| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Civil and Environmental Engineering, 220 Hinds Hall, Syracuse Univ., Syracuse, NY 13244
* Corresponding author (cejohns{at}mailbox.syr.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CPMAS, cross-polarization magic-angle spinning • DOC, dissolved organic carbon • DOM, dissolved organic matter • DON, dissolved organic nitrogen • HBEF, Hubbard Brook Experimental Forest • NMR, nuclear magnetic resonance • OC, organic carbon • OM, organic matter • TOC, total organic carbon • W1–W9, Watersheds 1 through 9
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The forest canopy and forest floor layers are the primary sources of DOM in many forest soils as reflected by large concentrations of DOC in seepage waters beneath the forest floor (McDowell and Likens, 1988; Johnson et al., 2000). Dissolution and movement of organic carbon (OC) in the forest floor and upper mineral horizons results in the transfer of OM, metals, nutrients, and pollutants through the soil to surface and ground waters (Chiou et al., 1986; Pohlman and McColl, 1988; Berggren et al., 1990; Qualls and Haines, 1991; Liu and Gary, 1993). Mobilization of OC, Al, and Fe in upper horizons and its subsequent transport to the subsoil plays an important role in podzolization, which forms Spodosols (Petersen, 1976; Dawson et al., 1978).
The mobility of DOM in soils is controlled by its sorption to mineral surfaces (McDowell and Likens, 1988). Concentrations and fluxes of DOC in soil solution typically decrease with increasing soil depth (Guggenberger and Zech, 1993; Johnson et al., 2000). The reduction in DOC concentration is accompanied by a change in DOM composition, with a general preferential decrease in hydrophobic DOM fractions (Guggenberger and Zech, 1993; Kaiser et al., 1996). Thus the nature of DOM itself affects sorption, with hydrophilic compounds being less likely to sorb to mineral surfaces than hydrophobic compounds (Kaiser et al., 1996; Kaiser and Zech, 1997). Sorption of DOM to soil mineral surfaces is also influenced by soil properties such as soil OC concentration, Al and Fe oxide and hydroxide content, and the mineralogy of the clay fraction (McDowell and Wood, 1984; Jardine et al., 1989; Donald et al., 1993). Also, DOM sorption may be influenced by adsorbed SO2–4 (Zech et al., 1994) and surface area (Kaiser et al., 1996).
Effects of pH on OM sorption have shown varying results. The adsorption of OM to iron oxides (Gu et al., 1994), Al hydroxides (Parfitt et al., 1977; Davis, 1982), and bulk soils (Jardine et al., 1989; Kennedy et al., 1996) increased with decreasing pH as a result of increasing positive charges on the hydroxides. However, David and Zech (1990) observed decreasing DOC adsorption in the B horizon of acid soils with decreasing pH, and in a batch adsorption experiment, Vance and David (1992) saw no effect of pH between pH 3 and 6 on DOC adsorption in mineral soil samples.
The purpose of this study was to investigate the adsorption of OM fractions to Spodosol mineral horizon soils from the HBEF in New Hampshire, USA. Our objectives were (i) to determine the OM sorption properties of Spodosol mineral horizons of the HBEF; (ii) to evaluate the effects of pH on OM sorption to different horizons; (iii) to determine the effect of OM type and source on OM sorption; and (iv) to determine the properties of OM and soils responsible for OM sorption.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
1400 mm of precipitation, of which 30% falls as snow (Federer et al., 1990). Prominent tree species at the HBEF are American beech (Fagus grandifolia Ehrh.), sugar maple (Acer saccharum Marshall), yellow birch (Betula alleghaniensis Britton), balsam fir [Abies balsamea (L.) Mill.], red spruce (Picea rubens Sarg.), and paper birch [Betula papyrifera var. cordifolia (Marsh.) Regel]. Conifers are the dominant vegetation type on W9, with <5% pure hardwood stands. Soils from five horizons (E, Bh, Bs1, Bs2, and C) were sampled from a conifer-dominated stand in W9. The soils were air-dried and sieved to pass a 2-mm sieve. The soil pH in deionized water was determined with a combination glass electrode at a soil to solution ratio of 1:1, after 30 min of equilibration at room temperature (22 ± 2°C). Total C, N, and H concentrations of soils and isolated OM were determined on ground subsamples with a CHN elemental analyzer (ThermoQuest, CE Instruments, Rodeno, Italy). Iron and Al in amorphous and crystalline oxides/hydroxides was determined by the Na-dithionite-citrate method, while organic-matter-bound Fe and Al was estimated separately by sodium pyrophosphate extraction (Loeppert and Inskeep, 1996). Aluminum and Fe concentrations in the extracts were determined by graphite furnace and flame atomic absorption spectrophotometry, respectively (Perkin Elmer Corp., Norwalk, CT). Pyrophosphate-extractable C was determined by the UV persulfate oxidation method with a total organic carbon (TOC) analyzer (Dohrman Phoenix 8000, Rosemount Analytical, Santa Clara, CA).
Organic Matter Solutions
Natural OM was extracted from O- (Oi + Oe + Oa) and Bh-horizon soil from W9. After collection, field-moist samples were sieved through a 2-mm sieve and stored at 3°C. The extraction was carried at room temperature immediately before the sorption experiments. NaOH (0.1 M) was added to an aliquot of soil (1 g soil:20 mL NaOH for O-horizon soil; 1 g soil:10 mL NaOH for Bh-horizon soil) and shaken for 30 min. The suspension was allowed to stand for 18 h, and then centrifuged. The suspension was acidified to pH 2 by adding 1 M HCl, centrifuged, and the supernatant filtered through a 0.45-µm membrane filter. The concentration of DOC in the extracts was measured with a TOC analyzer.
Soil OM is a complex mixture of compounds of varying molecular weights and structure with different physicochemical and biological properties. It contains a range of organic compounds from simple sugars to complex humic substances (Herbert and Bertsch, 1995). Since only a small fraction of C in the OM can be chemically identified, soluble OM is often fractionated chromatographically based on its solubility properties, including hydrophobicity and acidity. One commonly used method is based on the sorption of soluble OM to nonionic and ion exchange resins, thereby fractionating OM into hydrophobic and hydrophilic bases, acids, and neutrals (Leenheer, 1981). We fractionated the extracted OM into hydrophilic and hydrophobic fractions by the method of Leenheer (1981), as modified by Vance and David (1991). Briefly, the extracted soluble OM, acidified to pH 2, was pumped through a column packed with acid-washed Amberlite XAD-8 resin (Rohm and Haas Co., Philadelphia, PA). The C load on the resin was 1.0 mg C cm–3. The soluble OM adsorbed at pH 2 was considered hydrophobic, whereas the OM passing the XAD-8 column was considered hydrophilic. The acidic components of the hydrophobic fraction were eluted from the XAD-8 column with 0.1M NaOH. Hydrophobic neutrals were recovered from the XAD-8 resin with methanol. The methanol eluate was diluted with DI water, methanol was removed by means of a rotary evaporator, and the residue was added to the hydrophobic acid fraction. The hydrophilic and hydrophobic fractions were pumped through a column of strongly acidic cation exchange resin (AG-MP 50, BioRad Laboratories, Richmond, CA), to remove cations other than H+ before the adsorption experiment, and kept at 4°C until use in the sorption experiments.
Semiquantitative solid-state CPMAS 13C NMR analysis was conducted on freeze-dried subsamples of each fraction with a Bruker AMX 300 spectrophotometer at 75.47 MHz. Samples were spun at 5 kHz in a zirconia rotor within a MAS probe. The number of transients required for an acceptable signal-to-noise ratio ranged from 3918 to 99224. The contact time for each spectra was 1 ms and recycle time was 1 s. Spectral width was 33112 Hz, and the acquisition time and number of decay curves were 61 ms and 4096, respectively. Chemical shifts were externally referenced to glycine resonance at 176 ppm.
Studies of soils and humic substances have suggested that the fraction of C in the aromatic and carbonyl C regions of 13C CPMAS NMR spectra may be underestimated (Mao et al., 2000; Smernik and Oades, 2000). Consequently, 13C CPMAS NMR spectra must be interpreted with some caution. Kinchesh et al. (1995), however, suggested that the results can be used reasonably for comparisons among samples.
Sorption Experiments
Initial solutions ranging from 0 to 10 mmol OC L–1 were prepared by diluting the fractionated OM with deionized water. This range of OC concentrations spans the observed range of DOC concentrations measured in tension-free lysimeters at the HBEF between 1984 and 1992 (Johnson et al., 2000) and 1999 to 2001 (Wellington 2002). The sorption experiments were conducted at pH 3, 4, and 5. Fifty milliliters of OM solution were added to 5 g of soil. The suspension pH was adjusted to the appropriate pH by drop-wise addition of either 0.1 M HCl or 0.1 M NaOH with stirring during a period of 0.5 to 2.0 h. The suspension was shaken for 24 h, centrifuged, and then filtered through a 0.45-µm membrane filter. The DOC concentrations in the equilibrium solutions were determined by TOC analyzer.
A subsample of the equilibrium solution was fractionated into hydrophobic and hydrophilic fractions by columns filled with XAD-8 resin as described above, and the DOC concentration in each fraction was determined. In addition, total Fe and Al concentrations were determined in selected subsamples of the equilibrium solutions. Aluminum was determined by graphite furnace, while Fe was determined by flame atomic absorption spectrometry.
Because of the potential release of indigenous OM from the soil, the results of the sorption experiments could not be analyzed by standard isotherm models (e.g., Freundlich or Langmuir). Therefore, the initial mass isotherm approach of Nodvin et al. (1986) was used to describe the sorption. In this approach, the amount of C adsorbed or released per unit mass of soil (R) was plotted as a function of the initial DOC concentration (Xi), expressed as mmol C per kg soil:
 | [1] |
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
The CPMAS 13C NMR characteristics of the OM solutions used for the adsorption experiments are presented in Fig. 1 . The OM spectra had peaks in the resonance areas of alkyl C (0–50 ppm), O-alkyl C (50-110 ppm), aromatic C (110–160 ppm) and carbonyl C (160–220 ppm). The carbonyl C region in soils is dominated by carboxyl C, ketone C, and C associated with amides. Ketone-C resonates in the 190–220 ppm region (Baldock et al., 1997) and was negligible in our samples (Fig. 1). To account for amide C, we used the method of Qualls et al. (2003), who assumed that the molar content of amide-C was 90% of the molar content of N present in the sample. On the basis of this approach, 21–45% of the C in carbonyl C region was amide-C.
|
The signal intensity of alkyl C (0–50 ppm) was relatively strong for the OM fractions extracted from the Bh horizon, compared with those from O-horizon soils. In contrast, O-alkyl C (i.e., carbohydrates; 50–110 ppm) was lower in the fractions extracted from the Bh horizon (Fig. 1, Table 2). Similar patterns have been observed for humic substances extracted from soils of Watershed 5 at the HBEF, in which alkyl C increased while O-alkyl C decreased with soil depth (Dai et al., 2001; Ussiri and Johnson, 2003). Baldock and Preston (1995) and Baldock et al. (1997) suggested the ratio of alkyl C to O-alkyl C as an index of the extent of decomposition. This degree of decomposition index was greater for the OM fractions extracted from Bh horizon, as indicated by the ratio of alkyl C to O-alkyl C (Table 2).
The OM extracted from the Bh horizon also exhibited greater signal intensities for carboxyl C than the OM extracted from O horizon (Fig. 1, Table 2), indicating that OM from the Bh horizon has higher carboxyl group content and acidity than OM derived from the Oa horizons.
Sorption of Organic Carbon to Different Soil Horizons
Figures 2 and 3
show the sorption of OC fractions as a function of initial DOC concentration for OC extracted from Oa and Bh horizons, respectively. We observed similar trends in the sorption of OC by different soil horizons for OC extracted from Oa- and Bh-horizon soils. With the exception of the E horizon, all isotherms showed linear relationships between OC retained or released by the soils and OC added at the start of the batch experiment (Fig. 2, 3). In all of the experimental isotherms, the values of the intercepts were less than zero (b > 0 in Eq. [1]), and the slopes (m) were <1, as required by theoretical considerations (Table 3). The coefficients of determination (R2) of the linear regression between the initial OC addition and the amount of total OC released or retained ranged from 0.19 to 0.99 (Tables 3, 4). The lower R2 values were observed in E-horizon soils, which showed nonlinear adsorption isotherms for both hydrophobic and hydrophilic OC fractions (Fig. 2).
|
|
|
|
The intercept term (b) of the initial mass isotherms ranged from 1.45 to 78.2 mmol C kg–1 (Table 3). The amount of OC released after the addition of an OC-free solution was highest for Bh horizons, which contain high concentrations of OM (Table 1). The released OC (b) was lower for pH 4 suspension in all horizons; and increased with both decrease and increase in pH (Table 3, 4), indicating that OC was less soluble at pH 4. The released OC mostly belonged to the hydrophilic fraction (65 to 97%), though the contribution of the hydrophobic fraction to the equilibrium solution increased with increasing pH (Fig. 4) .
|
The partition coefficients (m) of B- and C-horizon soils ranged from 0.29 to 0.96 and 0.05 to 0.87 for hydrophobic and hydrophilic OC fractions, respectively (Tables 3, 4). With the exception of the E horizon, the partition coefficients (m) of the initial mass isotherm relationships indicated that in general the hydrophobic fraction was more strongly adsorbed than the hydrophilic fraction in all horizons. In addition, OC extracted from the Bh horizon had higher partition coefficients than that extracted from the Oa horizon (Tables 3, 4). The partition coefficient values were larger for the experiments performed at pH 4 for hydrophobic and hydrophilic fractions in Bh, Bs1, Bs2, and C horizons (Tables 3, 4), indicating stronger sorption of OC at pH 4 for both hydrophobic and hydrophilic fractions. With some exceptions, the partition coefficient increased from the Bh to Bs2 horizon, suggesting increasing affinity for OC sorption in lower horizons (Tables 3, 4). In the E horizon, sorption isotherms for pH 3 indicated a strong binding of hydrophobic and hydrophilic OC at lower concentrations, reaching a maximum at 20 mmol kg–1 (Fig. 2). At OC additions exceeding 20 mmol kg–1, OC sorption leveled off. At pH 4 and 5 in the E horizons, we observed only net release of OC, indicating a lack of adsorption. Indeed, at pH 5, the isotherms for E-horizon soils (Fig. 2) showed negative slopes.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Since we used OM extracted with NaOH and acidified to pH 2, our results may differ from the behavior of DOM. However, it is worth noting that the fractional composition of the OM we extracted from soils was similar to tension-free lysimeter soil solutions collected under Oa and Bh horizons in this watershed (Wellington, 2002).
The results of this study show that sorption characteristics of hydrophobic and hydrophilic OM can be adequately described by the linear initial mass isotherm method (Nodvin et al., 1986). The partition coefficients (m) calculated for NaOH-extracted hydrophobic and hydrophilic OM ranged from 0.05 to 0.93 in B- and C-horizon soils, and were comparable with those reported in other studies in a variety of soils. For example, Moore et al. (1992) reported m values ranging from 0.15 to 0.78 for Canadian soils, while Kaiser et al. (1996) obtained values ranging from 0.01 to 0.86 for carbonate-free temperate European soils. Nodvin et al. (1986) reported an m value of 0.60 for Hubbard Brook soils (Bhs horizon, similar to our Bs1) for unfractionated DOC. Our intercept (b) estimates ranged from 1.45 to 78.2 mmol kg–1, similar to other studies which used water-extractable DOC (Nodvin et al., 1986; Moore et al., 1992; Kaiser et al., 1996). The similarity in sorption affinities between NaOH-extractable OM and soil solution/water-extractable OM also suggests that the patterns we observed for hydrophobic and hydrophilic OM may be similar to DOM fractions in Hubbard Brook soil solutions.
Effects of Soil Properties on Organic Matter Sorption
The ability of Hubbard Brook soils to sorb OM appears to be related primarily to the C and extractable Fe content of the soil. The affinity of these Spodosol horizons for OC sorption followed the order Bs2 > C > Bs1 > Bh > E. Mineral soils from deeper horizons, low in OM and OC content (Bs2 and C horizons), strongly retained OC. Little or no sorption of OC was observed for soils with high OM contents (Bh and Bs1 horizons; Table 1; Fig. 2, 3). Other researchers have noted similar patterns in a variety of soils (Jardine et al., 1989; Dahlgren and Marrett, 1991; Moore et al., 1992; Kaiser et al., 1996; Kaiser and Zech, 1997). Jardine et al. (1989) observed a nearly four-fold increase in DOC sorption when indigenous OC was removed from soils by hot H2O2 treatment, suggesting that high OC content inhibits DOC adsorption. Kaiser and Guggenberger (2000) observed that an increase in the hydrophobic fraction coating on a Spodosol (Entic Haplorthod) subsoil resulted in a decrease in the surface area of soil and reduced DOC sorption. They suggested that the OM coatings mask soil mineral surfaces and therefore block active sorption sites, resulting in decreased OM sorption. Lack of sorption in the Bh horizon could be due, in part, to the high concentration of indigenous OC already present in the solid phase. The Bh horizon contains 89 g kg–1 of C (Table 1), which may impede further adsorption of added OC. The chemical characteristics of the soil OM also appear to influence DOC sorption in spodic horizons. The b-values observed in Bh horizons were 5 to 18 times greater than the corresponding values in Bs1 and Bs2 horizons despite the fact that total and pyrophosphate-extractable C were only 36 to 41% larger in the Bh horizon (Tables 1, 3, 4).
Despite very low C content, soils from the E horizon exhibited the lowest affinity for OC sorption because of low concentrations of amorphous Fe and Al oxides and hydroxides in this horizon (Table 1). Iron and Al oxides and hydroxides are known to play a significant role in DOC sorption (Jardine et al., 1989; Moore et al., 1992; Kaiser et al., 1996). For example, Jardine et al. (1989) demonstrated that removal of Fe oxides/hydroxides by dithionite treatment of soils resulted in a significant decrease in DOC sorption to an Ultisol B horizons. In addition, Moore et al. (1992) and Kaiser et al. (1996) have shown high correlations between DOC sorption and Fe and Al content in a variety of soils from Canada and Europe.
The importance of Fe and Al oxides and hydroxides also helps to explain the patterns observed in OM sorption to B horizons. We observed net release of OC from Bh horizons regardless of OM source, OM type, or input OC concentration (Fig. 2, 3). In contrast, we observed net sorption of OM to Bs1 and Bs2 horizons, except at the lowest input OC concentrations. This striking difference is partly because of the higher C content of the Bh horizon, but also may be attributable to lower extractable Fe and Al concentrations (Table 1). Indeed, the ratio of pyrophosphate-extractable C to dithionite-extractable Fe (Cp/Fed) appears to be a good indicator of the likelihood of OM sorption in these soils. In the E and Bh horizons, which showed little or no net sorption of OM, Cp/Fed was 2.3 and 2.8, respectively. In the Bs1 and Bs2 horizons, where OM sorption was large, Cp/Fed was 1.6 and 1.5, respectively. Sorption of OM appears to occur only when the Cp/Fed ratio is low. High Cp/Fed ratios probably reflect the saturation of OC binding sites in the oxide and hydroxide minerals.
The addition of extracted OM to soils resulted in the mobilization of Fe and Al regardless of the type of OC added (Fig. 5) . The resin treatment of the extracted OM resulted in OC solutions with Fe concentrations < 4.5 µmol L–1, and Al concentrations below detection (data not shown). Therefore, the added OM, initially undersaturated with respect to metal-complexation capacity, initiated the solubilization of Fe and Al from labile solid-phase pools by forming soluble complexes with these metals, leading to increasing solubilization of Fe and Al with increasing OC concentration (Fig. 5; Dahlgren and Marrett, 1991).
|
10 to 20 (Petersen, 1976; Buurman, 1985). The ratios found in this study ranged from 30 to 138 for most of the experiments. According to these results and the mechanisms proposed above, immobilization of DOC by precipitation of organo-metallic complexes is not likely to have occurred in our experiments.
Effects of Organic Matter Properties on Sorption
The sorption of an organic solute will depend on the charge associated with the solute and with the surface. The hydrophobic OM fraction exhibited higher affinity to soil than the hydrophilic fraction in all horizons and at all pH values evaluated in this study. Similar results have been reported by other investigators in a variety of soils (Jardine et al., 1989; Vance and David, 1989; Vance and David, 1991; Kaiser et al., 1996) and Fe and Al oxides (Gu et al., 1995; Kaiser and Zech, 1997). The binding of hydrophobic OM was accompanied by net release of hydrophilic OM, suggesting that hydrophobic OM displaced indigenous hydrophilic substances from the soil. Evidence for preferential sorption of the hydrophobic acid fraction in forest soils was given by Vance and David (1989) and Guggenberger and Zech (1993). They noted that when soil solution passes the spodic B horizons the proportion of hydrophobic acid declined with an accompanied increase in the percentage of hydrophilic acids and neutrals.
The compounds that comprise the hydrophobic and hydrophilic acid fractions are polyfunctional acids differing in degree of aromaticity and carboxylic acid contents (Table 2; McKnight et al., 1985). Carbon-13 CPMAS NMR spectra indicated that the hydrophobic fraction contained nearly twice as much alkyl C and was higher in aromatic and carboxyl C than the hydrophilic fraction (Table 2). The greater carboxyl C content of the hydrophobic fractions indicates a greater concentration of carboxylic functional groups. Dissociation of these groups is largely responsible for the negative charge of OM. Other factors may also contribute to stronger sorption of hydrophobic OM. Higher molecular weight measured for hydrophobic OM may be related to greater sorption (Davis and Gloor, 1981; Ochs et al., 1994; Gu et al., 1995). Davis and Gloor (1981) reported preferential sorption of DOM with molecular weight > 1000 Da to Al oxide. Also, the greater content of aromatic moieties in hydrophobic OM also favors stronger sorption (Table 2; McKnight et al., 1992). An additional factor that may have contributed to stronger sorption of hydrophobic OM may be greater concentration of ortho-O-substituted C (13C resonance 80–90 ppm) in hydrophobic OM, which is preferentially adsorbed by Fe oxide (Gu et al., 1995). This peak appears only as a shoulder in our spectra (Fig. 1), so it is difficult to evaluate this mechanism with our data.
We also noted greater adsorption of OM derived from Bh-horizon OM than OM from the O horizon. This difference may also be related to the structural chemistry of the OM. The OM derived from Bh horizons had 47 to 50% greater carboxyl C than OM from O horizons (Table 2). With a greater fraction of C in carboxylic functional groups, OM from the Bh horizon will have greater charge per unit mass than O-horizon OM at the same degree of dissociation, resulting in greater attraction to positively charged sites on soil surfaces. The greater adsorption of Bh-horizon OM is not surprising, since OM is present in the Bh horizon largely because of sorption processes.
The OM sorption was strongly pH dependent. Maximum sorption of both hydrophobic and hydrophilic OM occurred at pH 4 and decreased at both lower and higher pH values in Bh-, Bs1-, Bs2-, and C-horizon soils. These results are comparable with those of Jardine et al. (1989), who observed maximum OM sorption by an Ultisol Bt1 horizon at pH 4.5. Our results are also comparable with Davis and Gloor (1981) and Davis (1982), who observed adsorption maxima at pH 5 with decreased sorption at lower and higher pH for pure Al-oxide minerals. Vance and David (1991) reported mean pKa values for hydrophobic and hydrophilic acids ranging from 3.8 to 4.8 for DOM originating from a Spodosol forest floor in a northern hardwood forest in Maine. This could help explain the higher OM sorption affinity observed at pH 4 for hydrophobic and hydrophilic fractions. Lower OM adsorption at pH 3 could be the result of: (i) soluble complex formation with Fe and Al, which are more soluble at lower pH; and (ii) lower negative charge of OM at lower pH. Lower OC sorption at pH 5 may be because of decreasing positive charge of oxide surfaces at higher pH.
Specific sorption of carboxyl groups by replacing hydroxyl groups on oxide and hydroxide surfaces is the mechanism most authors assume to be responsible for the interaction between natural organic substances and soil mineral phases (Parfitt et al., 1977; Wershaw et al., 1996; Qualls, 2000). The pH effects observed in our study did not appear to conform to this ligand exchange mechanism. Ligand exchange should release OH– ions to the solution because of the displacement of hydroxyl groups at the oxyhydroxide surface by carboxyl groups. This should result in an increase in the pH of the equilibrium solution. In our experiments, however, the pH of the equilibrium solution either decreased or remained unchanged. The pH effects we observed were consistent with a sorption mechanism governed by electrostatic attraction between negatively charged carboxylic structures and positively charged oxide surfaces. However, because of the complexity of soil surfaces in nature, it is difficult to experimentally control the factors influencing DOC sorption. Therefore, other sorption mechanisms cannot be ruled out.
Implications for Ecosystem Biogeochemistry
The DOC retention processes in the spodic horizons of Spodosols largely control the DOC quantity and quality in stream water (Nelson et al., 1993). Studies of DOC in soil solutions obtained from tension-free lysimeters installed in Hubbard Brook watersheds (W1 and W6) have shown distinct horizonal patterns (Mo, 1997; Dai et al., 2001). These studies revealed that the leachate collected beneath the Oa horizon had the greatest DOC concentration, and was predominantly hydrophobic (60–70%). As soil solutions percolate through the spodic Bh and Bs horizons, there is a reduction in the hydrophobic fraction, with a corresponding increase in the hydrophilic fraction (Mo, 1997). In Bs horizon soil solution, the hydrophilic fraction became the dominant fraction, accounting for up to 73% of DOC (Dai et al., 2001). These patterns are consistent with the selective adsorption of the hydrophobic fraction that we observed in the B and C horizons.
The hydrophilic OM was generally richer in N content than the hydrophobic fractions (Table 2), suggesting that the hydrophilic fraction is an important N carrier in these soils. Qualls and Haines (1991) also indicated that a larger portion of DON was in hydrophilic DOM. Since hydrophobic OM is selectively retained in the soil, DON should be less strongly adsorbed than DOC. This is consistent with the conclusions of Kalbitz et al. (2000), and suggests that the DOC:DON ratio in soil solution should decrease with increasing depth.
Results of this study suggest that the flow-path of soil solutions will largely determine both the concentration and composition of DOC, and the concentration of Fe and Al in streams and lakes draining the watersheds. Drainage water from predominantly shallow flow-paths will tend to have higher DOC concentrations, rich in the hydrophobic fraction because of the lateral movement of water through the O and E horizons, and limited sorption of DOM in the Bs horizons. This can also occur in deep, well-drained soils during storm events when soils are saturated and lateral transport of soil water is enhanced. When drainage waters are derived primarily from deeper flow-paths, streams tend to have lower DOC concentrations, dominated by the hydrophilic fraction. This has been observed for the W1, W5, and W6 streams at Hubbard Brook (Dai et al., 2001).
Leaching of OM in O-horizon soils results in soil solutions with high DOC and low metal concentrations. The high concentrations of Fe and Al we observed in the equilibrium solutions (Fig. 5) suggest that complexation with DOM is the dominant mechanism of transport of Al and Fe to the lower mineral horizons (Bh and Bs1) and, ultimately, export of these metals in stream water. With increased podzolization, the capacity for DOC retention in the Bh horizon is reduced, resulting in a downward shift of the illuvial horizons and increasing DOC output from the solum. The net release of DOC that we observed from Bh horizons at all input DOC concentrations implies that the Bh horizon should be treated as a source of DOC in the ecosystem rather than a DOC sink in HBEF Spodosols.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Consistent with this mechanism, we observed preferential sorption of the hydrophobic fraction, which contained higher carboxyl-C content within the hydrophilic fraction. Additional factors that may have contributed to the preferential adsorption of the hydrophobic fraction include the greater abundance of alkyl-C and aromatic-C, which are associated with lower solubility of DOM and higher molecular weight.
|
ACKNOWLEDGMENTS |
|---|
Received for publication August 30, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. D. Strahm and R. B. Harrison Controls on the Sorption, Desorption and Mineralization of Low-Molecular-Weight Organic Acids in Variable-Charge Soils Soil Sci. Soc. Am. J., November 1, 2008; 72(6): 1653 - 1664. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Gjettermann, M. Styczen, S. Hansen, O. K. Borggaard, and H. C. B. Hansen Sorption and Fractionation of Dissolved Organic Matter and Associated Phosphorus in Agricultural Soil J. Environ. Qual., April 5, 2007; 36(3): 753 - 763. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. M. Jardine, M. A. Mayes, P. J. Mulholland, P. J. Hanson, J. R. Tarver, R. J. Luxmoore, J. F. McCarthy, and G. V. Wilson Vadose Zone Flow and Transport of Dissolved Organic Carbon at Multiple Scales in Humid Regimes Vadose Zone J., March 8, 2006; 5(1): 140 - 152. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Thiffault, D. Pare, N. Belanger, A. Munson, and F. Marquis Harvesting Intensity at Clear-Felling in the Boreal Forest: Impact on Soil and Foliar Nutrient Status Soil Sci. Soc. Am. J., February 27, 2006; 70(2): 691 - 701. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||
