The study of soil N cycling processes has been, is, and will be at the centre
of attention in soil science research. The importance of N as a nutrient for
all biota; the ever-increasing rates of its anthropogenic input in
terrestrial (agro)ecosystems; its resultant losses to the environment; and
the complexity of the biological, physical, and chemical factors that
regulate N cycling processes all contribute to the necessity of further
understanding, measuring, and altering the soil N cycle. Here, we review
important insights with respect to the soil N cycle that have been made over
the last decade, and present a personal view on the key challenges of future
research. We identify three key challenges with respect to basic N cycling
processes producing gaseous emissions:
quantifying the importance of nitrifier denitrification and its main
controlling factors; characterizing the greenhouse gas mitigation potential and
microbiological basis for N characterizing hotspots and hot moments of denitrification disentangling gross N transformation rates using advanced
linking functional diversity of soil fauna to N cycling processes beyond
mineralization; determining the functional relationship between root traits and soil N
cycling; characterizing the control that different types of mycorrhizal symbioses
exert on N cycling; quantifying the contribution of non-symbiotic pathways to total N
fixation fluxes in natural systems
Furthermore, we identified a key challenge with respect to modelling:
Finally, we propose four key challenges related to how ecological
interactions control N cycling processes:
We postulate that addressing these challenges will constitute a
comprehensive research agenda with respect to the N cycle for the next
decade. Such an agenda would help us to meet future challenges on food and
energy security, biodiversity conservation, water and air quality, and
climate stability.
Humankind's relationship with soil nitrogen (N) has been a long and troubled one. For most of agricultural history, farmers have struggled to maintain soil fertility levels in their fields, relying mostly on biological N fixation (BNF), decomposition of soil organic matter and redistribution of organic materials to provide N to their crops. With the onset of large-scale application of mineral fertilizers in the 1950s, the main focus in large parts of the world has gradually shifted towards minimizing harmful losses to the environment resulting from the large amounts of N entering the global food production system (Galloway et al., 2013).
The history of research on the soil N cycle reflects this shift. The study of
N cycling processes started after Carl Sprengel's discovery (popularized by
Justus Von Liebig) of the importance of N as a factor limiting the growth of
crop plants in the mid-19th century (Gorham, 1991). More than 150 years of
research has demonstrated that this element limits ecosystem productivity
over large areas of the globe and is highly sensitive to changes in
temperature, precipitation, atmospheric CO
The need for more information on soil N cycling process rates is highlighted
by large amounts of “missing N” that dominate N balances at all scales.
Inputs of N through fertilization, BNF, atmospheric deposition, and human and
animal waste have been found to be substantially higher than hydrological
outputs of N in many studies and at many scales (Howarth et al., 1996; Boyer et
al., 2002; Groffman, 2008). There is much uncertainty about the fate of this
excess N (Van Breemen et al., 2002). Is it stored in soils or vegetation? Is
it converted to gas, and if so, in which forms? This uncertainty is
particularly compelling in agricultural systems which receive high rates of N
input. The air and water quality impacts of the N exports in these systems
are cause for great concern (Davidson et al., 2012). In other ecosystems,
on the other hand, there is concern about missing N inputs. Unexplained
accumulation of N in aggrading forests (Bernal et al., 2012; Yanai et al.,
2013) and in vegetation exposed to elevated levels of atmospheric CO
A particularly pressing need in N cycling research has been in the area of
gaseous emissions, especially of those that contribute to global warming. The
role of soil biogeochemists is to generate field data on terrestrial
greenhouse emissions, but high uncertainties in soil N
One of the reasons that it has been so difficult to quantify and characterize N cycling processes is that they are to a large extent controlled by indirect, biotic interactions. It is becoming increasingly clear that ecological interactions play a major role in the terrestrial N cycle. The realization that global change may alter the nature and timing of biotic interactions and thereby their effects on the N cycle only increases the need for their study (Díaz et al., 1998; Chapin et al., 2000). In some ecosystems, N inputs to terrestrial ecosystems are dominantly mediated by mutualistic associations between plants and specific N-fixing microbial groups (Batterman et al., 2013a). More generally, plant species have an overarching impact on soil N cycling by directly mediating energy and material fluxes to soil microbial communities and/or by altering abiotic conditions that regulate microbial activity. For example, the type of mycorrhizal fungi that colonizes the plant root has been shown to correlate with organic N depolymerisation as fungal groups produce a specific set of enzymes. Also, soil fauna have both a direct and indirect role in the soil N cycle as grazing may strongly affect microbial N release as well as alter soil physical properties. All these ecological interactions have a high degree of specificity and sensitivity to global change, which increases the probability that a change in plant, microbial, or faunal community composition will have cascading effects on the rest of the system and on the overall soil N cycle (Chapin et al., 2000).
New insights and key challenges with respect to the soil N cycle, as identified in this paper. These include three N cycling processes (Sects. 2.1–2.3), a modelling challenge (Sect. 3), and four pathways through which ecological interactions might affect N cycling processes (Sects. 4.1–4.4).
Here, we review important insights with respect to the soil N cycle that have
been made over the last decade and present our view on the key challenges
of future soil research (Fig. 1). The approach adopted in this paper is
three-fold:
to identify and critically review specific N transformation pathways
related to the production of N to present methodological developments on to review mechanisms on how ecological interactions impact soil N
cycling. Specifically, we focus on soil faunal effects (Sect. 4.1), plant
root controls (Sect. 4.2), mycorrhizal symbioses (Sect. 4.3), and biological
N fixation (Sect. 4.4). Although other nutrient cycles can have strong
effects on all aspects of the N cycle (e.g. Baral et al., 2014), we consider
stoichiometric relationships to be mostly outside the scope of this paper and do
not exhaustively review them.
Although all authors agree with the contents of the final paper, some freedom
has been given to express a somewhat personal view on developments within our
respective fields of expertise (see Author Contributions). This
paper is not meant as a comprehensive literature review of soil N cycling
research in the past. Instead, we have tried to be judicious with respect to
referencing older studies, only citing some key papers and focusing instead
on more recent work. As such, we hope that our paper will spark discussion
and inspire further research on the elusive aspect of soil N cycling.
The study of nitrifier denitrification as a significant biogeochemical
N
With respect to terminology, it took a landmark paper (Wrage et al.,
2001) to clearly identify nitrifier denitrification as a distinct pathway for
N
Different pathways of N
Despite the similarity with classical denitrification, there are good reasons
to assume that nitrifier denitrification is controlled by different factors
and should therefore be considered as a distinct source of N
The process was described by early pure culture studies in the 1960s and
1970s (Hooper, 1968; Ritchie and Nicholas, 1972). Since then, it has been
reported several times (e.g. Poth and Focht, 1985; Schmidt et al., 2004), but
always in pure cultures. Despite suggestions that nitrifier denitrification
could be an important contributor to soil N
The main challenge to evaluating the importance of nitrifier denitrification
in soils is methodology. As the N in N
Various efforts have been undertaken to employ advanced stable isotope
analysis to determine the contribution of nitrifier denitrification as an
N
Wrage et al. (2005) proposed an alternative method based on artificially
enriched stable isotope tracing. They combined
Both net atmospheric and in situ N
Based on recent evidence from the literature we have identified three
possible routes for N
Second, some bacteria that perform dissimilatory nitrate reduction to ammonia
(DNRA) are capable of N
Third, there is evidence that both direct assimilatory N
The N
In conclusion, five possible pathways for N
Denitrification, the anaerobic microbial conversion of the nitrate
(NO
Addressing the challenge of denitrification requires advances in three main areas: (1) improved methods for quantifying N gas fluxes (see also Sect. 2.2); (2) experimental designs that incorporate hotspot and hot moment phenomena; and (3) approaches for temporal and spatial scaling that account for hotspot and hot moment phenomena at multiple scales.
Denitrification has always been a challenging process to measure (Groffman et
al., 2006), primarily due to the difficulty of quantifying the flux of
N
Our understanding of the N
The new soil core incubation systems, along with new soil O
As our ability to quantify denitrification has improved, our understanding of
the factors that control the occurrence of hotspots and hot moments of
activity has also increased. Riparian zones have been studied in this regard
for several decades (e.g. Lowrance et al., 1997; Mayer et al., 2007). This
has resulted in efforts to protect and restore riparian zones to decrease N
delivery to receiving waters in many locations. Still, there is great
uncertainty about just how much N is denitrified in riparian zones and
through other N control practices, and how much N remains in the soils and
vegetation of these areas where it is susceptible to later conversion back into
NO
There has long been recognition of the potential for hotspot and hot moment
denitrification to occur within crop fields or pastures. Periods of transient
saturation low in the soil profile can support significant amounts of
denitrification that are missed in sampling programs that focus on surface
soils (Werner et al., 2011; Morse et al., 2014). Areas of wet soil, low soil
O
Experiments incorporating new ideas about hotspots and hot moments can benefit from recent studies that have characterized diversity in denitrifying phenotypes that reflect adaptation to prevailing environmental conditions with consequences for denitrification activity (Bergaust et al., 2011). These ideas have the potential to improve these experiments by allowing for more mechanistic, hypothesis-driven approaches that underlie more “black-box” ideas based on proximal drivers of denitrification.
Estimates of denitrification produced by direct measurement in soil cores can
be validated using isotope measurements and models. Shifts in
The time is thus ripe for ecosystem-, landscape-, and regional-scale studies of denitrification. We have new methods capable of producing well-constrained estimates of denitrification at the ecosystem scale and new ideas about the occurrence of hotspots and hot moments at ecosystem and landscape scales. In combination with independent approaches for validation of denitrification estimates, our estimates of this important process are likely to improve markedly over the next decade.
This section will focus on how
The stable isotope
In addition to quantification of gross N transformation rates,
The influence of soil fauna on soil N processes and loss pathways.
Conventionally
Until recently, the influence of soil fauna on the soil N cycle in agroecosystems has been mostly neglected. Nitrogen transformation processes and nitrogen loss pathways have almost exclusively been related to the interplay between microbial dynamics in the soil and abiotic factors. At first glance this seems logical: microorganisms dominate the biomass of soil life to a large degree, and many conversions in the N cycle (e.g. nitrification, denitrification, nitrifier-denitrification, N fixation, DNRA) are the exclusive domain of microorganisms. Biochemical and physical processes, such as nitrification and N leaching are controlled by abiotic factors (e.g. pH, porosity and temperature). In turn, both microbial dynamics and abiotic factors can be changed by human influences such as N deposition in natural systems and fertilization, liming, soil tillage, and animal husbandry in agricultural systems (Fig. 4a).
What important role do soil fauna then have in the N cycle? Like the effect of humans, their role can be dramatic but is essentially indirect: through trophic interactions and burrowing activities they may strongly affect microbial dynamics in the soil and soil physical properties (Fig. 4b).
The only part of the soil N cycle where the role of soil fauna has been reasonably well established is N mineralization and subsequent plant uptake. Soil fauna affects N mineralization by a combination of activities, including trophic interactions (grazing on microorganisms, predation), fragmentation of organic matter, mixing organic matter into the soil, excreting nutrient-rich compounds, and dispersing microbial propagules (Bardgett and Chan, 1999).
In a literature study across natural and agricultural systems, Verhoef and
Brussaard (1990) found a relatively stable faunal contribution to N
mineralization of around 30 %. Different functional groups of soil fauna,
however, contribute to N mineralization differently, with the largest
contributions provided by bacterial-feeding microfauna (nematodes and
amoeba), followed by earthworms and potworms, and minor contributions by
fungal-feeding nematodes and microarthropods (De Ruiter et al., 1993). Among
meso- and macrofauna, the role of earthworms has been most extensively
studied (e.g. Postma-Blaauw et al., 2006; Van Groenigen et al., 2014). As
“ecosystem engineers”, they are well known to affect soil structure and
litter redistribution, thereby affecting many aspects of the N cycle and other soil processes (Shipitalo and Le Bayon, 2004; Blouin et al., 2013).
In a recent meta-analysis, Van Groenigen et al. (2014) showed that in
agricultural systems earthworms increase crop yield on average by 25 %.
This effect was consistent between different functional groups of earthworms,
but increased with earthworm density and crop residue application rates.
Because this beneficial effect disappeared with adequate N fertilization, it
was mainly ascribed to increased N mineralization from crop residue and soil
organic matter. In tropical ecosystems, soil-feeding termites are known to
have a similarly large impact on N mineralization (Ji and Brune, 2006).
Termites are also able to volatilize ammonia from their gut and faeces. However, this has only been shown to lead to high NH
The effect of faunal diversity rather than single faunal groups is complex. Combinations of functionally dissimilar soil fauna can increase the N mineralization rate due to facilitative interactions (Heemsbergen et al., 2004). These include one group benefitting from the activity of another group, for example through changes in soil structure or litter shredding by isopods promoting microbial growth (Wardle, 2006). However, competitive interactions may also positively influence mineralization rates (Loreau, 1998). For instance, predatory mites in the soil feed on fungivorous mites and potworms, springtails, and nematodes (De Ruiter et al., 1995), and can thereby influence microbial activities through trophic cascades (induced positive effects on microbes by feeding on microbial feeders). Even though empirical evidence of such trophic cascades in soil food webs is scarce (Mikola and Setälä, 1998; Bardgett and Wardle, 2010), the presence of predatory mites can potentially influence the behaviour of fungivorous mites and potworms in terms of their feeding rate and spatial distribution. Such interactions (both facilitative and competitive), within and across trophic levels, have not yet been explored for most N cycling processes, including N loss pathways.
Among the relatively few studies that have focused on processes other than N
mineralization, earthworms are again by far the most studied group. They have
been shown to affect microbial N immobilization (Brown et al., 1998) as well
as nitrification and denitrification (e.g. Parkin and Berry, 1999; Rizhiya et
al., 2007). A growing body of literature shows that earthworms can
considerably increase N
Evidence for involvement of other faunal groups in these processes is scarce.
Potworms, phylogenetically related to earthworms and with similar foraging
and burrowing habits (albeit at a smaller scale), have been recognized as
vectors for microbial colonization (Rantalainen et al., 2004) and may
influence both nitrification and denitrification processes (Van Vliet et al.,
2004). High soil NO
Changes in soil structure (porosity, aggregation) by faunal activity can affect soil physical processes as well. Burrowing activities of earthworms may create preferential flow pathways that increase leachate volume and consequently the total leaching loss of inorganic N and dissolved organic N (e.g. Dominguez et al., 2004). Interactions between other soil faunal species have received little attention with regard to their effects on soil physical properties. Smaller fauna such as potworms, springtails, mites, and nematodes are often assumed to have negligible direct effects on larger-scale soil structure, because they are usually confined to pre-existing voids in litter or soil (Lee and Foster, 1991; Whalen and Sampedro, 2010). However, these small fauna can significantly alter soil microstructure by producing faecal pellets, and potworms can also increase soil porosity and pore continuity by their burrowing activity (Topoliantz et al., 2000; Van Vliet et al., 2004).
Overall, soil biota are essential for maintaining healthy soils and
providing ecosystem services, such as N mineralization and plant uptake for
food, fuel, and fiber production. However, it is not clear whether they are
able to do so without creating detrimental effects on N loss pathways such
as N leaching and N
Soil microbial communities depend almost exclusively on plant-derived resources for their energy and nutrient supply. For a long time, it was presumed that plant litter was the most relevant organic matter input for the soil food web and that plant effects on soil biogeochemistry were mainly mediated via the indirect impacts of plant inputs on relatively inert soil properties. Therefore, most of our initial understanding of soil biogeochemistry was based on experiments with root-free soils.
The impact of spatially and temporarily dynamic processes occurring in the rhizosphere on N cycling has rarely been considered (Frank and Groffman, 2009; Rütting et al., 2011b). Nevertheless, an important share of the energy for microbial metabolism is delivered by belowground plant parts through root exudation, cell sloughing, and root and mycorrhizal fungal turnover (Nguyen, 2003). Healthy growing roots pass a large proportion of the C they receive to the soil as root exudates. This includes a range of materials, but soluble compounds, consisting of organic acids, carbohydrates, and amino acids, comprise the largest component (Farrar et al., 2003). The total amount and composition of root exudates varies between plant species and genotypes, and is influenced by plant phenology and environmental conditions (Nguyen, 2003). Moreover, fine root turnover, caused by the production, mortality, and decay of short-lived C-rich roots, is another key pathway of significant nutrient flux in terrestrial ecosystems that may equal or even exceed that of aboveground litter fall in certain ecosystems (Gill and Jackson, 2000; Yuan and Chen, 2010).
There are several mechanisms through which plant roots can affect rhizosphere
N cycling (reviewed in Paterson, 2003; Dijkstra et al., 2013; Cheng et al.,
2014). Rhizodeposition may enhance microbial growth and activity and
stimulates production of microbial exoenzymes that “mine” for more complex soil
organic N compounds, a process often referred to as “priming” (Paterson,
2003). Nitrogen immobilized by the microbial community may temporarily reduce
soil N availability, but immobilized N can become available in the
rhizosphere due to microbial turnover and the grazing of rhizosphere
microorganisms by soil microfauna (see Sect. 4.1). The quality of
rhizodeposition is an important determinant for soil microbial communities;
any shifts in their composition may affect decomposition processes through
the production of distinct sets of extracellular enzymes (Dennis et al.,
2010; Kaiser et al., 2010). Nevertheless, under conditions of low N
availability, plant N uptake may limit microbial substrate N availability and
reduce microbial growth and decomposition activity (Dijkstra et al., 2010;
Blagodatskaya et al., 2014). Moreover, the production of specific metabolites
that act as signaling molecules could accelerate or retard soil N cycling if
they act upon certain functional microbial taxa (De-la-Pena and Vivanco,
2010). Finally, specific N cycling processes, such as denitrification or N
fixation, could be altered in the rhizosphere due to altered microbial
substrate conditions, encompassing C, O
Although the quality and quantity of rhizodeposits clearly influence rhizosphere N cycling, a major challenge lies in determining to what extent plant community characteristics explain the observed variations of rhizosphere impacts (Cheng et al., 2014). Considering the great difficulties in assessing rhizodeposition under field conditions (Pausch et al., 2013a), a prospective approach may involve measuring “soft” plant traits that are relatively easy to observe and quantify (Fry et al., 2014). There are several traits that are good candidates due to their putative intimate relationship with rhizodeposition. For example, root exudation is linked to the intensity of canopy photosynthetic activity and photo-assimilate supply (Kuzyakov and Cheng, 2001). Fast-growing, acquisitive plants with high specific leaf area and short lifespan are thus thought to be associated with a larger rhizosphere effect (Wardle et al., 2004). Because root exudation is concentrated at the apices of the roots and at the nodes where lateral roots emerge (Jaeger et al., 1999), root architectural traits determine the expansion of the rhizosphere and exudate fluxes per unit of root biomass. A densely branched root system with high biomass and a rapid turnover thus contributes large quantities of exudates (Van der Krift et al., 2001). The chemistry of rhizodeposits is a key controlling variable of rhizosphere dynamics, as microbial communities may shift their N use efficiency in response to substrate stoichiometry, leading to changes in soil N cycling fluxes (Moorshammer et al., 2014).
Several studies have examined presumed relationships between N cycling parameters and plant traits, especially of aboveground plant organs (e.g. Wedin and Tilman, 1990; Orwin et al., 2010; Garcia-Palacios et al., 2013; Grigulis et al., 2013). Soil N cycling processes appear to be primarily driven by traits of the most abundant species (the biomass ratio hypotheses; Grime, 1998), although complex effects may arise due to interspecies interactions and non-additive species effects (Grigulis et al., 2013; Pausch et al., 2013b). These studies confirm that plant characteristics, including under-investigated root traits, exert a key control over soil microbial communities and modify the fundamental physiologies that drive soil N cycling. Nevertheless, the lack of clear-cut relationships between specific plant traits and N cycling parameters indicates the necessity for more research on plant communities to establish consistent links between plant traits and N cycling variables, especially under field conditions.
This section will focus on the extent to which the main types of mycorrhizal symbioses, arbuscular mycorrhiza and ectomycorrhiza, differentially affect the soil N cycle. Early conceptual models linked the replacement of arbuscular mycorrhizal plants by ectomycorrhizal plants to succession (Read, 1991) or to latitudinal and altitudinal gradients from warmer to colder climates (Read and Perez-Moreno, 2003). This was considered to be driven by shifts from P to N limitation, where simultaneously an increasing fraction of the N and P was present in organic forms to which ectomycorrhizal fungi were supposed to have better access than arbuscular mycorrhizal fungi. However, Dickie et al. (2013) noted a poor fit between these models and actual data on primary succession and suggested that nutrient limitation shifts from N to P limitation in retrogressive succession. Although a new model of general applicability has not yet been proposed, the underlying idea of a fundamental difference between arbuscular mycorrhiza-dominated ecosystems, with more open, inorganic nutrient cycles, and ectomycorrhiza-dominated ecosystems, with more closed, organic nutrient cycles has persisted, especially for forests in temperate regions (Phillips et al., 2013; Bradford, 2014). We note that the same distinction was proposed between bacterial- and fungal-dominated agroecosystems by De Vries and Bardgett (2012). Their conceptual model is apparently not applicable for the tropics, where both arbuscular mycorrhizal and ectomycorrhizal forests are characterized by an open N cycle (Kuyper, 2012; Tedersoo et al., 2012). This geographical contrast raises the question of to what extent the nature of the mycorrhizal symbiosis is causally relevant for differences in forest ecosystem functioning, or if plant traits other than the mycorrhizal symbiosis cause these differences. Arguments that the mycorrhizal symbiosis is causally relevant for soil N cycling are connected to the claim that ectomycorrhizal fungi, contrary to arbuscular mycorrhizal fungi, possess extensive saprotrophic activity and are therefore able to make N available in the soil (“mining”) (Koide et al., 2008; Talbot et al., 2008), and therefore could access organic sources of N and phosphorus.
Several authors have compared uptake of various amino acids by arbuscular and ectomycorrhizal plants. The ability to depolymerize large N-containing molecules (proteins) into smaller fragments that can be taken up (Schimel and Bennett, 2004) and the ability to increase access to these large molecules, which are often bound to phenolics and other recalcitrant compounds, have been mainly studied for ectomycorrhizal fungi. Talbot and Treseder (2010) demonstrated the widespread ability among ectomycorrhizal fungi to take up amino acids and noted that the relative benefit of the symbiosis was largest for the most common amino acids. Arbuscular mycorrhizal fungi also have widespread ability to take up amino acids (Whiteside et al., 2012). Arbuscular mycorrhizal plants took up significantly larger amounts of eight amino acids (phenylalanine, lysine, asparagine, arginine, histidine, methionine, tryptophan, and cysteine) than non-mycorrhizal plants and significantly smaller amounts in the case of aspartic acid. Contrary to the hypothesis of Talbot and Treseder (2010) for ectomycorrhizal plants, the authors noted that the mycorrhizal effect on uptake was inversely related to the abundance of that amino acid in the database of all known proteins. The authors speculated that preferential use of rare amino acids by arbuscular mycorrhizal plants may reduce competition with ectomycorrhizal plants for amino acids. However, the extent to which this form of niche differentiation would reduce competition depends on the rate at which amino acids become available in the soil solution and hence to what extent the two preceding steps (increased access to protein–polyphenol complexes; depolymerization of proteins) are rate-limiting. It is therefore necessary to assess the mycorrhizal role in those two steps.
Lindahl et al. (2007) showed an increased C : N ratio in deeper humus layers, and this effect was attributed to selective N mining by ectomycorrhizal fungi. Several studies have provided explicit support that ectomycorrhizal fungi can mine humus layers for N and have identified the relevant ectomycorrhizal fungi (Hobbie et al., 2013; Rineau et al., 2013; Bödeker et al., 2014). Wu (2011) on the other hand claimed that direct access by ectomycorrhizal fungi to N from the protein–polyphenol complex is likely limited and attributed a major role for interactions between saprotrophic and ectomycorrhizal fungi. Current evidence suggests that arbuscular mycorrhizal fungi have neither the ability to degrade humus for N-rich compounds nor the ability to depolymerize proteins into amino acids. The widespread ability of arbuscular mycorrhizal fungi to take up amino acids may therefore not be related to closed nutrient cycles with a major role for uptake of organic nutrients, but may rather function as a scavenging mechanism to re-absorb exudates, including amino acids. More information about the role of arbuscular mycorrhiza in the uptake of organic N is provided in recent reviews by Veresoglou et al. (2012) and Hodge and Storer (2015).
The stable isotope
A corollary of the conceptual model of Phillips et al. (2013) and of earlier
models is that arbuscular mycorrhizal and ectomycorrhizal plants differ in
their carbon and nutrient cycling traits (decomposability and nutrient
release). Data by Cornelissen et al. (2001) were consistent with this
prediction, showing that the mycorrhizal trait is a predictor for the
so-called “fast–slow” spectrum (Reich, 2014). However, the comparison
involved plant species that are not only different with regard to the
mycorrhizal trait but also with regard to a number of other traits. Koele et
al. (2012) applied phylogenetic correction, by comparing sister clades that
differed only in their mycorrhizal habit. Their data, based on 17 pairs of
taxa, indicate no differences in leaf N or phosphorus status after
phylogenetic correction and imply that the mycorrhizal trait is correlated
rather than causally related with these functional differences. Other claims
about differences in N cycling between arbuscular mycorrhizal and
ectomycorrhizal forests in the northern temperate zone may similarly indicate
problems of establishing whether mycorrhizal status is a causally relevant or
only a correlated trait. Thomas et al. (2010) showed a larger positive
response to N deposition by arbuscular mycorrhizal than ectomycorrhizal
trees, suggesting that the ability of the latter group to acquire organic N
was traded off against the possibility of benefitting from increased
inorganic N. Midgley and Phillips (2014) reported higher NO
Averill et al. (2014) reported that competition between ectomycorrhizal fungi/plants and decomposer microbiota results in N limitation for the latter group, which retards litter breakdown and hence results in increased C storage. They noted 70 % more C storage per unit N in ectomycorrhizal forests than in forests dominated by arbuscular mycorrhizal trees and suggested that mycorrhizal status exerts a much larger control over soil C than climatic variables at the global scale. However, this effect appears to be mainly driven by boreal trees (there is a dominance in the database of ectomycorrhizal trees belonging to the Pinales and Fagales, both orders that are characteristic of nutrient-poor soils), and the effect is only marginally significant when the analysis is performed on temperate and tropical forests (Averill et al., 2014). Therefore, plant traits that are inherently associated to mycorrhizal status should further be considered when assessing the key drivers of the differential C : N stoichiometry and C storage.
Nitrogen immobilization in the mycorrhizal mycelium may also have a large impact on the N cycle by reducing mineral N availability for plants. The general claim that mycorrhizal symbioses are beneficial for the plant and that cases of a negative plant performance in the mycorrhizal condition are explained by C costs of the symbiosis was refuted by Côrrea et al. (2012), who concluded that smaller plant size was caused by lower N uptake. Lower N content of the ectomycorrhizal plant could be due to mycorrhiza-driven progressive N limitation (Luo et al., 2004). Alberton et al. (2007) showed this to be the case as plant N content was significantly negatively correlated with hyphal length. Näsholm et al. (2013) showed that immobilization of N in the ectomycorrhizal mycelium can aggravate plant N limitation. They modelled competition between plants and fungi for N in a market model, and concluded that at N limitation the symbiosis does not alleviate plant N limitation but in fact even reduces plant growth (Franklin et al., 2014; Kuyper and Kiers, 2014). Yet, despite this negative effect on plant performance, a non-mycorrhizal strategy is competitively inferior, and therefore trees are trapped as they cannot terminate the association. Because the biomass of the arbuscular mycelium is usually one or two orders of magnitude smaller than that of the ectomycorrhizal mycelium, the amount of N immobilized by the arbuscular mycorrhizal mycelium is sometimes hypothesized to be quantitatively unimportant from the plant's perspective. However, recent studies (Hodge and Fitter, 2010; Grman and Robinson, 2013) indicate that N uptake and immobilization by arbuscular mycorrhizal fungi can also reduce plant performance.
Other pathways through which the mycorrhizal symbiosis may affect soil N cycling are modification of root exudation, root architecture, and fine root turnover (Churchland and Grayston, 2014). It is important to determine which of these differences are caused by the symbiosis and which by other root trait differences among species. For example, Comas et al. (2014) found that, after accounting for phylogenetic relationships, ectomycorrhizal plants have thinner roots and greater branching intensity than arbuscular mycorrhizal plants. It is therefore still a matter of debate whether differences with respect to the mycorrhiza-associated nutrient economy (Phillips et al., 2013) are controlled by the mycorrhizal trait, or whether the mycorrhizal trait is instead correlated with causally relevant plant and climate traits.
An important share of bioavailable N enters the biosphere via biological
fixation of atmospheric N
Nitrogen demand in young successional tropical forest is high. The large
fraction of leguminous plant species that forms symbiosis with N
However, a plant-level physiological perspective counters this assumption, as
numerous experiments have shown that symbiotic S-BNF by leguminous species is
mostly facultative and down-regulated when located in an N-rich environment.
Tropical leguminous species thus have the potential to fix atmospheric
N
The leaky nitrostat model adapted from Hedin et al. (2009). This
model indicates the importance of symbiotic (S-BNF) and free-living (F-BNF)
biological N
A recent spatial sampling method to assess total BNF indicated that tropical
forest BNF is likely much lower than previously assumed (Sullivan et al.,
2014). These authors reported mean rates of total BNF in primary tropical
forests of 1.2 kg N ha
In boreal forests, symbiosis between cyanobacteria and feather mosses
provides an important N input (DeLuca et al., 2002; Gundale et al., 2012). In
peatlands, which contain approximately 30 % of global soil carbon,
While large uncertainties exist regarding the temporal and spatial variability, dominant determinants, and the magnitude and impact of BNF on terrestrial ecosystems functions and services, even less is known regarding its future trajectories in view of global change.
This is an exciting time to study the soil N cycle. Years of surprising findings on unanticipated pathways and mechanisms have expanded the horizons of researchers. These findings have stimulated efforts to develop and test new methods for quantifying these processes. This has resulted in a better understanding of the complexity of soil N cycling processes and in powerful tools for future exploration.
Critical challenges remain. Many processes are still difficult to quantify
and variability and heterogeneity hamper our ability to provide well-constrained estimates relevant to water and air quality issues. We postulate
that addressing the issues formulated above would constitute a comprehensive
research agenda with respect to the N cycle for the next decade.
Particularly, we urge the following blueprint for action:
abandoning the long-disproved but persistent assumption that gaseous N
production in soils is the exclusive result of the interplay between
nitrification and denitrification, and to focus on a better assessment of
alternative pathways; dedicating scientific efforts to the continuing development of improved
techniques for the characterization, quantification, and modelling of
alternative N transformation pathways, eventually in conjunction with
state-of-the-art molecular techniques to determine the functional microbial
communities involved; and considering ecological interactions and trophic cascades as indirect but
essential drivers of soil N cycling, in particular in response to global
change.
Success will require interactions between soil science and other disciplines
that address both smaller (e.g. molecular and microbial) and larger
(ecosystem, landscape, and regional) scales. We believe that such an agenda
would help us meet future challenges of food and energy security,
biodiversity conservation, and climate stability.
All authors contributed to selecting the topics addressed in this
manuscript. P. Boeckx wrote the sections on BNF and N
The authors would like to thank the editors of SOIL for the invitation to write this review. J. W. van Groenigen and I. M. Lubbers were financially supported by an Netherlands Organization for Scientific Research (NWO; grant no. 823.01.016.). P. M. Groffman was partially supported by US National Science Foundation grant (grant no. NSF DEB 0919131). D. Huygens and P. Boeckx acknowledge the EU's Seventh Framework Program for Research (grant no. PIOF-GA-2011-301443) and the Fund for Scientific Research – Flanders (FWO). T. Rütting is financially supported by the Swedish strategic research area Biodiversity and Ecosystem services in a Changing Climate – BECC. Finally, we are thankful to three anonymous reviewers for their extensive and constructive comments on an earlier version of our paper. Edited by: K. Kalbitz