Various environmental factors influence DOM concentrations and fluxes in soils. These environmental controls may be natural or anthropogenic. We first review the effects of the most important natural environmental factors, temperature, precipitation, soil water fluxes, soil moisture, and snowmelt. We then discuss important anthropogenic controls on DOM dynamics in soil. Because only a few studies concerning environmental controls on DOM dynamics address DOM fluxes, we will focus on DOM concentrations.
Mulholland et al. (1990) postulated that temperature will always be a factor regulating microbial production of DOC, and "a better understanding of these factors may help to identify upper limits on DOC formation and explain the range of DOC concentrations observed in terrestrial and aquatic environments." Strong correlations between soil organic C pools and climate (Jenny, 1980; Post et al., 1982, 1985) imply temperature may affect DOM release from soils. Although the amount of soil organic matter is negatively correlated with temperature at both the global and the local scale (Post et al., 1982; Kirschbaum, 1995), the relationship between temperature and DOM release is equivocal. DOM production in catchments can increase in warmer climates, but the effect of climate on DOM is probably small because decomposition rates, which remove DOM, also increase (Kramer et al., 1990). Liechty et al., (1995) estimated that the differences in soil temperatures (2.1 °C) could be responsible for as much as 16% increase in DOC concentrations in forest floor solutions at the warmer compared with the colder site.
Soil drainage conditions play an important role in understanding climate effects on DOM release. For well drained and moderately drained soils, there is often an inverse correlation between average soil temperature and DOC concentration in surface soil leachates (Cronan, 1990). Thus, DOC concentrations generally increase in cooler environments. In poorly drained areas, DOC concentrations in surface horizons often reach very high levels, regardless of the climate, from cool northern peatlands to warm southern blackwater swamps. These inconsistent data indicate no general climatic effect on DOC.
DOM production may be affected by diurnal temperature variation as postulated by Kramer et al. (1990), but there are no field studies available that address this hypothesis. On the other hand, numerous field studies show seasonal variability in DOM concentrations and fluxes. In general, DOC concentrations in soil solution are higher in summer than in winter (Collier et al., 1989; Dalva and Moore, 1991; Chittleborough et al., 1992; Federer and Sticher, 1994; Heikkinen, 1994; Guggenberger and Zech, 1994; Liechty et al., 1995; Tegen and Dörr, 1996; Scott et al., 1998; McDowell et al., 1998; Guggenberger et al.; 1998; Tipping et al., 1999). Cronan and Aiken (1985) found mean DOC concentration increased by 26 to 32% in shallow soil solution in summer, but DOC concentrations in the B horizon remain relatively constant. The stability of DOC concentration in deeper soil horizons in summer was also corroborated by Qualls and Haines (1991) and Chapman et al. (1995). Differences in temperature effects between upper and lower soil horizons support the hypothesis that DOC concentration in the topsoil, with higher microbial activity than subsoil, is controlled mainly by microbial processes (Guggenberger et al., 1998) which are, in turn, partly controlled by temperature. This is also supported by the observation that hydrophilic compounds (derived mainly from microbial sources) (Bourbonniere, 1989; Vance and David, 1991b; Scott, et al., 1998) and carbohydrates (Guggenberger and Zech, 1994) are released preferentially into the soil solution during the growing season rather than in the dormant season.
However, studies on the seasonality of DOC concentration also show variable results. Vance and David (1991b) and Dosskey and Bertsch (1997) found no seasonal effects on DOC concentration in soil solution. The lack of seasonality shown by Dosskey and Bertsch (1997) may be due to relatively short, mild winters (South Carolina) with neither concentrated growth seasons in summer, nor litterfall periods in autumn (Dosskey and Bertsch 1997). In other words, this lack of seasonality was caused by a lack of seasons. Evans et al. (1996) report that DOC peaks can occur throughout the year.
Michalzik and Matzner (1999) found a correlation between temperature and both DOC and DON concentrations in the forest floor of a Norway spruce stand. However, fluxes of DOC and DON were not correlated to temperature. They concluded that abiotic processes supercede biological controls on DOM release under field conditions. In contrast, Liechty et al. (1995) found a significant positive correlation between soil temperature and DOC flux in forest floors of two hardwood stands. Unfortunately, studies on the effects of temperature on DOM fluxes in soil are rare. However, it seems reasonable that high DOC concentrations in summer are linked with low water fluxes because of much higher evapotranspiration in the growing than in the dormant season and that temperature effects on microbial activity may be a secondary factor.
Compared with DOC, there are fewer data available describing temperature effects on DON or DOP dynamics. According to Currie et al. (1996) and McDowell et al. (1998) DOC and DON show the same seasonal pattern, with higher concentrations in summer than in winter. In a study by Michalzik and Matzner (1999), the DON concentrations in the Oe layer were affected more by temperature than were DOC concentrations. Huang and Schoenau (1998) found decreasing concentrations of water-extractable organic nitrogen and phosphorus in summer because of high mineralization rates.
Some studies have addressed the effects of temperature on DOC in the laboratory, but we have found no such studies for either DON or DOP. The results of laboratory studies are more consistent than field investigations, with nearly all showing that an increase in temperature results in increased DOC concentrations. Only Seto and Yamagiya (1983) and MacDonald et al. (1999) did not find temperature effects on DOC concentrations in soil solution. Decomposition of DOC was probably enhanced to the same extent as DOC production. Tipping et al. (1999) concluded that warming and drying can accelerate the production of potential DOM in organic soil. Gödde et al. (1996) reported Q10 values of 2.0 and 1.5 for DOC from eight different red spruce stands between the ranges of 3 and 10 °C and 10 and 20 °C, respectively. Christ and David (1996b) showed relatively constant Q10 values (near 2.0) over a broad temperature range (3-28 °C) for DOC from one red spruce stand. The DOC response to temperature seems to depend on site conditions. In both studies, the Q10 values for DOC release are lower than those for respiration. Thus, laboratory studies suggest microbial respiration is more susceptible to increasing temperature than DOC release, implying that DOC release with increasing temperature may be caused partly by physical leaching rather than being entirely dependent on microbial activities (Moore, 1998).
Summarizing the effects of temperature on DOM in soils, there is a trend of increasing DOM concentrations with increasing temperature that is more obvious in laboratory experiments than in field studies. It seems unlikely that the release of DOM (DOM fluxes) under field conditions depends entirely on temperature. Climatic and hydrological conditions, litterfall and litter quality, and soil texture and other soil properties can modify and even mask the temperature response of DOM in the field. It is not clear at present how the release of DON and DOP is affected by temperature.
One of the most consistent findings in both field and laboratory studies is that DOC concentrations increase following rewetting after dry periods (McDowell and Wood, 1984; Zabowski and Ugolini, 1990; Haynes and Swift, 1991; Chittleborough et al., 1992; Kalbitz and Knappe, 1997; Lundquist et al., 1999b; Tipping et al., 1999; Zsolnay et al., 1999). It is likely that reduced rates of decomposition in dry soils cause microbial products to accumulate. This, together with cell death and lysis, can contribute to high DOC concentrations in the soil leachate after dry periods (rewetting effect).
Lundquist et al. (1999b) gave three possible explanations for the increase in DOC during rewetting cycles: (i) reduced microbial utilization of DOC in dry periods, (ii) enhanced turnover of microbial biomass and condensation of microbial products by rewetting, and (iii) disrupted soil structure making previously sequestered carbon more available as DOC. According to Christ and David (1994, 1996b), the high proportions of both hydrophilic neutrals and bases suggest a disruption of microbial biomass after soil drying, whereas wetter conditions favor the release of hydrophobic acids. Field studies indicate that the duration of dryness, temperature, and water fluxes are all correlated positively to the amount of DOC released at the onset of soil leaching after a dry period (Chittleborough et al., 1992; Tipping et al., 1999).
In contrast, in another field study, Guggenberger and Zech (1994) found no effects of soil moisture on DOC concentration and composition. Studies on the overall relationship between soil moisture and DOC concentration have shown varying results. Falkengren-Grerup and Tyler (1993) reported higher DOC concentrations at higher water contents in forest floor samples. This positive correlation between the DOC concentration and moisture in well drained soils might be attributable to enhanced biological activity with increasing moisture, subsequently increasing the formation of water-soluble organic compounds. Christ and David (1996b) support these results by observations in the laboratory and conclude that soils would release the most DOC during wet summers.
In most cases, anaerobic conditions caused by saturation increase the release of DOM from soils (Mulholland et al., 1990; Sedell and Dahm, 1990). Several studies (summarized by Moore 1998) have shown that differences in DOC concentrations among streams are strongly related to the amount of wetland (peatland) drainage contributing to streams (Eckhardt and Moore, 1990; Hemond, 1990; Koprivnjak and Moore, 1992; Hope et al., 1994; Mulholland 1997). Anaerobic decomposition of organic matter is less efficient than decomposition under aerobic conditions, with a higher proportion of water-soluble intermediate metabolites released (Otsuki and Hanya, 1972; Mulholland et al., 1990). Laboratory experiments also show that waterlogged soils release higher levels of water-extractable organic carbon than controls. (Homann and Grigal, 1992; Kalbitz et al., 1997).
In summary, increased concentration of dissolved organic matter (i) after dry periods and (ii) as a result of anaerobic conditions seems to be the most important effect of soil moisture on the dynamics of DOM. This conclusion can be drawn from both field and laboratory studies.
Precipitation and Water Fluxes
In addition to temperature and soil moisture, the dynamics of precipitation and water fluxes are also greatly responsible for seasonal changes in concentrations and fluxes of dissolved organic matter in soils. An inverse relationship between DOC concentration and water fluxes in organic soil horizons suggests that a simple leaching model might explain some of the seasonal changes in DOC (McDowell and Wood, 1984). Following this up, Easthouse et al. (1992) proposed that there would be considerable leaching and dilution of DOC during high precipitation events in O horizons. The contact time between the soil and the soil solution is decisive (McDowell and Wood, 1984; Michalzik and Matzner, 1999). Thus, in the spring, with more water passing through the forest floor and short contact times, DOC concentrations are lower; in summer, however, low soil water content and longer contact times may lead to higher DOC concentrations (McDowell and Wood, 1984; Bourbonniere 1989). Cronan (1990) and Heikkinen (1994) report increasing DOC concentrations from long contact times in soil horizons with high organic matter content. This concept, derived from field studies, is supported in the laboratory where, for instance, DOC adsorption is more pronounced at lower than at higher pore water velocity (Weigand and Totsche, 1998).
Storm events can significantly alter DOC concentrations and fluxes throughout the year by shifting dominant flowpaths toward preferential flow through macropores, runoff, and lateral flow. High pore water velocity leads to low contact times between soil solution and the solid matrix and creates conditions of chemical and physical nonequilibrium. Thus, adsorption of DOC is diminished in mineral soil horizons (Yavitt and Fahey, 1985; McDowell and Likens, 1988; Moore, 1989; Meyer, 1990). Flushing of DOC adsorbed on aggregate surfaces and concentrated in subsoil horizon micropores also contributes to increasing DOC concentrations and fluxes at the beginning of storm events (Jardine et al., 1990; Chittleborough et al., 1992).
For most storm events, the highest DOC concentrations were observed in the initial breakthrough phase of soil solution at any depth, with DOC decreasing rapidly to a baseline level as flow continues (Jardine et al., 1990; Easthouse et al., 1992). Boyer et al. (1996), modeling DOC dynamics as a function of water flow, suggested that DOC in the upper soil might build up during periods of low flow and be transported into streams during periods of high flow. Tipping et al. (1997) interpreted the observation that the highest DOC concentrations and fluxes occur in the fall in terms of soil hydrology and humification. They argued that in summer, the microbial processing of organic matter produces "potential dissolved organic carbon." At this time, percolating water penetrates relatively deeply into the soil, where DOC adsorption maintains low soil solution concentrations. As the soil wets up in fall, lateral flow becomes more significant. The proportion of the drainage water that passes through only surface horizons, and is therefore rich in DOC, increases.
Field data showing that DOC release increases after large rainfall events are confirmed in soil column studies in the laboratory as both high water fluxes (Kalbitz and Knappe, 1997) and high leaching frequency (Christ and David, 1996a; Gödde et al., 1996) increase the amounts of DOM released.
In contrast, the level of rainfall had no effect on the concentration of DOC in the soil solution in sandy soils (Dosskey and Bertsch, 1997). Here, preferential water and solute flow through macropores may be much reduced and, therefore, water flux intensity may be less important to DOC release than for more clayey soils. The lack of correlation between DOC/DON concentrations and water fluxes has been shown for forest floor of a coniferous stand (Michalzik and Matzner, 1999), for both forest floor and mineral soil solution (only DOC) (Guggenberger and Zech, 1994), and for forest floor leachates of both a hardwood and a coniferous forest (Michalzik et al., 1998). Since the effects of rainfall events on DOM concentrations and fluxes occur on a time scale of hours to days, a relatively high sampling frequency is necessary to detect possible effects (Michalzik et al., 1998).
The most significant effect of precipitation and water fluxes on DOM is the release of DOM at the beginning of large rainfall events. Initial DOM concentrations and fluxes in mineral subsoil horizons are highest under the following conditions: distinct preferential flowpaths, large accumulations of potential water-soluble organic compounds in the soil (caused, for example, by elevated temperature or dry periods without soil leaching), and very intensive rainfall. In addition to flushing out stored or adsorbed DOM at the beginning of storms, simple dilution of DOM during high precipitation events in O horizons is often observed. Temperature-dependent seasonal changes in DOM release from the forest floor and A horizon may be masked by these effects of precipitation and water flux (Mulholland et al., 1990).
Freeze/Thaw Cycles and Snowmelt
Laboratory studies have shown that simulated freeze/thaw cycles also cause DOM release (reviewed by Zsolnay, 1996), with the amount released a function of the water content of the soil before freezing. DeLuca et al. (1992) suggest that freeze/thaw events in soil disrupt microbial tissues (similar to the effects of drying and re-wetting or chloroform fumigation).
During snowmelt, the highest concentrations of DOM in forest floor leachates (Yavitt and Fahey, 1985) and upper soil horizons (Boyer et al., 1997) are observed in the early stages, with rapidly declining concentrations as melting progresses. Most of the DOM leached at snowmelt probably accumulated under the winter snowpack by decomposers continuously releasing soluble organic matter (Fahey, 1983; Fahey et al., 1985; Yavitt and Fahey 1985). In addition to this long-term accumulation of potential water-soluble organic compounds, organics could have become soluble in a rapid flush in spring just before snowmelt, as microbes responded to the initial wetting front reaching the forest floor (Yavitt and Fahey 1985). Another possibility for increased DOM concentration at the beginning of snowmelt is disrupted soil structure (caused by freeze/thaw cycles) making previously-stabilized organic matter more available as DOM.
As usual, there are far fewer studies for DON or DOP than there are for DOC. However, one study did show differences in the behavior of DOC and DON in forest floor leachate during snowmelt (Yavitt and Fahey, 1985). DOC concentrations decreased more than 3-fold during the snowmelt period, whereas DON concentrations decreased by 8-fold, resulting in an increasing DOC/DON ratio from 20:1 to 75:1. The authors gave no explanation for this phenomenon.
Most studies show an increase in DOC leaching during the early stages of snowmelt. If the forest floor has good percolation throughout the year, no decomposition products can accumulate and there will be no significant increase in concentrations or fluxes of DOM during snowmelt (Currie et al., 1996).
Although various aspects of the effects of elevated N deposition and N fertilization have been studied, little is known about the effects of N deposition on DOM turnover. Since it is hypothesized that large amounts of labile C are required to drive N immobilization (Aber, 1992), DOC concentrations and fluxes should, in theory, be high under N-limited conditions and low under N-saturated conditions (Gundersen et al., 1998). In a field N-addition experiment (Currie et al., 1996; McDowell et al., 1998), concentrations and fluxes of DOC from the forest floor remained unchanged, whereas DON concentrations showed a 200 to 300% increase with the highest rate of N amendment (150 kg N ha-1 yr-1) and lesser but significant DON increases at lower-N treatments. At present, it is speculative to assess the ecological significance of this additional DON production (McDowell et al., 1998).
In a field study, Cronan et al. (1992) showed that the DOC release rate decreased by 20% after N fertilization (NH4Cl) of a forest soil. This could be due, in part, to decreasing pH and increasing ionic strength (see sections pH, Ionic Strength). Stuanes and Kjønass (1998) and Emmett et al. (1998) found no changes in DOC or DON concentrations and fluxes in either the forest floor or the A horizon after N addition to conifer forests in Sweden and Wales, respectively. In a comparison of five sites (NITREX project), Gundersen et al. (1998) found no consistent long-term response in DOC leaching to N addition and no clear relationship between DOC flux and the N status of control sites. The original hypothesis that DOC should decline under conditions of N saturation was rejected (Aber et al., 1998; Gundersen et al., 1998; McDowell et al., 1998), and there was no evidence that available C decreases with N availability (Gundersen et al., 1998).
In contrast to the previously discussed studies, Zech et al. (1994, 1996) suggested, based on chemical structural studies, that elevated N deposition may increase in DOC mobilization rates in the forest floor by stimulated microbial activity and N-induced suppression of ligninase, which results in an accumulation of moderately degraded and highly water-soluble soft-rot products.
For arable soils, Chantigny et al. (1999) report that decreasing soil mineral N content was consistently associated with an increase in water-soluble organic carbon. In a field study, however, Rochette and Gregorich (1998) found no effect of inorganic N addition on DOC concentrations in an arable soil.
Many laboratory studies have addressed the effects of nitrogen on organic matter decomposition. A few laboratory experiments show that DOC production in soils can be sensitive to N status, but the results are not consistent. Nitrogen addition as urea (Homann and Grigal 1992, who used O and A horizons of forest soils) resulted in increased release of water-soluble organic carbon. Gödde et al. (1996), however, found significantly lower DOC production in samples with high total N and low C/N ratio in soil organic matter, whereas Michel and Matzner (1999) could not find an effect of the C/N ratio in soil organic matter on the release of DOC and DON.
At present, a clear conclusion about the effects of an increased N-input cannot be drawn. In most cases, DOM dynamics appear to be independent of N-inputs. In some cases, however, both increased and decreased DOM concentrations are reported.
Land Use and Management Effects
Land use changes, such as clear-cutting of forest stands, afforestation, liming and fertilization, converting forests into arable sites, and other management activities, influence the dynamics of DOM by (i) changing the input of organic matter, (ii) changing the substrate quality, and (iii) altering the rates, extent, and pathways of microbial degradation and synthesis of organic matter (Cronan et al., 1992).
Clear cutting and afforestation can have diverse effects on DOM concentrations and fluxes (Sollins and McCorison, 1981; Moore, 1989, Hope et al., 1994; Neal and Hill; 1994). Some studies (e.g., Johnson et al., 1995) have found an increase in DOC concentrations and fluxes in the forest floor and mineral soil horizons after clearcut, some (e.g., Meyer and Tate 1983) have found a decrease, and others (e.g., McDowell and Likens, 1988; Moore and Jackson, 1989) have found little change in DOC. Lepistö et al. (1995), finding no relationship between DOC concentration and clear cutting, postulated that the effect of clear cutting on DOC fluxes was caused mainly by increased runoff volume. Quideau and Bockheim (1996, 1997) reported increased DOC concentration in the soil solution after the afforestation of a former prairie site. However, another study has shown that the afforestation of a wetland did not affect DOC concentration in soil solution (Collier et al., 1989).
Peatland drainage also has variable effects on DOC concentrations, although DOC fluxes generally increase because of higher runoff (summarized by Moore, 1998).
Concentrations of C in solutions from soils in northern Saskatchewan (Cryoboralfs) decreased in the order aspen forest > recently cleared forest > wheat/fallow field (Ellert and Gregorich, 1995). In general, forest soils have higher topsoil DOM concentrations than arable soils (Seto and Yui, 1983). However, the cultivation of a forest soil led to a 2- to 5-fold increase in DOC concentration due to intensified mineralization during the first stage of cultivation (Delprat et al., 1997). During the second stage, once the original organic matter was stabilized, DOC concentrations decreased with increasing time of cultivation (Delprat et al., 1997). After 26 years of corn cultivation, the DOC concentration was less than that of the neighboring forest soil (Delprat et al., 1997). Reviewing long-term field experiments, Campbell et al. (1999) found that water-soluble organic C levels in arable soils were highest in the most fertile treatments and in fields under crop rather than fallow. In a field study, flow-weighted mean concentrations of total organic nitrogen in watershed discharge (streams) increased as the proportion of cropland in the watershed increased (Jordan et al., 1997). Studies comparing effects of different crops on DOM concentrations and fluxes in the field are not available, although there are considerable differences between the water-extractable organic carbon contents of various crop plants (Zsolnay, 1996). More information concerning plant species effects on dissolved organic matter is given in the section titled Substrate Quality.
Much information is available about the effects of liming on the dynamics of DOM in forest soils. In most cases liming results in increased DOC concentrations in upper soil horizons (Grieve, 1990; Göttlein and Pruscha, 1991; Göttlein et al., 1991; Andersson et al., 1994; Gensior, 1995; Kreutzer, 1995; Erich and Trusty, 1997). This may be caused by the higher solubility of DOM at higher pH (Thurman, 1985), stimulation of microorganism activity (Kreutzer, 1995), and a combination of these two factors (Andersson et al., 1994). Exceptions to these studies include Cronan et al. (1992), who reported no effects of liming on DOC concentration of forest soils, and Simard et al. (1988), who found that DOC flux from the A horizon of an arable soil decreases with increasing lime amendment, possibly because of an enhanced decomposition of DOM and/or a reduced solubility by increased ionic strength and Ca2+ concentration (see Ionic Strength and Specific Cations and Anions sections).
Only a few studies are available on the effects of liming on DON. Both Göttlein et al. (1991) and Kreutzer (1995) demonstrated that in addition to observed increases in DOC, DON leaching from forest floors also increased in limed plots. Since both DOC and DON increased, the DOC/DON ratio was unaffected by liming (Göttlein et al., 1991).
In agricultural soils, fertilization with organic compounds has been shown to increase the content of water-extractable organic carbon by a factor of 2.7 to 3.2, depending on the kind of fertilizer added (Zsolnay and Görlitz, 1994; Rochette and Gregorich, 1998; Gregorich et al., 1998). This effect is probably based on the direct addition of water-soluble organic carbon to the soil (Zsolnay, 1996). In manured soils, soluble C increased sharply after manure application, followed by a gradual decrease (Rochette and Gregorich, 1998). Cronan et al. (1992) reported an increase of DOC concentration in the forest floor after a sawdust amendment. Organic farming results in higher water-extractable organic carbon content than conventional farming, probably because of the higher input of organic matter by animal manure (Lundquist et al., 1999b).
No information is available about the effects of tillage on DOM dynamics in agricultural soils, although one would expect DOM release to increase after soil tillage as the result of enhanced mineralization of organic matter.
Although the effects of forest clear cutting and afforestation on DOM dynamics are still unclear, both liming and organic fertilization consistently result in enhanced release of DOM due to stimulated mineralization and the addition of water-soluble organic matter, respectively. Both field and laboratory studies support this conclusion.