The influence of biotic interactions on soil biodiversity

Authors


* E-mail:
david.wardle@svek.slu.se

Abstract

Belowground communities usually support a much greater diversity of organisms than do corresponding aboveground ones, and while the factors that regulate their diversity are far less well understood, a growing number of recent studies have presented data relevant to understanding how these factors operate. This review considers how biotic factors influence community diversity within major groups of soil organisms across a broad spectrum of spatial scales, and addresses the mechanisms involved. At the most local scale, soil biodiversity may potentially be affected by interactions within trophic levels or by direct trophic interactions. Within the soil, larger bodied invertebrates can also influence diversity of smaller sized organisms by promoting dispersal and through modification of the soil habitat. At larger scales, individual plant species effects, vegetation composition, plant species diversity, mixing of plant litter types, and aboveground trophic interactions, all impact on soil biodiversity. Further, at the landscape scale, soil diversity also responds to vegetation change and succession. This review also considers how a conceptual understanding of the biotic drivers of soil biodiversity may assist our knowledge of key topics in community and ecosystem ecology, such as aboveground–belowground interactions, and the relationship between biodiversity and ecosystem functioning. It is concluded that an improved understanding of what drives the diversity of life in the soil, incorporated within appropriate conceptual frameworks, should significantly aid our understanding of the structure and functioning of terrestrial communities.

Introduction

Over the past several decades, many ecologists have focused on trying to understand why different communities or ecosystems differ in the diversity of organisms that they contain (e.g. Hutchinson 1961; MacArthur & Wilson 1967; Grime 1973; Tilman 1982). Historically, most of the effort devoted to addressing this question in terrestrial ecosystems has focused on aboveground plant and animal species. However, it is well recognized that in most terrestrial ecosystems the belowground biota supports a much greater diversity of organisms than does the aboveground biota. While the vast majority of species of soil biota have yet to be described, several key groups (bacteria, fungi, nematodes and insects) almost certainly contain several hundreds of thousands to millions of species globally (De Deyn & Van der Putten 2005). This diversity is also apparent at local scales; for example, a few grams of soil may contain a few 1000 species of bacteria and several 100 species of invertebrates (Torsvik et al. 1994; Wardle 2002). Although soil ecologists have long recognized the need to understand the mechanisms by which soil biodiversity is maintained (Anderson 1975; Ghilarov 1977; Bruns 1995), it is only comparatively recently that questions relating to the maintenance and functional significance of soil biodiversity have attracted widespread interest from ecologists. As such, there has been a recently growing body of theory about how soil biodiversity may respond to local extrinsic stress and disturbance factors (Wardle 2002; Bardgett et al. 2005b), and on how belowground biodiversity interacts with aboveground biodiversity (Hooper et al. 2000; De Deyn & Van der Putten 2005).

This review addresses the issue of how biotic drivers (both above- and belowground) influence the biodiversity of soil organisms, and explores the mechanisms involved. Although there has been increasing recent research activity on this topic, this work is spread diffusely through two separate bodies of literature (i.e. the soil biology literature and the ecological literature) and there have been few attempts to conceptually synthesize it. Yet, an improved understanding of this topic would greatly enhance our understanding of terrestrial communities and ecosystems in several ways. First, it would assist our knowledge of the factors that influence soil food webs at a finer-scale levels of resolution (e.g. species and genera) than at the broad-scale functional or taxonomic scales that have characterized most soil food web studies to date (e.g. Wardle & Yeates 1993; De Ruiter et al. 1995). Second, it would enhance our understanding of feedbacks between aboveground and belowground biota at the community level and the degree to which the two communities drive each other (Wardle et al. 2004a). Third, it should contribute usefully to our understanding of how soil biodiversity (and loss of biodiversity) affects ecosystem functioning (Hooper et al. 2005), given that an adequate understanding of how diversity affects ecosystem functioning in a ‘real world’ context requires us to understand the factors that regulate biodiversity in the first instance (Grime 1998). Fourth, it should help us to better understand how global change phenomena influences real communities and ecosystems, especially when global change factors alter those biotic factors that influence soil biodiversity.

In this review, I will draw together literature on how biotic factors influence biological diversity within components of the soil community, and in doing so will explore the mechanistic basis by which soil biodiversity is regulated by biotic interactions. For the purposes of this review, taxonomic richness, taxonomic evenness, and functional diversity, will all be recognized as useful measures of soil biodiversity. In this light it is recognized that different measures of diversity, and different levels of resolution for quantifying diversity, may be useful for answering different questions (e.g. some will be more relevant to understanding what structures communities; others will be more relevant for understanding function). It is also recognized that different levels of taxonomic resolution are generally applied to different groups; for example, most studies on larger invertebrates consider diversity at the species level, whereas species level characterization is usually intractable for procaryotes and other measures of diversity are used. Components of the soil community (within which diversity will be considered) will include broad functional and/or taxonomic groups that are widely used in studies of soil ecology (e.g. bacteria, saprophytic fungi, arbuscular mycorrhizal fungi and oribatid mites). In reviewing this topic I will first explore how soil biotic drivers resident in the soil affect soil biodiversity. I will then consider the effects of aboveground biotic drivers on soil biodiversity. Finally I will explore how a conceptual understanding of the influence of biotic drivers on soil biodiversity can enhance our understanding of community and ecosystem ecology, and in doing so will identify key gaps in knowledge and areas in which future research could usefully contribute.

Biotic drivers resident in the soil

Within trophic group interactions

Biotic drivers of soil biodiversity operate over a range of spatial and temporal scales (Fig. 1). At the most local of these scales, diversity of soil organisms can potentially be regulated by interactions that occur among taxa within the same trophic grouping. Such interactions are most likely to have a role in structuring community attributes for those groups of organisms that are resource regulated rather than strongly consumer regulated in the soil food web, as the community structure of these groups of organisms is most likely to be regulated by competition. Different groups of soil organisms differ markedly in the extent to which they are regulated by resource availability as opposed to by their consumers, although most groups are regulated to some extent by both factors (De Ruiter et al. 1995; Wardle 2002). However, most of the evidence for strong competitive interactions in the belowground environment involves fungi and the fungal-based energy channel of the soil food web (Wardle 2002).

Figure 1.

 The hierarchical nature of biotic factors that may influence alpha-diversity within a given group of soil organisms. Here, ‘within group’ can be applied to key trophic groups (e.g. microbe-feeding nematodes) or taxonomic groups (e.g. fungi and oribatid mites). ‘Interference mechanisms’ apply primarily to interactions among fungi, among bacteria, and between bacteria and fungi; ‘predation’ by other soil organisms includes soil faunal consumption of bacteria, fungi and other soil fauna. Note that the right-hand axis is presented as a gradient, because while some organisms and processes operate exclusively in the aboveground or belowground compartments, others operate in both compartments (e.g. plants that have both aboveground and belowground structures; animals that spend time both above- and belowground). Note that the relative positions of these drivers in terms of temporal and spatial scale are only intended to be illustrative not exact, and that some of these drivers operate across a range of scales. Note also that several of these factors may interact to influence diversity (not shown).

Although there is ample evidence in the literature for competitive interactions in soil fungal communities, direct evidence for competition as a determinant of fungal diversity (i.e. through competitive exclusion) is scarce. Indirect evidence, however, is apparent for studies that reveal that the presence of a given saprophytic fungal species (or group of species) on a substrate may effectively exclude colonization of new saprophytic fungal species on that substrate; this has been long recognized for both litter substrates (Garrett 1963) and dead wood (Rayner & Todd 1979; Boddy 2000). Although this issue has seldom been addressed for mycorrhizal fungal species (but see Wu et al. 1999; Kennedy & Bruns 2005), competitive exclusion may also be important in determining ectomycorrhizal fungal diversity at the scale of the individual host plant root tip, at least if the majority of root tips are colonized by a single fungal species. For example, in a study for which tree seedlings were inoculated with eight fungal species either singly or in multiple species combinations, several fungal species persisted only in the monoculture treatment, presumably because they were competitively excluded from colonizing seedlings when in combination with fungal species with superior competitive abilities (Jonsson et al. 2001). Further, there is evidence that nitrogen fertilization can greatly reduce the diversity on plant roots of both arbuscular mycorrhizal fungi (Egerton-Warburton & Allen 2000; Egerton-Warburton et al. 2001) and ectomycorrhizal fungi (Lilleskov et al. 2002), probably in part because greater resource availability favours competitive fungal species that suppress subordinate species.

Another related within-trophic level effect, namely the production by microbes of antibiotics, may also have a role in influencing microbial diversity. Although the role of antibiotics in structuring microbial communities has been long recognized, their role in regulating diversity has seldom been considered. However, there is some theoretical evidence, founded on game theory, which predicts a role for antibiotic production in promoting bacterial diversity (Czárán et al. 2002). Conversely, antibiotic production by resident fungal species may serve to reduce colonization by new species, and it has long been recognized that soils dominated by antibiotic-producing fungi can resist invasion by new fungal species, including those that might operate as plant pathogens (Wicklow 1981).

Resource regulation, and therefore within-trophic level interactions (such as competition), is less likely to have a role in regulating the diversity within groups of soil animals, e.g. soil microarthropods; soil nematodes. Soil animal diversity generally does not show a hump-backed response to increases in disturbance intensity or resource availability (Freckman & Ettema 1993; Wright & Coleman 1993; Wardle 2002), indicating that factors that maximize soil animal biomass or density do not promote dominance of competitive species that reduce subordinate species by competitive exclusion. As discussed later, the relatively low intensity of competition among soil faunal taxa, and dearth of evidence for competitive exclusion, may contribute to the hyperdiverse nature of soil faunal communities.

Interactions within a given trophic level need not necessarily be negative, and there has been increasing recognition over the past decade of the role of positive interactions such as facilitation in structuring ecological communities (Callaway & Walker 1997). Facilitative interactions in soils have most frequently been identified for soil fungi and the fungal-based energy channel. For example, during decomposition of fresh dead plant material, those fungal species that colonize first will break down recalcitrant carbohydrates such as cellulose into simpler forms, thereby enabling subsequent colonization by sugar fungi (Frankland 1969). The net result is greater fungal diversity on the substrate over time. Further, Tiunov & Scheu (2005) found that mixtures of fungal species greatly promoted conversion of cellulose to CO2 despite some components of the mixture being unable to utilize cellulose. This outcome can only be explained by cellulose-degraders breaking cellulose into simpler carbohydrates that can be used by other species, and represents a type of facilitation that would serve to promote the diversity of fungal taxa present.

Interactions across trophic groups

Regulation of major groups of soil biota through predation is widespread in soil food webs (De Ruiter et al. 1995) and there are many examples of regulation of densities of both soil animals and soil microbes by their consumers (reviewed by Wardle 2002). Further, consumption of microbes by soil fauna is likely to be an important driver of soil microbial community structure. Fungal-feeding fauna show a distinct preference for some fungal taxa or hyphal types above others; this has frequently been shown for saprophytic fungi (e.g. Klironomos et al. 1992; Bonkowski et al. 2000), but there is also evidence of this for both mycorrhizal (Klironomos & Kendrick 1996) and plant-pathogenic (Sabatini & Innocenti 2000) fungi. Further, fungal-feeding fauna frequently prefer fungal species with certain functional attributes, such as primary colonizers of plant litter (Klironomos et al. 1992) and those with pigmented hyphae (Maraun et al. 2003). This is analogous to aboveground herbivores showing preference for plant species that are adapted for earlier successional sites and that have suites of traits that differ from those of less-palatable later successional species (Grime 1979; Coley et al. 1985). If selective feeding by fungal-feeding invertebrates can affect fungal community composition and therefore the relative dominance of different fungal taxa, then effects on soil fungal diversity are also likely. The influence of fungal grazers on fungal diversity has seldom been tested, although McLean et al. (1996) found a neutral effect of grazing by mites and collembolan on fungal diversity in microcosms. Whether shifts in fungal community composition resulting from soil invertebrates has generally positive or negative effects on fungal diversity remains little understood, but effects in either direction may be predicted depending on whether the invertebrates selectively suppress competitive dominant or subordinate, species (Holt & Lawton 1994). As a parallel, both positive and negative effects of aboveground grazers on plant diversity have been commonly reported in the literature (Proulx & Mazumder 1998).

Although regulation by consumers of soil dwelling groups other than fungi is well recognized, the implications of this for the community structure or diversity within these groups have seldom been addressed. However, bacterial-feeding nematodes preferentially feed on some bacterial taxa relative to others (Ettema 1998), so there are likely implications for bacterial community composition and diversity. Further, while how the predatory protozoa inhabiting the aqueous component of the soil affect the diversity of their prey remains unknown, there are some relevant studies involving protozoa in aqueous laboratory experiments; such studies have shown that predatory protozoa can exert a variety of effects on the diversity of both prey protozoa (Cadotte & Fukami 2005; Jiang & Morin 2005) and bacteria (Steiner 2005) depending on context. However, the extent to which predation of the major soil faunal groups by their predators may influence their community structure and diversity remains largely unknown.

Non-trophic interactions

Over larger spatial scales, soil invertebrates serve as agents of dispersal of smaller soil organisms, principally microbes. These include bacteria, and propagules of saprophytic, pathogenic and mycorrhizal fungi (Visser 1985; Dromph 2003; Müller et al. 2003; Rantalainen et al. 2004a). Soil microbes may be dispersed either through ingestion and subsequent egestion or through physical attachment to the body surface of the invertebrate. In the case of saprophytic soil fungi at least, those taxa that are most likely to attach to the animal's body surface are often those that are fast growing and establish readily in new microhabitats (Visser 1985), suggesting that dispersal may promote fungal colonization and enhanced fungal diversity of fresh organic matter. This provides an interesting parallel to the colonization–competition trade-off observed across species in plant communities (Rees et al. 2001). The role of detritivorous invertebrates in promoting fungal colonization and diversity through assisting their dispersal has also been recognized in relation to woody substrates (Müller et al. 2003).

While small-bodied soil animals interact with microbes primarily directly through predator–prey interactions, larger bodied soil animals also influence microbes (and small bodied soil animals) through altering the physical environment and creating structures that serve as habitats (Brown 1995; Lavelle & Spain 2001). For example, many microarthropods transform litter into faecal pellets, earthworms create middens and casts, termites build mounds, and ants produce mound nests. These structures all favour some components of the soil biota at the expense of others and therefore potentially influence soil biodiversity. For example, earthworm casts have been shown to promote diversity of fungal species (Tiwari & Mishra 1993; McLean & Parkinson 1998a) and oribatid mites (Loranger et al. 1998; McLean & Parkinson 1998b), although neutral effects of earthworms on diversity of other faunal groups has also been detected (Brown 1995). The most likely mechanism through which earthworms promote diversity of other groups (notably fungi) is through reducing the intensity of competitive interactions, probably by suppressing dominant competitive species within those groups (McLean & Parkinson 1998a). Subterranean ant nests are known to reduce the densities of soil-borne plant pathogens and fungal-feeding nematodes (Blomqvist et al. 2000), with possible consequences for the diversity of organisms within these groups, and for diversity within other groups that may benefit through reduced pathogen and fungal feeder densities. Ant mounds have also been recognized as greatly altering the composition of the soil food web (Laakso & Setälä 1998), and have been shown to promote microbial functional diversity, probably by creating a heterogeneous mosaic of organic matter and promoting habitat spatial variability (Dauber & Wolters 2000).

Plants and their aboveground consumers as biotic drivers

Effects of plant species and species combinations

Although there has been a long history in soil ecology of studying the community structure of soil organisms in relation to soil properties, over the past decade or so there has been a greatly increasing interest in understanding how plant identity, composition and diversity affects the soil community. This has been instigated in part by increasing recent interest in the so-called diversity-function issue (i.e. how organism diversity affects community and ecosystem processes), an issue that requires explicit integration of aboveground–belowground feedbacks (Wardle et al. 2004a; De Deyn & Van der Putten 2005; Hooper et al. 2005).

Plant species differ markedly in the belowground communities that they support, and this has important functional consequences (Hunt et al. 1988; Ayres et al. 2006). As such, soils from under monocultures of different plant species frequently support vastly different levels of diversity of various groups of soil organisms, e.g. nematodes (De Deyn et al. 2004; Vitecroft et al. (2005), mites (Badejo & Tian 1999), endomycorrhizal fungi (Eom et al. 2000; Johnson et al. 2003) and saprophytic microbes (Wardle et al. 2003b). Further, monospecific litters from different plant species have been shown to support differing diversities of decomposer invertebrates (Hansen 1999; Wardle et al. 2006). It is not well understood why different plant species differ in the diversity of organisms that they support, although this is likely to be driven by trait differences between plant species. Plant species with different suites of traits return organic matter of differing qualities to the soil (Díaz et al. 2004), which in turn has likely consequences for attributes of the soil food web (Wardle et al. 2004a), including taxonomic or functional diversity within major groups of soil organisms.

Given that plant species usually coexist in mixtures rather than occur singly, the issue then emerges as to whether multiple plant species combinations support greater diversity of soil biota. It has long been recognized that increasing plant diversity may potentially influence foliar herbivore diversity (Southwood et al. 1979; Siemann 1998) but the issue of how diversity of soil organisms are affected by plant diversity has attracted only relatively recent attention (Hooper et al. 2000; Wardle 2002). In this light, several recent studies have considered how live plant diversity affects diversity within major groups of soil organisms (Table 1). Most experimental studies have found diversity within major belowground groups to be unrelated to live plant diversity even when other community- and ecosystem-level properties are related (e.g. Wardle et al. 1999, 2003b, 2006; Korthals et al. 2001; Porazinska et al. 2003; Carney et al. 2004). This is also supported by observational evidence across gradients of plant diversity (Broughton & Gross 2000; Hooper et al. 2000). This points to above- and belowground biodiversity being somewhat uncoupled, despite aboveground and belowground community structure frequently being linked (Wardle et al. 1999; Hooper et al. 2000). Only two studies have found an effect of live plant diversity on soil diversity, but one of these (Stephan et al. 2000) incorporates an experimental design in which ‘sampling effect’ (Huston 1997) (i.e. the increased probability of including plant species with particular attributes in more diverse mixtures) may explain the outcome. In the other, De Deyn et al. (2004) did find greater soil nematode diversity in diverse plant species mixtures than in corresponding monoculture treatments, presumably because of a greater range of plant-derived resource types present. In contrast, three of four studies that have involved the manipulation of plant litter species richness through litter mixing (Hansen 1999; Kaneko & Salamanca 1999; Armbrecht et al. 2004) have found a positive effect of litter diversity on soil invertebrate diversity, with only one (Wardle et al. 2006) giving inconsistent results. All three studies used full monoculture litter treatments, and the results therefore cannot be explained solely by sampling effect. In particular, in the study of Armbrecht et al. (2004), the use of rarefaction analyses and explicit testing for patterns of spatial dependencies among experimental units makes this work one of the most rigorous demonstrations of an ecological effect of plant species diversity in the literature.

Table 1.   Summary of studies that have explicitly tested whether plant species diversity affects diversity of soil organisms
ReferencePlant diversity manipulationRange of species richnessMonoculture treatments?Ecosystem typeGroup(s) of soil organisms and level of resolutionResponse of diversity of soil organism group(s) to plant diversity
  1. PLFA, microbial phospholipid fatty acids.

Live plant manipulation
 Wardle et al. (1999)Removal of plant species3–8NoGrassland (Waikato, New Zealand)Microbes (PLFAs), nematodes (genera), root feeding arthropods (species)No consistent relationships
 Barros in Hooper et al. (2000)None (plant diversity gradient)10–43NoTropical pasture (Manaus, Brazil)Macrofauna (species)No relationship
 Stephan et al. (2000)Plant species richness1–32NoGrassland (Lupsingen, Switzerland)Microbes (catabolic diversity)Positive response
 Korthals et al. (2001)Plant species richness4 or 15NoGrassland (Planken Wambuis, Netherlands)Nematodes (genera)No response to vegetation treatments
 Porazinska et al. (2003)Plant species richness1 or 2YesGrassland (Konza Prairie, KS, USA)Microbes (PLFAs), nematodes (genera)No consistent trends
 Wardle et al. (2003b)Plant species richness1–9YesGrassland (greenhouse study)Microbes (PLFAs and catabolic diversity), nematodes (genera)Response to composition but not diversity
 Carney et al. (2004)Plant species richness1–5NoTropical vegetation (Costa Rica)Ammonium-oxidizing bacteria (operational taxonomic units)No relationship
 De Deyn et al. (2004)Plant species richness1–8YesGrassland (the Netherlands)Nematodes (genera)Positive effect of diversity, but composition more important
Plant litter manipulation
 Kaneko & Salamanca (1999)Species richness of leaf litter1–3YesMixed forest (Matsue, Japan)Oribatid mites (species)Positive effect of diversity
 Hansen (2000)Species richness of leaf litter1–7PartialDeciduous forest (North Carolina, USA)Oribatid mites (species)Positive effect of diversity
 Armbrecht et al. (2004)Species richness of twig litter1 or 8YesCoffee agroecosystem (Colombia)Twig-nesting ants (species)Positive effect of diversity
 Wardle et al. (2006)Species richness of leaf litter1–8YesMixed forest (Banks Peninsula, New Zealand)Nematodes (genera), macrofauna (species)No consistent trends

For plant diversity to influence soil diversity, the most likely mechanism would involve: (i) a more diverse plant community returning a more heterogeneous mixture of resources to the soil; and (ii) this more diverse resource mixture in turn influencing soil biodiversity, e.g. through promoting greater resource partitioning among the component soil organisms (Hooper et al. 2000; Wardle 2002). It is unclear as to whether a more diverse mix of plants necessarily produces a more heterogeneous mix of substrates, as a single plant species might be able to produce as wide a range of tissue types (in terms of quality and chemical composition) as would a whole range of plant species (Hooper et al. 2000). Further, in the only study to date to specifically investigate the ecological consequences of substrate chemical diversity (Orwin et al. 2006), it was found that increasing the diversity of carbohydrate types added to the soil from one to eight compounds had no consistent effects on either soil processes or on soil microbial functional diversity. This was despite the composition of carbohydrates having important belowground effects. Thus, even if live plants in a more diverse herbaceous community released a wider diversity of compounds to the rhizosphere than did a less diverse community, then this would not necessarily promote soil biotic diversity. Conversely, the diversity of physical types of microhabitats in forested systems has been shown to promote diversity of both soil invertebrates (Anderson 1978; Barker & Mayhill 1999) and ectomycorrhizal fungi (Dickie et al. 2002). Different invertebrate species can preferentially colonize litters with different physical structures (Hansen 1999) and different chemical compositions (Wardle et al. 2006), meaning that in plant communities that support a thick litter layer (notably in forested systems), a greater diversity of litter types might indeed promote invertebrate diversity. If this is the case, then plant diversity would promote soil diversity mostly in those systems in which a thick layer of litter was maintained and in which plant species drove soil organisms through litter production rather than through rhizosphere exudation.

There are also other possible mechanisms through which plant diversity may affect soil diversity, although these have not been addressed to date. For example, increasing plant diversity has been shown in several instances to promote plant productivity, particularly at low diversity levels (Hooper et al. 2005). It is recognized that plant production, and therefore the quantity of resources entering the soil, promotes subsets of the soil biota, particularly those that are regulated primarily by resource availability (Mikola & Setälä 1998). The implications of this for soil biodiversity are not well understood, although Degens et al. (2000) found that soils with a greater amount of basal resources present also supported microbial communities with a greater functional diversity. Further, there is recent evidence that the diversity of bacteria in aquatic systems (some of which also occur in the aqueous component of soils) can show hump-backed relationships with ecosystem productivity (Horner-Devine et al. 2003; Kassen et al. 2004).

Effects of foliar herbivores and their predators

Foliar herbivores are being increasingly recognized as important indirect drivers of the belowground subsystem, and both positive and negative effects of herbivores have been reported, depending on the mechanism through which the herbivores affect the decomposer subsystem (Bardgett & Wardle 2003). Further, there are several examples of studies that have shown both positive and negative effects of herbivory on plant diversity (Proulx & Mazumder 1998). The indirect effects of foliar herbivores on soil biodiversity remains relatively unstudied, although a growing handful of studies have addressed this question or at least provided relevant data (Table 2). Most have focused on browsing mammals, although one has investigated the effects of localized invertebrate herbivory on the diversity of soil organisms (Wardle et al. 2004c). Here, experimental planted mesocosms were set up with eight aphid species, added singly or in various multiple species combinations. Although aphid species identity had important effects on some belowground groups, the diversity of only one such group (microbe-feeding nematodes, which are secondary consumers) was affected by the addition of aphids or aphid species identity. Further, no group of soil organisms was consistently affected by aphid species diversity. The mechanistic basis by which aphid species identity might indirectly influence the diversity of a group of soil-dwelling secondary consumers in the soil is unclear, but presumably involves alteration by aphids of the nature of plant-derived resources entering the decomposer food web.

Table 2.   Summary of studies that have explicitly tested whether foliar herbivores indirectly influence the diversity of soil organisms
ReferenceFoliar herbivore typeEcosystem type consideredGroup(s) of soil organism and level of resolutionResponse of diversity of soil group(s) to presence of herbivores
  1. PLFA, microbial phospholipid fatty acids.

Browsing mammals
 Leetham & Milchunas (1985)CattleShortgrass prairie (Colorado, USA)Microarthropods (species)No response
 Wall-Freckman & Huang (1998)CattleShortgrass prairie (Colorado, USA)Nematodes (genera)No response
 Suominen (1999)ReindeerBoreal forest (northern Finland)Gastropods (species)Reduction by herbivores
MooseBoreal forest (Finland and Sweden)Gastropods (species)Enhancement of relative diversity
 Stark et al. (2000)ReindeerBoreal forest (Muonio, Finland)Nematodes (genera)Enhancement by herbivores
 Dombos (2001)SheepGrassland (southern Hungary)Collembola (species)Reduction by herbivores
 Wardle et al. (2001)Range of mammals, mostly deerRainforest (throughout New Zealand)Nematodes (genera)No response
Gastropods (species), beetles (species), diplopods (family)Generally reduction by herbivores
 Clapperton et al. (2002)CattlePrairie (southern Alberta, Canada)Mites (family)Reduction by herbivores
 Gehring et al. (2002)Range of herbivoresTropical forest (northern Queensland, Australia)Arbuscular mycorrhizal fungi (species)Reduction by herbivores of spore diversity
Invertebrates
 Wardle et al. (2004c)AphidsGrassland (mesocosm study)Microbes (PLFAs and catabolic diversity), herbivorous nematodes (genera)No effect
Microbe-feeding nematodes (genera)Sometimes positive herbivore effects

Studies involving herbivore exclusion plots in the field (Table 2) have most often found negative effects of mammals on the diversity of various groups of soil biota, e.g. arbuscular mycorrhizae (Gehring et al. 2002), microarthropods (Dombos 2001; Clapperton et al. 2002) and macrofauna (Suominen 1999; Wardle et al. 2001). However, two of the three studies that have considered how browsers influence soil nematode diversity have found no detectable effect (Wall-Freckman & Huang 1998; Wardle et al. 2001) while the third found some positive effects (Stark et al. 2000). The mechanism by which mammals reduce soil biotic diversity is unclear, but could involve changes in vegetation composition and the relative dominance of different plant species (Gehring et al. 2002) or physical disturbances to the soil caused by the mammals (Dombos 2001; Clapperton et al. 2002). In this light, Wardle et al. (2001) found for a range of forested sites throughout New Zealand that browsing mammals consistently reduced the abundance and diversity of various groups of soil macrofauna, apparently as a result of mammal-induced soil disturbance. These effects occurred even when the mammals did not adversely affect vegetation density, vegetation diversity or diversity of litter types present on the forest floor. Previous work suggests that soil faunal diversity is generally only adversely affected by soil disturbance (in contrast to predictions of the ‘intermediate disturbance’ hypothesis) (Wardle 2002; Bardgett et al. 2005b), and in this light, consistent adverse effects of mammal-induced soil disturbance on diversity of meso- and macrofauna may be expected. These effects are less likely to adversely affect the diversity of smaller-bodied soil organisms such as nematodes that inhabit water films in small pore spaces because these microhabitats are less likely to be influenced by physical disturbance (Stark et al. 2000; Wardle et al. 2001).

Although indirect effects of aboveground secondary consumers (predators of herbivores) on the belowground subsystem have been seldom studied, aboveground trophic cascades (where predators indirectly affect plants by influencing herbivore densities) can determine the nature of organic materials entering the soil, with implications for the soil community (Croll et al. 2005). In this light, two studies have provided relevant data on how aboveground predators may influence decomposer diversity through cascading trophic effects. In the first, Dyer & Letourneau (2003) found that manipulation of a top invertebrate predator (which feeds upon herbivorous and possibly detritivorous invertebrates), associated with a tropical shrub, did not influence the diversity of shrub-associated decomposer fauna in any of three trophic levels. Diversity within each level was instead affected by the amounts of resources made available for the decomposer subsystem. In the second, Wardle et al. (2005) established mesocosms with plants and aphids, and with or without predators of the aphids. The presence of top predators influenced the plant community, and this in turn induced a cascading effect on the abundances of organisms in each of three consumer trophic levels of the soil food web. However, there was no consistent effect on organism diversity within any of these three trophic levels, indicating that soil biodiversity can be highly buffered against interactions that occur aboveground.

Effects of aboveground changes over successional time

At larger temporal scales, e.g. in the order of decades to millennia, changes in plant community composition due to vegetation succession also cause shifts in soil communities. These changes result from plant species replacing each other over time through both biological (e.g. facilitation and interference) and abiotic (e.g. changes in nutrient supply from parent material) mechanisms. Feedbacks between aboveground and belowground communities at these larger scales have important implications for both community- and ecosystem-level properties (Bardgett et al. 2005a). Few studies have considered how soil communities change during vegetation succession, but there is evidence that during the first few decades of succession increases in plant diversity and community productivity are closely matched by increases in soil food chain length (Verhoeven 2002), and the diversity of soil microbes (Sigler & Zeyer 2002; Tscherko et al. 2003), soil invertebrates (Dunger et al. 2004; Hodkinson et al. 2004) and mycorrhizal fungi (Jumpponen et al. 2002). Further, these changes in soil community attributes appear to be deterministic and consistent across successions, at least at a regional scale (e.g. hundreds of kilometres) (Hodkinson et al. 2004). However, after the initial phases of succession soil diversity need not continue to increase; Chauvat et al. (2003) found for forest rotations that over time the diversity of microbes and springtails actually declined, apparently as a result of declining habitat diversity and food abundance.

In the order of millennia or more, ecosystems without significant disturbance can enter a state of ecosystem ‘retrogression’, in which vegetation succession ultimately proceeds to a highly unproductive state in which available nutrients, principally phosphorus, become very limiting (Walker & Syers 1976). Although the implications of this for the decomposer subsystem have been seldom explored (but see Vitousek 2004; Wardle et al. 2004b), there is evidence that microbial community structure may undergo substantial changes during retrogression, such as from fungal to bacterial domination (Wardle et al. 2004b). Little is known about how soil microbial or faunal diversity changes during retrogression, although Williamson et al. (2005) found microbial functional diversity to decline and nematode diversity to remain invariant across a 600 000 year retrogressive succession on terraces varying in time since uplift from the ocean.

Conceptual insights derived from understanding biotic drivers

It has previously been recognized that densities of soil organisms are regulated by a diverse range of factors that operate in a hierarchical manner across a range of spatial (Ettema & Wardle 2002) and temporal (Bardgett et al. 2005a) scales. It is apparent from this review that the diversity of life in the soil is also driven by a hierarchy of biotic factors, ranging in spatial scale from localized interspecific interactions at one end to vegetation succession on landscapes at the other end (Fig. 1).

The wide spectrum of biotic factors driving soil biodiversity may help explain the hyperdiverse nature of soil communities. Over 30 years ago, Anderson (1975) drew attention to the ‘enigma of soil biodiversity’ [a belowground analogue of Hutchinson's (1961)‘paradox of the plankton’], which highlights the issue as to how it is possible for soils to maintain a high diversity of organisms without biotic mechanisms such as competitive exclusion reducing diversity. It has previously been proposed that this diversity can be explained in part by the diversity of niche axes in the soil promoting considerable resource partitioning amongst trophically equivalent organisms (Faber & Joose 1993; Wardle 2002), coupled with a very low intensity of competition within large subsets of the soil biota (Wardle 2002). The present review also addresses whether biotic drivers (both below- and aboveground) may also serve to promote soil biodiversity across a range of spatial and temporal scales (Fig. 1). Although there is a dearth of data with which to explicitly or quantitatively test the importance of the various mechanisms in Fig. 1 in driving soil biodiversity, the available evidence suggests that many of these mechanisms serve to maximize soil biodiversity at a range of spatial and temporal scales, thus contributing to the high biotic diversity present in the soil subsystem. Further, the importance of many drivers that operate aboveground is likely to vary across ecosystems, and an enhanced understanding of them should contribute usefully to determining how and why soils differ in the biodiversity that they support.

Aboveground–belowground feedbacks

This review has focused on drivers of soil biodiversity, including those that are based aboveground. However, the aboveground and belowground components strongly influence each other, suggesting that there may be important feedbacks between the two. Just as there are several examples of how aboveground biota (plants and their consumers) affect soil communities (Tables 1 and 2), there are also examples of how soil biota affect aboveground communities, and these effects may propagate through several trophic levels (reviewed by Van der Putten et al. 2001; Wardle 2002). Further, components of the soil biota may influence plant diversity, either positively (e.g. Van der Heijden et al. 1998; De Deyn et al. 2003) or negatively (e.g. Brown & Gange 1989; Hartnett & Wilson 1999), depending on whether or not they promote subordinate plant species relative to dominant species (Ureclay & Díaz 2003). One study has also shown a positive response of plant species diversity to arbuscular mycorrhizal fungal diversity (Van der Heijden et al. 1998), although the mechanistic basis of that result remains unclear. Nevertheless, given the range of patterns reported in the literature on how belowground biota affects plant diversity and how aboveground biota affects belowground diversity, both positive and negative feedbacks between above- and belowground biodiversity are theoretically possible.

It is expected that the strength of feedbacks between above- and belowground biodiversity will vary depending on which groups of soil biota are considered. There is increasing recent recognition that those soil groups most directly associated with plant roots (e.g. mycorrhizal fungi, root pathogens and herbivores) show a higher degree of specificity than has been previously supposed (Wardle et al. 2004a). This means that a greater range of plant species should be able to support a greater range of root-associated biota, although the evidence for or against this prediction at the plant community level is scarce. However, if this is the case, then it provides evidence for niche partitioning among components of the soil biota, and may provide a useful test of niche-based models and approaches developed for aboveground and other systems (e.g. Schoener 1974; Tilman 1982). Further, if components of the root-associated biota show high specificity, then the community structure (including diversity) of that biota should in turn strongly affect community structure aboveground by affecting the relative densities of different plant species (e.g. dominant vs. subordinate) (Stampe & Daehler 2003; Ureclay & Díaz 2003), potentially influencing plant diversity. A different pattern is expected when decomposer soil biota are considered. While decomposers do show some resource specialization, this is driven more directly by differences among plant-derived substrate types than by differences among plant species, and decomposers are much greater generalists than root-associated biota in their associations with plant-derived resources. Further, different species combinations of decomposer biota are only likely to indirectly differ in their effects on the plant community, simply through differentially influencing the supply of plant-available nutrients from the soil. For this reason, it is reasonable to predict stronger links and feedbacks between plant diversity and the diversity of root-associated biota than between plant diversity and decomposer diversity, although insufficient data currently exists to adequately test this prediction.

A greater understanding of linkages between aboveground composition and decomposer diversity requires an understanding of the importance of regulation by resource availability, relative to regulation by consumers in higher trophic levels, in structuring soil food webs. The relative importance of these factors has frequently been considered in the context of decomposer food web theory, and most groups are structured by both to varying degrees (e.g. De Ruiter et al. 1995; Mikola & Setälä 1998; Moore et al. 2003). However, most of this work has focused at a coarse level of resolution (i.e. major functional or taxonomic groups) rather than at the finer degrees of resolution needed for understanding drivers of soil biodiversity. For example, significant responses of the diversity of a group of organisms to plant community attributes might be detectable at the species (or even the genetic) level even when diversity at a coarser level of resolution is unresponsive. The question of whether plant community attributes drive decomposer diversity, and the extent to which this is propagated through the decomposer food web, first requires knowledge of the extent to which the diversity in each decomposer trophic level is regulated by resource or food availability. However, it is likely that the diversity within some trophic levels is strongly resource regulated while that of others is less so, and this explains why the diversity of different trophic groupings may greatly vary in their response to aboveground biotic drivers (Wardle et al. 2003b; De Deyn et al. 2004) (Table 1).

Biodiversity and ecosystem functioning

Understanding the biotic drivers of soil biodiversity is directly relevant to the so-called diversity-functioning issue, which is focused on determining whether organism diversity influences key ecosystem properties such as decomposition, nutrient flow rates, productivity, and resistance and resilience to disturbances. This is because such information informs on whether changes in biodiversity caused by these drivers at the ecosystem level are sufficient to impact upon these functions. A growing number of studies have investigated how the biodiversity of belowground organisms may impact upon key ecosystem functions (Hättenschwiler et al. 2005; Hooper et al. 2005). There is evidence indicating possible effects of mycorrhizal fungal diversity and/or composition on ecosystem productivity at the species level (e.g. Van der Heijden et al. 1998; Jonsson et al. 2001) and even at the genetic level (Koch et al. 2006). Plant productivity has also recently been shown to respond to the community structure (although not species richness) of root herbivores (Brinkman et al. 2005). Further, it is likely that those processes that can only be carried out by a small number of specialized taxa (e.g. nitrification and symbiotic nitrogen fixation) are susceptible to losses in the diversity of these taxa (Wardle 2002). Meanwhile, decomposer and nutrient mineralization processes are carried out by a diverse range of taxa, suggesting a significant level of functional redundancy in the decomposer biota. In this light, loss of diversity of soil biota through experimental disturbances has been shown to have little effect on decomposer processes (Degens 1998; Griffiths et al. 2000), and belowground processes driven by soil fauna have been shown to be driven mainly by species identity and functional dissimilarities among faunal species than by species richness (Laakso & Setälä 1999; Heemsbergen et al. 2004). Some studies have shown decomposer processes to be influenced by species diversity of both saprophytic fungi (Setälä & McLean 2004; Tiunov & Scheu 2005) and invertebrates (Liiri et al. 2002), but these effects are likely to become saturated at low levels of diversity, probably well below those normally likely to be encountered in real communities. Shifts in soil biodiversity, such as those caused by the drivers listed in Fig. 1, are most likely to significantly influence ecosystem functioning for only those processes that are performed by a restricted number of taxa, and are unlikely to strongly affect decomposition or nutrient mineralization.

One habitat in which the diversity of decomposer organisms may potentially influence ecosystem functioning is in plant litter rather than in mineral soil. The issue of how mixing of litter from different plant species in turn influences decomposer processes has been attracting recent attention (Gartner & Cardon 2004; Hättenschwiler et al. 2005) and, as is apparent from Table 1, increasing litter diversity may increase the diversity of microhabitats and therefore decomposer faunal diversity. The presence of decomposer fauna has in turn been shown in two recent studies to exert important effects on how litter mixing affects litter decomposition (Hättenschwiler & Gasser 2005; Schädler & Brandl 2005). Whether the effects of litter diversity on invertebrate diversity in turn influence litter decomposition and nutrient dynamics remains unstudied, but such effects remain possible at least if litter mixing promotes significant resource partitioning (and hence resource use complementarity) amongst different invertebrate taxa.

Concluding remarks

Although there are a growing number of studies that have published data on how soil biodiversity may be influenced by biotic drivers, much remains to be understood with regard to the mechanistic basis of these effects. Such an understanding is required for developing general principles about what drives soil diversity. As discussed earlier, one area that remains poorly known but that would strongly contribute to this understanding at local spatial and temporal scales involves the relative importance of resource regulation and regulation by consumers in higher trophic levels in regulating soil community structure within trophic groups. At larger scales, further research is also needed about relative role of different mechanisms through which larger organisms (e.g. plants and earthworms) affect the diversity within groups of smaller organisms, including through altered resource supply (quantity and quality), modifications of habitat structure (including habitat heterogeneity) and changes in the soil disturbance regime. Coupled with this is a need for understanding of how abiotic factors that can be altered by larger organisms may in their own right affect the diversity of soil organisms.

There is a large and growing body of ecological theory about the maintenance and regulation of community diversity that has been developed mainly for aboveground and aquatic systems. One of the key challenges in soil ecology continues to be to better understand the extent to which these theories apply to soil; to achieve this would in turn provide a useful test of the generalities of these theories. For example, well-known theories about diversity responses to disturbance and stress gradients (e.g. Grime 1973; Connell 1978; Huston 1979) have seldom been explicitly addressed for soils (Wardle 2002; Bardgett et al. 2005b). Similarly, as described earlier, there is much to be found about how soil organisms and their diversity conform to niche-based theories (e.g. Grime 1979; Tilman 1982; Rees et al. 2001) or the extent to which niche partitioning may explain the hyperdiverse nature of soil communities. There have also been few attempts to apply the principles of island biogeography theory to understanding soil biodiversity (but see Wardle et al. 2003a). Further, while trait-based approaches in plant ecology (and notably the issue of tradeoffs among species between those with conservative vs. acquisitive resource strategies) are increasingly recognized as important for understanding community and ecosystem processes (e.g. Díaz et al. 2004), the extent to which similar principles apply to major groups of soil organisms has been seldom addressed (but see Faber 1991, with regard to soil faunal communities).

There are also several topics that have been gaining significant recent attention in the ecological literature, and to which a greater understanding of soil biodiversity may usefully contribute. Some are discussed earlier; other examples are as follows. First, there has been increasing interest in how invasive organisms influence communities of native organisms in the invaded habitat. A handful of studies have shown that the diversity within key groups of soil organisms may be influenced by community invasions of alien plants (Belnap & Phillips 2001; Belnap et al. 2005), earthworms (McLean & Parkinson 1998a,b) and browsing mammals (Wardle et al. 2001). Although such studies point to important above- and belowground consequences of biological invasions (Wolfe & Klironomos 2005), too few have been performed to develop general principles about how alien organisms influence soil biodiversity or the underlying mechanisms through which this happens. Second, a growing number of studies have reported responses of belowground community structure to key atmospheric drivers of global change such as elevated CO2 (Niklaus et al. 2003; Yeates et al. 2003), temperature (Bardgett et al. 1999) and nitrogen deposition (Egerton-Warburton et al. 2001; Wiemken et al. 2001). There are several mechanisms by which these global drivers may affect soil biota, both directly (through affecting the soil organisms themselves) and indirectly (by influencing plant growth characteristics, plant community structure, resource input to the soil) and therefore potentially soil biodiversity. However, it is unclear as to which of these mechanisms are important in driving soil biodiversity and when a better understanding of this would assist prediction of how biological communities respond to global change. Third, a better knowledge of what drives soil biodiversity would contribute to greater understanding of the ecological consequences of habitat fragmentation and habitat size. Recent studies point to low turnover of fungal diversity across large geographical distances (Green et al. 2004) as well as to inconsistent effects of habitat size and isolation on diversity of decomposer biota (Wardle et al. 2003a; Rantalainen et al. 2004b) and arbuscular mycorrhizal fungi (Mangan et al. 2004). However, whether the size and isolation of fragments containing groups of organisms known to drive soil biota (e.g. vegetation) or habitat patch size, influences soil biodiversity, remains little understood. Fourth, specifically in relation to bacteria, it is recognized that habitat properties (e.g. heterogeneity and resource availability) are important drivers of bacterial diversification through adaptive radiation, and that that this radiation can occur over the matter of days (Kassen et al. 2004). The issue of how biotic drivers of resource and habitat properties (e.g. plants and earthworms) in turn drive the evolution of bacterial populations and thus their diversification remains unexplored, but investigations in this area would greatly aid our understanding of the interface between ecology and evolution in the soil (Crawford et al. 2005).

Finally, our understanding of what drives soil biodiversity has lagged well behind that for aboveground and many aquatic systems, in part because of the physical complexity of the soil environment and sheer numbers of types of soil organisms present. The main impediment to an improved understanding of soil biodiversity is in actually characterizing this diversity. For soil fauna and some microbial components (notably fungi), this characterization requires access to specialist taxonomic skills, and there are a diminishing number of individuals worldwide who have the necessary expertise. Bacteria pose a particular challenge, as they do not lend themselves as easily to the classic ‘species’ concept as do eucaryotes. Yet, they are undoubtedly an important group to understand from both a community-level and an ecosystem-level perspective, because their diversity probably rivals that of most groups of eucaryotes, and because of their key role in a range of biogeochemical processes. Despite the obvious difficulties in quantifying soil microbial diversity (especially at fine levels of taxonomic resolution), methodologies that might yield useful insights in the future are rapidly evolving. In particular, although currently available molecular approaches have limitations in reliably assessing the relative abundance of different microbial taxa (e.g. across systems or experimental treatments), over time these methods will probably develop to a level where they can more successfully achieve this. This will in turn greatly increase the range of ecological questions to which they can be applied. In any case, an improved knowledge of what drives the diversity of life in the soil, incorporated within appropriate conceptual or theoretical frameworks, should substantially enhance our knowledge of the structure and functioning of real terrestrial ecological communities.

Acknowledgements

Ian Dickie, Tadashi Fukami and Gregor Yeates made helpful comments on earlier drafts, and three anonymous referees and J. Chase made helpful suggestions on a later version.

Ancillary