Carbon and progressing microbial reduction
In our bottle test, the redox-states, pH, and the concentrations of Mn, Fe, SO
4, and H
2S prevailing at the end of incubations were typical of those found in the sediments of brackish systems, such as the open and coastal Baltic Sea (see e.g., Lehtoranta and Heiskanen
2003; Reed et al.
2011). In sediments, most of the organic C consists of compounds which are not readily available for mineralization. It is likely that the mineralization is controlled by the ability of the sediment to produce labile C compounds (e.g., by fermentation) rather than by the concentration of C itself. In our experiment, acetate was used to simulate a system that was able to produce labile C for microbial mineralization using Mn, Fe, and SO
4 as electron acceptors. By manipulating the supply of labile organic C to brackish soil suspensions, we initiated progressing redox evolution, analogous to that encountered by the eroded soil when it settles on anoxic sediments with different abilities to produce labile C.
When no acetate was added, the organic C present in the soil was able to produce Mn and Fe in the solution, but SO
4 remained at the initial level. Low additions of acetate maintained the accumulation of Mn and Fe and resulted in a minor decrease in SO
4 (Fig.
2a–c). Thus, a low C level could trigger the microbial reductions of Mn and Fe, but the system did not proceed into SO
4 reduction with such an efficiency that it would have markedly decreased the level of SO
4, even though
Incubation 1 took more than 2 years. However, SO
4 reduction was activated with higher additions of acetate. At moderate C levels, the concentration of SO
4 decreased, but no H
2S was detected. The lack of H
2S was possibly due to coprecipitation of sulfides and Fe
2+ in the solution (Chapelle et al.
2009), as suggested by the fact that the concentration of dissolved Fe was lower in these units.
A high C addition was required to exhaust the reserves of reactive Mn and Fe oxides in the soil, giving room for the dominance of SO
4 reduction. When the C/soil ratio was at least 6, soils in all units turned black, as a result of the transformation of Fe(III)oxides into black Fe sulfides. Furthermore, the concentration of dissolved Fe became nondetectable and H
2S accumulated into the water. The presence of H
2S indicates that even the abiotically reducible Fe oxides may have precipitated as Fe sulfides. The microbial and chemical processes in the experiment produce many simultaneous reactions with an opposite effect on pH (Soetaert et al.
2007), but in general, the pH was the lowest when dissolved Fe was present (low C additions) and the highest in units with presence of H
2S (high C additions, Fig.
4a–d).
Organic C and P release
Of the elements monitored, P was the only nonredox-dependent element that increased in concentration as a function of organic C (Table
2). However, after exceeding a threshold in C/soil ratio, further C additions caused lower additional P release, and the release could be modeled well with the amount of soil alone (Fig.
6d). The diminishing increase in P release from soil (Fig.
5b) could be due to, for example, P readsorption on Fe(II) hydroxides, Al oxides, and mineral surfaces (Patrick and Khalid
1974; Roden and Edmonds
1997; Gächter and Müller
2003).
When no C was added, the release of P was accompanied by the release of Fe, with the molar Fe/P ratio in solution ranging from 0.7 to 2.8 (calculated from data in Fig.
2d). This ratio indicates that microbial Fe reduction results in an accumulation of both Fe and P, and presupposes that Fe has the capacity to bind most of the released P, when oxic conditions are encountered (Blomqvist et al.
2004). Therefore, the eroded soil may carry enough Fe to capture the P released through this coupling of the Fe and P cycles in the brackish sedimentary systems, but our results indicate that this ability is maintained only in environments low in C.
In the units with C/soil ratio <6, the molar Fe/P ratio ranged from 0.02 to 0.50 and in the units with C/soil ratio ≥6, the ratio was at maximum 0.1 and mostly zero. Thus, organic C had the ability to decouple the cycles of Fe and P, through SO
4 reduction, by deteriorating the capacity of Fe to retain P. The pattern found follows the theory presented by Lehtoranta et al. (
2009). Those authors hypothesized that anoxic oligotrophic brackish and marine sediments may be able to maintain the coupled Fe and P cycling, but if eutrophication proceeds, the ability of Fe to retain P is lost through the efficient reduction of SO
4. Our results also endorse the previous study by Lehtoranta et al. (
2008) which states that the ability of the surface sediment to retain P is related to the difference in the dominance of microbial Fe and SO
4 reductions in the bottom areas of the Baltic Sea. However, we have only studied the behaviors of Fe, Mn, and P in a clayey soil, and our study gives little information about the fates of Fe and P bound, e.g., to humic matter in SO
4-rich sedimentary systems. The forms of C, Fe, and P and their respective loadings from rivers to the subbasins of the Baltic Sea vary significantly from one region to another. The recipient sedimentary system may also respond differently to each load type, and the release of P may depend on the binding and the molar ratios of C, Fe, and P in the terrestrial material.
The P release in the units with high C/soil ratio exceeded the amount of P extracted by buffered dithionite that has earlier been used to estimate the mobile P in sulfidic sediments. The difference may be explained by the mineralization of organic P present in significant amounts in the test soil, and also by the duration of the anoxic period—15 min in the dithionite extraction as opposed to over two years in the incubations. During the incubation, microbes and sulfides may be able to liberate P also from the less readily accessible metal oxides (such as P in interiors of Fe oxides). In addition, the organic anions, such as acetate, may have competed with PO
4 for adsorption sites (on organic anion competition, see, e.g., Froelich
1988), and the increased pH may have promoted P desorption via anion exchange with OH
− and through the decreasing positive charge of oxide surfaces (Hingston et al.
1967; Caraco et al.
1989; Hawke et al.
1989; Hartikainen and Yli-Halla
1996). The addition of acetate was likely to elevate pH both directly due to its weak basicity (pKb ≈9.3) and indirectly via SO
4 and Fe reductions (Ben-Yaakov
1973; Lamers et al.
1998; Soetaert et al.
2007; Boudreau and Canfield
1993).
Our experimental approach may give insight into the mobility of soil-bound P under conditions where P-release is controlled by anaerobic microbial processes, serving as an additional method besides algal assays and chemical extractions. The long incubation may in part reflect realistic dynamics of microbial processes affecting the P release in brackish and marine sediments. In future tests, a shorter incubation time may be compensated by increasing the temperature. However, particulate organic P may behave differently compared to P bound to reactive Fe oxides. Studies from shallow coastal sediments have shown that the efficient reduction of Fe and release of P can be measured within weeks (Jensen et al.
1995; Kristiansen et al.
2002), but that the mineralization of organic P may take years in anoxic sediments (Ahlgren et al.
2006). Therefore, a short incubation time may reveal better the release of Fe bound P than that of organic P.