Introduction

Many terrestrial soils have aquic features and are saturated for periods of weeks to months each year. In the US Midwest, for instance, about 20% of the terrestrial soils have aquic features and are classified as Aquolls and Aqualfs according to the US Soil Taxonomy and Classification System. In addition to the historical aquic soil conditions, nowadays many well drained terrestrial soils may experience temporally flooding as a consequence of increasing moisture extremes due to climate changes (Lins and Slack 1999). Flooding of terrestrial soils may induce biogeochemical changes in those soils, such as changes in the carbon and nutrients cycling, and that can lead to great environmental and agriculture concerns. For instance, Fiedler and Kalbitz (2003) found that dissolved organic carbon (DOC) production increased in flooded forest soils and soluble DOC might reach underground water thus increasing the risk of groundwater contamination. Phosphorus can dissolve upon flooding of terrestrial soils (Hutchison and Hesterberg 2004). Elevated soluble P concentrations in surface waters have caused problems, such as algal blooms and eutrophication in streams and lakes.

Terrestrial soils under flooding and reducing conditions may behave like wetland soils that are often saturated for a more prolonged period of time during the year. Wetland soils have been extensively studied and it is known that many redox (reduction–oxidation) sensitive chemical species that are stable under aerobic (oxidizing) conditions, such as NO3 −, Fe, and Mn, often undergo reduction and can be released from the soil to the solution phase during periods of flooding (e.g., Ponnamperuma 1972; Reddy and Patrick 1976; Kirk et al. 2003). Typically, large amounts of organic carbon accumulate in wetland soils, and carbon mineralization drives important redox reactions and subsequent chemical transformations in these soils (e.g., Bridgham et al. 1998; Yao et al. 1999).

In contrast to wetland soils, knowledge about the biogeochemical changes that occur under periodic flooding and anoxic conditions for terrestrial soils and the effects of these changes on soil properties is limited. Previous studies (Peters and Conrad 1996; Scalenghe et al. 2002; Hutchison and Hesterberg 2004; De-Campos et al. 2009) showed that reducing conditions were established immediately after anoxic incubation of terrestrial soils. They observed sequential reduction processes similar to those reported for wetland soils, i.e., reduction of NO3 − followed by reduction of Mn, SO4, and Fe. Hutchison and Hesterberg (2004) found that the onset of redox reactions in a non-dextrose amended terrestrial soil was much slower than in a dextrose amended soil under flooding. This finding indicates that the amount of readily available carbon in terrestrial soils under flooding can influence redox reactions. Hutchison and Hesterberg (2004) also found a strong correlation between the dissolution of phosphorus and DOC upon the onset of reducing conditions and the CO2 production indicated the potential for microbially mediated reduction reactions toward the changes in the solution chemistry.

In a companion paper, we examined the influence of short-term flooding on aggregate stability of the same suite of soils studied here (De-Campos et al. 2009). The focus of the earlier publication was on aggregate stability and we found that the aggregate stability of terrestrial soils is highly affected by changes in their biogeochemistry upon flooding. Overall, the results showed that strongly reducing conditions followed by increases in the solution concentrations of redox-sensitive metals (Mn and Fe), alkaline metals (Ca, Mg, and K), and DOC are well correlated with decreasing soil aggregate stability. These previous findings stress the needs to investigate biogeochemical processes that occur in terrestrial soils upon flooding to better understand their effects on soil properties.

As reported for wetland soils, the rates of redox reactions and associated biogeochemical transformations in terrestrial soils upon flooding depend on the presence of organic carbon. Larger amounts of readily available carbon sources, such as added dextrose, may enhance redox reactions and nutrient cycling (Hutchison and Hesterberg 2004). Terrestrial soils have a wide range of organic matter content and that may exhibit different degrees of redox reactions and biogeochemical processes under reducing conditions. Furthermore, soil management can impact the source of carbon in soil organic matter (Schnitzer and Khan 1989). Forest soils with more readily available source of carbon, such as fulvic acids, should induce faster redox reactions upon flooding than agriculture soils.

Soil organic matter and management practices can play important roles in redox reactions in terrestrial soils. However, it is not clear to what extent the short-term reducing conditions and biogeochemical changes in these soils are affected by the amount and type of soil organic matter or by the history of soil management. In this study, we systematically investigated the influence of short-term water saturation on terrestrial soils of varying organic matter concentrations and management history using a biogeochemical reactor. We hypothesized that flooding induced redox reactions and chemical release in terrestrial soils will increase with the amount of soil organic matter and that cultivation decreases redox reactions and nutrient cycling.

Materials and methods

Soils collection and characterization

The surface horizons (0–10 cm depth) of three cultivated and three uncultivated/forest soils were sampled from the Purdue University Research Farm (ACRE) and studied in the laboratory. The sampled areas were chosen from the criteria that these soils were close enough to have the same climate, landscape position, drainage, settlement history, management history, and represented continuous soil series on the soil map; to avoid any uncertainty in comparisons between cultivated and uncultivated soils. Cultivated soils are soils that have been used for crop production (mainly corn-soy bean rotation) for the last 20 years. Uncultivated soils are soils under forest that have not been disturbed by any agricultural activity at least in the past 20 years, according to the Purdue farm records. The uncultivated soils are under the same type of vegetation cover and conservation history. They are located in forest patches of about 1 km2 in size that are surrounded by agriculture areas. The density of trees in the patches is similar and some trees, such as oaks, are tall enough to indicate that the forest patches have not been disturbed for some time. These forest patches can be remnant from native vegetation.

The collected soils were thoroughly mixed and immediately stored at 4°C. Prior to the experiment, the soils were air-dried, sieved and the soil aggregates ranging from 0.5 to 1 mm diameter were used for incubation. This soil fraction was chosen for a purpose of coupling the results presented in this paper with the study on soil aggregation conducted in the companion paper (De-Campos et al. 2009). Air drying and sieving the soils did not affect the observed biogeochemical changes reported in this manuscript. A companion study (De-Campos 2006) using undisturbed soil cores from the same soils and sites where the disturbed soils were sampled, showed that the results on the measured biogeochemical parameters of the undisturbed samples were highly correlated with the results of the disturbed samples presented in this paper. Properties of the studied soils are listed in Table 1. Results on DTPA extractable Mn and Fe procedure mainly extracts chiefly free or amorphous ferrous (hydr)oxides (Loeppert and Inskeep 1996), are reported in Table 1. Those results are used for comparison between soil extractable DTPA and soluble Mn and Fe measured in the soil solution.

Table 1 Selected soil properties of the original soils
Full size table

Incubation and sampling method

The experiments were conducted in an anoxic biogeochemical reactor. The reactor is basically the same as described in the earlier study (De-Campos et al. 2009) coupled with a CO2 trap system. It consisted of a modified, sealed 250 ml polyethylene centrifuge bottle with fixed pH and Eh electrodes that was connected to a CO2 trap system. A total of 18 reactors were used in parallel in this study.

For incubation, 50 g of soil aggregates (0.5–1 mm) and 75 ml of degassed deionized-distilled Millipore® water (18 MΩ) were placed in each of the centrifuge bottles and incubated for 1, 3, 7, and 14 days under anoxic conditions. The cells were sealed immediately after the degassed water was added and CO2-free argon (Ar) gas was flowed through the reactor continuously at a rate of 15 cm3 min−1. Residual CO2 in the Ar gas was scrubbed by flowing the gas through a NaOH trap. The gas was subsequently passed though a degassed water bubbler to increase the relative humidity of the Ar gas. Under these conditions the headspace volume was recycled every 10 min. The experiments were performed in the dark at a temperature of 25°C (±1°C).

Redox potential (Eh) and pH values were recorded daily in the soil solution in the cells and the Eh values were corrected using the equations proposed by Rowell (1981).

Evolved CO2 from the reactor was trapped by flowing the Ar gas in a series of 2 vials containing a solution of 0.07 M NaOH. The concentration of 0.07 M NaOH was chosen based on the amount of CO2 evolved during the course of the 14 day experiment. The vials were replaced daily during the incubation.

After each incubation period, the centrifuge bottles were transported to an Ar-flushed glove bag and the soil solution was collected under an oxygen-free condition. Inside the glove bag, the sealed bottles were opened and 1 ml of non-filtered aliquot was removed immediately for Fe+2 analysis. After collecting the samples for Fe+2 analysis, the centrifuge bottles were sealed again under the oxygen-free conditions and centrifuged at 8,000×g for 10 min. After centrifugation, the supernatant was obtained in the glove bag using a 30 ml syringe and filtered through an Ar-gas-purged 0.45-μm filter. The supernatant was divided into three equal aliquots. Two of the aliquots were acidified using two drops of 4.5 M H2SO4 solution and later analyzed for nitrogen and metals. The non-acidified aliquot was analyzed for DOC.

Chemical analysis

Iron (Fe2+) in the supernatant was analyzed by measuring the absorbance at 562 nm in a UV spectrophotometer (Fisher Sci. Inc. Thermo Genesis 10uv) following the method of Kirk et al. (2003). One of the acidified filtered aliquots was analyzed colorimetrically for NO3 − and NH4 + using the Quickchem 1100 system (Lachat Instruments, Milwaukee) following the methods of Wendt (1999) and Prokopy (1997), respectively. The remaining acidified filtered aliquot was analyzed for metals (total dissolved Mn and Fe) and soluble phosphorus using inductively coupled plasma with optical emission spectroscopy (ICP-OES-Perkin Elmer Optima 2000DV, Norwalk). DOC and inorganic carbons (DIC) were analyzed using a total organic carbon TOC analyzer from the non-acidified filtered supernatant (Shimadzu TOC-Vws, USA). The samples were analyzed directly in the ICP and TOC without any further processing besides acidification. The measured concentration values of metals, phosphorus and DOC and DIC were recalculated in a soil mass basis.

The evolved CO2 that was trapped in each of the 0.07 M NaOH vials was determined by titration with a 0.07 M HCl after adding BaCl2 following the method of Zibilske (1994). Most of the evolved CO2 was captured in the first vial and the total amount of evolved CO2 was determined by adding the amount of CO2 captured from each of the vials in series. Total soil carbon and nitrogen was analyzed by the combustion method using a CNH analyzer (LECO CNH 2000 Analyzer, LECO Corp., St. Joseph, MI).

More detailed data on DOC, Fe, and Mn have been presented previously by De-Campos et al. (2009) when the short-term water ponding effect on aggregate stability was quantified. Reference will be made when these results are used in this manuscript to support the current analysis.

Statistical analysis

Analysis of variance was used to determine significant effects of management systems (cultivation vs. forest), saturation duration (1, 3, 7, and 14 days), and concentration of soil organic matter on the solution chemistry of the six soils (SAS Institute Inc.). Significant effects were identified at the 5% probability level. Simple linear and Spearman’s correlations were also used.

Results

Changes in Eh and pH

The changes in Eh and pH for the cultivated and uncultivated soils after incubation are shown in Fig. 1a, b. As a group, the Eh values for the uncultivated soils decreased rapidly to an average value of ~−120 mV within 1 day of incubation and were stabilized to ~−100 mV during the 14-day incubation (Fig. 1a). The response of the cultivated soils was distinct and showed a more gradual decrease in Eh from an initial value of 450 mV decreasing linearly over time to ~50 mV after the 14-day incubation (Fig. 1a).

Fig. 1
figure 1

a Changes in redox potential (Eh) for the averaged soils. b Trends in pH during anoxic incubation for the six soils. The soils are cultivated: (A) Peotone, (B) Drummer, (C) Miami; uncultivated: (D) Drummer, (E) Whitaker, (F) Miami. Error bars represent ±1 SE

Full size image

Similar to the observed changes in Eh, the uncultivated soils showed a greater change in pH, with an average increase of 1.2 pH units after 14 days, than the cultivated soils that only increased 0.75 pH units for the same time frame (Fig. 1b). Unlike the rapid decrease in Eh, however, the pH changes for the uncultivated soils was more gradual.

Carbon production

The cumulative amount of CO2 evolved during the 14-day incubation for the average of each group of the three cultivated and three uncultivated soils is plotted in Fig. 2. On average, the amount of CO2 evolved from the uncultivated soils was 295 ± 19 μg C CO2 g −1soil after 14 days. This amount was roughly twice the CO2 production from the cultivated soils (average of 163 ± 10 μg C CO2 g −1soil after 14 days). For both types of soils, CO2 concentration increased linearly over time indicating a fairly constant CO2 evolution throughout the short-term anoxic incubation (Fig. 2).

Fig. 2
figure 2

Trends in evolved CO2 and DOC concentrations through time for the averaged cultivated (cult.) and uncultivated (uncult.) soils. Error bars represent ±1 SE

Full size image

The DOC concentrations are plotted in Fig. 2 for the average of each group of soils. In the uncultivated soils, the DOC concentration increased rapidly after 1 day to an average value of 170 ± 11 μg C g −1soil followed by a more gradual increase to 515 ± 47 μg C g −1soil after 14 days. The amount of carbon mobilized as DOC by the uncultivated soils was roughly 67% greater than the corresponding amount of carbon evolved as CO2 (Fig. 2). In the case of the cultivated soils, the production of DOC was completely different with considerably less production of DOC (20% of the total amount of carbon mobilized as DOC + CO2) relative to the amount of CO2 evolved (Fig. 2).

The partition of DOC, DIC, and evolved CO2 for each uncultivated and cultivated soils are shown in Fig. 3. On average, the total amount of carbon mineralized and solubilized during the 14-day incubation was 905 ± 84 μg C g −1soil for the uncultivated soils whereas only 240 ± 31 μg C g −1soil was released from the cultivated soils. The carbon solubilized as DIC was approximately the same for the two soil groups relative to the total amount of carbon released for each group of soils.

Fig. 3
figure 3

Total cumulative evolved CO2 and DOC and DIC for all soils after 14 days of incubation in a soil basis. Numbers between parentheses are the original amount of organic carbon for each soil. The soils are cultivated: (A) Peotone, (B) Drummer, (C) Miami; uncultivated: (D) Drummer, (E) Whitaker, (F) Miami. Error bars represent ±1 SE

Full size image

The amount of carbon mineralized and solubilized as DOC, DIC, and evolved CO2 normalized to carbon content is shown in Fig. 4 for the six soils. As shown, less carbon was released relative to the total amount of carbon in the higher organic matter content soils than for the soils with the lower organic matter concentrations (Fig. 4; Table 1). In other words, the relative amount of carbon mobilized as DOC and evolved CO2 normalized to carbon content increased as the carbon content for each soil decreased, as indicated by the simple linear correlations between the amount of carbon mobilized as DOC and evolved CO2 normalized to carbon content and the percentage of carbon decomposed for each soil (R 2 = 0.99 for both cultivated and uncultivated soils). Similar patterns were evident for both the cultivated and uncultivated soils. This figure also shows that relative to the carbon content the amount of carbon mineralized as evolved CO2 was larger for the cultivated soils, whereas DOC represented the largest fraction of carbon released from the uncultivated soils.

Fig. 4
figure 4

Total cumulative evolved CO2 and DOC for all soils after 14 days of incubation in a carbon basis. Numbers between parentheses are the original amount of organic carbon for each soil. The soils are cultivated: (A) Peotone, (B) Drummer, (C) Miami; uncultivated: (D) Drummer, (E) Whitaker, (F) Miami. Error bars represent ±1 SE

Full size image

Release of metals, nitrogen, and phosphorus

The soluble iron measured as total dissolved Fe and Fe2+ were highly correlated and the total dissolved Fe is reported here.

The amount of total soluble Mn and Fe from each of the six soils after 14-day incubation and the carbon content of the studied soils are shown in Fig. 5. The amount of Fe and Mn in the soil solution increased for both cultivated and uncultivated soils with time of incubation, as described previously in the companion publication (De-Campos et al. 2009), and was greater for the uncultivated soils. The reduction of Fe occurred more slowly compared to that of Mn. In general, the combined dissolution of Mn and Fe was the inverse of the organic carbon content of the two soil groups, as indicated by the inverse correlations between the sum of Mn and Fe dissolution and the percentage of carbon of each soil (R 2 of 0.91 and 0.90 for the cultivated and uncultivated soils, respectively). After the 14-day anoxic incubation, there was no clear pattern between the dissolution of Mn and Fe and the amount of DTPA-extractable Mn and Fe (Table 2). On average, the percentage of Mn dissolved during incubation compared to the DTPA-extractable Mn was greater than that observed for Fe (Table 2).

Fig. 5
figure 5

Total Fe and Mn dissolved for each soil after 14 days of incubation. Numbers between parentheses are the original amount of organic carbon for each soil. The soils are cultivated: (A) Peotone, (B) Drummer, (C) Miami; uncultivated: (D) Drummer, (E) Whitaker, (F) Miami. Error bars represent ±1 SE

Full size image
Table 2 Total Fe and Mn dissolved from the original soils after 14 days of incubation
Full size table

The reduction of NO3 − occurred rapidly and within 1 day of incubation its concentration in the soil solution was below detection limits (data not shown).

The concentrations of soluble ammonium and phosphorus increased in the soil solution during incubation. Soluble ammonium concentrations are shown in Fig. 6. More ammonium was solubilized from the uncultivated soils relative to that solubilized from the cultivated soils. On average, 2.6 μmol g −1soil of ammonium was solubilized from the uncultivated soils after 14 days as compared to 0.2 μmol g −1soil from the cultivated soils. Similar to the dissolution of Mn and Fe, less ammonium was solubilized from the soils with the highest organic matter concentrations (soils A and D; Fig. 6). Mineralized ammonium increased as DOC was increased (Fig. 7). Since relatively small amounts of DOC and ammonium were solubilized from the cultivated soils (Figs. 2, 7), the positive correlation in Fig. 7 is mainly attributed to the production of DOC and ammonium from the uncultivated soils.

Fig. 6
figure 6

Changes in NH4 + through time for all soils. Error bars represent ±1 SE

Full size image
Fig. 7
figure 7

Simple linear correlation between NH4 + and DOC. Data points are average of triplicates for each cultivated and uncultivated soil

Full size image

Soluble phosphorus from each of the six soils over the 14-day incubation is shown in Fig. 8. The pattern of soluble phosphorus from these soils was complex. The soluble phosphorus concentration from the cultivated soils is positively correlated with the production of DOC to the soil solution (Fig. 9). This pattern is not evident for the soluble phosphorus concentration from the uncultivated soils (Simple linear correlation R 2 value of 0.02).

Fig. 8
figure 8

Changes in total P through time for all soils. Error bars represent ±1 SE

Full size image
Fig. 9
figure 9

Simple linear correlation between total dissolved phosphorus and DOC for the cultivated soils. Data points are average of triplicates for each cultivated soil

Full size image

Discussion

Both the soil solution chemistry and the amount of CO2 mineralized to the atmosphere were strongly influenced by flooding. Increasing anoxic incubation time resulted in strongly reducing conditions with an associated increase in pH, and production of CO2, DOC, soluble Fe and Mn, NH4 +, and, to a lesser extent, soluble phosphorus. A clear distinction between the cultivated and uncultivated soils was observed for all of the biogeochemical parameters studied, which were significantly affected by soil management (Table 3). In contrast, Eh, CO2, total dissolved Mn and Fe, and NH4 +, were not significantly affected by the soil organic matter concentrations (Table 3). Another factor that could influence the observed biogeochemical changes between the six soils is the initial soil microbial population, particularly the Mn and Fe reducing organisms, which we did not measure in this study. A complementary study related to the determination of the initial soil microbial population and how it changes through time of anoxic incubation would be worthy. Overall, our results are in the range of other results reported in the literature for terrestrial and wetland soils (Table 4).

Table 3 Analysis of variance for the effects of soil organic matter for cultivated (S1) and uncultivated (S2) soils, management system (M), and days of anoxic incubation (D) on the measured biogeochemical variables
Full size table
Table 4 Observed changes after 14 days of incubation and a comparison to the literature values
Full size table

Influence of short-term saturation on redox potential and pH

A remarkable result from this study is the difference in the pathways of redox potential and pH showed by the uncultivated and cultivated soils (Fig. 1). The onset and development of strongly reducing conditions within 24 h for the uncultivated soils observed in this study (Fig. 1a), and in the companion study by De-Campos et al. (2009) using the same soils, reactor, and similar anoxic incubation procedures; was quicker than reported in previous studies for terrestrial and wetland soils under flooding (Table 4).

One reason for the rapid drop in Eh for both the cultivated and uncultivated soils is that oxygen-free H2O was used in the incubation experiments. In a normal field setting, some dissolved oxygen would be present in the H2O and this would provide some residual electron acceptors in the system that would slowly be depleted. In our study, oxygen was removed from the H2O prior to incubation resulting in a more rapid transition from oxidizing to reducing conditions. Regarding the sharp contrast between the cultivated and uncultivated soils, the underlying chemical mechanisms are not fully understood at the present time and deserve further investigation. One would conclude based upon the induction of reducing conditions and the attendant biogeochemical changes that the nature of the organic matter in the uncultivated soils is more labile, as suggested in our earlier paper (De-Campos et al. 2009). The presence of more labile carbon sources can cause faster decrease in Eh, as reported by Hutchison and Hesterberg (2004) for a dextrose amended experiment, and that may explain the fast decrease in Eh observed in our experiment for the uncultivated soils.

The rapid onset of reducing conditions for the uncultivated soils can then be attributed to a reactive carbon component that is either absent or occluded in the cultivated soils. Agronomic practices on the cultivated soils may have changed the soil aggregates and caused the distribution of soil organic matter within aggregates being distinct from that of the uncultivated soils. The companion study about the anoxic effects on aggregation of these six studied soils (De-Campos et al. 2009) showed a more significant reduction of the aggregate stability on cultivated soils than that of uncultivated soils. In addition, it was found that the initial concentration, nature, and solubilization of soil organic matter were important factors controlling the overall aggregate stability of these terrestrial soils after short-term flooding.

The results on redox potential (Fig. 1a) and nature of soil organic matter (De-Campos et al. 2009) suggest that the quality of soil organic matter rather than the quantity of organic matter is the main soil carbon attribute that determines rates of redox reactions in terrestrial soils upon flooding. This hypothesis is further supported by the analysis of variance that shows the redox potential was not significantly affected by the soil organic matter concentrations (Table 3).

Carbon dynamics

An important finding of this study is that the management system, along with anoxic incubation time, are major factors that determine the quantity of carbon decomposed in terrestrial soils under flooding, as shown by the analysis of variance (Table 3). The carbon mineralization and solubilization (sum of CO2, DOC, and DIC) was about three to four times greater for the uncultivated soils as compared to the cultivated soils (Fig. 3), despite the fact that these soils represented a wide range of soil organic carbon concentrations. This suggests that the amount of soil organic matter did not limit carbon mineralization and solubilization.

However, when the release of carbon is normalized to carbon content, the influence of the amount of organic carbon present in the soil becomes apparent (Fig. 4). In this case, a greater fraction of carbon is mineralized and solubilized from the soils with lowest organic carbon concentrations. During the 14-day anoxic incubation, 2.3% of the total carbon present was mobilized from the soil with lowest organic carbon content (Soil C; Fig. 4) against 0.1% for the soil with the highest organic matter content (Soil D; Fig. 4). This pattern suggests that soils with high organic matter concentration exhibit mobilization of a lower fraction of their total carbon concentration upon flooding than soils with low organic matter concentration.

The observed changes in DOC concentration for the uncultivated and cultivated soils were remarkable (Figs. 2, 3). We expected an increase in the DOC concentration in both soil groups as organic matter was decomposed and CO2 production increased (Fig. 2) because this pattern has been reported by other studies and is common in soils under reducing condition (Phillips and Greenway 1998; Hutchison and Hesterberg 2004). For instance, Hutchison and Hesterberg (2004) found that most of the carbon decomposed from the silt + clay size fraction of a sandy clay loam under anoxic incubation was in the form of DOC. In their study, DOC loss after a 14-day of incubation accounted for 82% of the total carbon decomposed whereas in our study it accounts for 57 and 17% in the uncultivated and cultivated soils, respectively (Fig. 3). It is also known that the concentration of DOC produced by forest soils is higher than that of arable soils mostly due to vegetation types and concentration of organic matter (Chantigny 2003). Therefore, higher DOC concentrations in the uncultivated soils and lower but increased DOC concentration in the cultivated soils were expected. Possible explanations for what happened with the fate of DOC in those soils are as follow.

The DOC molecules in the cultivated soils consisted of high molecular weight organic molecules whereas in the uncultivated soils low molecular weight organic molecules predominated in the DOC fraction, as reported in our earlier paper (De-Campos et al. 2009). Molecules with high molecular weight are preferentially adsorbed to mineral surfaces compared with low molecular weight molecules because of their higher hydrophobicity, lower solubility, and presence of aromatic rings and carboxylic acids (Jardine et al. 1989; Gu et al. 1995; Kalbitz et al. 2000). Thus, it is likely that a large quantity of high molecular weight organic molecules generated by organic matter decomposition in the form of DOC in the cultivated soils was adsorbed by the mineral phases of those soils, and that caused the observed limited increase in DOC in the solution phase of those soils (Fig. 2).

Iron reduction has been also described as an important mechanism of dissolved organic matter (DOM) solubilization in soils under anoxic conditions because it promotes the dissolution of iron oxides thus DOM desorption (Fiedler and Kalbitz 2003). Iron reduction and dissolution was highly correlated with the increase in DOC concentration in the uncultivated soils (Spearman’s correlation coefficient of 0.78) and this suggests that DOC production was affected by iron reduction in those soils.

The observed changes in DOC concentration may be also related to changes in pH. It is reported that increases in pH lead to desorption of DOC from oxide mineral surfaces (Gu et al. 1994; Kalbitz et al. 2000). The pH in the uncultivated soils increased to 7.7 after 5 days of incubation (Fig. 1b). This may have facilitated desorption of DOC from mineral surfaces and caused the observed increase in DOC production in the uncultivated soils.

In summary, it appears that DOC was more tightly adsorbed by the cultivated soils considering that these soils had higher contents of preferentially adsorbed higher molecular weight organic molecules, lower iron (hydr)oxides dissolution rates, and lower changes in pH. Thus the fate of DOC in the uncultivated soils led to a high solubilization of organic molecules whereas adsorption of soluble organic molecules was favored in the cultivated soils.

The carbon data showed here, based mainly on CO2 and DOC, present an interesting and consistent picture of how these two groups of soils respond to flooding. During the anoxic incubation CH4 should have been produced as a consequence of microbial respiration and CO2 reduction, as observed by Peters and Conrad (1996) for terrestrial flooded soils, and that could add information to further explain the cycling of carbon in those soils. Methane was not measured in this study because we were unable to trap CH4 in our reactor and this would be of interest in a complementary study.

Nutrients cycling

Overall, the turnover of N and P compounds was greater in the uncultivated soils (Figs. 6, 7, 8, and 9). Nitrate was rapidly consumed (data not shown). It is likely that the small concentration of this compound in the original soil (Table 1) coupled with lower concentration of oxygen due to the degassing procedure adopted in this experiment were responsible for this fast consumption of nitrate.

An interesting finding of this study is the changes in ammonium concentrations. Increase in NH4 + concentration in the solution of soils under anoxic condition has been reported (Ponnamperuma 1972; D’Angelo and Reddy 1994) and was expected to happen in both soil groups. Such increase in NH4 + concentrations was observed for the uncultivated soils (Fig. 6). Conversely, the ammonium concentration in the cultivated soils did not increase with the incubation time and it seems to follow a different pathway similar to DOC (Figs. 2, 6).

Sorption mechanisms may have controlled the fate of ammonium in the cultivated soils. Fernando et al. (2005) reports that DOC can chemically react with free ammonium in solution and fix it through formation of DOC–NH4 + complexes that would be sorbed on soil colloids. As discussed early in this paper, the DOC was potentially adsorbed to the cultivated soils. Therefore, it is likely that the DOC molecules retained in the mineral surfaces of cultivated soils formed complexes with ammonium present in solution, in agreement with Fernando et al. (2005). The formation of these complexes would be responsible for the observed low concentration of ammonium in the solution phase of the cultivated soils.

Our results on P dissolution to the soil solution during anoxic incubation are comparable with literature data (Table 4). Prior incubation studies were focused on terrestrial soils with much higher total P concentrations (Scalenghe et al. 2002; Hutchison and Hesterberg 2004). In these studies, more total P was dissolved during incubation (Table 4) and this was a reflection of the total amount of P present in the soil initially.

There are two possible sources of soil P dissolution during anoxic incubation. One possible explanation is due to decomposition of organic matter and that would be consistent with the production of DOC. A more likely explanation, however, is that the combined decomposition of organic matter from the soil in the form of DOC and the reducing conditions that were established during incubation, caused P dissolution in association with Fe-oxides. However, the soluble P concentrations were not correlated with the dissolution of Fe (data not shown; R 2 value of 0.26). Hutchison and Hesterberg (2004) reported that increased DOC concentration leads to increased P dissolution similarly to what we found for the cultivated soils (Fig. 9). They found a strong correlation between DRP and DOC and postulate that increases in DOM concentration from organic matter decomposition during microbial reduction induced dissolution of PO4 by mechanisms of competitive adsorption between DOM and PO4 for mineral surfaces. In contrast, Scalenghe et al. (2002) in a P fractionation study found that in slightly acid soils up to 80% of the total extractable P was associated with Fe and Al. Thus P dissolution in their study was strongly controlled by Fe and Al-minerals dissolution. The mechanisms that are responsible for P dissolution from our soils during anoxic incubation are not clear at present and warrant further study.

Metals cycling

Nitrate, manganese, and iron were the measured redox electron acceptors in this study. Reduction of manganese as expected occurred very early and was related to the redox state of the soil since the beginning of the experiment. The redox potential of both soil groups was in the range of 400 mV after one hour and below 400 mV after 1 day of incubation (Fig. 1a). Manganese reduction is likely in the redox range of 400 mV (Evangelou 1998) and the increase in soluble manganese concentration in soils under anoxic conditions due to reduction of manganese (IV) is extensively reported (e.g. Peters and Conrad 1996; Gao et al. 2002; Shenker et al. 2005).

Reduction of iron in soils under anoxic conditions is also well documented (e.g. Ponnamperuma 1972; Patrick and Jugsujinda 1992; Peretyazhko and Sposito 2005). In general, most of the soluble iron comes from free or amorphous iron (hydr)oxides that are reduced under redox potential below 100 mV (Ponnamperuma 1972; Petruzzelli et al. 2005). Based on the DTPA test for the whole soil (Table 2), the concentrations of soluble iron, and the range of Eh observed in this study (Fig. 1a), it is likely that most of the measured iron in our experiment came from the reduction of free or amorphous ferrous (hydr)oxides.

An interesting finding of this study was the amount of Mn and Fe dissolved from each soil after 14-day incubation that was inversely proportional to the organic carbon content of the soils (Fig. 5). This may indicate soil organic matter protecting oxide surfaces against microbial manganese and iron reduction. Thus, the microbes were unable to reduce large amount of manganese and iron in the soils with higher organic carbon content.

Concluding remarks

From our controlled reactor study, we demonstrated the complexity of biogeochemical processes associated with anoxic incubation of terrestrial soils. Reducing conditions developed very quickly and were strongly influenced by soil management. As shown, short-term flooding conditions lead to faster chemical transformations in the soils under forest and that can impact carbon budgets and the cycling of nutrients and metals.

Furthermore, the data illustrate that the soil carbon concentration alone does not determine the rates of biogeochemical transformations. We found a clear linkage between the nature of soil carbon and the production of C, N, P, Fe, and Mn from the soil during short-term saturation events. In future work a more detailed examination on N transformation, including the rapid loss of NO3 −, the evolution of nitrous oxide along with loss of methane and possibly H2S, and the influence of reducing organisms; would be of interest. The rapid onset of strongly reducing conditions and the concomitant changes that occur to the soil solution, point to dynamic behavior of soil organic matter.