Introduction

In unmanaged and remote ecosystems, nutrient allocation within plants and soil microbial communities have developed into long-lasting climatically driven equilibria, which have been repeatedly re-established after natural ecosystem disturbances during the pre-industrial period (e.g., Vitousek 2004). In the pre-industrial period, atmospheric deposition of reactive nitrogen (N) \(\left( {{\text{Nr}} = {\text{NH}}_{ 4} {\text {-}} {\text{N}} + {\text{NO}}_{ 3} {\text {-}} {\text{N}}} \right)\) and sulphate (\({{\text{SO}}_{ 4}}^{ 2-}\)) were low and stable, and it is likely that all available N forms in soils were immobilized by microbial and plant biomass. The soil C:N ratio decreased only slowly because the net ecosystem N gain from N2 fixation and low Nr deposition was to a large extent balanced by terrestrial exports of recalcitrant dissolved organic N (DON), whereas NO3 − leaching was probably small, similar to the current situation in N-limited terrestrial ecosystems (Perakis and Hedin 2002; Vitousek 2004). Since the beginning of the industrial period, however, soil biogeochemistry and fertility of ecosystems adapted to these steady state conditions have begun to change due to rapidly increasing SO2, NO x , and NH3 emissions from energy and food production, and consequent \({{\text{SO}}_{ 4}}^{ 2-}\) and Nr deposition (Reuss and Johnson 1986). A new steady-state has not yet been established during the industrial period, because S and Nr emission/deposition rates have not stabilized and the ecosystem response is slow (Smith et al. 2011; Kopáček and Posch 2011).

As Nr is deposited on terrestrial ecosystems, ecosystem requirements for additional N are reduced. An early ecosystem response is a reduction in rates of more energy demanding N2 fixation (DeLuca et al. 2008). A major subsequent symptom accompanying elevated Nr inputs is the alleviation of the natural N limitation of unmanaged ecosystems, triggering a cascade of changes in both terrestrial and aquatic ecosystems (Galloway et al. 2004). Consequences of long-term N enrichment include changes in ecosystem biodiversity (Emmett 2007; Bobbink et al. 2010), microbial composition (Treseder 2004; Högberg et al. 2007a, b), internal N cycling (Aber et al. 1998), and NO3 − leaching (Stoddard 1994). These responses are all biologically mediated, and are often assumed to be associated with the general role of N as a limiting nutrient in terrestrial ecosystems.

In this synthesis we suggest that changes in the ability of the terrestrial ecosystem to retain N are due not just to the role of N as a nutrient, but also to mechanisms associated with NO3 − as a strong acid anion in soil solution, and as an electron acceptor in microbial energy metabolism. Elevated Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) inputs change the physico-chemical equilibria and redox conditions of soil solutions, and thereby affect C cycling and pools of bioavailable dissolved organic carbon (DOC) for soil microorganisms (Gauci et al. 2004; Alewell et al. 2008). The microbial communities in the ecosystem thus may transit from N limitation (C saturation) to C limitation (N saturation), and the ecosystem may start to leach NO3 − in runoff. We synthesise information on trends in soil and water chemical composition in natural areas during industrial period, links between NO3 − leaching and terrestrial C cycling, N saturation hypotheses, and microbial and chemical factors affecting DOC bioavailability in soils. Then, we estimate the potential effect of \({{\text{SO}}_{ 4}}^{ 2-}\) and NO3 − as electron acceptors in microbial processes on soil pools of bioavailable DOC. Finally, we link N saturation to C limitation of soil microbial processes in a new conceptual scheme combining the N, C, and S cycles.

Observed trends in Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition and surface water chemistry

Elevated emissions and deposition of S and N compounds in Europe and eastern North America since the early 1800s caused significant acidification of soils and waters in many acid sensitive areas. Chemical changes included elevated leaching of strong acid anions (mostly \({{\text{SO}}_{ 4}}^{ 2-}\) and NO3 −) and the acid cations H+ and ionic aluminium (Aln+); reduced soil pools of base cations (calcium, magnesium, potassium, and sodium); their partial replacement by H+ and Aln+; increased ionic strength of soil solutions; and changes in N cycling (Reuss and Johnson 1986; Aber et al. 1989, 1998). Disentangling the individual effects of Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition on terrestrial N cycling is difficult, because their deposition fluxes are strongly spatially and temporally correlated in many regions. This situation has changed somewhat in Europe and North America since the 1980s, when the impacted ecosystems started to recover from acidification due to efficient S emission controls (Smith et al. 2011). The resulting reduction in S deposition has caused a general decline in surface water \({{\text{SO}}_{ 4}}^{ 2-}\) in these areas (Stoddard et al. 1999; Evans et al. 2001). Many soils have switched from being a sink to a source of S (Driscoll et al. 1998; Prechtel et al. 2001). Fluxes of \({{\text{SO}}_{ 4}}^{ 2-}\) in runoff from catchments with thin soils have decreased rapidly (Evans et al. 2001) in comparison to catchments with thick soils and high \({{\text{SO}}_{ 4}}^{ 2-}\) retention capacity (Alewell 2001).

In contrast to S, global Nr emissions continue to increase, and have been effectively reduced only in Europe (Kopáček and Posch 2011). The response of NO3 − leaching to Nr deposition is less straightforward than in the case of S. The maximum (~50 %) decline in Nr emission and deposition occurred in the central European region (Kopáček et al. 2012), where it resulted in a prompt decrease in NO3 − leaching (Veselý et al. 2002; Kopáček et al. 2005). Other regions of Europe with more modest declines in Nr deposition have resulted in either no change or only small decreases in NO3 − leaching (e.g. the United Kingdom; Monteith et al. 2012). In parts of northeastern North America, NO3 − leaching has decreased despite unchanged Nr deposition (e.g., Driscoll et al. 2003). Interestingly, some of the observed decreases in NO3 − leaching were more pronounced than the decline in Nr deposition (Veselý et al. 2002; Driscoll et al. 2003; Oulehle et al. 2008). These observations are from areas rapidly recovering from acidification due to reduced \({{\text{SO}}_{ 4}}^{ 2-}\) deposition, and many of the same regions also exhibit increasing DOC leaching (Monteith et al. 2007). Goodale et al. (2005) hypothesized that the decline in NO3 − leaching could result from increased N immobilization and/or denitrification driven by increased availability of terrestrial DOC. Monteith et al. (2007) showed that increasing DOC concentrations coincide in time and space with decreasing \({{\text{SO}}_{ 4}}^{ 2-}\) deposition. The apparent co-occurrence of these trends at numerous sites suggests that N, C, and S should be considered together.

Further strong evidence of close links between the N and C cycles come from spatial and seasonal variability of NO3 − leaching, and its apparent links to pools and concentrations of organic C. (1) The characteristics of NO3 − losses from natural terrestrial ecosystems (rate of the response, proportion of the Nr input, and seasonal trends) are not uniform and reflect differences in vegetation, pools and composition of soil, landscape characteristics, hydrology (water residence time in soil), and land use history (e.g., Dise and Wright 1995; Gundersen et al. 1998; Goodale et al. 2000; Kopáček et al. 2005). Among all factors, the most important characteristics determining the capacity of these systems to accumulate N in the long term are related to the pool size of organic C (Fig. 1a; Evans et al. 2006), i.e., the ability of the soil to retain N through a stoichiometric sink (Curtis et al. 2011). Systems with high pools of organic C have usually higher N accumulation capacities and lose less NO3 − than C-poor soils (Evans et al. 2006). (2) Tight inverse relationships between NO3 − and DOC concentrations are typically observed in mountain lakes and streams both on spatial and temporal scales (Fig. 1b; Goodale et al. 2005; Helliwell et al. 2007; Camarero et al. 2009). Similar negative nonlinear correlations between NO3 − and DOC concentrations exist along a hydrologic continuum from soils, through freshwater systems to the open ocean, indicating that resource stoichiometry strongly couples DOC and NO3 − cycling (Taylor and Townsend 2010). (3) Straightforward effects of DOC availability on NO3 − immobilization have been demonstrated by experimental additions of bioavailable DOC sources to soils and near-stream environments, which stimulated NO3 − assimilation and denitrification (Azam et al. 1988; Hedin et al. 1998), as well as to streams, which stimulated NO3 − assimilation and bacterial growth (Bernhardt and Likens 2002).

Fig. 1
figure 1

Concentrations of DOC and \({{\text{NO}}_{ 3}}^{-}\) in the Tatra Mountain lakes: a Data are divided according to the soil carbon pools in catchments (F > 80 % is deep forest soil covering >80 % area; M > 70 %, M = 30–70 %, and M < 30 % are shallow soils of dry alpine meadows covering >70 %, 30–70 %, and <30 % area, respectively). b \({{\text{NO}}_{ 3}}^{-}\) versus DOC concentrations. Data are derived from Kopáček et al. (2000)

Full size image

Considering organic C and N cycles together seems to be necessary to identify factors governing ecosystem capacity to accumulate N (Gärdenäs et al. 2011), because heterotrophic microbes, relying on bioavailable dissolved organic C, are responsible for most N transformations in soils (e.g., Demoling et al. 2008; Barton and Northup 2011). This implies that environmental changes affecting organic C solubility and bioavailability in soils could also affect N cycling. Several processes have the potential to induce a shortage in bioavailable DOC in soils. Two of them—decreased allocation of plant photosynthates to soils, associated with elevated Nr availability, and chemical suppression of DOC, resulting from soil acidification—have been well established in the literature. In contrast, a role of \({{\text{SO}}_{ 4}}^{ 2-}\) and NO3 − as electron acceptors in anaerobic respiration has only been considered in a few studies (e.g., Hedin et al. 1998; Alewell et al. 2008) as a factor significantly increasing DOC mineralization and, hence, reducing pools of bioavailable DOC. These studies show that diffusion of \({{\text{SO}}_{ 4}}^{ 2-}\) and NO3 − to anaerobic soil micro-sites can increase mineralization of bioavailable DOC and further reduce its availability for N immobilizing microorganisms.

Established mechanisms for Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) impacts on biogeochemical processes

Conceptual models of N saturation

The original conceptual model of N saturation (Aber et al. 1989) postulates a series of changes that occur in N-limited terrestrial ecosystems during long-lasting exposure to elevated N inputs. The ecosystem is assumed to progress through a sequence of consecutive changes: N enrichment of plant tissues and litter; elevated soil N pools; stimulated N mineralization and nitrification; and finally NO3 − leaching. A number of fertilization experiments, however, showed high and rapid retention of Nr in soil organic matter (SOM) without first passing through the aboveground plants (Aber et al. 1998). This led to the conclusion that mycorrhizal fungi (and not vegetation directly) are the primary Nr sink (Aber et al. 1998) and that suppression of mycorrhizal functioning at elevated Nr availability (see next section) may be followed by a change in fungi:bacteria ratio (Tietema 1998), increased \({{\text{NH}}_{ 4}}^{ + }\) availability for nitrifiers, and NO3 − mobility and leaching.

The hypothesis of Aber et al. (1998) assumes that N is accumulated in the vegetation and soil until retention capacity of these sinks is reached. The ecosystem response to elevated Nr inputs should thus be relatively slow, with a relatively long phase between the beginning of Nr input and onset of NO3 − leaching. However, most experimental studies with either elevated or reduced Nr inputs to terrestrial ecosystems (e.g., Emmett 2007; Moldan and Wright 2011) and stable isotope experiments (for review see Curtis et al. 2011) have shown that NO3 − leaching responds rapidly to Nr inputs. Lovett and Goodale (2011) have assumed in a new conceptual model that the added N can flow simultaneously to all sinks in the system, and that its fate and losses depend on the size of the sinks and the factors that control them. N losses from the system may thus be the results of either capacity or kinetic saturation. Capacity N saturation occurs when the N sinks in the vegetation and soil are filled up, while kinetic N saturation occurs when a sink accumulates N at the rate lower than the rate of N input (Lovett and Goodale 2011). This suggests that NO3 − leaching can occur even if the N retention capacity has not yet been saturated, as is in fact observed at many sites.

This conceptual model is in agreement with a general picture of N cycling in terrestrial ecosystems formulated by Schimel and Bennett (2004). Their scheme incorporates three important assumptions for N cycling in soils: (1) The primary N source for both plants and microbes in N-limited ecosystems is low molecular weight organic monomers which result from the depolymerization of N-containing polymers by microbial extracellular enzymes. (2) Plants take up first organic N, then \({{\text{NH}}_{ 4}}^{ + }\) and finally NO3 −, depending on the relative abundance of these in soil (Nordin et al. 2001). (3) Soil micro-sites with different availability of substrate (electron donors) and energy (electron acceptors), redox conditions, and N mineralization and immobilization rates provide large small-scale heterogeneity and gradients of soil properties (e.g., Schimel and Bennett 2004; Alewell et al. 2006, 2008). This heterogeneity is followed by the heterogeneous distribution of microorganisms, connected with resources by water (diffusion and convection) (e.g., Ekschmitt et al. 2008). Altogether, this means that all microbial processes can occur simultaneously in individual micro-sites within a relatively small distance in the soil, provided they are thermodynamically and physiologically possible under the actual substrate and energy availabilities, and the environmental conditions (e.g., moisture, temperature, pH, ambient toxicity, etc.).

Based on these assumptions, Schimel and Bennett (2004) described four gradual stages of N cycling in terrestrial ecosystems along a gradient of N availability for microbial and plant communities: (1) In extremely N-limited ecosystems, production of inorganic N forms is negligible, with low litter input and slow decomposition rate, and plants and microbes compete primarily for N at the organic-monomer stage. (2) In moderately N-limited ecosystems, microbes begin to mineralize N particularly in N-rich micro-sites. The \({{\text{NH}}_{ 4}}^{ + }\) produced diffuses away and is mostly used by the N-limited microbes in the N-poor sites, with negligible amounts remaining for nitrifiers and plants. (3) As N availability increases further, mineralization rates and \({{\text{NH}}_{ 4}}^{ + }\) diffusion from N-rich micro-sites increase, alleviating the competition between plants and microbes for N. \({{\text{NH}}_{ 4}}^{ + }\) becomes the dominant available soil N pool. Plants can utilize \({{\text{NH}}_{ 4}}^{ + }\) remaining in the system after microbial demands are met, which can promote net primary production and increase N concentrations in plant tissue. The supply of the surplus \({{\text{NH}}_{ 4}}^{ + }\) to nitrifiers is still low and \({{\text{NO}}_{ 3}}^{-}\) is localized only in N-rich micro-sites. (4) Finally, at high N availability, more \({{\text{NH}}_{ 4}}^{ + }\) remains for nitrifiers, the system becomes progressively more \({{\text{NO}}_{ 3}}^{-}\) dominated, and plant preferences for \({{\text{NO}}_{ 3}}^{-}\) increase with soil \({{\text{NO}}_{ 3}}^{-}\) concentrations.

The N saturation (Lovett and Goodale 2011) and the general N cycling (Schimel and Bennett 2004) concepts assume that all processes of N transformation occur in the soil simultaneously, and can explain the rapidity of ecosystem responses to added Nr. Terrestrial N losses (amounts and forms) depend on the stage of N limitation and the N retention capacity of the ecosystem prior to the elevated Nr input. If the system is extremely N-limited (i.e., in stage 1 of the concept by Schimel and Bennett 2004) both the amount and duration of Nr input necessary to cause \({{\text{NO}}_{ 3}}^{-}\) leaching should be higher than in systems, which have reached stage 3. Stage 3 is typical for moderately fertile temperate forests (Pastor et al. 1984).

Effects of Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) on soil microbial processes and C cycling

Fungi dominate decomposition of SOM in the litter layer, where oxygen availability enables enzymatic reactions involved in the utilization of phenolic substrates (including lignin). These reactions exude substantial amounts of soluble low molecular weight organic matter (LMW OM), which serves, together with plant photosynthates exuded by roots, as a C and energy source for other saprotrophic organisms (Ekschmitt et al. 2008; Janssens et al. 2010). Fungal mycelia can transfer the necessary C to nutrient-rich micro-sites or nutrients back to C-rich micro-sites, and only a small fraction of nutrients is released in inorganic form and lost (Lindahl et al. 2002; Boberg et al. 2010). Hyphal fungi are thus more effective in competition for nutrients in nutrient-poor soils than bacteria, which generally rely on solute transport. This causes bacteria to be more dependent on water content and the presence of soluble compounds in soil than fungi (Ekschmitt et al. 2008; Schmidt et al. 2011).

Nr enrichment reduces the fungal contribution to the soil microbial biomass, decreases the fungi:bacteria ratio, and affects the activity of the rhizosphere, particularly of mycorrhizal root symbionts (Treseder 2004; Högberg et al. 2007b; Boyle et al. 2008; Ramirez et al. 2010; Zak et al. 2011). Mycorrhizal immobilization of N is driven by plant photosynthesis. Mycorrhizal fungi use photosynthates from the host plants as the dominant carbon source, strongly remove available N in soils, and in N-limited ecosystems supply plants with most of the necessary N in the form of organic N monomers (Högberg et al. 2007a). Under conditions of elevated Nr availability, plants can utilize mineral N (unused by or inaccessible to microorganisms) and their dependence on the mycorrhizally-derived organic N monomers decreases (Johnson et al. 1997; Schimel and Bennett 2004). The reduced need for symbiotic nutrient exchange results in the promotion of above-ground biomass production at the expense of below-ground carbon allocation (Treseder 2004; Janssens et al. 2010). The elevated Nr supply to soils thus directly regulates supply of energy-rich plant exudates to mycorrhizal fungal symbionts, and reduces the biomass of mycorrhizal fungi (Högberg et al. 2007b).

In addition to this mechanism, mycorrhizal biomass can also be reduced due to decreases in fine root biomass, which in turn results from two effects of Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition: (1) As N limitation decreases, plants reduce their fine-root biomass (Clemensson-Lindell and Persson 1995; Nadelhoffer 2000). (2) The toxic Aln+ forms produced by soil acidification inhibit plant root growth and branching (Göransson and Eldhuset 1991; Heim et al. 2000). The total area of fine roots (habitats for mycorrhizal colonization), and consequently mycorrhizal biomass, decreases (Demoling et al. 2008).

The changes in functioning of the soil microbial community, resulting from decreasing fungi:bacteria ratio, also affect litter decomposition and mineralization of soil N pools. (1) Decomposition rates of easily-degradable organic C (e.g., leaves of most broadleaf trees and herbs, fine roots) increase, while those of recalcitrant organic C, like conifer needles or woody biomass, decline (Knorr et al. 2005; Janssens et al. 2010) due to low activity of the lignin-degrading enzymes produced mainly by fungi (Waldrop et al. 2004; Frey et al. 2004). (2) Provided that fungi dominate the decomposer community (e.g., in natural forest ecosystems and grasslands), net N mineralization and leaching are usually low due to the active fungal C translocation and lower N demand (Boberg et al. 2010; De Vries et al. 2006). The reduced fungi:bacteria ratios at higher N availability in soils result in higher N mineralization and N losses (De Vries et al. 2006; Högberg et al. 2007a). We can deduce that fungi play a dominant role in N cycling in N-limited ecosystems, i.e., in the first two stages of Schimel and Bennett (2004). In contrast, bacteria are distributed heterogeneously in small-scale habitats, physically connected by water, and “wait” for DOC and nutrients to be transported to their surfaces by diffusion (Ekschmitt et al. 2008), or along preferential flowpaths, where hotspots of microbial activity have been inferred from high rates of N uptake associated with macropore flows during 15N tracer studies (Hagedorn et al. 1999; Evans et al. 2008). With increasing N availability in soil solutions, more N is transported to bacteria and their role in N cycling increases.

Elevated Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition thus affects both N and C cycling in soils via changes in composition of soil microbial communities. On the basis of the above separately demonstrated effects of Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) on the C cycle, we propose the following sequence of possible changes in pools and fluxes of soil organic C: Compared to the preindustrial period (Fig. 2a), the biomass of fine roots and mycorrhizal fungi decreases, their supply with photosynthates decreases, and fungi:bacteria ratio in soils decreases (Fig. 2b), which continuously changes the original soil ecosystem functioning. Decreasing fungal biomass results in reduced litter decomposition (higher litter accumulation), decreased production of LMW OM for other decomposers, and reduced fungal SOM production and CO2 efflux. These changes in fungal C cycling affect N cycling (N distribution between fungi and bacteria), and consequently, bacterial C cycling. As fungi:bacteria ratio decreases, more N is turned over by bacteria. Originally N-limited free bacteria can utilize available DOC in the C-rich micro-sites (not affected by lower LMW OM production) and transform it to the bacterial SOM. This leads to a more rapidly decreasing soil C:N ratio (because bacterial C:N is lower than fungal C:N) and to elevated CO2 efflux by bacteria (Fig. 2b).

Fig. 2
figure 2

Conceptual scheme of C cycling in soils a under natural conditions, b after decreased allocation of plant photosynthates due to elevated Nr deposition and soil acidification, c after suppression of DOC solubility and bioavailability as soil acidifies, and d after elevated microbial DOC mineralization due to elevated concentrations of electron acceptors (\({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\)). Numbers identify specific carbon fluxes changing due to elevated Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) inputs: 1 Litter decomposition by fungi, 2 plant exudation of photosynthates, 3 fungal respiration, 4 production of soil organic matter (SOM) in fungal biomass, 5 fungal production of DOC, including low molecular weight organic matter available for bacteria, 6 bacterial uptake of DOC, 7 DOC production by fermentation, 8 production of SOM in bacterial biomass, 9 bacterial CO2 production due to elevated availability of electron acceptors, and 10 net change in DOC solubility and bioavailability associated with acidification-induced chemical suppression. Red arrows and boxes show fluxes and pools changed compared to the previous stage (a through d), but the processes may occur together. Elevated Nr availability and soil acidity decrease plant exudation, biomass of mycorrhizal fungi, litter decomposition, and fungi:bacteria ratio, affecting fluxes (1–5). Soil acidification decreases DOC solubility (flux 10). Elevated availability of Nr and electron acceptors increases the microbial DOC mineralization and decreases fermentation, affecting fluxes (6–9). All these changes contribute to lower concentrations of bioavailable DOC in soils and lower DOC leaching

Full size image

Effects of Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) on chemical suppression of DOC solubility and bioavailability

The negative relationship between DOC concentration in soil solutions and soil acidification by acid rain has been indicated by numerous field and laboratory observations (e.g., Kalbitz et al. 2000; Clark et al. 2006; Hruška et al. 2009; Ekström et al. 2011; Evans et al. 2012), as well as modelling (Tipping and Hurley 1998). The acidity effect on DOC solubility acts as a partial buffer against acidification (Evans et al. 2008). The chemical effect is caused by elevated fluxes of strong acid anions, which reduce soil water concentrations of DOC. As soil water acidifies and ionic strength increases, DOC solubility is reduced by a combined effect of protonation of functional groups and coagulation of individual molecules to insoluble aggregates (Stumm 1992; Nierop et al. 2002). The negative effects of acidity on DOC solubility in organic soils have recently been demonstrated in field manipulation experiments and shown to be associated with pH-dependent changes in the dissociation of weak organic acids (Ekström et al. 2011; Evans et al. 2012).

Progressive acidification also increases soil ability to adsorb anions (including organic acids) because the total positive surface charge of Al and Fe oxyhydroxides increases as pH declines to ~3–3.5 (Stumm 1992). DOC sorption to mineral surfaces thus further reduces its mobility (Kalbitz et al. 2000, 2005) and possible diffusion to bacteria. Moreover, aggregation and adsorption to the solid phase protects DOC against enzymatic depolymerisation to low molecular weight molecules (Scheel et al. 2007), which can be transported across cell membranes. Additionally, the soil pH itself can affect microbial enzyme activity; phenol oxidase and peroxidase activities generally decrease as soil pH declines (Toberman et al. 2008; Sinsabaugh 2010).

Both laboratory and field experiments have documented the suppressive effect of soil acidification not only on DOC solubility, but also on its bioavailability. Scheel et al. (2007) showed that organic particulates, resulting from precipitation of soil DOC with Aln+, became substantially stabilized against microbial decay, so that its mineralization was an order of magnitude lower than that of the original DOC. Moreover, recalcitrance of these particles increased as the pH at which the precipitate formed decreased. Similarly, Mulder et al. (2001) observed that artificial addition of Aln+ (as AlCl3) to the forest floor caused a pronounced decrease in DOC leaching and a 30–40 % reduction of decomposition rate of SOM. Hence, we can assume that soil acidification caused by atmospheric deposition of \({{\text{SO}}_{ 4}}^{ 2-}\) and Nr contributes to DOC immobilization in soils, as well as to its lower bioavailability for bacteria (Fig. 2c).

Effects of \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) as electron acceptors on soil pools of biovailable DOC

Microbial metabolism depends on the use of electron donors and electron acceptors in redox reactions that generate energy. Generally, bioavailable DOC dominates among electron donors and O2 dominates among electron acceptors in soils. The role of \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) as electron acceptors for microbial processes is important in anoxic soils, as well as in micro-sites with periodic or negligible O2 in otherwise relatively well aerated soils (e.g., Hedin et al. 1998). Hence, we might expect that more bioavailable DOC has been mineralized in soils after their enrichment with electron acceptors by elevated deposition of \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) (Fig. 2d), as flows from studies by Hedin et al. (1998) and Alewell et al. (2008). The question, however, is to what extent this process could increase C mineralization and decrease the overall pool of bioavailable DOC in soils, and if this change could play an ecologically significant role? This undoubtedly depends on the \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) fluxes and their microbial reductions. The efficiency of the latter process generally reflects the physico-chemical properties of soils, pools of bioavailable DOC, water residence time, and hydraulic connectivity, which enables diffusion of electron acceptors and DOC to microbes (Hedin et al. 1998; Alewell et al. 2006, 2008). Nitrate is preferentially used over \({{\text{SO}}_{ 4}}^{ 2-}\) as an electron acceptor during microbial metabolism in relatively DOC-poor systems, with insignificant \({{\text{SO}}_{ 4}}^{ 2-}\) reduction unless \({{\text{NO}}_{ 3}} {\text {-}} {\text{N}}\) is first reduced to concentrations <0.2 mg L−1 (Hedin et al. 1998). However, Alewell et al. (2008) showed that these processes (even though attributed to different redox conditions) can occur simultaneously on a soil micro-scale under DOC-rich conditions.

Microbial sulphate reduction

Stable isotope (δ34S) studies and budget calculations suggest that biological S turnover in forest soils is an important contributor to \({{\text{SO}}_{ 4}}^{ 2-}\) leaching (Torssander and Mörth 1998; Novák et al. 2000). Organic cycling of atmospherically-derived S plays an important role in S polluted coniferous forests in central Europe (Novák et al. 2000, 2005) and probably also in Scandinavia (Mörth et al. 2005) and the United Kingdom (Bartlett et al. 2009). Sites near industrial S emission sources have higher S pools in the forest floor than unpolluted sites; 75–95 % of this soil S is present as organically-bound S, and this proportion increases with atmospheric \({{\text{SO}}_{ 4}}^{ 2-}\) loads (Novák et al. 2005). Sulphur isotope and mass budget studies in highly \({{\text{SO}}_{ 4}}^{ 2-}\) polluted areas indicated that 30–70 % of soil and stream water \({{\text{SO}}_{ 4}}^{ 2-}\) was organically cycled (Novák et al. 2000, 2003), i.e., reduced, eventually incorporated in biomass, and oxidized back to \({{\text{SO}}_{ 4}}^{ 2-}\). Moreover, Novák et al. (2004) observed that 95 % of the deposited 35S was cycled or retained in a severely polluted spruce forest soils for several months and H2S volatilization represented ~20 % of atmospheric \({{\text{SO}}_{ 4}}^{ 2-}\) input in summer months. This suggests that the proportion of the organically cycled atmospheric \({{\text{SO}}_{ 4}}^{ 2-}\) could be even higher than that observed in the surface waters (Novák et al. 2000, 2003).

Dissimilatory \({{\text{SO}}_{ 4}}^{ 2-}\) reduction, when S serves as the terminal acceptor of electrons released in microbial oxidation of bioavailable DOC, can be described by:

$$2{\text{CH}}_{2} {\text{O}} + {\text{SO}}_{4}^{2 - } + 2{\text{H}}^{ + } = {\text{H}}_{2} {\text{S}} + 2{\text{CO}}_{2} + 2{\text{H}}_{2} {\text{O}}$$
(1)

This reaction requires 1 mol of S for oxidation of 2 mols of C in the form of glucose. The reduced S may be further used by S-oxidizing bacteria (and oxidized back to \({{\text{SO}}_{ 4}}^{ 2-}\)) as an electron donor (energy source) after aeration of the originally anoxic micro-site, e.g., after its erosion or during dry periods (Clark et al. 2006), or in the presence of \({{\text{NO}}_{ 3}}^{-}\) (An and Gardner 2002). This means that the deposited \({{\text{SO}}_{ 4}}^{ 2-}\) can be repeatedly recycled in the reduction–oxidation process.

Undoubtedly, the extent of dissimilatory \({{\text{SO}}_{ 4}}^{ 2-}\) reduction may differ between sites, reflecting differences in soil physico-chemical properties. Generally low \({{\text{SO}}_{ 4}}^{ 2-}\) reduction can be expected in C poor and well aerated systems (such as sandy soils), with a low capacity to reduce \({{\text{SO}}_{ 4}}^{ 2-}\). In very C-rich systems (such as peatlands), the effect on DOC concentrations is probably difficult to detect, because the resulting change in DOC leaching may be masked by the quantitatively more important chemical effects on SOM solubility (Evans et al. 2012). The pH increase accompanying \({{\text{SO}}_{ 4}}^{ 2-}\) reduction (Eq. 1) may increase SOM solubility and mitigate the decrease in DOC concentrations associated with the mineralization of bioavailable DOC. The relatively highest effects, however, can be expected in mountain forest soils, with intermediate-size organic C pools, high richness of anoxic micro-sites, and usually high wet and dry SO4 2− deposition, e.g., in the spruce forest areas where the studies by Novák et al. (2000, 2003) were performed.

For simplicity of further estimates of the ecological importance of S cycling on soil DOC, we assume that (i) the repeated use of SO4 2− is not quantitatively important, (ii) 50 % of the deposited SO4 2– undergoes a microbial reduction/oxidation cycle (the middle of the range observed in the central European forest soils by Novák et al. (2000, 2003), and (iii) effects of other S transformations (assimilatory reduction and oxidation of the reduced S forms) on the C cycle can be neglected. Under such conditions, we can expect that 1 mol of bioavailable DOC is oxidized per mol of deposited SO4 2−.

Microbial nitrate reduction

Curtis et al. (2011) have evaluated results of numerous isotopic studies employing natural abundances of 15N and 18O in precipitation and surface water \({{\text{NO}}_{ 3}}^{-}\), and conclude that the majority of \({{\text{NO}}_{ 3}}^{-}\) leaching to waters has been microbially transformed (i.e., reduced to \({{\text{NH}}_{ 4}}^{ + }\) and then oxidized again via nitrification) even in C-poor systems. We can consider two major pathways:

  1. (1)

    \({{\text{NO}}_{ 3}}^{-}\) assimilation by microbes (Eq. 2; Mara and Horan 2003):

    $${\text{CH}}_{2} {\text{O}} + 0.2{\text{O}}_{2} + 0.114{{\text{NO}}_{3}}^{ - } + 0.114{\text{H}}^{ + } = 0.114{\text{C}}_{5} {\text{H}}_{7} {\text{O}}_{2} {\text{N}} + 0.429{\text{O}}_{2} + 0.657{\text{H}}_{2} {\text{O}}$$
    (2)

    followed by mineralization of organically-bound N and nitrification of the liberated \({{\text{NH}}_{ 4}}^{ + }\):

    $${{\text{NH}}_{4}}^{ + } + 2{\text{O}}_{2} = {{\text{NO}}_{3}}^{ - } + 2{\text{H}}^{ + } + {\text{H}}_{2} {\text{O}}$$
    (3)
  2. (2)

    Dissimilatory nitrate reduction to ammonium (DNRA):

    $$2{\text{CH}}_{2} {\text{O}} + {\text{NO}}_{3}^{ - } + 2{\text{H}}^{ + } = {\text{NH}}_{4}^{ + } + 2{\text{CO}}_{2} + {\text{H}}_{2} {\text{O}}$$
    (4)

    followed by nitrification (Eq. 3).

Another important process involved in \({{\text{NO}}_{ 3}}^{-}\) cycling is denitrification:

$$5{\text{CH}}_{2} {\kern 1pt} {\text{O}} + 4{\text{NO}}_{3}^{ - } + 4{\text{H}}^{ + } = 2{\text{N}}_{2} + 5{\text{CO}}_{2} + 7{\text{H}}_{2} {\text{O}}$$
(5)

Similarly to \({{\text{SO}}_{ 4}}^{ 2-}\), all these microbial \({{\text{NO}}_{ 3}}^{-}\) reductions require energy gained from oxidation of bioavailable DOC. Both \({{\text{NO}}_{ 3}}^{-}\) assimilation and DNRA involve the reduction of N from an oxidation state of +V to −III, but they differ in the necessary DOC demand. Reaction (2) includes both DOC needed for assimilative N reduction and the cell growth and maintenance. The net DOC amount for \({{\text{NO}}_{ 3}}^{-}\) reduction is thus lower than suggested by the stoichiometry of Eq. (2) and can be estimated as the difference between C demand for assimilation of 1 mol N in the form of \({{\text{NO}}_{ 3}}^{-}\) and \({{\text{NH}}_{ 4}}^{ + }\), which is 3.76 and 3.02 mol CH2O, respectively (Mara and Horan 2003). The difference of 0.74 mol C mol−1 N represents the minimum C (and energy) demand for electron transfer during \({{\text{NO}}_{ 3}}^{-}\) reduction. However, the energy transfers by the organisms are not 100 % efficient. Bacteria promote electron transfer using adenosine triphosphate (ATP) and the efficiency of the energy transfer in ATP formation is assumed at ~55 % (Mara and Horan 2003). Consequently, the amount of bioavailable DOC required for bacterial assimilation of \({{\text{NO}}_{ 3}}^{-}\) can be estimated at ~1.4 (=0.74/0.55) mol mol−1 N. A similar estimate can be obtained via an alternative calculation of the energy required for the transfer of eight electrons during \({{\text{NO}}_{ 3}}^{-}\) reduction. Ideally, three ATPs are used per electron pair transferred to the terminal electron acceptor (i.e., 1.5 ATP electron−1; Mara and Horan 2003). The free energy released by ATP hydrolysis is 31.8 kJ mol−1 (Barton and Northup 2011). Hence, the minimum amount of energy for \({{\text{NO}}_{ 3}}^{-}\) reduction during N assimilation is 382 kJ mol−1 (= 8 × 1.5 × 31.8). Again assuming 55 % efficiency of the energy transfer in ATP formation, the total amount of necessary energy is ~694 kJ mol−1 N (= 382/0.55), which is 1.38 times more energy than is released by aerobic respiration of 1 mol CH2O in the form of glucose (501 kJ mol−1 C; Hedin et al. 1998). We assume that similar DOC demand is associated also with fungal \({{\text{NO}}_{ 3}}^{-}\) assimilation. However, total amount of bioavailable DOC removed from the system during assimilation of 1 mol \({{\text{NO}}_{ 3}}^{-}\) is higher in the case of fungi than bacteria due to higher C:N ratios of fungal biomass (Šantrůčková et al. 2009).

The amount of DOC consumed by \({{\text{NO}}_{ 3}}^{-}\) reduction during DNRA is more straightforward (Eq. 4) than \({{\text{NO}}_{ 3}}^{-}\) assimilation. Based on stoichiometry, two mols of bioavailable DOC are oxidized per each mol of the \({{\text{NO}}_{ 3}}^{-}\) reduced in DNRA.

The DOC demand of denitrification depends on the terminal oxidation state of N, i.e., proportion of the produced N2, N2O, and NO, which is pH dependent. For simplicity we assume that most of \({{\text{NO}}_{ 3}}^{-}\) is denitrified to N2. Then, 1 mol of \({{\text{NO}}_{ 3}}^{-}\) enables oxidation of 1.25 mol of DOC (Eq. 5).

Consumption of bioavailable DOC due to microbial \({{\text{NO}}_{ 3}}^{-}\) and \({{\text{SO}}_{ 4}}^{ 2-}\) reduction

The total amount of DOC removed from soil solutions by microbial \({{\text{NO}}_{ 3}}^{-}\) reduction depends on the relative proportion of denitrification, assimilation, and DNRA in the \({{\text{NO}}_{ 3}}^{-}\) cycling in soils. Gundersen et al. (1998) reported denitrification losses of 3–4 kg N ha−1 year−1 at \({{\text{NO}}_{ 3}}^{-}\) doses of 8–17 kg N ha−1 year−1 to coniferous forest stands (i.e., 20–50 % of the input). This is similar to global estimates of 10–40 % (average 25 %) of the N input to terrestrial systems, which is assumed to be denitrified (Galloway et al. 2004). Consequently, we may assume that ~25 % of the deposited \({{\text{NO}}_{ 3}}^{-}\) is denitrified and the remaining 75 % is reduced either via \({{\text{NO}}_{ 3}}^{-}\) assimilation or DNRA, and then further microbially processed. The potential for significant DNRA exists in most soils, but is more important under anoxic conditions in micro-sites rich in low molecular weight C sources, and at soil bioavailable DOC to \({{\text{NO}}_{ 3}}^{-}\) ratios >12 (Yin et al. 1998; Rütting et al. 2011). The relative proportion of DNRA in total gross \({{\text{NO}}_{ 3}}^{-}\) consumption ranges from negligible (<1 %) to dominant (~99 %) values in temperate forests and grasslands, depending on annual temperature and soil moisture (Rütting et al. 2011). For an approximate estimate of the effect of the microbial \({{\text{NO}}_{ 3}}^{-}\) reduction on DOC cycling, we assume that the proportion of \({{\text{NO}}_{ 3}}^{-}\) assimilation and DNRA is equal. The average DOC removal associated with microbial reduction of deposited \({{\text{NO}}_{ 3}}^{-}\) can thus be roughly estimated at ~1.6 mol C mol−1 N (=0.25 × 1.25 + 0.75/2 × 1.4 + 0.75/2 × 2). The individual components of this equation represent the proportions of denitrification (0.25), assimilation (0.75/2), and DNRA (0.75/2) in the total microbial \({{\text{NO}}_{ 3}}^{-}\) reduction; the values 1.25, 1.4, and 2 mol C mol−1 N represent the carbon demands of these processes, respectively. Other processes involved in \({{\text{NO}}_{ 3}}^{-}\) cycling in soils are neglected. The selected proportions of individual processes in \({{\text{NO}}_{ 3}}^{-}\) reduction in this trial have only limited effects on the final coefficient, which varies between 1.25 and 2 mol C mol−1 N, with minimum and maximum for 100 % proportion of denitrification and DNRA, respectively. This estimate does not include \({{\text{NO}}_{ 3}}^{-}\) produced in soils by nitrification of the deposited \({{\text{NH}}_{ 4}}^{ + }\), which partly underestimates the effect of the elevated Nr deposition on the consumption of bioavailable DOC.

The approximate estimates of bioavailable DOC mineralized to CO2 by microbial reduction of deposited \({{\text{SO}}_{ 4}}^{ 2-}\) (50 %) and \({{\text{NO}}_{ 3}}^{-}\) (100 %) in soils at the average consumption of 2 and 1.6 mol C mol−1 of the reduced S and N, respectively, can be calculated on a molar basis as \(\Updelta {\text{DOC}} = \Updelta {\text{CO}}_{2} = 0.5 \times 2 \times \left( {{\text{SO}}_{4}^{2-} {\text{deposition}}} \right) + 1.6 \times \left( {{\text{NO}}_{3}^{-} {\text{deposition}}} \right)\). Applying this estimate to Čertovo Lake catchment, a site in the Bohemian Forest, Czech Republic, indicates that elevated atmospheric deposition of electron acceptors could play an important role in the overall concentrations of biovailable DOC in soil water (Fig. 3). The Bohemian Forest is similar to the sites at which Novák et al. (2000, 2003) conducted the soil–water δ34S studies. During the peak of acidic emissions (Fig. 3a) and deposition (Fig. 3b) in the 1980s, the fluxes of atmospherically deposited \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) in this area averaged 126 and 100 mmol m−2 year−1, respectively (Majer et al. 2003). The associated hypothetical DOC oxidation to CO2 amounts to 286 mmol m−2 year−1 (Fig. 3c), with almost equal contributions from microbial \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) reduction. This CO2 flux would be equivalent to a decrease of ~2.5 mg L−1 in DOC concentrations at an average runoff of 1.35 m year−1 (286 mmol m−2 year−1 × 12 mg mmol−1/1350 L m−2 year−1 = 2.5 mg L−1). In other words, microbial reduction of deposited \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) would be able to reduce DOC leaching by ~2.5 mg L−1, compared to the pre-industrial period of negligible deposition of electron acceptors. Similarly calculated, the current (2008–2011) DOC leaching is still by 0.9 mg L−1 lower than its hypothetical value at zero deposition of electron acceptors (Fig. 3d). A comparison of the modelled change in DOC concentrations with the observed (and adjusted) DOC concentrations in the Bohemian Forest lakes shows that the proposed mechanism could be responsible for ~55 % of the observed change since 1991. The modelled increase in DOC concentrations was 0.7 mg L−1, whereas actual lake DOC concentrations increased by 0.5–1.8 mg L−1 (1.3 on average) in four monitored lakes during this period (Fig. 3e).

Fig. 3
figure 3

Historical trend in a anthropogenic SO2 and NO x emissions in the area of the Czech Republic and Slovakia and b \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) deposition in the Bohemian Forest (Czech–German–Austrian border) (Majer et al. 2003). c CO2 emission from microbial reduction of atmospherically-deposited SO4–S and NO3-N, calculated on a molar basis as CO2 = S + 1.6 × N, assuming 50 % reduction of deposited \({{\text{SO}}_{ 4}}^{ 2-}\) at the DOC consumption of 2 mol mol−1 S (Eq. 1), and 100 % reduction of deposited \({{\text{NO}}_{ 3}}^{-}\) by denitrification (25 %), DNRA (37.5 %), and during N assimilation (37.5 %) at the DOC consumption of 1.25, 2.0, and 1.4 mol mol−1 N, respectively. d Modelled decrease in DOC concentrations resulting from reduction of deposited \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) compared to preindustrial period with natural S and N emissions (zero value). In this calculation, one mol of CO2 production equals one mol of DOC removed. Concentrations were calculated using an average runoff in the area of 1.35 m year−1 (Majer et al. 2003). e Detail of recovering trend in DOC concentration after reduced \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{ - }\) deposition. Modelled values from panel (d) are compared to observed changes in the autumn DOC concentrations in the Bohemian Forest lakes (CN Černé, CT Čertovo, PL Plešné, and PR Prášilské) and their average value (red line). The observed concentrations were adjusted to their 2008–2011 averages of 1.6, 3.4, 3.9 and 5.0 mg L−1 in CN, CT, PL, and PR, respectively. Adjusted values \(\left( {C_{i}^{*} } \right)\) were calculated for each year (i) using the 2008–2011 average observed (C avg ) and modelled \(\left( {C^{M}_{avg} } \right)\) concentrations as \(C_{i}^{*} = C_{i} - (C_{avg} - C^{M}_{avg} )\)

Full size image

Despite the considerable uncertainty of the real proportions of \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) reduced in soils by individual microbial processes, this assessment suggests that elevated fluxes of mobile electron acceptors can have ecologically important effects on DOC cycling, and that they could significantly consume DOC in soils. The supported anaerobic respiration probably further accelerates the shift from a fungal- to bacterial-dominated microbial community in soils. The increased opportunity to use \({{\text{SO}}_{ 4}}^{ 2-}\) and \({{\text{NO}}_{ 3}}^{-}\) as electron acceptors also brings a competitive advantage to mineralising over fermenting bacteria (Madigan et al. 2010). This may further increase the total bacterial CO2 efflux, reduce pool of bioavailable DOC in soil solution for the outcompeted fermentation bacteria, and reduce their DOC transformation and contributions to DOC leaching (Fig. 2d).

Linking N saturation to carbon limitation of soil microbial processes in acidified soils: a new conceptual hypothesis

The ecological importance of individual processes lowering soil pools of bioavailable DOC (decreased allocation of plant photosynthates, chemical suppression, and elevated mineralization at increased availability of electron acceptors; Fig. 2) varies with the physico-chemical and biological properties of the plant-soil system, and the levels of Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition. The combined effects of these processes could be sufficiently strong to continuously reduce bioavailable DOC to such an extent that bacterial processes switch from N limitation to C limitation. The declining pool of bioavailable DOC is accompanied by progressively-increasing N availability in soils (Taylor and Townsend 2010), which can result in ecosystem transition from stage 3–4 in the conceptual model of Schimel and Bennett (2004), and to a breakpoint between Nr immobilization and \({{\text{NO}}_{ 3}}^{-}\) leaching. Some ecosystem responses to changing Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition affect the soil C and N cycles independently, while others influence both cycles, and moreover, the cycles influence each other. To explain a possible interconnectedness of the associated processes, we put forward the following conceptual hypothesis, linking the N and C cycles with S cycling in terrestrial ecosystems exposed to elevated Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition.

Compared to the preindustrial period, when the system was N-limited and terrestrial N export was dominated by DON (Fig. 4a), elevated input of Nr alleviates N limitation, and together with \({{\text{SO}}_{ 4}}^{ 2-}\) deposition causes soil acidification and increases availability of electron acceptors for soil microbial processes. The chemical and microbial responses to these changes can affect C and N cycling through a chain of both sequential and parallel effects, as follows:

Fig. 4
figure 4

Conceptual scheme of the coupled C and N cycling in soils a at the beginning of industrial period (1850s) and b during high Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition (1980s). Black arrows indicate N fluxes, brown arrows are DOC fluxes, blue arrows are CO2 fluxes, and red arrows are fluxes of electron acceptors (EA = SO4–S + NO3–N). Numbers identify specific changes caused by elevated availability of Nr and EA: 1 Decreasing fungi:bacteria ratio, lower litter decomposition, lower exudation of plants’ photosynthates, higher N input via N-richer litterfall, and higher mineralization at elevated availability of Nr, 2 lower bacterial \({{\text{NH}}_{ 4}}^{ + }\) immobilization in microbial biomass at lower DOC, 3 higher \({{\text{NO}}_{ 3}}^{-}\) production due to promoted nitrification at lower DOC, 4 lower denitrification and dissimilatory nitrate reduction to ammonium at lower DOC, and 5 bacterial \({{\text{NO}}_{ 3}}^{ - }\) immobilization in microbial biomass, which may be potentially lower due to lower concentrations of bioavailable DOC, or higher due to higher \({{\text{NO}}_{ 3}}^{ - }\) concentrations in the system. Changes in \({{\text{NH}}_{ 4}}^{ + }\) leaching are expected to be small and are not shown. (Color figure online)

Full size image
  1. (1)

    Elevated Nr availability diminishes plant dependence on mycorrhizal supply of organic N monomers, which leads to reduced exudation of photosynthates (Fig. 3b) and to increased N concentrations in the plant tissue, and therefore to increased N fluxes in litterfall back to the soils (Fig. 4b).

  2. (2)

    The reduced supply of photosynthates and elevated availability of N change soil microbial community and both the fine root biomass and fungi:bacteria ratio decrease. Fine root biomass also decreases due to Aln+ toxicity in strongly acidified soil.

  3. (3)

    Due to decreasing fungal biomass, litter decomposition and fungal production of LMW OM slow down, but bacterial DOC mineralization increases because of elevated availability of N (the limiting nutrient) and electron acceptors for anaerobic respiration (Fig. 3b, d).

  4. (4)

    As Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition increase, bacteria become less and less N and energy (electron acceptor) limited, and can mineralize progressively more bioavailable DOC, while less remains for the outcompeted fermenting bacteria and the leaching of DOC decreases (Figs. 3d, 4b).

  5. (5)

    Elevated N mineralisation occurs at higher Nr availability in soils, producing more \({{\text{NH}}_{ 4}}^{ + }\). In parallel, the pool of bioavailable DOC decreases due to the elevated mineralization. Consequently, relatively less bioavailable DOC remains for \({{\text{NH}}_{ 4}}^{ + }\) assimilation and immobilization in microbial biomass, leaving more \({{\text{NH}}_{ 4}}^{ + }\) for nitrifiers. As a result, internal \({{\text{NO}}_{ 3}}^{-}\) production increases.

  6. (6)

    The higher bacterial demand for DOC under elevated availability of Nr and electron acceptors comes into conflict with increasing chemical suppression of DOC solubility and bioavailability in the progressively acidifying soils (Fig. 3c), and finally results in the C limitation of microbial metabolism.

  7. (7)

    Due to reduced pools of bioavailable DOC, \({{\text{NO}}_{ 3}}^{-}\) removal by denitrification and DNRA decrease (Fig. 4b).

  8. (8)

    Bacterial assimilation of \({{\text{NO}}_{ 3}}^{-}\), which depends on the stoichiometric \({{\text{NO}}_{ 3}}^{-}\) to DOC ratio in the substrate, may initially increase with increasing \({{\text{NO}}_{ 3}}^{-}\) concentrations, but then decrease due to lower pool of bioavailable DOC.

Altogether, the system moves from stage 3 to stage 4 of the conceptual model of Schimel and Bennett (2004), leading to increased concentrations of \({{\text{NO}}_{ 3}}^{-}\) in soil and leaching of unused \({{\text{NO}}_{ 3}}^{-}\). Even though no study has confirmed the proposed sequence of effects so far, the individual key processes and their sequences have been demonstrated in numerous studies (Table 1).

Table 1 Key processes and observed changes in C and N cycling in natural terrestrial ecosystems due to elevated Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition
Full size table

Hypothesis versus observations

Our hypothesis is not in conflict with previous schemes explaining the development of N-cycling in and \({{\text{NO}}_{ 3}}^{-}\) leaching from terrestrial ecosystems (Aber et al. 1989, 1998; Schimel and Bennett 2004; Lovett and Goodale 2011), but combines and develops them, and suggests additional mechanisms for several observed responses of upland and forest ecosystems to changing Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) deposition:

  1. (1)

    The higher N availability in soils and lowered supply of plant photosynthates reduce fungi:bacteria ratios, which results in decreased decomposition rates of recalcitrant SOM (Waldrop et al. 2004; Frey et al. 2004; Janssens et al. 2010). This process provides an alternative (or additional) explanation for the observed litter accumulation that occurred in central European spruce forests during the acid rain period, as well as to its rapid mineralisation during the recovery phase, which were ascribed primarily to changes in soil acidification by \({{\text{SO}}_{ 4}}^{ 2-}\) deposition (e.g., Meiwes et al. 2009; Oulehle et al. 2011).

  2. (2)

    The hypothesis causally links elevated \({{\text{NO}}_{ 3}}^{-}\), as well as \({{\text{SO}}_{ 4}}^{ 2-}\) concentrations to reduced DOC leaching during the acidification stage (Gjessing et al. 1998), and their reversed trends during the recovery phase (Monteith et al. 2007).

  3. (3)

    It also reasonably explains the maximum \({{\text{NO}}_{ 3}}^{-}\) leaching from C-poor ecosystems exposed to elevated atmospheric inputs of Nr and \({{\text{SO}}_{ 4}}^{ 2-}\) (Evans et al. 2006; Curtis et al. 2011). Even though microbial activity remains high in these soils (Curtis et al. 2011), there is small capacity (i.e., pools of bioavailable DOC) to further accumulate N, and more \({{\text{NH}}_{ 4}}^{ + }\) remains for nitrifiers. Experiments on highly organic soils show far slower responses to additional Nr application than similar rates of application at mineral soil sites, despite similarly high ambient Nr deposition (e.g. Pilkington et al. 2005), consistent with the suggestion above that the former sites are less susceptible to C limitation and therefore to \({{\text{NO}}_{ 3}}^{-}\) leaching.

  4. (4)

    The hypothesis assumes increased CO2 efflux from bacterial mineralization at the expense of DOC leaching (Fig. 2D), which is seemingly in contrast to negligible changes in total microbial CO2 production observed in N addition experiments (Aber et al. 1998). The hypothesis, however, assumes a parallel decrease in the fungal CO2 efflux due to lower SOM mineralization (Fig. 2b). The decreasing fungi:bacteria ratio thus may result in a negligible net change in the total microbial CO2 efflux (Šantrůčková et al. 2009), and the net change in the total soil C pool may be that SOM accumulates while DOC leaching decreases. In addition, the estimated increase in CO2 production due to higher anaerobic respiration in the Bohemian Forest (0.29 mol m−2 year−1) was <1 % of total soil respiration of spruce forest ecosystems (25–68 mol m−2 year−1; Borken et al. 1999; Davidson et al. 2006), i.e., probably lower than variability of field measurements.

  5. (5)

    Our hypothesis suggests that elevated atmospheric input of electron acceptors to soils and their acidifying effects could cause the abrupt depletion of available DOC and rapidly decrease the N retention capacity of soil. The biogeochemical implication of this process is that \({{\text{NO}}_{ 3}}^{-}\) leaching could be caused not only by the elevated Nr inputs, but also by \({{\text{SO}}_{ 4}}^{ 2-}\) deposition high enough to decrease pools of bioavailable DOC to the extent that nitrification is promoted (Fig. 4b). The reverse situation has probably occurred during the recovery phase; decreased \({{\text{SO}}_{ 4}}^{ 2-}\) deposition (i.e., lower availability of electron acceptors in soils) led to lower removal of DOC from the system, alleviating C limitation of bacteria, and increasing N immobilisation in bacterial SOM, as well as \({{\text{NO}}_{ 3}}^{-}\) reduction via denitrification and DNRA. This mechanism could perhaps explain why decreasing \({{\text{SO}}_{ 4}}^{ 2-}\) deposition was accompanied by decreasing \({{\text{NO}}_{ 3}}^{-}\) leaching in some regions, despite no change in Nr deposition (Driscoll et al. 2003; Goodale et al. 2005; Oulehle et al. 2008). In contrast, other sites (e.g., in the UK; Monteith et al. 2012) exhibit no clear changes in \({{\text{NO}}_{ 3}}^{-}\) export despite reduced \({{\text{SO}}_{ 4}}^{ 2-}\) deposition. Explanatory factors here could include local land-management factors (e.g., forest management or moorland burning), or the excess of bioavailable DOC from the highly organic soils of many UK acid-sensitive catchments. In such soils, it is less likely that C limitation would have led to reduced N assimilation and greater nitrification in the past, and thus less likely that \({{\text{NO}}_{ 3}}^{-}\) leaching would be decreasing now in response to reducing \({{\text{SO}}_{ 4}}^{ 2-}\) deposition.

  6. (6)

    Most experiments with Nr addition to terrestrial ecosystems, including those with the described kinetic saturation, were done at sites already affected by acidic deposition (Lovett and Goodale 2011). Their kinetic saturation can be, undoubtedly, explained to some extent by natural ecosystem properties and limits of existing microbial population to take up a pulse of N input. There may be partial physical bypass, or the microbial population may not be able to grow fast enough to assimilate all the extra N. According to our hypothesis, it seems also plausible that these sites were close to C limitation already prior to the artificial Nr additions, and consequently, responded to the Nr dose with a rapid transition from stage 3 to stage 4 of the conceptual model of Schimel and Bennett (2004). In such cases kinetic saturation could reflect a shortage of bioavailable DOC (relative to the available N) for microbial N immobilization. This could occur not only due to excess Nr inputs, but also due to reduced amounts of bioavailable DOC (e.g., during the winter decline in below-ground allocation of plant photosynthates; Högberg et al. 2010). Temporal trends in \({{\text{NO}}_{ 3}}^{-}\) leaching thus could be partially associated with seasonal changes in bioavailable DOC concentrations in soil solutions.

  7. (7)

    Finally, some of the assumed causalities in our hypothesis have been supported by observed biogeochemical changes during recovery from acidification. Increasing pH, lower ionic strength and Aln+ toxicity, lower Nr concentrations, and lower availability of electron acceptors in soil solutions after reductions in acidic deposition have increased soil pools of bioavailable DOC and established conditions for recovery of fine roots of plants (Navrátil et al. 2007) and their re-colonization with mycorrhizal fungi (Boxman et al. 1998). Observed increases in DOC leaching (Monteith et al. 2007), more rapid litter mineralisation (Meiwes et al. 2009; Oulehle et al. 2011), and reduced \({{\text{NO}}_{ 3}}^{-}\) leaching (Veselý et al. 2002; Oulehle et al. 2008; Driscoll et al. 2003; Goodale et al. 2005) during the recovery phase probably reflect a partial ecosystem transition back towards preindustrial conditions.

Further research to test the hypotheses

Studies to improve or refute our hypothesis might include: (1) More detailed description of changes and differences in microbial communities and their activities in manipulated versus control plots and understanding of their effects on bioavailability and recalcitrance of soil DOC, (2) more detailed assessment of the fate of \({{\text{NO}}_{ 3}}^{-}\) and \({{\text{SO}}_{ 4}}^{ 2-}\) in soils, especially the proportion of individual microbial process involved in their transformations, and (3) laboratory and in situ manipulation experiments with additions of bioavailable DOC, \({{\text{NO}}_{ 3}}^{-}\), and \({{\text{SO}}_{ 4}}^{ 2-}\), which should indicate how important this is as a constraint on the processes. These experiments should be as comprehensive as possible, including studies on element mass budgets, as well as changes in microbial communities.