sign in sign up
<<
Home , Ecosystem

CHAPTER 3.5 The of Microbial Communities and Ecosystem Processes: Implications for and Ecosystem Models

M ark A. B radford and Noah F ierer

Although we might have “faith” that the true eco- 3.5.1 Predicting environmental response lies somewhere within the predic- responses of soil processes tions from the models, predictions from We are entering a period of rapid and pronounced mathematical models that rely on empirical (or sta- environmental change. Understanding how this tistical) relationships may be much more similar. change will in! uence soil communities is essential Empirical models comprise the majority of ecosys- for accurate prediction of future , biogeo- tem models used to predict soil biogeochemical chemical cycles, and human well-being. Such relia- processes (see Box 3.5.1 ). Our con" dence in the pre- ble prediction is the fundamental test of scienti" c dictions by empirical models of future ecosystem understanding (i.e. can we use knowledge of under- response relies on the assumption that observa- lying mechanisms to predict accurately how phe- tions of regulatory variable-process rate relation- nomena change across and ). Predicting hold across space and time (Reynolds et al. how respond to environmental change will be 2001 ). So, we are faced with a dilemma. Do we central to adaptive management of ecosystems, develop and employ mechanistic models that may given our reliance on soils for sustained pro- be more accurate but also more uncertain or do we duction, water puri" cation, storage, and follow the more traditional approach, relying on nutrient retention. This ability to predict with cer- empirical models that may predict ecosystem proc- tainty how environmental change will in! uence esses with more certainty but less accuracy? There soil processes requires mechanistic understanding is likely no single right answer to this question of how soils work. Developing this mechanistic when developing models to predict how ecosys- understanding is perhaps the “Grand Challenge” tems will respond to environmental change. A fam- for soil ecologists if we are to advance our basic ily of models that covers this continuum of understanding of soils and apply this knowledge to approach—between empiricism and mechanism effective environmental management. –will permit us to determine where model predic- Mechanistic understanding improves our con " - tions agree. Where they disagree likely indicates dence in predictions of the future that extrapolate the greatest uncertainty in model predictions, and observed relationships between regulatory varia- it is in addressing these disagreements that future bles (e.g. temperature) and process rates (e.g. soil research might yield most insight. respiration) (Reynolds et al. 2001 ). Perhaps para- The soil predominately relies doxically, there may be much less agreement in on more empirical models and this reliance has predicted outcomes between ecosystem models repercussions beyond the discipline and can in! u- when more mechanistic understanding is included. ence global policy. For example, the soil submodels

Soil Ecology and Ecosystem Services. First Edition. Edited by Diana H. Wall et al. © 2012 Oxford University Press. Published 2012 by Oxford University Press. 190 AND ECOSYSTEM SERVICES

Box 3.5.1 Comparison of mechanistic and empirical models

Most commonly it is the perspective of an individual, not ecological and biogeochemical sciences—using regression model , which determines whether a model is relationships derived from data. When we extrapolate these classed as mechanistic or empirical. In reality, these two relationships we are assuming they are robust across space terms distinguish opposite ends of a continuum. Mechanistic and time, even if the physical, chemical, and biological models are designed to explain outcomes with formulations conditions under which the data were collected differ that represent actual physical, chemical, and biological markedly to those in the new location or time for which we processes, such as increased catalytic rates of respiratory make the prediction. With mechanistic models the argument enzymes as temperature increases. Empirical models are is that—if we understand and represent the underlying often referred to as “correlative or statistical models” and mechanisms accurately in the model formulations—then we are designed to predict—as opposed to explain—outcomes. can make robust predictions for other locations and . To enable this prediction, formulations are used that best The challenges with mechanistic models are thus twofold: describe relationships between two measured variables, we have to both understand the mechanisms and then be such as temperature and rates, with no able to represent them accurately when formulating models. necessary regard for the underlying mechanism. Empirical These challenges are not trivial, and may delimit the very models are typically parameterized—at least in the edges of our understanding of soils.

embedded within the coupled atmosphere– actually at odds with the fact the models are usu- models used by the Inter- ally also implicitly rooted in fundamental biologi- governmental Panel on (IPCC) to cal understanding. But what it does mean is that predict future climate are largely empirical. The we assume we can omit (and within “biol- IPCC has high uncertainty in some of the assump- ogy” we include ecology) from the models as a tions inherent to the embedded soil models, and force that could shape responses of soil processes this leads to uncertainty in climate prediction. For to perturbation. That is, we assume physicochemi- example, there is marked uncertainty as to whether cal controls on soil processes are unaffected by warmer temperatures will lead to the net loss of soil phenotypic or genotypic changes in biota, and carbon that then generates a positive to biotic interactions, across time and space. Yet at global warming ( Denman et al . 2007 ). Given the vast the same time we’re aware that soil —at published literature on this topic (e.g. Davidson & least in part—mediate many soil processes (e.g. Janssens 2006 ; Bradford et al . 2008 ; Allison et al . respiration and nitri" cation). The omission of biol- 2010 ), it is unsettling that we—the soil ecology com- ogy from soil models is perhaps best illustrated munity—cannot reach agreement on the question through example. Take the approach of using a of how stocks will respond to tempera- single, " rst-order decay equation to estimate how ture change, especially given the societal need for the rate of soil carbon varies with resolution. One plausible reason for this disagree- temperature (e.g. Kirschbaum 2004 ; Knorr et al . ment is that biological and ecological mechanisms 2005 ). The equations are typically parameterized are typically represented implicitly—not by data collected across a large range in one or explicitly—in soil models ( Bahn et al . 2010 ). more controlling factors, in this case temperature. Variation in temperature is generated by use of 3.5.2 Misplaced physics envy broad spatial gradients (e.g. elevation, latitude) or in soil models short temporal changes (e.g. incubators, daily vari- ation). Using these statistical relationships to Soils are not merely physical and chemical entities, project how soil carbon decomposition responds yet we generally model soil processes by assuming to environmental change assumes that the biology they can be represented simply as a product of is invariant across space and time. In other physics and . This assumption is not words, understanding of biological mechanism is THE BIOGEOGRAPHY OF MICROBIAL COMMUNITIES AND ECOSYSTEM PROCESSES 191 unnecessary for accurately predicting how soil exhibit identical -history strategies (e.g. Clark processes will respond to environmental change. 2009 ; Rosindell et al . 2010 ). Yet recent work high- Social scientists are well aware of the shortcom- lights that those patterns attributed to this assump- ings of using principles from classical physics to tion of identical strategies can equally well be predict response of human (and hence biological) explained by statistical artefact introduced when to perturbation (for an excellent commen- dealing with large data sets, and that the same data tary see Bernstein et al . 2000 ). Application of princi- sets used to make this “probability argument” can ples such as 1) invariance, 2) probability, and actually only have arisen through complex (i.e. eco- 3) simplicity “promise” a level of certainty in pre- logical niche) processes ( Warren et al . 2011 ). In other diction that will not be realized where biological words, the assumption of probability might accu- processes in! uence outcomes. With invariance (1) rately predict patterns (or process rates) that we we assume that history does not , with observe, but for the wrong mechanistic reasons. responses a " xed of controlling variables. Finally, with simplicity (3) the assumption is that In human systems, we can understand that “experi- only a few variables in! uence outcomes and ence” in! uences decision-making, so outcomes of that these variables can be measured with accuracy. perturbations differ if a person is naïve or familiar Yet even if simplicity holds then phenomena that with the event. Soil organisms may not exhibit should be predictable from deterministic parame- higher learning but directional selection in gene fre- ters can be inherently unpredictable because of spa- quencies provides a mechanism whereby history tial complexity. For example, Bernstein et al . ( 2000 ) and past conditions can in! uence outcomes of per- observe that the arrangement of balls on a pool turbations. For example, the development of resist- table, following their break, should be predictable ance of microbial chronically exposed from a few physical laws. Yet, when we introduce to antibiotics in! uences whether populations are spatial complexity across different pool tables, such extirpated when antibiotics are applied in acute as in the exact lay of the table, the nap of the felt, the doses. Likewise, microbial communities that have curvature of each ball, and so on, we introduce " ne- developed under conditions of moisture stress are scale complexity that means the resulting distribu- likely to have accumulated osmoregulants and tion of the balls is unpredictable (approximating other physiological strategies to cope better with one that is random). The development of the sub- future events. " eld of “spatial soil ecology” is testament to the fact With probability (2) the assumption is that all that rates of biogeochemical processes in soils are individuals are equally likely to respond identi- in! uenced—at least to some degree—by spatial cally—hence we can predict rates of radioactive complexity ( Ettema & Wardle 2002 ). For example, decay even if we can’t predict which actually the presence of anaerobic microsites may not be undergoes decay. Yet we know that soil organisms captured in models with insuf" cient spatial com- sort out along an r to K life-history continuum plexity, meaning that denitri" cation might be ( Fierer et al . 2007 )—or at least exhibit metabolically observed but not predicted for soils that are, on alert strategies—so whereas r -selected individuals average, aerobic. might rapidly increase their intrinsic No soil ecologist would argue that soils follow growth rates if resources are made available, the rules of invariance, probability, or simplicity. K -selected organisms will a lagged response. When we assume these rules apply (i.e. through our In practice this could mean that, for example, the modeling efforts that “black box” the biology) we decomposition of labile versus more recalcitrant recognize that we are abstracting reality. There are carbon pools could exhibit different temperature two primary reasons for such an abstraction. The responses if microbes with r versus K strategies, " rst is pragmatic; you have to start somewhere and respectively, decompose these different carbon the best place to start is with simplifying assump- pools. Admittedly, there is burgeoning debate about tions. Mathematical models force us to formalize whether we can assume—at least when explaining our understanding, highlighting to us the assump- distributions—that species tions we make, the evidence supporting these 192 SOIL ECOLOGY AND ECOSYSTEM SERVICES

Box 3.5.2 Messages soil ecologists can communicate when we “black box” soil communities

When we, as a community of soil ecologists, submit 1. Black boxing is valid because the effect of soil com- mathematical models of soil biogeochemical processes to munities on ecosystem function aggregates to a the wider scienti! c and policy community we communicate common, invariant response across environmental important information about the state-of-the-art of our variation (e.g. temperature) in time and space. You understanding of how soil communities in" uence ecosystem can have high certainty in the projections of our function. The majority of our models “black box” soil models. communities and, in doing so, they do not account for how 2. We black box because the complexity of how soil changes in soil communities across time or space might communities in" uence function is beyond our cur- in" uence biogeochemical pools and " uxes. These models are rent modeling capabilities. This is the best we have, embedded in management and policy efforts of global so be skeptical of the certainty of our model signi! cance, such as projections of future climate. It is projections. therefore imperative that we make a statement concerning 3. We black box because it is the simplest assumption how the wider scienti! c community should interpret the we can make, and we recognize the need for research black boxing of soil communities. We consider we have to test this assumption robustly because it might be three main options and that numbers two and three likely wrong. This is the best we have, so be skeptical of best represent the state of our science: the certainty of our model projections.

assumptions, and the need for empirical research to Andrén & Balandreau 1999 ), we restrict our discus- test them. In this regard they provide the most effec- sion to and fungi (collectively “microbes”). tive tool for integration of theory and empirical This is not to minimize the potential role of research to advance basic understanding. They also in shaping ecosystem processes (e.g. Schmitz 2008 ) permit us to make predictions that can inform pol- but because in modeling soil and ecosystem proc- icy and adaptive management. Even though we are esses, we typically discuss the microbes as the pri- aware of limitations in our knowledge and meas- mary agents of biogeochemical transformations. urements, such modeling efforts must be favored over inaction. What we have to be careful with is 3.5.3 Functional redundancy, similarity, the message we present to those outside of the soil equivalence, and biogeography ecology community—i.e. do we wish to communi- cate that we recognize our abstraction is likely Terms in ecology are often used to convey multiple wrong, or that we think the abstraction is a fair gen- meaning (e.g. consider “niche” when used by Grin- eralization (see Box 3.5.2 )? If the latter then this nel vs. Hutchinson), which creates confusion. The gives us the second reason for abstracting the biol- terms functional redundancy, similarity, and equiv- ogy in soil models: we believe the abstraction is alence are applied interchangeably in ecology to valid. That is, we can reliably predict how ecosys- comparisons between species, between communi- tem processes will respond to a perturbation ties, and as properties of a community ( Resetarits & because biological phenomena aggregate to a com- Chalcraft 2007 ; Allison & Martiny 2008 ; Strickland mon, invariant response (e.g. Rosindell et al . 2010 ). et al . 2009 ). Following Allison and Martiny ( 2008 ), This is the rationale often invoked for black boxing “functional redundancy” is the ability of one micro- the biology in soil and ecosystem models, and the bial taxon to carry out a process at the same rate as rationale is presented under the hypotheses of func- another under the same environmental conditions. tional redundancy, similarity, and equivalence. Species loss can occur without change in function if Although these hypotheses have been applied to a community has many taxa that are functionally soil communities in an inclusive framework encom- redundant ( Fig. 3.5.1 ). The same can even be true passing both animals and (e.g. where taxa are not functionally redundant but their THE BIOGEOGRAPHY OF MICROBIAL COMMUNITIES AND ECOSYSTEM PROCESSES 193

COMMUNITY FUNCTION IN A COMMON COMPOSITION ENVIRONMENT

At1 1 redundancy not redundant Up arrows represent change in a community across time At0 At0 At1 At0 At1 2 1 similarity not similar

Down arrows B represent 3 change in 4 At0 B At0 B communities across space equivalence not equivalent

C 2 1

At0 BCAt0 B C

Figure 3.5.1 across space and time of the terms functional redundancy, similarity, and equivalence when applied to the biogeochemical processes performed by soil microbial communities. Circles represent communities. Shown within each community is the abundance of each taxon (denoted by the numbers 1–4). For community A, a perturbation that shifts the community composition from time zero to time one (through the loss of taxon 2) but does not alter function—in a common environment—supports the hypothesis of redundancy. In comparing different communities (i.e. across space), the hypothesis of similarity is supported when a change in community composition does not alter function; again assuming the comparison is made in a common environment. The hypothesis of equivalence is less restrictive: there is no burden on the investigator to demonstrate a difference in richness, relative abundance, or phylogenetic structure of taxa in an assemblage (collectively referred to as “composition”) to falsify the hypothesis of equivalence, as there is with redundancy and similarity. Instead, a change in for example phenotypic abundance or within a community—without a necessary change in composition—that alters function in a common environment disproves the hypothesis of functional equivalence. Note that in the ! gure community C can be classed as functionally non-equivalent to both communities A t0 and B, despite having the same composition as A t0 . collective positive and negative responses to distur- composition of communities, whereas equivalence bance sum to zero net change in aggregate commu- includes the effects of these shifts and/or 1) changes nity function: a so-called “portfolio effect.” Allison in the phenotypes (including physiology) of the and Martiny ( 2008 ) de" ne “functional similarity” as members of a community, and 2) interactions the ability of two microbial communities to carry between individuals. Indeed, Allison and Martiny out a function at a similar rate under the same envi- ( 2008 ) restrict use of the terms redundancy and sim- ronmental conditions, regardless of differences in ilarity to differences in microbial community com- community composition. The same criterion of position—where they de" ne “composition” as the common environmental conditions applies to the richness, relative abundance, and phylogenetic de" nition of functional equivalence (sensu Strick- structure of taxa in an assemblage. The de" nition of land et al . 2009 ), which is the ability of two microbial functional equivalence is much less restrictive. communities to carry out a functional process at a Here we focus on functional equivalence because similar rate, regardless of differences in community we argue “equivalence” (as opposed to redundancy composition, physiology, and/or interaction or similarity per se) is the assumption that is made strengths (see Fig. 3.5.1 ). In this regard, mechanisms when using soil models, that black box microbial underlying functional similarity are a subset of the communities, to predict ecosystem process rates. mechanisms that might drive functional equiva- This is best illustrated by example. Imagine two lence. That is, similarity refers to the functional identical communities: one maintained at a constant effects of shifts in the taxonomic or phylogenetic temperature and one to which we apply a step 194 SOIL ECOLOGY AND ECOSYSTEM SERVICES increase in temperature. With temperature increase microbial communities exert proximate control on we would expect enhanced physiological rates (e.g. ecosystem process rates. Yet, because the environ- respiration) because of the positive effect of tempera- ment ultimately shapes the community, its role is ture on the rate of enzyme-mediated reactions. implicit in the model when we regulate the function Although physiological rates increase, there is not of the black box with environmental variables such necessarily a change in organismal physiology and if as temperature and moisture. temperature is returned to control conditions then We only require the second part of Baas Becking’s both communities function equivalently. If the tem- hypothesis to hold (i.e. “the environment selects perature treatment alters community composition, [function]”) for microbial communities to exert only then functional redundancy and similarity are only proximate control on ecosystem processes. This is falsi" ed if the compositional change elicits functional pertinent to the “black-box debate” because evi- differences when control conditions are restored. If dence is accumulating that microbes are not all glo- functional differences are observed, say through bally dispersed—at least not at rates that obviate altered physiology (e.g. shifts in enzyme expression), the role of historical contingencies in shaping com- but there is no measurable change in composition, munities—even at spatial scales of only a few meters then there is no evidence to reject the redundancy ( Ramette & Tiedje 2007 ). Even with limited disper- and similarity hypotheses. However, functional sal we can, however, still invoke a number of prop- equivalence is falsi" ed. Allison et al . ( 2 0 1 0 ) i n c o r p o - erties of microbial communities that might serve to rate microbial physiology in a soil model and show homogenize microbial functional potentials across that by doing so respiration rates in warmed soils, communities; namely high taxon abundance and that are returned to ambient conditions on reaching diversity, rapid evolutionary , and pro- steady-state, are lower than in control soils given the li" c growth rates ( Allison & Martiny 2008 ; Green role of physiological response in reducing microbial et al . 2008 ). We can apply the same arguments to jus- and extracellular enzyme abundance. tify disbelief in the functional equivalence hypoth- In subsuming the hypotheses of redundancy and esis. Under limited dispersal, these same properties similarity, the hypothesis of functional equivalence of microbial communities might facilitate rapid has broad application to the question over whether genetic differentiation. When populations are iso- the biogeography of microorganisms can be caus- lated geographically, they can solve the same sur- ally linked to variation in ecosystem function. The vival problem different ways, which might have two overarching factors that shape biogeographic different functional implications. Such phenomena patterns are environment and history. The role of introduce contingencies in biological systems that environment is nicely summarized by Baas Beck- make them “unique” (e.g. Jacob 1977 ; Levin 1998 ). ing’s evocative paradigm that “everything is every- In this regard they differ fundamentally from physi- where, but , the environment selects” ( de Wit & cal and chemical entities in that the rules change Bouvier 2006 ). The mechanism underlying this () across time, making replication dif" cult hypothesis is global dispersal where propagules of and history fundamental to understanding their all microbial taxa are ubiquitously dispersed across current structure and function. Arguments that we the globe ( Martiny et al . 2006 ). Growth and survival can ignore the identity of microbial taxa (e.g. is then determined (non-randomly) by environmen- Falkowski et al . 2008 ) and instead focus on the 500 tal conditions and there is at least some evidence or so enzyme systems underpinning biogeochemi- that this factor microbial assemblages at cal cycling—have not considered that many enzyme coarse levels of phylogenetic resolution (e.g. Fierer systems (e.g. aerobic respiration) are conserved & Jackson 2006 ; Lauber et al . 2009 ). In soil models, across the domains of life ( Hochachka & Somero we can reliably black box microbial communities if 2002 ) but yield markedly different process rates Baas Becking’s hypothesis can be extrapolated to under the same environmental conditions. Indeed it function and if we can identify the key environmen- is now well supported that rates of biogeochemical tal factors driving the microbial process of interest. processes such as litter decomposition, measured in Under such a scenario, we can still recognize that the " eld, are not only a product of the environment THE BIOGEOGRAPHY OF MICROBIAL COMMUNITIES AND ECOSYSTEM PROCESSES 195 but also adaptation/specialization of the soil com- necessary relationship between microbial commu- munity ( Ayres et al . 2009 ). nities and ecosystem function. They focus on The question that remains is how to test the approaches that consider “whole communities” (i.e. hypothesis of functional equivalence if we are to those that we " nd extant at " eld sites). The three justify a microbial-explicit modeling approach. approaches have increasing power to tease out Before reviewing approaches in the next section, we causal relationships between microbial communi- justify the methodological advantage of testing for ties and function. The " rst approach is long-term functional equivalence, as opposed to redundancy environmental manipulation. These types of stud- or similarity. The advantage is that to test for equiv- ies identify correlative relationships between alence there is no requirement to demonstrate dif- microbial communities and function, such as the ferences in microbial community composition. predominance of more r -selected phylotypes under When attempting to falsify functional redundancy N fertilization (e.g. Ramirez et al . 2010 ). Common or similarity there is the requirement to prove shifts garden experiments, where a microbial community in the structure of microbial communities are asso- is manipulated through environmental treatment at ciated with shifts in function. This is not trivial—the a common location, permits short-term investiga- structure of microbial communities can be analyzed tion of links between microbial communities and at many different levels of resolution (from the phy- function. In the longer-term the change in the envi- lum to the strain level), and we do not know which ronment is considered itself to play a role in regulat- level of phylogenetic resolution is most closely ing biogeochemical rates (i.e. the idea of the related to function. And with macro-organisms, environment as the ultimate control) meaning that such as angiosperms, we now know that different as with long-term environmental manipulations, within a single population can markedly correlative and not causative relationships are eval- in! uence ecosystem process rates ( Whitham et al . uated. The third approach—and that approach with 2008 ). Such a genetic shift in the composition of soil the most power to identify causation—is the use of communities is unlikely to be detected even with reciprocal transplants where the transplant is the the most advanced molecular techniques available, community (e.g. a soil monolith) and the environ- given that the researcher—before quantifying the ment is considered common to the location where genotypic diversity of a population—must " rst the different communities are brought together. identify the taxon within the many thousands Reed and Martiny ( 2007 ) highlight that even with present. There are numerous other issues that com- this approach caveats include the fact that the envi- plicate the association of community shifts with ronment of the transplanted units may not fully functional shifts, including microbial equilibrate with the new environment. We might (who is active and when), horizontal gene transfer, ask how long it would take for a soil core trans- and the fact that we often lack a fundamental under- planted from a pine to an oak to assume the standing of which taxa are likely responsible for a environment of the oak forest. For example, total speci" c process. So, if we follow philosophical argu- soil carbon and soil texture are parameters that are ments that a scienti" c hypothesis must be falsi" a- unlikely to change in such an instance, and even if ble, the methodological issue of being able to controlled for there is likely to be immigration from demonstrate difference in composition could be the surrounding community which obscures clear used to question the validity of even posing the relationships between microbial communities and hypotheses of functional redundancy and similarity function that are separate from the environmental for soil communities. conditions. All three of the approaches evaluated by Reed 3.5.4 Experimental tests of functional and Martiny ( 2007 ) reduced—to at least some equivalence extent—the issue we have in " eld observation where it is not possible to move beyond correlation Reed and Martiny ( 2007 ) comprehensively evaluate (to causation) between microbial communities three approaches used to test whether there is a and function, given the confounding issue of 196 SOIL ECOLOGY AND ECOSYSTEM SERVICES

environmental variation. Admittedly, observations crossed from multiple sites) approaches and meas- are the classical starting point for scienti" c investi- ured carbon mineralization rates from the micro- gation—and such approaches can identify poten- cosms across 300 days. The microbial community tially important environmental and microbial inocula explained as much as 86% of the variation in factors that might regulate (e.g. mineralization rates, providing strong support for Strickland et al . 2010 ). Yet to link microbial commu- the hypothesis that functional dissimilarity, not nity composition unquestionably to function functional equivalence, can be important and bioge- requires tightly controlled experiments. These are ochemically-relevant in soil communities. provided in the form of experimental assemblies of The common caveat to tests of functional equiva- known isolates. Strict control of environmental con- lence is that we expect the community to modify the ditions (e.g. in a bioreactor) permits identi" cation of environment as the experiment progresses which, ultimate causation through the microbial commu- in turn, modi" es the community. The crux here is nity in terms of regulating process rates. Such stud- modi" cation of the environment. We only falsify ies unambiguously show that microbial composition equivalence if microbial community effects are and diversity in! uence and nutrient compared “under common environmental condi- cycling ( Bell et al . 2005 ). Where the environment is tions.” This need to hold the environment constant slightly more realistic (e.g. wood disks), studies explicitly recognizes that functional rates vary with natural isolates have even shown that the directly as environmental factors change. Indeed, assembly history of the community in! uences covariation between functional rates and control- decomposition and carbon release rates ( Fukami ling factors is central to all soil models, whether et al . 2010 ). The major limitation of these app- they treat the microbial community as a black box roaches is, of course, the dif" culties associated with or not. To identify microbial communities as ulti- assembling the complex communities found in the mate controllers of biogeochemical process rates " eld. If we assume that a gram of soil may contain requires differences under constant environmental as many as 10,000 taxa, the drastic simpli" cation of conditions. In Strickland et al . ( 2009 ) there was high the experimental assemblages (~10 taxa) is certainly certainty that the initial environments were essen- not representative of the enormous taxon diversity tially the same, but with time the environments in natural microbial communities. This means that diverged as decomposition proceeded and the com- mechanisms—such as the portfolio effect and func- munities shaped the environmental conditions (e.g. tional redundancy—are strongly selected against in the carbon pools available, nutrient levels, pH, etc.). isolate experiments as agents creating functional It then us with the conclusion that initial equivalence. In addition, if ~1% of taxa are cultura- functional differences in the community led to over- ble and these taxa likely represent more r -selected all differences in function that themselves might organisms, then experiments with isolates select for have been a product of the community and/or a narrow slice of the ecological strategies observed altered environment. Separation of the effects of in natural communities. community composition and the environment is S t r i c k l a n d et al . (2009 ) present an experimental challenging, and clearly we need the suite of experi- approach for testing functional equivalence that is a mental approaches outlined in this section to compromise between whole community and cul- robustly challenge the hypothesis of functional tured isolate approaches. Recognizing the joint equivalence. Criticism of one approach to support needs of establishing a highly diverse microbial another fails to recognize the advantages each community and to have a common but realistic brings to the discourse (e.g. Fukami et al . 2010 ) and environment, they established experimental micro- will only reinforce the hypothesis in the absence of cosms with milled and sterilized litter (the appropriate falsi" cation. Given the common environment), and inoculated these with whole assumption of the functional equivalence hypothe- communities through introduction of a small mass sis in ecosystem modeling there is a scienti" c and of soil. They combined common garden (a single societal need to test it appropriately. If we do so we leaf litter) with reciprocal transplant (litters and soils will advance basic understanding of soil ecology THE BIOGEOGRAPHY OF MICROBIAL COMMUNITIES AND ECOSYSTEM PROCESSES 197 and permit reliable evaluation of modeled ecosys- in arguing that by " nding evidence for the sub- tem responses and to environmental strate-depletion hypothesis they had falsi" ed alter- change. nate hypotheses that microbial community responses explained observed patterns of respira- 3.5.5 Putting ecology into soil models tion to warming. Demonstrating that one mecha- nism can explain an observed pattern does not The soil (and ecosystem) models applied widely to falsify alternate mechanisms that might equally rec- address questions related to feedbacks to global reate the same pattern (i.e. the absence of evidence warming, carbon stocks across regional gradients, is not evidence of absence). To evaluate the compet- and dynamics across agricultural manage- ing hypotheses of functional equivalence and ment regimes, all “black box” the microbial com- redundancy, we require soil models that open-up munity. They assume functional equivalence in the black box by explicitly modeling microbial their parameterization, validation, and prediction. dynamics to evaluate the role of microbes in driving Indeed, we often parameterize ecosystem models biogeochemical processes. using data collected across space, and then apply Allison et al . ( 2010 ) present one of an emerging these parameterizations to make predictions across family of microbe-explicit models (e.g. Lawrence time for a system. This approach assumes microbial et al . 2009 ). They compare conventional multi-pool, communities do not ultimately in! uence biogeo- SOC models to an enzyme-based approach that rep- chemical cycling either in time or space. So what resents solubilization of SOC by extracellular happens when we construct models that open up enzymes, microbial assimilation of dissolved the black box and permit the identity of taxa, their organic carbon compounds, and the expected nega- physiology, and biomass to in! uence ecosystem tive relationship between temperature and micro- process rates? An in-depth model evaluation and bial growth ef" ciencies. Model predictions were review is beyond the scope of this chapter, but most sensitive to this latter parameter. In response below we present a brief discussion to identify to sustained warming, microbial biomass was future modeling needs by highlighting the implica- reduced because less of the carbon assimilated by tions of relaxing the “black box” assumption. the microbes was allocated to growth. This served Using multi-pool, soil organic carbon (SOC) to decrease the abundance of microbial extracellu- models—which black box soil communities— - lar enzymes that solubilize SOC, creating a negative Kirschbaum ( 2004 ) and Knorr et al . ( 2005 ) showed feedback to warming-induced losses of SOC. By that depletion of labile SOC pools could explain the explicitly modeling the microbial dynamics they ephemeral augmentation of soil respiration under found no evidence for to climate simulated warming. Once the warmed systems warming through loss of SOC to the atmosphere, a reached a steady-state—i.e. the labile pools had " nding contrary to most black box soil models. The been depleted to a constant value—respiration rates " nding is signi" cant because when we relax the matched carbon input rates, which were unchanged assumption of functional equivalence, model pre- from pre-warming conditions. That inputs equal dictions that in! uence policy may well differ from outputs is expected for any steady-state system. Yet those derived from more conventional modeling the studies demonstrated that the conventional way approaches. we model soil carbon can predict observed respira- Soil models that relax the assumption of func- tion responses to soil warming. Indeed, there is tional equivalence are not a recent phenomenon empirical support for this substrate-depletion (e.g. Hunt et al . 1987 ) but, like their earlier counter- mechanism ( Bradford et al . 2008 ). However, at the parts, they have not yet been incorporated into same site, there is also empirical support that the modeling efforts that might in! uence policy and microbial communities adjust to the thermal regime practice on environmental issues such as global in a manner that in! uences respiration rates ( Brad- climate change and carbon emissions (e.g. those ford et al . 2008 ). So where the conclusions of Kirsch- models used in Denman et al . 2007 ). They are baum ( 2004) and Knorr et al . ( 2005 ) went too far was also deterministic, and so do not permit context- 198 SOIL ECOLOGY AND ECOSYSTEM SERVICES dependent histories that are likely essential to gen- equivalence. Few ecosystem ecologists would erating functional dissimilarity to shape functional argue—if we held composition and the environ- outcomes that are not fully reversible. For example, ment constant—that microbial biomass was irrele- in the Allison et al . ( 2010 ) model returning the sys- vant to ecosystem process rates in soils (nor would tem to a pre-warming state will eventually permit any ecologist dare argue that plant biomass microbial biomass, extracellular enzyme abun- and community type are irrelevant to predicting dance, and carbon stocks to recover to pre-warming photosynthetic rates). Yet we have to accept that as conditions. Given context-dependent histories, we soil ecologists the dominant paradigm we espouse rarely expect changes in ecological systems to be through our ecosystem modeling is that soil micro- fully reversible upon restoration of original envi- bial communities are homogenously functioning ronmental conditions ( Levin 1998 ). Maybe such units across space and time, which exhibit invari- context-dependency is too dif" cult to incorporate functional responses to changes in controlling into current modeling efforts. However, the tracta- factors such as temperature, even where microbial bility of including microbial dynamics in determin- biomass differs. istic soil models has been demonstrated (e.g. Recent models (e.g. Lawrence et al . 2009 ; Allison Lawrence et al . 2009 ; Allison et al . 2010 ) and pro- et al . 2010 ) challenging the functional equivalence vides a likely productive direction for exploring the paradigm have not yet been coupled with efforts implications of assuming functional equivalence vs. that provide the scienti" c basis for policy to miti- dissimilarity when projecting biogeochemical gate and adapt to environmental problems (e.g. response and feedbacks to environmental change. Denman et al . 2007 ). Even within academic circles, appears to have had little in! uence in 3.5.6 Revisiting the functional paradigm shaping general ecological knowledge ( Barot et al . in soil ecology 2007 ). If our " eld is to advance knowledge and application outside of its own perimeters, then we Soil models—including those used in the coupled must take a fresh look at the paradigms of func- carbon cycle models to project climate change— tional redundancy, similarity, and equivalence typically assume that soil communities are func- and—and if we " nd them lacking—challenge appli- tionally equivalent. To put this in colloquial terms, cation of these paradigms where soil ecological “it doesn’t matter who is there, nor in what form, knowledge is applied outside of our " eld. This number, or location,” because every soil commu- includes application in areas of high societal impor- nity is essentially a black box that functions the tance, such as the coupled atmosphere-biosphere same way under the same environmental condi- carbon cycle models for projecting feedbacks to cli- tions when we look across space or time (see mate change ( Denman et al . 2007 ). Plant ecologists S c h i m e l 2 0 0 1 ) . A p p l i c a t i o n o f t h i s h y p o t h e s i s have engaged with atmospheric modelers in these assumes that functional responses to a realms, and soil ecologists must now do the same if (e.g. temperature change) can be described with a we wish to make a robust claim that soil biology single, mathematical equation. In making this need be considered when managing ecosystems assumption “history”—in its broadest sense—is and climate in the face of environmental change. disregarded as a force that in! uences the function- ing of microbial communities through changes in biomass, composition, or the physiology of soil References taxa. For example, even the decrease in total micro- Allison, S.D., & Martiny, J.B.H. (2008) , resil- bial biomass observed by Allison et al . (2010 )— ience, and redundancy in microbial communities. Pro- when they warmed a system and assumed a ceedings of the National Academy of Sciences of the United negative effect of temperature on microbial growth States of America 105 : 11512–19. ef" ciencies—would fail to elicit an initial difference Allison, S.D., Wallenstein, M.D., & Bradford, M.A. (2010) in respiration rates between a system pre and post Soil-carbon response to warming dependent on micro- a warming disturbance if we assume functional bial physiology. Geoscience 3 : 336–40. THE BIOGEOGRAPHY OF MICROBIAL COMMUNITIES AND ECOSYSTEM PROCESSES 199

Andrén, O., & Balandreau, J. (1999) and soil Fukami, T., Dickie, I.A., Wilkie, J.P., et al. (2010) Assembly functioning—from black box to can of worms? Applied history dictates ecosystem functioning: evidence from Soil Ecology 13 : 105–8. wood communities. Ecology Letters 13 : Ayres, E., Steltzer, H., Simmons, B.L., et al. (2009) Home- 675–84. "eld advantage accelerates leaf litter decomposition in Green, J.L., Bohannan, B.J.M., & Whitaker, R.J. (2008) . Soil Biology & Biochemistry 41 : 606–10. : from taxonomy to traits. Sci- Bahn, M., Janssens, I.A., Reichstein, M., Smith, P., & Trum- ence 320 : 1039. bore, S. (2010) Soil respiration across scales: towards Hochachka, P.W., & Somero, G.N. (2002) Biochemical adap- and integration of patterns and processes. New Phytolo- tation: Mechanism and process in physiological evolution. gist 186 : 292. Oxford University Press, New York. Barot, S., Blouin, M., Fontaine, S., Jouquet, P., Lata, J.C., & Hunt, H.W., Coleman, D.C., Ingham, E.R., et al. (1987) The Mathieu, J. (2007) A tale of four stories: soil ecology, detrital in a shortgrass . Biology and Fer- theory, evolution and the publication system. PLoS One tility of Soils 3 : 57–68. 2 : e1248. Jacob, F. (1977) Evolution and tinkering. Science 196 : Bell, T., Newman, J.A., Silverman, B.W., Turner, S.L., & Lil- 1161–6. ley, A.K. (2005) The contribution of and Kirschbaum, M.U.F. (2004) Soil respiration under pro- composition to bacterial services. Nature 436 : 1157–60. longed soil warming: are rate reductions caused by Bernstein, S., Lebow, R.N., Stein, J.G., & Weber, S. (2000) acclimation or substrate loss? Global Change Biology 10 : God gave physics the easy problems: adapting social 1870–7. science to an unpredictable . European Journal of Knorr, W., Prentice, I.C., House, J.I., & Holland, E.A. (2005) International Relations 6 : 43–76. Long-term sensitivity of soil carbon turnover to warm- Bradford, M.A., Davies, C.A., Frey, S.D., et al. (2008) Ther- ing. Nature 433 : 298–301. mal adaptation of soil microbial respiration to elevated Lauber, C., Knight, R., Hamady, M., & Fierer, N. (2009) Soil temperature. Ecology Letters 11 : 1316–27. pH as a predictor of soil bacterial community structure Clark, J.S. (2009) Beyond netural science. Trends in Ecology at the continental scale: a pyrosequencing-based assess- & Evolution 24 : 8–15. ment. Applied and Environmental 75 : Davidson, E.A., & Janssens, I.A. (2006) Temperature sensi- 5111–20. tivity of soil carbon decomposition and feedbacks to Lawrence, C.L., Neff, J.C., & Schimel, J.S. (2009) Does add- climate change. Nature 440 : 165–73. ing microbial mechanisms of decomposition improve de Wit, R., & Bouvier, T. (2006) “Everything is everywhere , soil models? A comparison of four mod- but, the environment selects ”; what did Baas Becking and els using data from a pulsed rewetting experiment. Soil Beijerinck really say? Environmental Microbiology 8 : Biology & Biochemistry 41 : 1923–34. 755–8. Levin, S.A. (1998) Ecosystems and the biosphere as com- Denman, K.L., Brasseur, G., Chidthaisong, A., et al. (2007) plex adaptive systems. Ecosystems 1 : 431–6. Couplings between changes in the climate system and Martiny, J.B.H., Bohannan, B.J.M., Brown, J.H., et al. (2006) biogeochemistry. In: Solomon, S., Qin, D., Manning, M. Microbial biogeography: putting microorganisms on et al. (eds.) Climate change 2007: The physical science basis. the map. Nature Reviews Microbiology 4 : 102–12. Contribution of working group I to the fourth assessment Ramette, A., & Tiedje, J.M. (2007) Multiscale responses of report of the Intergovernmental Panel on Climate Change . microbial life to spatial distance and environmental het- Cambridge University Press, Cambridge. erogeneity in a patchy ecosystem. Proceedings of the Ettema, C.H., & Wardle, D.A. (2002) Spatial soil ecology. National Academy of Sciences of the United States of Amer- Trends in Ecology & Evolution 17 : 177–83. ica 104 : 2761–6. Falkowski, P.G., Fenchel, T., & Delong, E.F. (2008) The Ramirez, K.S., Lauber, C.L., Knight, R., Bradford, M.A., & microbial engines that drive ’s biogeochemical Fierer, N. (2010) Consistent effects of nitrogen fertiliza- cycles. Science 320 : 1034–9. tion on soil bacterial communities in contrasting sys- Fierer, N., Bradford, M.A., & Jackson, R.B. (2007) Toward tems. Ecology 91 : 3463–70. an ecological classi" cation of soil bacteria. Ecology 88 : Reed, H.E., & Martiny, J.B.H. (2007) Testing the functional 1354–64. signi" cance of microbial composition in natural com- Fierer, N., & Jackson, R.B. (2006) The diversity and bioge- munities. FEMS Microbiology Ecology 62 : 161–70. ography of soil bacterial communities. Proceedings of the Resetarits, W.J., & Chalcraft, D.R. (2007) Functional diver- National Academy of Sciences of the United States of Amer- sity within a morphologically conservative genus of ica 103 : 626–31. predators: implications for functional equivalence and 200 SOIL ECOLOGY AND ECOSYSTEM SERVICES

redundancy in ecological communities. Functional Ecol- Schmitz, O.J. (2008) Effects of predator hunting mode on ogy 21 : 793–804. grassland ecosystem function. Science 319 : 952–4. Reynolds, J.F., Bugmann, H., & Pitelka, L. (2001) How Strickland, M.S., Callaham, M.A., Jr, Davies, C.A., et al. much physiology is needed in forest gap models for (2010) Rates of in situ carbon mineralization in relation simulating long-term response to global to land-use, microbial community and edaphic charac- change? Limitations, potentials, and recommendations. teristics. Soil Biology & Biochemistry 42 : 260–9. Climatic Change 51 : 541–57. Strickland, M.S., Lauber, C., Fierer, N., & Bradford, M.A. Rosindell, J., Hubbell, S.P., & Etienne, R.S. (2010) Pro- (2009) Testing the functional signi" cance of microbial tracted speciation revitalizes the neutral theory of biodi- community composition. Ecology 90 : 441–51. ersity. Ecology Letters 13 : 716–27. Warren II, R.J., Skelly, D.K., Schmitz, O.J., & Bradford, Schimel, J.P. (2001) Biogeochemical models: implicit M.A. (2011) Universal ecological patterns in college bas- vs. explicit microbiology. In: Schulze, E.D., Harri- ketball communities. PLoS One 6 : e17342. son, S.P., Heimann, M., et al. (eds.) Global biogeo- Whitham, T.G., Difazio, S.P., Schweitzer, J.A., et al. (2008) chemical cycles in the climate system . Academic Press, Extending genomics to natural communities and eco- San Diego, CA. systems. Science 320 : 492–5.