Fe in soil; and (v) provides a potential source of carbon for microbial growth (Meyer et al., 1987). Characterization of DOM in soil and stream water is very difficult because it consists of a myriad of com- pounds, none of which is present in large proportions. Four general approaches have been used to charac- terize DOM: (i) cataloging individual compounds; (ii) analyzing for broad biochemical classes of compounds such as proteins, free amino acids, monosaccharides, pplysaccharides, lipids, pplyphenols, and tannins; (iii) dividing into molecular-size classes; and (iv) compre- R.G. Quail, School of Forestry and Environmental Studies, Duke Univ.. Durham, NC, 27706; and B.L. Haines, Inst. of Ecology and Botany Dep., Univ. of Georgia, Athens, GA 30602. Contribution from the Inst. of Ecology and Botany Dep., Univ. of Georgia. Re- ceived 22 Jan. 1990. 'Corresponding author. Published in Soil Sci. Soc. Am. J. 55:1112-1123 (1991). hensively fractionating into hydrophobic and hydro- philic acids, neutrals, and bases (Leenheer and Huffman, 1976). The first approach can be termed a bottom-up approach, and the last two can be termed top-down approaches. In choosing an approach, we must consider the mo- tives for characterizing DOM. Soil scientists, ecolo- gists. and geochemists are usually concerned with questions of its (i) mode of forma tipn, (ii) general types of chemical structures, (iii) mobility and behavior towards surfaces (cation-, anion-, and ligand-exchange sites, and hydrophobic surfaces), (iv) metal complex- ation and transport of N and P, (v) influence on the pH of the solution, and (vi) quality as a substrate for decomposers. A partial list of a small identifiable percentage of the DOM is inevitably biased. Most of the DOM is macromolecular, with varying structure (Thurman, 1985). For example, although numerous small car- boxylic acids have been identified in natural waters, the sum of all the identifiable small acids make up only a small percentage of the DOM (summarized in Thurman, 1985) and, therefore, are not representative of the carboxylic acid fraction as a whole. Analysis of broader biochemical categories has been used extensively to infer mode of formation, struc- tures, mobility, and substrate quality but, unfortu- nately, humic substances do not fit well into these traditional categories. Carbohydrates and amino acids are intimate parts of the humic molecules, but the surface properties of humic molecules do not resemble those of carbohydrates and proteins. Furthermore, they may not be as biodegradable as free carbohydrates and proteins are. For example, most of the amino acids and a large portion of the carbohydrates in DOM from an Oregon river were bound to anionic humic sub- stances (Lytle and Perdue, 1981; Sweet and Perdue, 1982). Fractionation of DOM according to molecular size has been very widely used and can account for all of the DOM in a given sample (the third approach). While molecular size is certainly a fundamental aspect of structure, the molecular-size distribution does not directly inform us about mode of formation, surface reactivity, acidity, or biodegradability. Large molec- ular size is sometimes tacitly associated with resistance to bipdegradation. The large-molecule fraction could contain proteins, complex carbohydrates, humic sub- stances, and tannins — components that differ radi- cally in their degradability. In the work reported here, we used the fractionation procedure of Leenheer and Huffman (1976) and Leen- heer (1981). This fractionation procedure is based on surface properties at various pHs. Substances found in the various fractions are listed in Table 1. The di- vision between hydrophobic and hydrophilic com- ponents is arbitrary and is based on the capacity factor for the compounds on nonionic XAD-8 resin (Sigma 1112
Transcript
Geochemistry of Dissolved Organic Nutrients in Water Percolating
through a Forest Ecosystem
Robert G. Quails* and Bruce L. Haines
ABSTRACT
Dissolved organic matter (DOM) is a major vehicle for the trans-location and loss of N and P from forest ecosystems. The chemicalproperties of DOM and its interactions with soil surfaces are crucialin determining the mobility of these organic nutrients. We fraction-ated DOM from throughfall; all soil horizons (Ultisols and Incep-tisols), and stream water from an Appalachian mountain forestecosystem into hydrophobic or hydrophilic acids, neutrals, and bas-es. \Ve analyzed each fraction for dissolved organic C (DOC), N(DON), and P (DOP). Most of the DOC was in the acid fractions,with the humic fractions (hydrophobic acids and phenols) comprising35 to 57% of the DOC in all samples except summer throughfall.Concentrations of all fractions declined with depth in the soil. As apercent of total DOC, the humics declined with depth, whereas thehydrophilic neutrals increased. Bases, which we expected to containcationic amino groups, were < 2.5% of the DOC. Instead, most DONwas in the humic. hydrophilic acid, and hydrophilic neutral fractions.Most DOP occurred in the hydrophilic acid, humic, and hydrophilicneutral fractions. The functional groups in which N and P occur hadlittle influence on the behavior of most of the DOM as a whole since:(i) cationic DOM was such a minor component, and (ii) P was simplytoo rare to influence the anionic behavior of many molecules. Never-theless, for those molecules in which P did occur, P may have in-fluenced their behavior since a large percentage of the DOP was inthe hydrophilic acid (i.e., anionic) fraction. The carboxylic and phe-nolic functional groups, or in some cases the neutrality, of the DOMmolecules appeared to be much more important than N-containinggroups in influencing the behavior of the N carried passively by theDOM.
DISSOLVED ORGANIC MATTER plays an importantrole in interrestrial and stream ecosystems be-
cause it: (i) is a major mode of export of N and P inmany ecosystems that are not experiencing severe ero-sion (Sollins and McCorison, 1981); (ii) plays a majorrole in determining the balance of soil N and P overthe time of soil development; (iii) affects soil structure,e.g., resulting in a deeper redistribution of soil organicmatter and the coating of clay particles with organicmatter, (iv) is the principal vehicle for movement ofAl and Fe in soil; and (v) provides a potential sourceof carbon for microbial growth (Meyer et al., 1987).
Characterization of DOM in soil and stream wateris very difficult because it consists of a myriad of com-pounds, none of which is present in large proportions.Four general approaches have been used to charac-terize DOM: (i) cataloging individual compounds; (ii)analyzing for broad biochemical classes of compoundssuch as proteins, free amino acids, monosaccharides,pplysaccharides, lipids, pplyphenols, and tannins; (iii)dividing into molecular-size classes; and (iv) compre-
R.G. Quail, School of Forestry and Environmental Studies, DukeUniv.. Durham, NC, 27706; and B.L. Haines, Inst. of Ecology andBotany Dep., Univ. of Georgia, Athens, GA 30602. Contributionfrom the Inst. of Ecology and Botany Dep., Univ. of Georgia. Re-ceived 22 Jan. 1990. 'Corresponding author.
Published in Soil Sci. Soc. Am. J. 55:1112-1123 (1991).
hensively fractionating into hydrophobic and hydro-philic acids, neutrals, and bases (Leenheer andHuffman, 1976). The first approach can be termed abottom-up approach, and the last two can be termedtop-down approaches.
In choosing an approach, we must consider the mo-tives for characterizing DOM. Soil scientists, ecolo-gists. and geochemists are usually concerned withquestions of its (i) mode of forma tipn, (ii) general typesof chemical structures, (iii) mobility and behaviortowards surfaces (cation-, anion-, and ligand-exchangesites, and hydrophobic surfaces), (iv) metal complex-ation and transport of N and P, (v) influence on thepH of the solution, and (vi) quality as a substrate fordecomposers.
A partial list of a small identifiable percentage ofthe DOM is inevitably biased. Most of the DOM ismacromolecular, with varying structure (Thurman,1985). For example, although numerous small car-boxylic acids have been identified in natural waters,the sum of all the identifiable small acids make uponly a small percentage of the DOM (summarized inThurman, 1985) and, therefore, are not representativeof the carboxylic acid fraction as a whole.
Analysis of broader biochemical categories has beenused extensively to infer mode of formation, struc-tures, mobility, and substrate quality but, unfortu-nately, humic substances do not fit well into thesetraditional categories. Carbohydrates and amino acidsare intimate parts of the humic molecules, but thesurface properties of humic molecules do not resemblethose of carbohydrates and proteins. Furthermore,they may not be as biodegradable as free carbohydratesand proteins are. For example, most of the amino acidsand a large portion of the carbohydrates in DOM froman Oregon river were bound to anionic humic sub-stances (Lytle and Perdue, 1981; Sweet and Perdue,1982).
Fractionation of DOM according to molecular sizehas been very widely used and can account for all ofthe DOM in a given sample (the third approach).While molecular size is certainly a fundamental aspectof structure, the molecular-size distribution does notdirectly inform us about mode of formation, surfacereactivity, acidity, or biodegradability. Large molec-ular size is sometimes tacitly associated with resistanceto bipdegradation. The large-molecule fraction couldcontain proteins, complex carbohydrates, humic sub-stances, and tannins — components that differ radi-cally in their degradability.
In the work reported here, we used the fractionationprocedure of Leenheer and Huffman (1976) and Leen-heer (1981). This fractionation procedure is based onsurface properties at various pHs. Substances foundin the various fractions are listed in Table 1. The di-vision between hydrophobic and hydrophilic com-ponents is arbitrary and is based on the capacity factorfor the compounds on nonionic XAD-8 resin (Sigma
1112
QUALLS & HAINES: GEOCHEMISTRY OF DISSOLVED ORGANIC NUTRIENTS 1113
Chemical Co., St. Louis, MO) (Leenheer, 1981), whichis related to water solubility (Thurman et al., 1978).The term hydrophobia acids is used because these sub-stances are hydrophobic at pH 2, when the carboxylicacid groups are protonated and the molecule is un-charged. It is important to keep in mind that thesesubstances are charged, and therefore hydrophilic, atneutral and alkaline pH. An exception is a fractiontermed weak (phenolic) hydrophobic acids, which maynot have carboxylic acid groups and are charged onlyat very alkaline pH (Leenheer and Noyes, 1934). How-ever, this fraction may also include polyphenols withless than one carboxylic acid group per 13 C atoms(Thurman, 1985). For simplicity, we will refer to thisfraction as phenols, with the understanding that thehydrophobic acids may also have phenolic hydroxylgroups. Aquatic humic substances have been opera-tionally denned as the hydrophobic acids (includingthe weak hydrophobic acids, i.e., phenols) isolated us-ing the XAD-8 resin (Thurman, 1985).
The list of substances to be found in each of thefractions (Table 1) is generally based on: (i) retentionby XAD-8 of a variety of simple compounds (Thur-man et al., 1978); (ii) infrared spectra of fractions ofDOC from river water (Leenheer, 1981); or (iii) hy-pothetical grounds based on the general relationshipsof retention by XAD to structure and aqueous solu-bility (Thurman et al., 1978), and the known chargeson compounds at various pHs. Kroeff and Pietrzyk(1978) determined the capacity factors of a number ofamino acids and peptides on XAD-4 and XAD-7 res-ins and found that they were in the range that wouldallow them to pass through the XAD resin as used inLeenheer's (1981) procedure. Lytle and Perdue (1981)found that XAD-7 (a resin similar to XAD-8) retainedonly 17% of an algal protein mixture and 12% of analgal protein hydrolysate at pH 2. Since they are hy-drophilic and positively charged at pH 2, most aminoacids, free peptides, and free proteins would be ex-pected to be in the hydrophilic base fraction.
The fractionation procedure developed by Leenheerand Huffman (1976) and Leenheer (1981) has severaladvantages. First, it accounts for all of the DOM inthe sample except for small amounts lost during theprocedure. The cationic, anionic, and neutral hydro-phobic resins used in the procedure more or less mimicsome properties of soil and sediment surfaces, and thehydrophobic and hydrophilic acid fractions are thosecapable of interacting with ligand-exchange sites. An-other advantage is that the acid fractions are isolateddirectly and their titration curves can easily be deter-mined (Cronan and Aiken, 1985). These same acidfractions are those responsible for metal cpmplexation.
The objectives of this study were to divide the DOMin throughfall, soil water, and stream water into classesbased on surface behavior in order to determine: (i)the classes that are the major carriers of C, N, and P;(ii) the functional groups and potential surface inter-actions that may control the movement of dissolvedN and P; and (iii) the classes of C, N, and P containingsubstances that are preferentially removed as waterpercolates through the soil profile, and, using this data,to hypothesize what processes are important in con-trolling their removal.
Table 1. Substances found in specific fractions of dissolved organicmatter.
Fraction CompoundsHydrophobic neutrals
Weak (phenolic) hydrophobicacids
Strong (carboxylic) hydrophobicacids
Hydrophilic acids
Hydrophilic neutrals
Bases
Hydrocarbons!ChlorophyllCarotenoidsPhospholipidsHumic substances with <1 ionic or
phenolic group per 13 C atoms!Tannins!FlavonoidsOther polyphenols (<1 carboxyl
group per 13 C atoms)!VanillinFulvic acid and humic acid!Humic-bound amino acids and
carboxyl group per 12 C atoms)!Long-chain fatty acids (>C5)§Huraic-like substances with lower
molecular size and higher COOH/C ratios
Oxidized carbohydrates with COOHgroups
Small carboxylic acidsflInositol and other sugar phosphatesSimple neutral sugarsNon-humic-bound polysaccharidesfAlcohols (<C4)§ProteinsFree amino acids and peptidesAromatic aminesAminc-sugar polymers (such as from
microbial cell walls)t Substances indicated by infrared spectra of fractions of river-water dissolved
organic matter (Leenheer, 1981).! Thurman, 1985.§ Determined by retention of model compounds by XAD-8 resin (Thurman
et al., 1978).11 Fractionation of model compound, oxalic acid (David et al., 1989).
MATERIALS AND METHODSSite Description
Samples were gathered from 12 plots located in an oak(Quercus) -hickory (Carya) forested watershed (WS-2) at theCoweeta Hydrologic Laboratory in the southern Appalachi-an Mountains of North Carolina. Plots were stratified by soiltype and slope position. Three soil types occur on the wa-tershed: a coarse-loamy, mixed, mesic Umbric Dystrochreptof the Tusquitee series in the riparian zone, a fine-loamy,micaceous, mesic Typic Hapludult of the Fannin series, anda coarse-loamy, micaceous, mesic Typic Dystrochrept of theChandler series.
Sample Collection and Preservation
Samples of throughfall were collected in troughs. Waterpercolating from the bottom of the forest floor, the Oa ho-rizon, was collected in zero-tension soil water collectors (Jor-dan, 1968). Soil water was collected from the upper Ahorizon, the bottom of the AB or A2 horizon, the B horizon,and the upper C horizon in porous-cup vacuum samplers at50 kPa. Stream water was collected at the base of the wa-tershed during baseflow. Long-term stream water chemistryis summarized by Swank and Waide (1988).
Samples were collected from 12 plots and composited inFebruary, May, August, November, and December 1987.The samples were composited in order to produce a budgetfor the entire watershed for another part of the study, andalso because DOC concentrations showed no significant dif-ferences by soil type. All samples from December and Au-
gust, and selected samples from other months, werefractionated. We fractionated throughfall and forest-floor so-lution samples from other seasons because we initially sus-pected their composition would be more variable. However,the fractions of forest-floor solutions were quite consistent.Concentrations of DOC in late winter throughfall were verylow, so the February sample was omitted.
Every effort was made to collect, filter, and preserve thesamples as quickly as possible. Throughfall and Oa water sam-ples were collected during storms and within 1 h after majorrain storms during one sampling week. Soil water collectorswere emptied every 24 h or less during the sampling period.Samples were placed on ice within 1 h after collection, filteredthrough a Whatman G/F glass-fiber filter within 8 h, com-posited and frozen in liquid N within 36 h in most cases, andkept frozen until analysis. While conventional slow freezingmay lead to flocculation in some water samples (Giesy andBriese, 1978), there was no significant particle formation inour frozen samples after thawing.
Sample Preparation
In samples that differ widely in DOC concentration, aconsistent division of hydrophobic and hydrophilic com-ponents depends on remaining within the linear range of theXAD-8 adsorption isotherm. To compare samples initiallyvarying from 0.5 to 52 mg C/L, we either diluted or con-centrated all samples into a range of 6 to 15 mg C/L. Toconfirm whether this range was satisfactory, one Decemberforest-floor water sample was diluted to 5, 15, and 20 mg C/L and fractionated. All fractions were essentially identicalexpressed as a percent of input concentration. Low-levelsamples were concentrated using a Virtis freeze concentrator(Virtis Co., Gardiner, NY) (Shapiro, 1961). This method waschosen because it is gentle, nonselective, and the sample isconcentrated at 0 °C, so microbial growth and evaporationof volatiles are prevented. Concentration of stream waterwas limited to levels below 9 mg C/L because an inorganicprecipitate (perhaps silica) began to form after excessive(*& 13-fold) concentration.
Test Compounds
The compounds bovine serum albumin, quercetin, rutin,sodium oxalate, and glucose monophosphate (all from SigmaChemical Co., St. Louis, MO) were subjected to the frac-tionation procedure using an initial concentration of 10 mgC/L. Albumin represented a model protein. Quercetin andrutin represented model flavonoids (and polyphenols with-out carboxylic acid groups).
After every two or three samples, reverse-osmosis-deion-ized water was run through the fractionation procedure asa system blank.
Fractionation Procedure
We used the fractionation procedure of Leenheer andNoyes (1984) with several modifications (Fig. 1). The pro-cedure was scaled down to use a 250-mL water sample. Thehydrophobic acids were first adsorbed on the XAD-8 columnas in Leenheer and Huffman (1976), rather than on the an-ion-exchange column. Because hydrophobic bases are an in-significant fraction of natural water samples (Leenheer, 1980;Antweiler and Drever, 1983), we omitted the step wherehydrophobic bases were eluted from the XAD-8 resin with0.1 M HC1, leaving both hydrophobic and hydrophilic basesin the base fraction. Because the hydrophobic neutrals weremeasured by difference after the initial sorption on XAD-8,the hydrophobic bases could potentially be counted in thehydrophobic neutral fraction also. However, the hydropho-bic bases were assumed to be negligible.
The hydrophilic acid fraction was eluted with the recyclingprocedure used by Leenheer and Noyes (1984) and imme-diately desalted by running through a cation-exchange resin.The anion-exchange resin bled significant amounts of whatLeenheer and Noyes (1984) believed to be ammo-phenols,a component of the resin. These were removed by runningthe anion-exchange eluate through the cation-exchange col-umn. Since the hydrophilic fraction had already passedthrough the cation-exchange column once during fraction-ation, this second pass should not have removed any ad-ditional material from the sample itself. The hydrophilicneutral fraction (the effluent from the anion-exchange col-umn) was concentrated 10-fold by rotary evaporation andrun through the cation-exchange column to remove resincontamination. All of the other desalting steps following thefractionation procedure by Leenheer and Noyes (1984) wereomitted, since they were not necessary for elemental analy-sis. Hydrophobic neutrals were eluted from the XAD-8 usingacetonitrile acidified to pH 2 with formic acid (J. Leenheer,1987, personal communication), but hydrophobic neutralswere calculated by difference (Fig. 1). The XAD-8 waswashed with methanol and 250 mL of water after use, sinceacetonitrile was a potential contaminant Acidic or basiceluates were immediately neutralized with NaOH or H2SO4.Chloride was not used in the procedure because it interfereswith K2S2O3 oxidation.
In the presence of the SOs' added as H2SO4 to the samples,the weak base anion-exchange resin (Duolite A-7) did notcompletely adsorb all orthophosphate and glucose phosphateused as test compounds. This probably occurred because ofthe relatively low affinity of the resin for PO4 compared withSOI'. However, the test compound oxalic acid was com-pletely adsorbed by the resin. Consequently, an extra stepwas added to the procedure for application just to the or-ganic-P results. The concentrated effluent from the DuoliteA-7 resin (Diamond Shamrock Corp., Houston, TX) was runthrough Dowex 1 X-8 strongly basic anion-exchange resin(Dow Chemical Co., Midland, MI) and the influent and ef-fluent were analyzed for POj~ and total P. The hydrophilicneutral DOP was considered to be the DOP in the effluentfrom the strongly basic anion-exchange resin. The hydro-philic acid DOP was considered to be the DOP removed byboth the weakly basic and the strongly basic anion-exchangecolumns. This difference in calculation was applied only tothe DOP results for the hydrophilic fractions and is not in-dicated in Fig. 1.
Elemental Analyses
Dissolved Organic Carbon was measured by automatedK2S2O8 oxidation followed by infrared detection of the lib-erated CO2 using an 01 Model 700 TOC Analyzer (OI Corp.,College Station, TX). The digestion efficiency was checkedusing a solution of desalted freeze-dried fulvic acid, the Ccontent of which was determined using a C-H-N analyzer(Perkin-Elmer Corp., Norwalk, CT). Recovery averaged near100% throughout the range of concentrations encountered.
Total dissolved N and P were measured in duplicate usingpersulfate oxidation (Koroleff, 1983). Nitrate, NHj, andPOJ", before and after digestion, were measured as detailedin Quails et al. (1991). Various tests for accuracy, precision,and interferences are also described in Quails et al. (1991).Dissolved organic N = total dissolved N minus NO3-Nminus NH4-N. Dissolved organic P = total dissolved P mi-nus PO4-P. Selected samples were digested in four replicatesfor an estimate of analytical error.
Total hydrolyzable carbohydrate was determined on frac-tions of selected samples using the phenol-H2SO4 method(Handa, 1966) with glucose as the standard.
QUALLS & HAINES: GEOCHEMISTRY OF DISSOLVED ORGANIC NUTRIENTS 1115
CalculationsFor the calculation of several fractions, we used the prep-
arative procedure (Leenheer, 1981) rather than the analyticalprocedure (Leenheer and Huffman, 1976) because the con-centrated eluates allowed better analyses of the small quan-tities of N and P in some fractions (see Fig. 1). The deionized-
water system blank values were subtracted from the DOC,DON, and DOP concentrations for each sample fraction togive the final concentrations. The mass of DOC, DON, andDOP recovered in each fraction was calculated and com-pared with the mass of each present in the sample beforefractionation.
STEP1250 ml FILTEREDSAMPLE AT pH 7
STEP 2ELUTEWrm0.1 M NaOH
STEP 3SAMPLEATpHZ
STEP 4ELUTE WITH0.1 M NaOH
-OOC, DON 1-
Adjust sample topH2withH2S04-
STEP5ELUTE WITH 1 M NaOH
-DOC, DON 2DOC, DON 3-*-
STEPSRECYCLING ELUTIONWITH 0.1 M NaOH
DOC, DON 5-
N
••••'• 1
:.....
'•.
r
DCf
x_
/-
3.75 ml XAD-8 RESIN
-+-DOC, DON 4
ADSORBSHYDROPHOBICNEUTRALS, ACIDS,PHENOLS
DOC, DON 8 •*-'
3.75 mL AG-MP-50H+ SATURATED ADSORBSCATION-EXCHANGE RESIN BASES
DOC, DON 6
7.5 mL DUOLITE A-7 ADSORBSANION-EXCHANGE RESIN HYDROPHILICFREE-BASE FORM ACIDS
HYDROPHILIC NEUTRALS-STEP 7 CONCENTRATE
AG-MP-50 RESIN (FOR REMOVINGAMINO CONTAMINANTS)
DOC, DON 9
CALCULATIONS
HYDROPHOBIC OOC, DON
PHENOLS = DOC, DON 3 x
STRONG ACIDS = DOC, DON 4 x
ELUATE VOL.SAMPLE VOL
ELUATE VOL.SAMPLE VOL.
NEUTRALS = OOC, DON 1 - DOC, DON 2 - PHENOLS
HYDROPHILIC DOC, DON
TOTAL BASES = DOC, DON 5 x ELUATE VOL.SAMPLE VOL
ACIDS (DOC) = OOC 6 - DOC 7
ACIDS (DON) = OON 8 x DOC ACIDSDOC 8
NEUTRALS = DOC, DON 9 x CONG. VOL.SAMPLE VOL.
Fig. 1. Procedure for dissolved organic C (DOC), N (DON), and P (DOP) fractionation. Places where a subsample is analyzed are indicated(e.g., DOC, DON 1). Calculations for the DOC and DON fractions are identical in form except for hydrophilic acids, in which case theyare shown separately. The procedure and calculations for DOP (not shown) are identical to those for DON, except in the case of thehydrophilic fractions (modified from Leenheer [1981]).
Table 2. Behavior of test compounds during fractionation for dissolved organic C (DOC), N (DON), and P (DOP).Hydrophobic behavior
Compound
QuercitinRutinOxalic acid
Albumin
Glucose-phosphate
a t p H 2
yesyesno
no
no
a t p H 7
yesyesno
no
no
Important functionagroups
phenolic-OHphenolic-OHCOOH
amino.iCOOH
phosphateester
Hydrophobic
15O
9
0
AcidsPhenols
96920
0
ND
Hydrophilic
NDfND93
5
tM. ne nnr ...
89
NeutralHydrophobic
330
0
0
Bases
NDND0
86
ND
t ND — not determinedj Carboxyl groups are protonated at pH 2, giving the protein a net positive charge due to the charged amino groups.
RESULTS
Methodological Tests
Over 90% of the rutin and quercitin (model flavon-oid compounds) appeared in the weak hydrophobiaacid (phenol) fraction as expected (Table 2). A smallportion adhered to the XAD-8 resin after base elutipn,probably by H bonding to the resin. This small portionwould be counted as part of the hydrophobic neutralfraction.
About 86% of the N in the albumin protein standardappeared in the base fraction as expected (Table 2). Ninepercent appeared in the hydrophobic acid fraction.
60-
cc•r-LUOOo
40
20
§
Sia, ^
llTT in a) r*.
*- - do" oo
120- iThroughtall Oa-- -A— AB —3 C— Stream100-
80-
60-
40-
20-
a aQ u-> o a o o <3o 532 Q < O Q < Q <
MONTHHYDROPHILIC ACIDSPHENOLS O HYDROPHILIC NEUTRALSHYOROPHOBIC AGIOS 0 HYDROPHOBIC NEUTRALS
oS
BASES
Fig. 2. Original concentration of dissolved organic C (DOC) in thesamples from the field (upper) and the percentage of total DOCin each of the fractions (lower). Note that the bars are arrangedso that the bottom three bars together comprise all the carboxylicand phenolic acids. A total that varies from 100% indicates var-iable recovery or analytical error. Error bars for the December Oahorizon sample indicate ± 1 standard error of the mean for n =3.
The December Oa horizon sample was fractionatedthree times at 20, 15, and 5 mg C/L to test whetherthe initial concentration had any effect on the distri-bution of fractions and to estimate the standard error.There were no trends evident as a function of con-centration. The standard error of the mean, expressedas a percent of the original sample, is shown in Fig.2. The precision of the DOC analysis alone was a mi-nor component of the error with a CV of ± 1.5% ofthe mean above 1 mg C/L and a standard deviation(SD) of ±0.03 mg C/L below 1 mg C/L. An averageof 94.0% (± 6.1% SD) of the DOC before fractionationwas accounted for, indicating some loss of materialduring the procedure.
Dissolved Organic Carbon
Generally, the relative abundance of the differentfractions of DOC was (from highest to lowest): hy-drophobic acids > hydrophilic acids > or < hydro-philic neutrals > hydrophobic neutrals > phenols >bases (Fig. 2). There were exceptions, however. Hy-drophilic neutrals were particularly abundant in thethroughfall and C horizon samples. The phenols onlycomprised from 0.3 to 2.4% of the DOC in the forest-floor water, soil water, and stream water, but inthroughfall the phenols comprised up to 6% of theDOC. Altogether, the acid fractions (the hydrophobicand hydrophilic acids and phenols) comprised from44 to 91% of the DOC. Both neutral fractions togethercomprised 11 to 42% of the DOC. Bases were a verysmall component, 2.5% at most.
Seasonal differences were only examined in detailin the upper, more biologically active strata. Inthroughfall, the only striking difference between thegrowing and dormant season was the abundance ofhydrophilic neutrals in the growing season. The dis-tribution of fractions in the forest-floor solution wassurprisingly consistent throughout the year, except inNovember just after litterfall. The November samplecontained an unusually high content of hydrophilicneutrals and an unusually low content of hydrophilicacids. Because of the general lack of strong seasonaldifferences in the forest-floor solution, and the simi-larity of the August and December soil water, we didnot pursue a more detailed seasonal survey for the soilwater.
There were some consistent trends with depth in the
QUALLS & HAINES: GEOCHEMISTRY OF DISSOLVED ORGANIC NUTRIENTS 1117
soil. The total DOC decreased about 100-fold as waterpercolated from the forest floor to the upper C horizon.Consequently, all fractions declined with depth interms of actual concentrations of DOC. Changes inthe relative percentages of the fractions simply meanthat concentration of some fractions declined withdepth less than others. The percentage of hydrophobicacids was lower in throughfall than in the forest floor,but it decreased consistently with depth (by factors of1.3 in August and 1.8 in December). The percentageof hydrophilic acids did not decline consistently withdepth throughout the profile, but there was a lowerpercentage in the C horizon than in the forest-floorwater. The percentage of hydrophilic neutrals in-creased dramatically (2.5-fold) and consistently withdepth. The percentages of total neutrals increased withdepth at the expense of total acids, but that patternsimply reflected the two largest fractions of each: thehydrophobic acids and the hydrophilic neutrals. Like-wise, an increase in the percent total hydrophilics re-flected those same two components.
Finally, the distribution of fractions in the streamwater resembled a mixture of water from the C andthe Oa horizons.
The range of pH values for the solutions collectedin the field was: throughfall. 5.4 to 6.2; Oa horizon,4.2 to 5.4; A horizon, 5.7 to 5.9; B horizon, 5.9 to 6.2;C horizon, 6.0 to 6.2; and stream water, 6.5 to 6.9.
of absolute concentrations as water percolated throughthe soil. There was a consistent decrease in the C/Nratio moving from the forest floor to the C horizon.This relative enrichment in N content of the DOM asa function of depth was primarily the result of a shiftin the distribution of the various fractions, and, sec-ondarily, the result of lower C/N ratios for the majorfractions. The N-poor (Table 3) hydrophobic acids de-clined as a percent of the total DOM (Fig. 3) with depthin the soil. While the C/N ratio of the hydrophobicacids did decline from the forest floor to the C horizon,it was only by about one third. The N-rich hydrophilicacids did not really change in a consistent way withdepth in terms of percent of total DON, but their C/N ratio tended to decline with depth. The relativeamount of DON in the hydrophilic neutral fractionconsistently increased with depth. The N-rich basesalso comprised a greater percent of the total DON withdepth (Fig. 3).
The variability in the DON results was considerablygreater than for the DOC results (Fig. 3). The analyt-ical error in determining DON was greater than thatfor DOC, and there were substantial quantities ofDON, NOj, and NHJ contamination in the elutionsof the blank runs. Contaminants originated from theresins (especially the anion-exchange resin), and fromthe large amounts of NaOH and H2SO4 necessary to
Dissolved Organic Nitrogen
Most DON was carried by hydrophobic acids, hy-drophilic acids, and, in some cases, hydrophilic neu-trals (Fig. 3). While the hydrophobic acids were themain DOC fraction, they were relatively poor in Ncontent (Table 3) and so were somewhat less impor-tant as a carrier of N. The hydrophilic acids were animportant N carrier, since they were generally richerin N content than the aggregate sample (Table 3). Hy-drophilic neutrals were sometimes richer and some-times poorer in N content than the aggregate sample.The bases were very rich in N content, with a C/Nratio similar to that of proteins. Since bases were sucha small component of the DOM, however, they com-prised <7% of the total DON in ail except the Augustthroughfall.
The scarcity of cationic DOM was shown not onlyby the smallness of the base fraction, but also by anexperiment in which the samples were run through thecation-exchange column first. Since, in the regular pro-cedure, the hydrophobic acids are removed before thecation-exchange column, there could conceivably besome cationic properties of the hydrophobic acids.However, removing cations first resulted in only mi-nor changes in the C/N ratio of the hydrophobic acidsin the December Oa horizon sample (Table 3), indi-cating that the N in the hydrophobic acids had littleinteraction with the cation-exchange resin. Similar re-sults were found for the August Oa horizon samplebut are not shown.
The DON concentrations increased after passingthrough the forest floor and then declined 35- to 70-fold as water percolated through the soil profile. Con-sequently, all fractions of the DON declined in terms
Fig. 3. Original concentration of dissolved organic N (DON) in thesamples from the field (upper) and the percentage of total DONm each of the fractions (lower). Note that the bars are arrangedso that the bottom three bars together comprise all the carboxylicand phenolic acids. A total that varies from 100% indicates var-iable recovery or analytical error. Error bars for the December Oasample indicate ± 1 standard error of the mean for n = 3.
Table 3. The dissolved organic C (DOQ/dissolved organic N (DON) and DOC dissolved organic P (DOP) ratios, by weight, for the initialsample and the six fractions of dissolved organic matter. Samples were taken from canopy throughfall (T-fall), the Oa, A, AB, B, and Csoil horizon solutions, and a stream.
f Insufficient N or P content to calculate a significant ratio.t Fractionated by running sample at pH 2 through cation-exchange column first, then XAD resin.§ ND — not determined.1 Unacidified sample run through cation-exchange column first, then acidified before XAD resin.
neutralize the samples. Substantial concentrations ofNHj and NO} in the throughfall samples became con-centrated in the base and the hydrophilic acid frac-tions, respectively, and tended to obscure the DONcontent of those fractions. Because of the very low Cand N content of many of the phenol fractions, the C/N ratio incorporates a great deal of error (Table 3).An average of 97% (± 9.7 SD) of the initial DON wasaccounted for.
Dissolved Organic Phosphorus
Most P occurred in three fractions: the hydrophilicacids, hydrophobic acids, and hydrophilic neutrals(Fig. 4). Hydrophobic acids were less important as aP carrier than as a C carrier, except for the DecemberA and August C horizon samples (Fig. 2, 3, and 4).Indeed, P was a rare atom in the hydrophobic acidfraction, since DOC/DOP ratios ranged from 400 to8400 (about 1040 to 22 000 in molar C/P ratios) (Table3). The hydrophobic acid fraction was less important
and the hydrophilic acid fraction more important inthe lower A and the B horizons in terms of the relativeDOP distributions (Fig. 4). The hydrophobic neutrals,which could potentially contain phospholipids, wereonly a minor carrier of DOP.
The original concentrations of DOP declined withdepth in the soil, as did DOC (Fig. 4). Phosphorus wasa rare, trace constituent of the DOM, with DOC/DOPratios ranging from about 400 to about 4200 in theinitial samples (Table 3). However, the DOM fromthe lower soil horizons was much richer in P, expressedin terms of the DOC/DOP ratio, than that from theforest floor and A horizon. This vertical trend in theDOC/DOP ratio is more pronounced than that for N(Table 3). Consequently, the depletion of DOP aswater moved down the soil profile was less than thatfor DON and DOC.
The shift in the relative distributions of the DOPfractions with depth in the soil (Fig.- 4) was more er-ratic than that for DOC (Fig. 2). The most importantfactor in the decrease of the aggregate DOC/DOP ra-
QUALLS & HAINES: GEOCHEMISTRY OF DISSOLVED ORGANIC NUTRIENTS 1119
HYDROPHILIC ACIDSPHENOLSHYDROPHOB(C ACIDS
d HYDROPHILIC NEUTRALS® HYDROPHOBIC NEUTRALS BASES
Fig. 4. Original concentration of dissolved organic P (DOP) in thesamples from the field (upper) and the percentage of total DOPin each of the fractions (lower). Note that the bars are arrangedso that the bottom three bars together comprise all the carboxylicand phenolic acids. A total that varies from 100% indicates var-iable recovery or analytical error. Error bars for the December Oahorizon sample indicate ± 1 standard of the mean for n = 3.
tios with depth in the soil was the P enrichment of thehydrophilic acid fraction with depth (Table 3).
Analytical precision for the DOP analyses was about±0.7 Mg P/L standard error (SE) (n — 2 for the frac-tions and n = 3 for initial samples) for concentrations<10 jig P/L. Because the initial DOP concentrationsin some samples were very low (3-16 /ig P/L), theanalytical error in determining the initial DOP was amajor contribution to the error in calculating percentrecovery. The low concentrations of DOP were themajor reason for eluting concentrated fractions, ratherthan using just the analytical version of the fraction-ation (as in Leenheer and Huffman, 1976). Still, anumber of fractions were below the limit of detection,as indicated in Table 3. Recovery averaged 90% (±19SD) (Fig. 4). Despite the low DOP concentrations,DOP exceeded dissolved inorganic P in all of theseoriginal samples (not shown).
DISCUSSION
Nitrogen, Phosphorus, and the Behaviorof Dissolved Organic Matter
The functional groups in which N and P occur hadlittle influence on the behavior of most of the DOM.Many of the functional groups in which N is found,such as amines and some heterocyclics, are basic andwould be charged at acidic pHs. Despite the commonoccurrence of N in the DOM, only a small portionwas cationic. The small base fraction had a C/N ratiosimilar to that of amino acids and proteins. This frac-
tion may sorb electrostatically to the cation-exchangesites of the soil. It may be argued that bases are scarcein soil solution not because there is little productionof free cationic DOM, but because it is quickly sorbedby the abundant cation-exchange sites on clays andsolid humus. The cationic organic matter, though, wasscarcest in the forest-floor water. Furthermore, thereseems to be little cationic soluble organic mattersorbed by cation exchange in our A horizon soil, sinceconcentrated KC1 and NaCl failed to desorb DOC orDON (Quails, 1989) and failed to suppress adsorption.Perhaps even the litter may have sufficient cation-ex-change capacity to retain potentially soluble cationicorganic matter. While the cation-exchange sites on soilclays are probably the most important surface propertyfor many aspects of nutrient cycling, they don't seemto play a very important direct role in sorption oforganic matter, except possibly through the bridgingof trivalent cations to carboxylic acid groups.
A substantial proportion of the N was associatedwith humic substances. While Lytle and Perdue (1981)found that 96% of the amino acids in river water werebound to humic substances, we found that a smallerproportion of the N was humic bound in our samples.The failure of the humic-bound N in the forest-floorwater sample to sorb to the cation-exchange resincould be because: (i) the N is primarily in amide bondsor uncharged heterocyclics, or (ii) the small numbersof positively charged amine groups are stericallyshielded within the macromolecules and preventedfrom close contact with the exchange site. The C/Nratios in the hyhdrophilic acids and neutrals were low-er than for the hydrophobic acids and the same ar-guments may be made for the N in those molecules.Consequently, the carboxylic acid groups and perhapsthe phenolic groups of the organic acids control themobility of a substantial portion of the DON that isassociated with the same molecules. These acids aresusceptible to adsorption by h'gand exchange on Fe andAl hydroxide surfaces (Tipping, 1981), trivalent cationbridg^g to cation-exchange sites, or electrostatic at-traction to anion-exchange sites. Protonation of the car-boxylic acids in very acid soils may cause these N-containing acids to be somewhat hydrophobic.
Phosphorus is simply too rare to be an integral com-ponent of a large proportion of the organic molecules.It is, therefore, not a factor in controlling the mobilityof a large portion of the DOM. However, for thoseorganic molecules that do contain ester phosphatefunctional groups, these groups may influence the in-teractions of the molecule. The ester phosphate func-tional group is the major form of P in soil organicmatter (Thurman, 1985) and is negatively charged atneutral pH. Consequently, hydrophilic ester phos-phates would occur in the hydrophilic acid (i.e., hy-drophilic anionic) fractions. Indeed, much of the P didoccur in that fraction. Inositpl phosphates are prob-ably the main form of organic P in soil, but tend toform insoluble complexes with Al, Fe, and Ca (Ste-venson, 1982). They have been identified in naturalwater (Eisenreich and Armstrong, 1977) and couldcompose much of the hydrophilic acid P fraction. Be-cause carbohydrates can be incorporated into humicsubstances, the presence of inositol phosphates could
explain the small but significant occurrence of P in thehydrophobic acid fraction (Thurman, 1985). As withN, other more numerous functional groups or lipidmoieties on the same large molecule could control themobility of some P-containing molecules. It was sur-prising that a substantial amount of P in some samplesoccurred in the hydrophilic neutral fraction. It is dif-ficult to explain why the P did not sorb to the anion-exchange column, unless the P was shielded within thematrix of a much larger carbohydrate molecule.
Throughfall
Throughfall represented a substantial input of DOC,DON, and especially DOP, to the soil. The detailedbudget for this watershed (Quails et al., 1991) revealedthat about 32% of the DOC and 65% of the DOP fluxesentering the A horizon originated in throughfall andstemflow.
In throughfall, there was a surprising abundance ofhydrophobic acids that contained carboxylic acidgroups and were highly colored, like humic substances.The weak hydrophobic acids (phenols) were a rela-tively small proportion of the DOC, except in the Au-gust' throughfall sample (they were still a smallerfraction than the hydrophobic acids). The forest can-opy may seem like an unlikely environment for thehumification process in the usual sense, but we wonderwhether these organic carboxylic acids represent com-pounds derived directly from plants. Most tannins andflavonoids are polyphenols without carboxylic acidgroups, and would be expected in the weak hydro-phobic acid (phenol) fraction, according to Thurman(1985) and our tests of model flavonoids. The poly-phenols, quercetin, and catechin have been found inwater dripping from live leaves (Gomah and Sakhar,1972). Malcolm and McCracken (1968) also identifiedcatechin in washings of oak leaves and noted a gen-erally high content of unidentified flavonoids. How-ever, they also noted a high carboxylate acidity of theDOC in the washings. Several small hydrophobic acids(vanillic acid, p-hydroxybenzoic acid, /7-coumeric acid,and gallic acid) as well as several small hydrophilicacids (citric acid, malic acid, and succinic acid) havebeen found in rain washings of plant leaves (Bruckertet al., 1971).
The following reasons might explain the abundanceof hydrophobic (carboxylic) acids in our throughfallsamples: (i) most of the hydrophobic acids were small,aromatic, phenolic acids such as gallic, ^-hydroxyben-zoic, vaniilic, or p-coumeric acids; (ii) rapid dissocia-tion of the ester bonds between monomers of tanninsand other polyphenols exposed carboxyl groups (seeSteelink, 1985); (iii) the organics were easily subjectto physicochemical or photochemical oxidation out-side the cellular environment; or (iv) microbes oxi-dized material exuded in the canopy. The means bywhich this substantial input of DOM may undergowhat superficially resembles humification deservesfurther attention.
The changes in the concentrations of the variousDOC fractions as the throughfall passed through theforest floor are illustrated in Fig. 5. This comparisonshows the sources of the materials more directly than
Fig. 5. Changes in the concentration of dissolved organic C (DOC)in each of the fractions as throughfall passes through the forest floor(collected at bottom of Oa horizon). Notice the differences betweenthe adjacent solid and light bars, which indicate the net increase ordecrease in concentration in each fraction as throughfall percolatesthrough the forest floor. Then compare this net gain or loss betweenthe various fractions. Two of the bars for bases are too small to beseen.
does the data on soil DOC, because the throughfalland forest-floor water for a particular time representthe changes in the same aliquot of water during a stormor storms. Water flux out of the forest floor was about98% of the throughfall input (Quails et al., 1991); there-fore, changes in concentration as the throughfallpassed through the forest floor were approximatelyproportional to fluxes during a particular week. Also,the adsorption was presumably much less in the forestfloor and the residence time of rain was short. By farmost of the hydrophobic acids and hydrophilic acidsoriginated from the forest floor. Much of the phenolfraction appears to have originated in the canopy. Insummer, a substantial proportion of the carbohydrate-rich hydrophilic neutrals, which could arguably be sug-ars and other carbohydrates from leaves, originatedfrom the canopy. There was actually a slight reductionin hydrophobic neutrals after passing through the for-est floor in summer. The hydrophobic surfaces in thelitter may have served to adsorb them. After the au-tumn litterfall, however, more hydrophobic neutralsoriginated from the forest floor. These may have beenthe fresher, easily decomposed lipids and pigmentsfrom the Litter. The small amount of bases in through-fall were essentially removed or complexed in the for-est floor in the summer.
Comparison with Other Studies
Hydrophobic and hydrophilic acids were the dom-inant fractions of dissolved organic matter in all thestudies we reviewed (Table 4). In many studies, in-cluding ours, the hydrophobic acids were the largestfraction. However, hydrophilic acids were a more im-portant constituent in several studies, e.g., in intersti-tial water in buried volcanic ash (Antweiler andDrever, 1983). The environment examined by Cronanand Aiken (1985) might be the most similar to ours,a forest soil with abundant rainfall and litterfall. How-ever, their soil was a Spodospl, in which one mightexpect less sorption of DOC in the A or E horizons.They found a shift from dominance by the hydropho-bic acids in the forest floor to a dominance by hydro-philic acids in the B horizon. Vance and David (1989)
QUALLS & HAINES: GEOCHEMISTRY OF DISSOLVED ORGANIC NUTRIENTS 1121
Table 4. Results of other studies in which identical or very similar fractionation procedures were used. The initial dissolved organic C (DOC)of the sample, and the percent of the total DOC in each fraction, is listed.
11, Cronan and Aiken, 1985; 2, Yavitt and Fahey, 1985b; 3, Yavitt and Fahey, 1985a; 4, David et al., 1989; 5 Antweiler and Drever, 1983; 7, Leenheer, 1980t 0/A horizon indicates the bottom of the forest floor.§ Neutral and base categories were lumped together.
and David et al. (1989) concluded from batch andcolumn leaching studies that hydrophobic acids wereselectively removed by Spodosol B horizon soil sam-ple. Yavitt and Fahey (1985b) found a decrease withdepth in the percent of the DOC in the hydrophobicacid fraction, quite similar to ours. However, the ratiosof hydrophobic/hydrophilic acids were 1.4 in the forestfloor, 1.1 in the root zone, and 1.2 in the subsoil intheir study.
These studies suggest that a selective net removalof hydrophobic acids, as DOM percolates through thesoil, may be a general phenomenon. However, thetrends in hydrophilic acids are more variable and noclear pattern of the ratio of hydrophobic/hydrophilicacids is apparent for the Coweeta soil profiles. Al-though the ratio of hydrophobic acids/hydrophilicacids (not shown) declined from the forest floor to theB horizon in December, the ratio for the C horizonwas as high as that for the forest floor. Hence, amongthe carboxylic organic acids, the hydrophilicity of theprotonated acid form did not seem to be a very im-portant interaction controlling mobility. If we assumethe mean pK^ (negative log of the acid dissociationconstant) of these carboxylic acids is about 3.85 (Cron-an and Aiken, 1985), then more than half of the car-boxyl groups would be charged at the typical pH ofour soil solutions (which ranged from 4.2 to about 6.2),and they may generally be sufficiently hydrophilic intheir charged form. Differences in the pH of the soiland the soil solutions might explain the differencesbetween studies in the apparent preferential removalof hydrophobic acids over hydrophilic acids, partic-ularly since both studies that indicated a shift in theratios involved Spodosols.
Humic substances in water have been operationally
defined as the hydrophobic acids isolated on the XAD-8 resin (Thurman, 1985). We separated the very smallfraction of weak hydrophobic acids that is usually in-cluded with the humic substances. As expected, ourhydrophobic acid fractions were highly colored andsuperficially resembled fulvic acid. Less is knownabout the character of the hydrophilic acid fraction,but it was still quite colored and could conceivablycontain humic-like substances. The small organicacids would occur in this fraction. Most of these smallorganic acids would be easily degraded by microbes.Since they both contain carboxylic acid groups, boththe hydrophobic acids and the hydrophilic acids maybe susceptible to adsorption to Fe and Al hydroxidesby ligand exchange (Parfitt et al., 1977), and we foundno evidence to rule out ligand exchange. Perhaps thepresence of multiple carboxylic acid groups on mol-ecules of both fractions and their participation in li-gand exchange might explain why there was nodramatic shift in the ratio of hydrophobic acids/hy-drophilic acids with depth.
Jardine et al. (1989) explored the mechanisms ofDOC adsorption in a Typic Paleudult and a TypicDystrochrept. Iron oxides and hydroxides were re-sponsible for 50 to 70% of the DOC adsorption. Theybelieved that the principal mechanism was physicaladsorption of hydrophobics driven by favorable en-tropy changes, rather than ligand exchange or cationbridging to cation-exchange sites. They supported thisconclusion by finding that much greater percentagesof the hydrpphobic fractions were adsorbed than ofthe hydrophilic fractions.
While the hydrophilic neutral fraction was not re-moved by any of the resins used in the fractionation,it did not pass through the soil unhindered. The ab-
solute concentration of hydrophilic neutrals declinedwith depth in the soil, but they were not removed tothe same extent as the other fractions. Some genera-tion of soluble hydrophilic neutrals by roots and mi-crobes might also have occurred in the soil. About halfof the DOC in the hydrophilic neutral fraction washydrolyzable carbohydrate.
CONCLUSION
We have described some properties of DOM thatregulates its interactions with surfaces, with an em-phasis on the fate of DON and DOP. We have alsodescribed the apparent shifts in distribution of variousfractions as water percolates down through the soilprofile. The fractions were defined by their behavioron nonionic and ion-exchange resins. The role of dom-inant functional groups in determining the averageproperties was further defined by their behavior as afunction of pH. It is important to realize that the largemacromolecules of DOC may have many functionalgroups and some may not have the opportunity tointeract with surfaces because they are overwhelmedby other groups or are sterically shielded. Such maybe the case for N-containing functional groups.
Although we have described some properties of theDOM, to demonstrate that particular mechanisms aredominant in the soil requires sorption and ion-com-petition experiments. These aspects are dealt with inother work (Quails and Haines, 1989, unpublisheddata). To demonstrate that hypothetical mechanismsare operating in the soil, however, it is necessary torefer to the changes in distribution as water percolatesthrough the horizons to see if the hypothetical mech-anisms are consistent with the observed patterns inthe field.
We found that most of the DOC was in the acidfractions with hydrophobic acids composing 35% to57% of the DOC in soil water and stream-water sam-ples. Concentrations of all fractions declined withdepth as water percolated through the soil, suggestingthat perhaps more than one mechanism was involved.But we found no clear evidence that hydrophobic acidswere removed to a greater extent than hydrophilicacids, as has been reported for some Spodosols. Freeproteins and amino acids are a very small percentageof the DON in soil solution. Instead, most N is as-sociated with hydrophobic and hydrophilic acids. Themobility of most of the N-cpntaining DOM is not in-fluenced by cationic properties of N. Instead, the car-boxylic acid groups, and perhaps phenolic functionalgroups, seem to be much more important in influenc-ing the mobility of the N carried passively by theDOM. A significant portion of the DON was in thehydrophilic neutral DOM.
Most P occurred in the hydrophilic acid and hy-drophobic acid-fractions. Phosphorus was simply toorare an atom to influence the behavior of a large por-tion of the organic molecules. But the am'onic P esters,perhaps in the form of sugar phosphates, may havebeen responsible for the surface behavior of most ofthe DOP-containing molecules.
ACKNOWLEDGMENTS
We sincerely thank the following people for their help:George Aiken and Jerry Leenheer generously provided help
and laboratory space at the U.S. Geological Survey Labo-ratory in Wheatridge, CO, for a pilot run, as well as contin-uing consultation. Gary Mills provided the resins. John Ertelloaned us some parts for the chromatography columns. KentTankersley and Patty Anderson helped in the analyses. Thisresearch was supported by NSF Dissertation Grant BSR-8501424. The facilities supported by the NSF Long-TermEcological Research Program at the Coweeta HydrologicLaboratory were indispensable in the field work. The exec-utive committee and laboratory personnel of the Institute ofEcology allowed R. Quails to use the TOC analyzer.
REFERENCESAmweiler, R.C., and J.I. Drever. 1983. The weathering of a late
Tertiary volcanic ash: Importance of organic solutes. Geochim.Cosmochim. Acta 47:623-629.
Bruckert, A., F. Toutain, J.T. Tchiecaya, and F. Jacquin. 1971. In-fluence des pluviolessivats de hetre et de pin sylvestre sur lesprocesses d'humificatiqn. Oecol. Plant. 6:329-339.
Cronan, C.S., and G.R. Aiken. 1985. Chemistry and transport of sol-uble humic substances in forested watersheds of the AdirondackPark, New York. Geochim. Cosmochim. Acta 49:1697-1705.
David, M.B., G.F. Vance, J.M. Fussing, and F.J. Stevenson. 1989.Organic carbon fractions in extracts of O and B horizons from aNew England Spodosol: Effects of acid treatment. J. Environ.Qual. 18:212-217.
Eisenreich, S.J., and D.E. Armstrong. 1977. Chromatographic in-vestigation of inositol phosphate esters in lake waters. Environ.Sci. Technoi. 11:497-501.
Giesy, J.P.. and L.A. Briese. 1978. Paniculate formation due tofreezing humic waters. Water Resour. Res. 14:542-544.
Gomah, A.M., and N.N.C. Sakhar. 1972. Heavy metal chelates inherbage and soil profiles from upland sites in Wales. Rep. WelshSoils Discuss. Group 13:17-39.
Handa, N. 1966. Examination on the applicability of the phenolsulfuric acid method to the determination of dissolved carbohy-drate in sea water. J. Oceanogr. Soc. Jpn. 22:79-85.
Jardine, P.M., N.L. Weber, and J.F. McCarthy. 1989. tVIechanismsof dissolved organic carbon adsorption on soil. Soil Sci. Soc. Am.J. 53:1378-1385.
Jordan, C.F. 1968. A simple, tension-free lysimeter. Soil Sci.105:81-86.
Koroleff, F. 1983. Simultaneous oxidation of nitrogen and phos-phorus compounds by persulfate. p. 168-169. In K. Grasshoff etal. (ed.) Methods of seawater analysis. 2nd ed. Verlag Chemie,Weinheimer, Germany.
Kroeff, E.P., and D.J. Pietrryk. 1978. Investigation of the retentionand separation of amino acids, peptides, and derivatives on po-rous copolymers by high performance liquid chromatography.Anal. Chem. 50:502-511.
Leenheer, J.A. 1980. Origin and nature of humic substances in thewaters of the Amazon River basin. Acta. Amazonica 10:513-526.
Leenheer, J.A. 1981. Comprehensive approach to preparative iso-lation and fracuonation of dissolved organic carbon from naturalwaters and waste waters. Environ. Sci. Technoi. 15:578-587.
Leenheer, J.A., and E.W.D. Huffman, Jr. 1976. Classification oforganic solutes in water by using macroreticular resins. J. Res.U.S. Geol. Surv. 4:753-751.
Leenheer, JA., and T.I. Noyes. 1984. A filtration and column adsorp-tion system for onsite concentration and fractkmation of organicsubstances from large volumes of water. U.S. GeoL Surv. WaterSupply Pap. no. 2230. U.S. Gov. Print. Office, Washington, DC.
Lytle, C.R., and E.M. Perdue. 1981. Free, prqteinaceous, and humic-bound amino acids in river water containing high concentrationsof aquatic humus. Environ. Sci. Technoi. 15:224-228.
Malcolm, R.L., and R.J. McCracken. 1968. Canopy drip: A sourceof mobile soil organic matter for mobilization of iron and alu-minum. Soil Sci. 137:23-32.
Meyer, J.T., R.T. Edwards, and R. Risley. 1987. Bacterial growthon dissolved organic carbon from blackwater river. Microb. Ecol.13:13-29.
Parfitt, R.L., A.R. Eraser, and V.C. Farmer. 1977. Adsorption onhydrous oxides. III. Fulvic acid and humic acid on goethite, gibb-site, and imogolite. J. Soil Sci. 28:289-296.
Quails, R.G. 1989. Geochemical and biological properties of dis-solved organic matter in the soil and stream of a deciduous forestecosystem: Their influence on retention of N and P. Ph.D. diss.Univ. of Georgia. Athens (Diss. Abstr. 90-03448).
Quails, R.G., B.L. Haines, and W.T. Swank. 1991. Fluxes of dissolvedorganic nutrients in a deciduous forest. Ecology 72:254-266.
Shapiro, J. 1961. Freezing-out, a safe technique for concentrationof dilute solutions. Science (Washington, DC) 133:2063-2064.
1
FISHER & STONE: IRON OXIDATION AT SATURATED ROOT SURFACES 1123
Sollins, P., and P.M. McCorison. 1981. Nitrogen and carbon solu-tion chemistry of an old growth coniferous forest watershed beforeand after cutting. Water Resour. Res. 17:1409-1418.
Steelink, C. 1985. Implications of elemental characteristics of humicsubstances, p. 457-476. In G.R. Aiken et al. (ed.) Humic sub-stances in soil, sediment, and water Geochemistry, isolation, andcharacterization. John Wiley & Sons, New York.
Stevenson, FJ. 1982. Humus chemistry. John Wiley & Sons, NewYork.
Swank, W.T., and J.B. Waide. 1988. Characterization of baselineprecipitation and stream chemistry and nutrient budgets for con-trol watersheds, p. 57-79. In W.T. Swank and D.A. Crossley, Jr.(ed.) Forest hydrology and ecology at Coweeta. Springer-Verlag,New York.
Sweet, M.S., and E.M. Perdue. 1982. Concentration and speciationof dissolved sugars in river water. Environ. Sci, Technol.16:692-698.
Thurman, E.M. 1985. Organic geochemistry of natural waters.Nijhoff/Junk Publ., Dordrecht, the Netherlands.
Thurman, E.M., R.L. Malcolm, and G.R. Aiken. 1978. Predictionof capacity factors for aqueous solutes on a porous acrylic resin.Anal. Chem. 50:775-779.
Tipping, E. 1981. The adsorption of aquatic humic substances by.iron oxides. Geochim. Cosmochim. Acta 45:191-199.
Vance, G.F., and M.B. David. 1989. Effect of acid treatment ondissolved organic carbon retention by a spodic horizon. Soil Sci.Soc. Am. J. 53:1242-1247.
Yavitt, J.B., and TJ. Fahey. 1985a. Chemical composition of in-terstitial water in decaying lodgepole pine bole wood. Can. J. For.Res. 15:1149-1153.
Yavitt, J.B., and T.J. Fahey. 1985b. Organic chemistry of the soilsolution during snowmelt leaching in Pinus contorta forest eco-systems, Wyoming, USA. p. 485-496. In D.E. Caldwell et al. (ed.)Planetary ecology. Van Nostrand Reinhold, New York.
i
Iron Oxidation at the Surfaces of Slash Pine Roots from Saturated Soils
H. M. Fisher* and E. L. Stone
ABSTRACT
Root aeration in saturated gleyed soils is indicated by Fe oxidecoatings on live-root surfaces. Oxygen diffusing through gas-perme-able tissues, from the atmosphere to submerged roots, ultimatelydiffuses into and oxidizes the rhizosphere. The objectives of thisstudy were to: (i) characterize the oxidizing conditions of slash pine(Pinus elliottii Engelm. var. elliottii) roots growing in saturated con-ditions, and (ii) test the hypothesis that submerged taproot systemsare internally aerated. One study examined Fe-oxide deposits on theroots of six saplings grown outdoors in drums filled with saturated,Fe-amended peat. Another study examined roots of two mature slashpines growing naturally on a poorly drained PaleaquulL Orange Fe-oxide precipitates coated cortical tissue of sapling taproots and rootbranches growing in saturated peat. Iron concentrations in theseroots were 6 to 96 times greater than in the surrounding peat Ingleyed horizons of the Paleaquuit, Fe-rich precipitates cementedsand to the surfaces of large-diameter roots and entirely ensheathedfine roots of pine and pine cypress (Tasodium ascendent Brogn.).Prominent mottles, all neoferrans, surrounded fine pores, remnantsof decayed-root channels. Despite the anoxic growing conditions, allof the examined slash pine root systems oxidized and accumulatedFe, supporting other evidence that the taproot systems of slash pinegrowing in wet soils are internally aerated.
IRON OXIDE DEPOSITS commonly occur on rootsgrowing in saturated soils (Armstrong, 1982). Such
precipitates are most evident on herbaceous andshrubby species endemic to wetland habitats (Arm-strong and Boatman, 1967; Armstrong, 1968), but arealso formed on the roots of some mesophytes thattolerate temporarily saturated soil conditions (Bartlett,1961). Iron-oxide formation is related to the internaltransport of O2 to submerged roots and to O2 leakageinto the rhizosphere.
There are relatively few reports of Fe oxides on treeroots. Swallow (1855) noted Fe-enriched zones around
H.M. Fisher, 1914 N. 24th St.. Boise. ID 83702; and E.L. Stone,Soil Science Dep., Univ. of Florida. Gainesville, FL 32611. Con-tribution from the Florida Agric. Exp. Stn., Journal Series no. R-00252. Research supported in part by CRIFF (Cooperative Researchin Forest Fertilization). Received 26 Oct. 1989. Corresponding au-thor.
Published in Soil Sci. Soc. Am. J. 55:1123-1129 (1991).
live and dead roots of white oak (Quercus alba L.).Dickson and Broyer (1972) observed that more re-duced soil conditions led to greater amounts of Fe3 +deposition on roots of water tupelo (Nyssa aquaticaL.) and bald cypress (Taxodium distichum [L.] Rich,var. distichum). The oxidative capacity of tupelo rootswas clearly related to adaptions that allowed O2 trans-port to the submerged roots of swamp tupelo, Nyssasylvatica var. biflora (Walt.) Sarg., (Hook et al., 1970,and 1974) and water tupelo (Hook and Brown, 1972).Flooded seedling roots of loblolly pine (Pinus taedaL.) also accumulated Fe3 + within epidermal and cor-tical tissues, with some lighter deposits within the stele(McKevlin et al., 1987). Deposits of Mn and Fe oxideswere similarly located in the flooded roots of otherconifer seedlings (black spruce, Picea mariana [Mill.]B.S.P., and red pine, Pinus resinosa Ait., Levan andRiha, 1986).
Iron oxides have not previously been reported onroots of slash pine. This species occurs naturally onpoorly drained soils of the lower Coastal Plain of thesoutheastern USA, growing with pond cypress alongthe margins of and within ponded depressions. Onsuch sites, the surface root system of slash pine is pe-riodically submerged. The massive taproots and ver-tical sinkers grow, however, to depths as much as 0.9m below seasonally low water-table levels into contin-ually saturated soil (Schultz, 1972). Previously re-ported work showed that a large connected volume ofgas-filled space in the secondary xylem of deep slashpine roots can conduct 02 (Fisher and Stone, 1990a),and that O2 transport to submerged sapling root sys-tems maintains aerobic nutrient-uptake functions(Fisher and Stone, 1990b).
Philipson and Coutts (1978, 1980) showed that O2movement through secondary xylem and lacunatesteles of lodgepole pine (Pinus contorta Douglas exLoudon) roots sustained a radial diffusion of O2 fromwoody and nonwoody roots, causing oxidation of therhizosphere. Thus, it seems likely that internally aer-ated slash pine roots growing in saturated soil wouldalso oxidize the rhizosphere. The objectives of ourstudies were to: (i) characterize the oxidizing condi-tions of deep slash pine roots by quantifying Fe-oxide
