1 Introduction

Fe is an essential micronutrient to all living things (Cullen 1999; Falkowski 2002). The soluble Fe in seawater is a limiting factor to the phytoplankton biomass in high nitrogen low chlorophyll maritime space (Martin and Fitzwater 1988; Martin 1990, 1992) while the land–ocean linkage by Fe transport through rivers is now believed to play a key role in the maintenance of ocean ecosystem (Capone et al. 1997; Karl et al. 1997; Narita et al. 2004; Shiraiwa 2005; Yoh 2004). Land surface condition affects the mobility of soil Fe, which might affect the transport of dissolved Fe to the river, giving definite effects on the biomass production in the ocean (Narita et al. 2004; Yoh 2004; Shiraiwa 2005).

Sanjiang Plain is an important part of Amur Basin and lies off the southwestern coast of Okhotsk Sea. It has a wide distribution of wetland being considered as a crucial Fe source of the Amur River and Okhotsk Sea (Hao et al. 2003). The wetland area in Sanjiang Plain changed from 19,450 km2 in 1980 to 9,069 km2 in 2000 (Chen et al. 2002; Liu and Ma 2000; Fig. 1). The reclamation of wetland always starts with the construction of drainage ditches to reduce excess water, and tillage can increase soil aeration increasing the O2 flow into and the CO2 release out of the soil (Reicosky and Lindstrom 1993). Under these conditions, the amount of dissolved soil Fe would be changed giving potential effects on the Fe input of Amur River and further, the phytoplankton growth of the Okhotsk Sea, a famous fishing ground in the world (Narita et al. 2004; Yoh 2004).

Fig. 1
figure 1

Changes of wetland area in Sanjiang Plain from 1980 to 2000

Full size image

In this paper, profile samples of soil in Sanjiang Plain were collected from natural wetland, lowland rice field, and upland soybean field to study the distribution patterns of soil Fe oxides, aimed to elucidate the vertical distribution characteristics of soil Fe under different land use with a focus on evaluating the effects of cultivation on the soil Fe mobility in Sanjiang Plain.

2 Materials and methods

2.1 Sampling sites

The Sanjiang Plain of Northeast China is a low-lying alluvial floodplain of about 66,600 km2. It has a typical continental monsoon climate, i.e., summer is short, warm, and rainy while winter is long and cold. The water and soil in the wetlands are completely frozen from October to next April, and begin to thaw from late April till July. The minimum, maximum, and mean annual temperatures are −18°C to −21°C, 21°C to 22°C, and 1.6°C to 1.9°C, respectively, and the mean annual precipitation is about 600 mm, in which approximately 60% is concentrated in June– September (Guo et al. 2008). The main soil types are Albic Luvisol, Mollic Planoso1, Mollic Glegsol, and Gleysol (FAO-Unesco 1997).

Twelve sites between two upper reaches of Amur River, i.e., Naoli River and Nongjiang River, were selected as sampling sites covering natural wetland, lowland rice field, and upland soybean field. The wetlands are seasonally waterlogged marshes with Calamagrostis angustifolia as the dominant vegetation species, and the farmlands have been cultivated for rice or soybean for 15–17 years (Fig. 2).

Fig. 2
figure 2

Sampling sites in the study area

Full size image

2.2 Soil sampling and analysis

2.2.1 Soil samples collection

In August 3rd–9th, soil samples were collected by a piston-corer consisting of a stainless steel cylinder 5 cm in diameter and sharpened at the end with duplicates at each sampling site. The soil cores had a length of 120 cm, which were wrapped with thin-wall plastic films and stored at 0–4°C to prevent oxidation. After being transported to laboratory, the soil cores were cut into segments (0–10, 10–20, 20–40, 40–60, 60–90, and 90–120; Todorova et al. 2005; Xi et al. 2007; Park and Burt 1999).

2.2.2 Analytical methods

Acid ammonium oxalate-extractable Fe-amorphous Fe oxides (Feo) and sodium hydrosulfite–sodium citrate–sodium bicarbonate (DCB) extractable Fe-free Fe oxides (Fed) are most frequently used to describe pedogenesis (Schwertmann and Murad 1990). Feo includes organically bound Fe and poorly crystallized Fe oxides (Campbell and Schwertmann 1984) accounting for most of the Fe involved in podzolization (Farmer et al. 1983) and partly in gleyization, and Fed includes better crystallized forms in addition to Feo (Schwertmann and Murad 1990). Fed/Fet is considered as a useful indicator of soil formation processes and pedogenic environments and of importance in differentiating soil horizon and weathering rate (McKeague and Day 1966; Blume and Schwertmann 1969). Feo/Fed (often termed the “active ratio”) is used as a measurement of the proportion of amorphous Fe in total Fe oxides and characterizes the Fe oxides crystallinity (Lair et al. 2009; Blume and Schwertmann 1969) and also proposed as a sensitive indicator of soil podzolization and gleyization (Park and Burt 1999). These parameters, together with Fe(II) and water-soluble Fe (Few), the good indicators of Fe mobility, were selected to interpret the effects of land use type on soil Fe in this study.

Fe(II), Few, and soil moisture content were measured with fresh soil. Fe(II) was extracted by 0.5 M HCl which could release any Fe(II) bound to soil particle surfaces. The extract was determined by phenanthroline colorimetry (Analtikjena AG, Germany; Thompson et al. 2006). Few was extracted by deionized water and determined by a flame atomic absorption spectrophotometer (Aanalyst 200, America; Loeppert and Inskeep 1996). The soils from each core were loosely disaggregated, air-dried at room temperature, and passed through 2-mm mesh sieve to determine pH (H2O) and water soluble Ca2+ and through 0.25-mm mesh sieve to determine total Fe (Fet), free-Fe oxides (Fed), amorphous Fe oxides (Feo), and organic C.

The Fet content was determined by the flame atomic absorption spectrophotometer (Aanalyst 200, America) after sodium carbonate fusion digestion. Fed was extracted by DCB at pH 7.0 and determined by phenanthroline colorimetry (AnaltikjenaAG, Germany; Mehra and Jackson 1960). Feo was extracted by acidified ammonium oxalate at pH 4.0 and determined by phenanthroline colorimetry (AnaltikjenaAG, Germany; Schwertmann 1973). Soil moisture content was measured by determining mass loss after heating in 105°C (Hendershot et al. 1993). Soil pH was measured in 1:2.5 soil–water suspension by using Elico Digital EC meter, and organic C was determined by dry combustion using TOC 5000A autoanalyzer (Shimadzu, Japan). Ca2+ was extracted by deionized water and determined by AAS.

2.3 Statistic analyses

One-way analysis of variance (ANOVA) was performed to compare the differences of test parameters between different land use types and between different soil layers. The separation of means was made according to Tukey's verified significant difference at P < 0.05. Pearson correlation analysis was performed to analyze the relationships among test parameters. Both the ANOVA and the Pearson correlation analysis were performed with the software SPSS for Windows, Version 11.0. The figures were drawn with Microsoft Excel 2003 software and Adobe Photoshop 7.1 software.

3 Results

3.1 Vertical distribution of soil organic C and pH

The soil organic C content in wetland and reclaimed lands had a marked decrease from 0–10 to 40–60 cm, and a lesser change downward. In 0–10 cm layer, there was a significant difference in the soil organic C content under different land use (P < 0.05) with the sequence of wetland > lowland rice field > upland soybean field. Compared with that in wetland, the soil organic C content in 0–10 cm layer in lowland rice field and upland soybean field was decreased by 75.3% (P < 0.05) and 82.6% (P < 0.05), respectively (Table 1).

Table 1 Some soil physical and chemical properties at sample sites
Full size table

Soil pH increased with depth in reclaimed lands and was higher below 40 cm in reclaimed lands than in wetland (see Table 1).

3.2 Vertical distribution of soil Fet

Soil Fet content in wetland increased with depth and was 76.2% higher (P < 0.05) in 90–120 than in 0–10 cm. The Fet content in the soil profiles was decreased in the sequence of upland soybean field > lowland rice field > wetland. In 20–40 cm, the Fet content increased by 65.3% (P < 0.05) in lowland rice field and by 71.4% (P < 0.05) in upland soybean field compared with that in wetland (Fig. 3).

Fig. 3
figure 3

Vertical distribution of Fet. Data shown were the means of 16 replicates and error bars

Full size image

3.3 Vertical distribution of soil Fed and Fed/Fet

The Fed in soil profiles was decreased in the sequence of upland soybean field > lowland rice field > wetland. In wetland, the Fed concentration in 90–120 cm soil layer was 306.9% higher (P < 0.05) than that in 0–10 cm. At the depth of 20–40 cm, Fed concentration increased by 311.6% (P < 0.05) in lowland rice field and 498.1% (P < 0.05) in upland soybean field compared with that in wetland (Fig. 4).

Fig. 4
figure 4

Vertical distribution of Fed. Data shown were the means of 16 replicates and error bars

Full size image

Soil Fed/Fet ratio had the similar variation trend with Fed. In 20–40 cm layer, it was increased by 150.0% (P < 0.05) in lowland rice field and by 249.7% (P < 0.05) in upland soybean field compared with that in wetland (Fig. 5).

Fig. 5
figure 5

Vertical distribution of Fed/Fet. Data shown were the means of 16 replicates and error bars

Full size image

3.4 Vertical distribution of soil Feo/Fed, Fe(II), and Few

Land use change led to a significant decrease of Feo/Fed along soil profiles (P < 0.05). In 0–20 cm soil layer, the Feo/Fed was higher in lowland rice field than in upland soybean field, but no significant differences were found below 40 cm (Fig. 6).

Fig. 6
figure 6

Vertical distribution of Feo/Fed. Data shown were the means of 16 replicates and error bars

Full size image

The Fe(II) content in the soil profiles was decreased in the sequence of wetland > lowland rice field > upland soybean field. Comparing with that in wetland, the Fe(II) content in 0–10 cm soil layer in lowland rice field and upland soybean field was decreased by 39.6% (P < 0.05) and 94.7% (P < 0.05), respectively (Fig. 7). The Few had the similar variation trend with Fe(II). At the soil depth of 0–10 cm, Few concentration decreased by 39.0% (P < 0.05) in lowland rice field and 99.5% (P < 0.05) in upland soybean field compared with that in wetland (Fig. 8).

Fig. 7
figure 7

Vertical distribution of Fe(II). Data shown were the means of 16 replicates and error bars

Full size image
Fig. 8
figure 8

Vertical distribution of Few. Data shown were the means of 16 replicates and error bars

Full size image

The correlations among soil Fe forms and soil physical and chemical properties showed that there was a positive relationship between Feo/Fed and Fe(II; Fig. 9), and both Feo/Fed and Fe(II) were positively correlated with soil organic C and moisture content but negatively correlated with soil pH (Fig. 10).

Fig. 9
figure 9

Relations between the Fe(II) and Feo/Fed of soil at 12 sample sites

Full size image
Fig. 10
figure 10

Relations between the Fe forms and physical and chemical properties of soil at 12 sample sites

Full size image

4 Discussions

4.1 Effects of land use change on the profile distribution of soil organic C and pH

The higher storage of organic C in surface soil layer was closely related to the accumulation of plant materials while the differences in the dynamics of organic C in this soil layer should have close relations with the amount and quality of plant residues as well as the environmental and soil conditions (Dick 1983; Wander et al. 1998; Needelman et al. 1999). The greater amount of soil organic C in wetland than in cultivated land was often accompanied by its concentration gradient from surface to subsurface layers (Dick 1983). Under cultivation, the organic C in upper soil layers reduced due to the oxidation resulted from a relatively well-mixing of soil body during farming practices (Thomas et al. 2007).

The higher soil weathering rate as a result of farming practices promoted the higher concentrations of base cations and higher bicarbonate alkalinity and hence maintained the pH near neutral. Near-neutral pH was an indicator of high soil Ca2+ concentration, which was desirable for crop growth, as opposite to low pH values. In all test sites, soil Ca2+ concentration was highly correlated with soil pH (r > 0.6, P < 0.01, data not shown) indicating that Ca2+ was a strong competitor with H+ and Al3+ for exchange sites on the soil particle surface (Sartori et al. 2007; Richter and Markewitz 1995; Reich et al. 2005).

4.2 Effects of land use change on profile distribution of Fet

The Fet concentration in wetland increased with depth and was greatly higher at the depth of 90–120 cm than in surface soil. Such a vertical distribution could be explained by podzolization which often occurred in flooded soil and implied that a significant amount of Fe was leached out from topsoil (Schwertmann and Murad 1990). The vertical distribution pattern of soil Fet in test sites corresponded with the results of Fiedler and Sommer (2004), who found the low level of Fet in the topsoil where permanent water saturation occurs.

The weathering and activation in surface soil was most likely enhanced by the presence of organic acids that would form complexes with Fe. Fe formed soluble organic complexes with organic acid anions and other organic functional groups, which might contribute significantly to the total soluble Fe in soil solution (Lindsay 1991). Some of the Fe-organic matter complexes in top soils were leached out via lateral and vertical migration resulting in the presence of “albic” horizon, and some were accumulated in the bottom (Giesler et al. 2000).

Under the conversion of wetland to farmland, a redox soil layer at lower positions occurred due to the artificial disturbances, and the horizons with high accumulation of Fe were characterized by highly variable redox conditions. Significant amounts of leached Fe from topsoil, which moved vertically within the soil profile, were deposited in subsurface soil (20–40 cm) where a good aeration occurred preventing the further loss of Fe as a solute.

Due to the scarcity of accurate mineralogical and geochemical data, we were not able to reveal all the processes involved in the mobilization and precipitation of soil Fe at the study sites, but the vertical distribution patterns of soil Fet in test sites gave clear evidence of the mobilization and precipitation of soil Fe.

4.3 Effects of land use change on profile distribution of Fed and Fed/Fet

The similar variation trend of soil Fed and Fet and their significant positive relationship (0.817, P < 0.01) suggested that different soil Fet content undoubtedly contributed to the observed differences in the mean concentration of soil Fed between wetland and reclaimed lands. Similar results were also found in other studies (Blume and Schwertmann 1969).

In our study, the higher Fed/Fet ratio in reclaimed lands revealed that reclamation increased the Fed concentration via increasing weathering rate (Collins and Jenkins 1996). Weathering was known to be related to soil temperature and moisture content and was likely to be increased by farming practices.

4.4 Effects of land use change on profile distribution of Feo/Fed, Fe(II), and Few

Comparing with that of soil Fed and Fed/Fet, the vertical distribution pattern of soil Feo/Fed could reflect more dynamic aspects of the removal processes of Fe oxides associated with podzolization and gleyization (Park and Burt 1999). Feo including oxidized Fe(III) and reduced Fe(II) was the most reactive and soluble Fe form in soils (Chen and Barak 1982). Amorphous Fe(III) oxides were relatively immobile while Fe(II) was readily soluble in water and thus, quite mobile.

In this study, the decreased Feo/Fed, Fe(II), and Few in the whole soil profile in reclaimed lands than in wetland indicated the decreased Fe mobility after reclamation while the higher values of these three indicators in upper soil layers in lowland rice field than in upland soybean field suggested the stronger effects of upland farming practices than lowland farming practices on the soil Fe mobility.

There was a positive relationship among Few, Fe(II), and Feo/Fed, which could be explained by the following scheme (Wahid and Kamalam 1993).

figure a

Reclamation would change the soil physical and chemical properties such as soil organic C, moisture content, and pH, which might decrease the production of Fe(II) and Few by restraining the activation of crystalline Fe(III) oxides and the reduction of amorphous Fe(III) oxides to Fe(II; see Fig. 10).

Soil organic C might have inhibitory effects on the crystallization of Feo and play important roles in the solubility and availability of Fe (Coates et al. 1998; Jansen et al. 2003; Ammari and Mengel 2006; Hyun et al. 2007). One reason was that the strong adsorption of Feo by organic anions could disturb and even prevent the rapid crystallization of Feo (Schwertmann 1966) and another reason was that most soils had the microsites where organic C was actively metabolized. If oxygen existed, it could be utilized as electron acceptor, and thus, soil organic C and microorganisms were responsible for the reduction and increased solubilization of Fe. It was generally considered that Fe(III) reducers must come into direct contact with Fe(III) oxides to reduce them. The need for this contact could be overcome in subsurface environment with soil organic C because the organic C might solubilize the Fe(III) prior to the reduction or function as a soluble electron-shuttling compound shuttling the electrons from the surface of Fe(III) reducers to Fe(III) oxide surface (Lovley 1997).

Feo/Fed and Fe(II) were positively correlated with soil moisture content, which was supported by related studies. Wahid and Kamalam (1993) suggested that, in flooded soil, crystalline Fe oxides were rapidly converted into Feo. This transformation appeared to be a prerequisite for the microbial reduction of Feo. Moreover, most soils contain microsites where organic matter was actively metabolized. When these sites became water-saturated, oxygen entry was restricted, and the sites became partially anaerobic. Parts of Fe oxides dissolved releasing Fe(III) and Fe(II) that combined into multivalence and less stable Fe hydroxides (Lindsay 1991).

Soil Feo/Fed and Fe(II) increased with decreasing pH in present study. It followed from reaction (1) that the activity of Fe increased with decreasing pH because the positive charges on Fe oxide surface increased with decreasing soil pH.

$$ {\text{FeOOH}} + {\text{e}}^{ - } + 3{\text{H}}^{ + } \to {\text{Fe}}^{{{\text{2}} + }} + {\text{2H}}_{\text{2}} {\text{O}} $$
(1)

Furthermore, chloride or sulfate ions might form a surface complex together with H+ at the Fe oxide surface, speed up the dissolution process of Fe by further weakening Fe–O bond. However, with the subsequent decrease in pH, the dissolution would ultimately slow down because the ligands were increasingly protonated and discharged resulting in lesser adsorption. Therefore, the maximum dissolution rate of Fe with organic ligands was often found at a certain pH (Schwertmann 1991).

5 Conclusions

The distribution of various Fe forms in soils under different land use type was examined in the Sanjiang Plain of Northeast China where a relatively uniform parent material allowed the direct comparison of the effects of land use type on the vertical distribution of soil Fe.

Our findings showed that in wetland, significant amount of leached Fe from topsoil was moved vertically without any enrichment in subsurface leading to the low levels of Fet and Fed in upper layers while in farmland, eluviated Fe moved vertically along soil profile and deposited in subsurface layers where a good aeration occurred preventing the further loss of Fe as a solute. Wetland reclamation promoted the production of soil Fed and retarded the formation of Feo, Fe(II), and Few. The soil Fe mobility decreased in the sequence of wetland > lowland rice field > upland soybean field suggesting the stronger effects of upland farming on the soil Fe mobility than lowland farming.

The results supported the ideas that, in the Sanjiang Plain, the conversion of wetland into farmland, especially into upland, could change the soil Fe vertical distribution giving potential effects on the mobility of soil Fe.

6 Recommendations and perspectives

Further researches would be made to better understand the functions of different land use in discharging soil-dissolved iron and to quantify the flux of dissolved iron from the Sanjiang Plain to the Amur River and the Sea of Okhotsk based on the high-resolution geographical distribution maps of land surfaces in the alluvial plain.