In recent years, the monitoring of groups of soil biota occupying different trophic levels has grown as a method of evaluating trophic interac- tions in soil following environmental impacts and remediation programs (Sochova et al., 2006). In fact, the study of how human activities and environmental change influence soil organisms and their functions drove, to a large extent, the development and evolution of soil ecology (Wallet al., 2005). In one example, Wardle (1995) provided an extensive review and synthesis of the literature pertaining to the effects of tillage and other weed management practices on detritivore food webs in agro-ecosystems; this synthesis is broad in that, in addition to demonstrating the differential sensitivities of soil biota to weed management, it contextualizes soil food web dynamics within the testing of ecological theories. A single soil sample can contain multiple trophic groups, which is appealing given recent sugges- tions that evaluating restoration projects by monitoring solely the plant community inadequately estimates ecosystem recovery (Gratton and Denno, 2006; Levinet al., 2006). In this section, I briefly summarize four topics of recent, general interest in which soil biota represents major response variables: biodiversity loss, invasive species, climate change, and genetically modified (GM) crops. Most of the studies mentioned in this section quantify effects at more than one trophic level, although the distinction is not always made between direct effects at multiple trophic levels and indirect effects that cascade through trophic interactions. These topics could be subjected to extensive literature reviews if expanded to studies where only a single trophic level is considered, and have been in some cases (Dunfield and Germida, 2004; Wardle, 2002; Wolfe and Klironomos, 2005).
3.1. Biodiversity loss
Ecosystem responses to the loss of biodiversity may be due to some intrinsic property of biodiversity (e.g., niche complementarity) or the loss of function- ally important species (sampling effect) (Hooperet al., 2005). Researchers have attempted to determine the potential effects of biodiversity loss on soil food
web dynamics by estimating the responses of soil organisms to different numbers of species in microcosms or field plots, then statistically evaluating the importance of plant species diversity, per se, relative to other factors, such as the traits of the plant species present. When Porazinskaet al. (2003) observed differences in organism abundance within trophic groups associated primarily with the identity of grass species rather than number of species present in monoculture, plant species richness had few effects on abundance within trophic groups. In another study, plant species identity, but not plant species richness, affected the abundance of bacterial-feeding and plant-parasitic nema- todes, but not fungal-feeding nematodes (De Deynet al., 2004). Furthermore, plant-parasitic nematodes were generally greater in the presence of Holcus lanatus,Leucanthemum vulgare, andCentaurea jaceaand reduced in the presence ofPlantago lanceolata, regardless if they were measured in monoculture or in the presence of other plant species, further suggesting the importance of plant species identity in relationships between plant and soil biodiversity (De Deyn et al., 2004).
Synthesizing these results, species identity effects were stronger than diver- sity effects per se. The particular traits that resulted in species identity effects are not clear. Plant development time may have been important since it repre- sented a significant source of variation for all trophic groups except fungal feeders in the study by De Deynet al. (2004). Resource quality homogeneity may be another important factor. Resource quality (C3vs C4photosynthetic pathway) and species origin (native vs exotic) did not represent significant sources of variation in the study by Porazinskaet al. (2003). Gastineet al. (2003) observed no effect of plant functional group (grasses, legumes, forbs) diversity on microbial activity, microbial-feeding nematodes, or predatory nematodes.
However, Wardleet al. (1999) observed complex responses of various groups of soil biota to removal of plant functional groups (C4grasses, C3annual and perennial grasses, legumes, forbs) from field plots, linking these responses to shifts in resource quality.
Caution is necessary when interpreting these results since the effects of plant biodiversity on soil food web dynamics are not clearly understood.
In one case, increased plant species diversity often negatively affected abundance within soil functional groups relative to their abundance in monoculture. Wardle et al. (2003) observed that plant species identity in monoculture had significant effects on abundance of enchytraeids and plant-parasitic, microbial-feeding, and predatory nematodes; however, when plant species and functional group diversity were manipulated, these trophic groups often did not achieve their abundance in the monocultures.
In addition, focusing on abundance within functional groups may underestimate responses when compared with responses within taxonomic groups. Korthalset al. (2001) studied the effect of low diversity (initially four plant species) versus high diversity (initially 15 plant species) on nematode
communities of abandoned agricultural land and observed positive and negative effects of diversity on several plant-, bacterial-, and fungal-feeding nematode genera responding to the treatment; however, the only trophic group showing an overall response was that of bacterial feeders, the abun- dance of which was reduced in the high diversity treatment. In the study by De Deyn et al. (2004), plant species identity influenced taxo- nomic diversity within functional groups but had little effect on overall functional diversity, as indicated by various indices of trophic structure.
Currently, biodiversity researchers are using increasingly complex exper- imental designs in an attempt to approximate the complexity of interactions occurring in reality. Recent studies attempted to identify the linkages, direct and indirect, between aboveground and belowground communities, and to determine the consequences of these linkages for the structure of plant communities and higher trophic levels (De Deyn and Van der Putten, 2005; Hooperet al., 2000; Wardle, 2002; Wardleet al., 2004). These linkages are manifested via general mechanisms, such as detrital inputs and effects on primary productivity, as well as specific mechanisms, such as the role of plant species identity on herbivory and decomposition. For example, above- ground influences on plant growth, such as herbivory, can influence the movement of energy through soil food webs; in an assembled grassland community, increasing defoliation intensity progressively increased shoot- to-root ratios, and reduced root mass, resulting in increased abundance of bacterial consumers and fungal consumers and reduced herbivore abundance in soil (Mikola et al., 2001). Linkages between aboveground and below- ground communities increase the potential for effects in one community as a result of biodiversity loss in the other community, and greater understanding of the strengths of these linkages may allow for more accurate predictions as to how ecosystems will respond to further biodiversity loss.
3.2. Invasive species
Increased human migration and economic trade has resulted in accelerated transcontinental and intercontinental movement of biota. Species introduc- tions into exotic environments can impact native species and ecosystems via a number of direct and indirect ecological mechanisms, including competi- tion, facilitation, and trophic cascades (Whiteet al., 2006). Species invasions in aquatic systems can impact food web structure resulting in diet shifts and trophic cascades (Townsend, 1996; Vander Zandenet al., 1999); however, effects on trophic cascades in terrestrial systems are poorly understood (White et al., 2006). Several studies have looked at the impacts of exotic plant invasions on soil biota and the mechanisms by which soil biota may mediate exotic plant invasions (Belnap and Phillips, 2001; Belnapet al., 2005;
Callawayet al., 2004a,b; Kourtevet al., 2002; Stinsonet al., 2006). Few of these studies, however, have measured responses in soil biota at more than one trophic level and effects are often plant and/or environment specific.
Belnap and Phillips (2001) observed several variable and inconsistent effects of cheatgrass (Bromus tectorum) invasion on various groups of soil biota; they compared invaded and uninvaded plant communities in Utah and observed inconsistent responses among the two native grassland types. Similarly, Yeates and Williams (2001) detected effects of wandering jew (Tradescantia fluminensis) and gorse (Ulex europaeus), but not wild ginger (Hedychium gardnerianum), on fungal-feeding, plant-feeding, and omnivorous nematodes at various sites in New Zealand; however, the direction of the effect often differed among sites. In one Colorado field experiment, phospholipids associated with Gram-negative bacteria were less abundant in the presence of an exotic grass (Caucasian bluestem, Andropogon bladhii) than native grasses, but nematode trophic structure or mite assemblages were not affected (Porazinskaet al., 2003; St. Johnet al., 2006). Pritekelet al. (2006) observed effects of invasive leafy spurge (Euphorbia esula) on mite assemblages in Colorado grasslands, with prostigmatid (in 1 of 2 years) and cryptostigma- tid (in both years) mites being less abundant in leafy spurge-invaded field plots, but not on collembolan abundance; various herbicides had been applied in prior years to invaded plots, but not uninvaded plots, in an effort to control leafy spurge, so effects on mites may not have been entirely or in part due to the invasive plants themselves.
Recent research on the role of phytochemistry in plant invasion suggests that many exotic plants exude bioactive compounds into the rhizosphere, including compounds with antiherbivore, antimicrobial, and antifungal activities (Cappuccino and Arnason, 2006). For instance, glucosinolates exuded from the roots of garlic mustard release isothiocyanates that have negative effects on AM fungi (Roberts and Anderson, 2001; Stinsonet al., 2006) and possibly other fungi (Smolinskaet al., 2003). Negative effects on fungal dynamics might result in cascading effects for trophic interactions involving fungal feeders and the partitioning of energy through the bacterial trophic pathway in stands of garlic mustard; however, more research is required to determine if phytochemicals associated with plant invasions can have such cascading effects on soil food webs.
There are several examples of humans introducing exotic soil organisms into new environments, with consequences for ecosystem functioning. For example, exotic earthworms alter nutrient cycling and plant community dynamics (Bohlenet al., 2004), nonnative ectomycorrhizal fungi facilitated the colonization of South Africa by exotic Pinus spp. (Richardson et al., 2000), and agronomists are concerned about the movement of plant- parasitic nematodes and other plant pathogens. Whether introductions of exotic soil organisms can affect trophic interactions in the invaded range is an open question. Soil food webs are considered to contain high levels of
functional redundancy, with many similar consumers feeding on many similar prey (Seta¨la¨et al., 2005). Such a generalist view of soil trophic interactions suggests that exotic organisms are likely to encounter predators in the invaded range similar to those in their native ranges, buffering the influence of the invader on the rate and direction of energy flowing through a soil food web.
However, if specialist relationships with parasites and pathogens dominate population regulation in the native environment, the invader may exert a significant influence on trophic structure and function in the invaded range.
For example, mole crickets (Scapteriscusspp., Gryllotalpidae) are particularly damaging pests of pasture and turf in the southeastern United States, inadver- tently introduced from South America (Walker and Nickle, 1981). Released from natural enemies in their native range, including an entomopathogenic nematode (Parkmanet al., 1993) and a carabid beetle (Weed and Frank, 2005), mole crickets attain high population densities in the absence of control measures (Adjeiet al., 2003). Populations follow an annual cycle, with adults emerging, laying eggs, and dying in the spring; synchronized mortality likely results in resource pulses, similar to pulses following periodic emergence and mortality of cicadas stimulating bacterial and fungal biomass and altering abundance of various detritivorous macroarthropods (Yang, 2004, 2006).
3.3. Climate change
Climate change is occurring in many forms, including changing temperature and precipitation regimes, prolonged exposure to UV-B radiation, and exposure to increasing concentrations of CO2. Soil biota may be affected by climate change as a direct result of these factors or, indirectly, due to effects on litter quality and quantity, evapotranspiration rates, and nutrient availability (Norby and Luo, 2004).
One approach to studying the effects of climate change on soil biota under field conditions is to artificially increase temperature. Ruess et al.
(1999) simulated climate change by manipulating temperature and nutrient availability at two sites in northern Sweden. After 6 years of manipulation, fertilization stimulated microbial biomass, and active fungal biomass at one of the sites; bacterial- and fungal-feeding nematodes appeared to increase in abundance under the elevated temperature treatment, although differences were not statistically significant. Sohlenius and Bostrom (1999) increased temperature by transplanting soil cores from northern Sweden to various warmer locations throughout Sweden, and then monitored changes in the nematode communities over the course of 1 year. This approach also takes into account that plant community responses to climate change may indi- rectly influence soil biota. Total nematode abundance increased in soils transplanted to warmer sites, and various effects on plant-, bacterial-, and fungal-feeding taxa and predatory taxa were observed. Effects were more
pronounced at sites with less plant cover, suggesting that plant and soil variables may be more important drivers of climate change induced shifts than increased temperature (Sohlenius and Bostrom, 1999).
The most common approach to studying the effects of climate change on soil biota is to measure their responses to elevated concentrations of CO2. Several field studies, at a number of locations worldwide, have attempted this. Sonnemann and Wolters (2005) exposed a German grassland to a 20% increase in [CO2] for 3 years. Bacterial biomass, but not fungal biomass, increased in the elevated CO2 treatment relative to the ambient CO2 treatment over the course of the experiment; trophic cascades were not observed, however, since the abundance of bacterial-feeding nematodes was not affected by the CO2treatment. Root hair-feeding nematodes were stimulated and predaceous nematodes were reduced in the first year of the CO2treatment, but no treatment effects were observed in subsequent years.
In a Michigan forest, Hoeksemaet al. (2000) observed shifts in nematode community composition following exposure of young aspen (Populus tremuloides) to twice-ambient [CO2]; however, responses were dependent on soil origin (low N and soil organic matter vs high N and soil organic matter). As a group, plant-feeding nematodes were more abundant under elevated CO2in the high N, but not the low N soil. Predaceous nematodes, on the other hand, were more abundant under elevated CO2 in only the low N soil.
At a Swiss calcareous grassland, Niklauset al. (2003) observed responses in various components of the soil food web exposed to elevated CO2for 6 years. They observed no significant effects of elevated CO2 on proto- zoans, collembolans, microbial-feeding nematodes, plant-feeding nema- todes, or mites. Predaceous and omnivorous nematodes were less abundant in the elevated CO2treatment.
Hungateet al. (2000) observed that responses to 4 years of elevated CO2 at sandstone and calcareous grasslands in California were sensitive to sea- sonal fluctuations, with stimulatory effects on fungi, flagellate protozoans, and plant-feeding nematodes (sandstone only) detected early in the growing season. Other groups, including bacteria, other bacterial-feeding protozo- ans, and microbial-feeding nematodes were not responsive to the CO2
treatment. In the same experiment, after 6 years of elevated CO2, Rillig et al. (1999a) observed that fungal biomass was greater in the elevated CO2 treatment (sandstone only), as were fungal-feeding microarthropods. No effects on bacterial biomass or the biomass of bacterial-feeding protozoans were detected.
Allenet al. (2005) estimated responses of bacteria, fungi, and their consumers to elevated CO2 in a Californian chaparral over a 3-year period. Biomass of bacteria and their consumers did not respond to elevated CO2. Biomass of fungi and their consumers also did not respond to elevated CO2, except for one
growing season in which mite biomass increased with increasing [CO2].
Fungal biomass was not affected by elevated CO2over this same sampling period; however, when the authors estimated fungal biomass consumed by mites and added this estimate to the standing fungal biomass estimates, it revealed a positive relationship between fungal biomass and [CO2] up to 550 ppm.
Neheret al. (2004) evaluated the abundance and energetics of nematode communities at two locations in the eastern United States. Biomass and respiration of fungal and bacterial feeders decreased in response to elevated CO2. Bacterial feeders decreased in abundance, in contrast to fungal feeders, which increased in abundance, in response to elevated CO2. Responses of predators, omnivores, and root feeders were significant but inconsistent at the two locations.
These results suggest that elevated CO2 stimulates various functional groups in soil, although effects were often inconsistent among studies. Effects on soil biota are usually attributed to changes in the abundance and quality of resources. Elevated CO2can increase net primary productivity, resulting in increased root growth and litter fall (Norby and Luo, 2004). Cotrufoet al.
(1998) reviewed the elevated CO2literature and estimated a 14% reduction of nitrogen concentration of plant tissues associated with elevated CO2; differences varied among plant types with C3plants demonstrating greater reductions than C4 and nitrogen-fixing plants. As a result, responses to elevated CO2 are linked to the availability of other nutrients that limit productivity and influence resource quality; for example, Klironomoset al.
(1996) demonstrated the dependence of elevated CO2effects on soil fertility.
Sagebrush (Artemisia tridentata) plants were grown in a low fertility soil either supplemented with nutrients or unfertilized. Soil food web responses to elevated CO2 were muted in the unfertilized soil and exaggerated in the fertilized soil. CO2 stimulated fungal and bacterial biomass and microbe- feeding microarthropods, but only in the fertilized treatment. Only nema- tode abundance was stimulated by elevated CO2in the absence of fertiliza- tion, and further stimulation was observed following nutrient addition.
Further research is necessary to determine the identity and relative impor- tance of the mediating factors resulting in inconsistencies among studies.
Ecosystem models that estimate soil food web dynamics and nutrient cycling in response to various components of climate change, such as the approach used by Kuijper et al. (2005), will likely be required to reconcile inconsistencies among studies in the observed effects. The model by Kuijper et al. (2005) predicts that fungal-feeding taxa will be more sensitive to climate change than bacterial-feeding taxa, a trend that is supported by many of the previously mentioned studies, and that omnivory and weak trophic links will limit effects of trophic cascades on higher trophic levels. Further experimental research is necessary to test these and other predictions.
When discussing functional group responses to elevated CO2, research- ers less frequently ascribe these responses to interactions with abiotic factors than to trophic cascades. In one example, elevated CO2 lead to reduced evapotranspiration rates in the study by Niklaus et al. (2003), resulting in increased soil moisture and reduced soil aggregate sizes; reductions in predatory and omnivorous nematodes were speculatively attributed to reduced mobility under the modified soil structure. Rillig et al. (1999b), however, observed the opposite effect of elevated CO2on soil structure in California calcareous and sandstone grasslands. Larger soil aggregates (1–2 mm) increased in abundance, as did the stability of aggregates down to 0.25 mm in diameter, in the elevated CO2 treatment; abundances of omnivorous and predatory nematodes were not reported (Hungate et al., 2000; Rilliget al., 1999a).
Extrapolating published results of experiments applying abrupt changes in [CO2] is problematic, however, as responses may be quite different from those under gradual increases in [CO2]. Klironomos et al. (2005) observed effects of elevated CO2on structural and functional properties of AM fungal communities; these effects were only observed following an abrupt increase in [CO2] and not when [CO2] was increased gradually, over a period of 6 years. More research is required to determine if responses in soil food web dynamics to artificial climate change represent adequate estimates of responses to actual rates of climate change over the coming decades.
3.4. GM crops
GM cropping systems have been the focus of recent attention, largely due to public concern about their safety. There are two general mechanisms in which GM cropping systems can affect organisms in and around fields:
(1) effects associated with the modification itself and (2) effects associated with the management of transgenic varieties.
Initially, attention focused on the modified traits themselves. Several studies document shifts in endophytic and rhizosphere microbial communities associated with GM crops (pleiotropic effects or other varietal changes), but few look at effects at higher trophic levels. Donegan et al. (1997) observed increased nematode abundance but reduced collembolan abundance asso- ciated with litter from transgenic proteinase inhibitor I producing tobacco in field soil. In another field experiment, Doneganet al. (1999) observed shifts in soil microbial communities and enzymatic activities associated with a recom- binant nitrogen-fixing bacterium (Sinorhizobium meliloti) and transgenic lignin peroxidase- (but not amylase-)producing alfalfa, but observed no effects on protozoa, nematodes, and microarthropods. Saxena and Stotzky (2001) observed no effect of CryIAb protein, either in corn root exudates or in litter added to field soil, on microbial, protozoan, or nematode abundance. Griffiths et al. (2006) observed that omnivorous nematodes and protists were more