Frontiers of Ecology

Ecological research is entering a new era ofintegration and collaboration as we meet the challenge

of understanding the great complexity of biological systemsEcological subdisciplines are rapidly combining and incor-porating other biological physical mathematical and soci-ological disciplines The burgeoning base of theoretical andempirical work made possible by new methods technologiesand funding opportunities is providing the opportunity toreach robust answers to major ecological questions

In December 1999 the National Science Foundation con-vened a white paper committee to evaluate what we knowand do not know about important ecological processeswhat hurdles currently hamper our progress and what in-tellectual and conceptual interfaces need to be encouragedThe committee distilled the discussion into four frontiers inresearch on the ecological structure of the earthrsquos biologicaldiversity and the ways in which ecological processes con-tinuously shape that structure (ie ecological dynamics) Thisarticle summarizes the discussions of those frontiers and ex-plains why they are crucial to our understanding of how eco-logical processes shape patterns and dynamics of globalbiocomplexity The frontiers are

1 Dynamics of coalescence in complex communities2 Evolutionary and historical determinants of ecological

processes The role of ecological memory3 Emergent properties of complex systems Biophysical

constraints and evolutionary attractors4 Ecological topology Defining the spatiotemporal do-

mains of causality for ecological structure and processes

Each of the four research frontiers takes a different ap-proach to the overall ecological dynamics of biocomplexityand all require integration and collaboration among thoseapproaches These overlapping frontiers themselves are notnecessarily new Within each frontier however are emerg-ing questions and approaches that will help us understandhow ecological processes are interconnected over multiplespatial and temporal scales from local community structureto global patterns

Research frontier 1 Dynamics ofcoalescence in complex communitiesWe use the term community coalescence to refer to the development of complex ecological communities from a

January 2001 Vol 51 No 1 BioScience 15

Articles

Frontiers of EcologyJOHN N THOMPSON O J REICHMAN PETER J MORIN GARY A POLIS MARY E POWER ROBERT WSTERNER CAROL A COUCH LAURA GOUGH ROBERT HOLT DAVID U HOOPER FELICIA KEESING CHARLES RLOVELL BRUCE T MILNE MANUEL C MOLLES DAVID W ROBERTS AND SHARON Y STRAUSS

Articles

John N Thompson (e-mail thompsonbiologyucscedu) is a professor in the Department of Ecology and Evolutionary Biology University of California

Santa Cruz CA 95064 O J Reichman is director of the National Center for Ecological Analysis Synthesis 735 State Street Santa Barbara CA 93101

Peter J Morin is a professor in the Department of Ecology Evolution and Natural Resources Cook College Rutgers University New Brunswick NJ

08901-8551 Gary A Polis (deceased) was professor and chair of the Department of Environmental Sciences and Policy University of California Davis

CA 95616 Mary E Power is a professor in the Department of Integrative Biology University of California Berkeley CA 94720 Robert W Sterner is

professor and head of the Department of Ecology Evolution and Behavior University of Minnesota St Paul MN 55108 Carol A Couch is a senior

research scientist in the Water Resources Division of the US Geological Survey Norcross GA 30360-2824 Laura Gough is an assistant professor in

the Department of Biological Sciences University of Alabama Tuscaloosa AL 35487-0206 Robert Holt is a professor in the Department of Ecology

and Evolutionary Biology Natural History Museum University of Kansas Lawrence KS 66045 David U Hooper is an assistant professor in the

Department of Biology Western Washington University Bellingham WA 98225-9160 Felicia Keesing is an assistant professor in the Department of

Biology Bard College Annandale NY 12504 Charles R Lovell is a professor in the Department of Biological Sciences University of South Carolina

Columbia SC 29208 Bruce T Milne and Manuel C Molles are professors in the Department of Biology University of New Mexico Albuquerque NM

87131 David W Roberts is an associate professor in the Forest Resources Department Utah State University Logan UT 84322 Sharon Y Strauss

is an associate professor in the section of Evolution and Ecology University of California Davis CA 95616

AS ECOLOGICAL RESEARCH ENTERS A NEW

ERA OF COLLABORATION INTEGRATION

AND TECHNOLOGICAL SOPHISTICATION

FOUR FRONTIERS SEEM PARAMOUNT FOR

UNDERSTANDING HOW BIOLOGICAL AND

PHYSICAL PROCESSES INTERACT OVER

MULTIPLE SPATIAL AND TEMPORAL SCALES

TO SHAPE THE EARTHrsquoS BIODIVERSITY

16 BioScience January 2001 Vol 51 No 1

Articles

regional species pool This coalescence depends on inter-actions among species availability physical environment evo-lutionary history and temporal sequence of assemblyEcologists have made important strides in understanding theprocess of community coalescence but it will take evengreater integration of approaches for us to be able to con-fidently predict the pathways or endpoints of community as-sembly (Belyea and Lancaster 1999 Gotelli 1999 Weiher andKeddy 1999) So far we cannot predict which species arelikely to invade or to be lost from particular natural com-munities although some patterns are beginning to emerge(Petchey et al 1999) We also have much to learn aboutcommunity responses to perturbations at different stages indevelopment

Much of what we do know comes from a handful of eas-ily studied systems whereas the functioning of communi-ties undoubtedly also depends heavily on little-studiedhidden players such as microbes fungi and soil invertebrates(de Ruiter et al 1995) Indeed many attempts to create orrestore communities (eg freshwater wetlands or saltmarshes) fail for reasons that remain poorly understoodUntil we can fill these fundamental gaps in our knowledgehuman impacts on community patterns will remain hard topredict

Research conducted over the last decade suggests whatkinds of studies are needed to fill the gaps Progress will al-most certainly depend on developing new ways to simplifythe study of complex communities Focusing on functionalgroups or other as yet undeveloped constructs may help Weare also aware that we need to learn far more about the sys-tematics of many smaller or cryptic organisms which mayform links that are crucial to our understanding of com-munity coalescence The difficulty of linking importantecosystem functions to key taxa and interspecific interactionsis a particular challenge

Research in five key areas described below could provideimportant insights into the community assembly process

Functional traits and community composition Aswe study communities that are increasingly a mix of nativeand introduced taxa we need to find out whether coevolvedspecies or local populations differ from coevolutionarily naivepopulations in the strength or nature of their interactions dur-ing or after community assemblyWe need to improve our abil-ity to pinpoint the traits of species that affect the probabilityof invasion or extinction within developing communities(Pimm 1989 Rejmaacutenek and Richardson 1996 Belyea andLancaster 1999) Moreover we need to further our under-standing of how the importance of those traits depends on ex-isting species composition critical thresholds (eg speciesrichness functional group composition) and history of acommunity (Rejmaacutenek 1989 Burke and Grime 1996 Levineand DrsquoAntonio 1999)

Pathways toward community coalescence For anypool of potential community members there are many

possible assembly sequences and endpoints The diversity ofinitial sequences is less problematic if many of the alternativepathways tend to converge on a limited subset of endpointsTo that end we need to find out whether developing com-munities move toward single or multiple points or states (at-tractors) Past some stage of assembly communitycomposition may become canalized but it may be highlysusceptible to perturbations in the early stages of coales-cence Anthropogenic change may limit pathways of com-munity development by changing the physicochemicalenvironment altering biogeochemical cycles or changingthe genetic structure of populations Global homogeniza-tion of the species pool may itself alter the pathways or end-points of community assembly reshaping both short-term andlong-term successional patterns across landscapes

Functional groups Ecologists recognize that the con-cept of functional groups is a valuable tool for simplifyingcommunity complexity to manageable levels Currently mostgroups are defined in a system-specific way (Wilson 1999)based on biochemical morphological or trophic criteria(Vitousek and Hooper 1993 Hooper and Vitousek 1997Naeem and Li 1997 Tilman et al 1997) although some at-tempts have been made to develop general classifications(Grime et al 1997) Each of these attempts has helped us iden-tify the advantages and disadvantages of different approachesTogether they are moving us toward the development ofrobust frameworks that work across multiple taxa and ecological communities Those advances should ultimatelyhelp us to understand how functional group structure shapesassembly dynamics as well as to evaluate how different con-figurations of functional groups affect the development ofcommunities or the dynamics of ecosystems (eg rates andtrajectories of community development with and without nitrogen-fixing species)

Hidden players Community assembly almost certainly de-pends upon cryptic invertebrate microbiological and my-cological groups which include a vast array of free-livingspecies parasites and mutualistic symbionts These groups andinteractions have historically been ignored or unrecognizedbecause of technological limitations and biases in the train-ing of ecologists Many microbial species cannot yet even becultured Nevertheless they may play a keystone role in com-munity development and function and their absence may beresponsible for some failed attempts at community restora-tion (Wall Freckman et al 1997 Brussaard 1998) The im-portance of these hidden players may become apparent onlywhen they become problems as sometimes happens when wealter community structure Lyme disease for example maybe the result of the emergence of a hidden player in responseto changes in community composition and landscape patterns(Jones et al 1998)

Part of our challenge will be to determine how the struc-ture of microbial diversity should be incorporated into dif-ferent kinds of analyses of the local regional and global

dynamics of ecological processes Traditional definitions ofspecies and traditional metrics of distribution and abundanceoften make a poor fit with the structure of microbial pop-ulations and communities

Our research choices have been biased not only towardparticular taxa but also toward particular communities Wehave detailed community assembly information on only asmall number of well-studied kinds of communities (egsuccessional fields temperate lakes rocky intertidal shores)Studies of underrepresented communities taxa and inter-actions will together provide important insights into gen-eral patterns in the assembly process (eg roles of particularkinds of microbial interaction or mutualistic interaction)

Restoration The best measure of our understanding of anycomplex system is whether we can reconstruct it from itsparts (Jordan et al 1987) Successful restoration of sustain-able communities and de novo creation of persistent com-plex systems that provide essential ecosystem services in novelenvironments are therefore the true tests of our under-standing of community dynamics Those tasks in short con-stitute the shared frontier of community ecology and ecologicalrestoration We need to know which aspects of communitystructure are restorable once disassembled and which arenot Although our understanding of what is restorable is lim-ited our societies are nonetheless moving apace to manipu-late community assemblages worldwide (Simberloff et al1997) We must therefore continue to work toward a more

comprehensive theory of community coalescence if we are tosucceed in efforts to mitigate the effects of invasive species onlocal communities and to make those communities less proneto disruption The deliberate introduction of species whichtakes place now in biological control efforts will surely be lessrisky once we understand how to enhance the control ofspecies through the assembly of persistent but self-limitingsubwebs Similarly bioremediation efforts will benefit fromunderstanding how to develop persistent self-limiting as-semblages of species that accumulate or degrade toxic wastes

Research frontier 2 Evolutionary andhistorical determinants of ecologicalprocessesThe genetic and evolutionary structure of organisms incombination with the recent history of environmentalevents continually shape the dynamic patterns of commu-nity coalescence outlined in frontier 1 Biology is a sciencestrongly influenced by historical events Evolution and en-vironmental history impose ldquoecological memoryrdquo on com-munities introduce time lags in ecological processes andconstrain the trajectories of community coalescence in waysthat are poorly understood Ecological memorymdashthe resultof past environmental conditions and subsequent selectionon populationsmdashis encoded in the current structure ofbiological communities and reflected in the genetic struc-ture of species It affects how communities assemble and itmay also affect the likelihood that they can be restored once


Articles

Figure 1 The Salmon River within the large wilderness core of Idaho The few remaining relatively pristinelandscapes worldwide will become increasingly important for testing hypotheses at the frontiers of ecology Theseunfragmented landscapes are our only way of testing hypotheses on the dynamics and trajectories of ecosystemand evolutionary processes acting on extant species over long time scales and across broad geographic areasPhoto John Thompson

disassembled We therefore need a much better under-standing of which aspects of phylogeny ongoing evolutionand recent history are most important in shaping currentecological processes and patterns across landscapes

The combined roles of phylogeny and ongoing evolution Although biologists commonly divide the tem-poral continuum into ecological time and evolutionary timeevolution shapes ecological processes across all time intervalsThe phylogenetic history of species creates large-scale patternsin the ecological relationships of taxa and combines withrapid evolution over the scale of decades to generate ongoingecological dynamics We need to answer six crucial questionsif we are to develop a theory of ecology that takes into accountthe genetics and evolution of organisms

Phylogenetic structure of ecological processes How does theshared phylogenetic history of species shape ecologicalprocesses Because of shared phylogenetic history speciescannot be treated as independent units in their ecological rolesWe know for example that ecological specialization is phy-logenetically constrained (Futuyma and Mitter 1996 Webb2000) A few studies have analyzed how shared speciesrsquo traitsand historical biogeography combine to constrain and shapecommunity and ecosystem structures For example analysisof the phylogeny of Caribbean anole lizards casts doubt on theoverall applicability of a widely cited ecological theory thetaxon cycle (Butler and Losos 1997) Recent analyses of rainforest in Borneo indicate that more closely related species werefound in the same microhabitats more often than could be ex-plained by chance alone (Webb 2000) We need similar analy-ses across taxa and ecosystems to further integrate phylogeneticnonindependence into community and ecosystem ecology

Rapid evolution and ecological dynamics To what extent aresuccession population dynamics and ecosystem dynam-icsmdashhistorically considered to be solely ecological processesmdashgoverned by rapid evolutionary change in species and theirinteractions There are dozens of examples in which ecolog-ically important traits of species are known to have evolvedduring the past century (Thompson 1998 1999) Rapid evo-lution of bill morphology in Galapagos finches wing color inpeppered moths and pesticide resistance in many taxa are well-known examples that highlight the speed with which naturalpopulations can evolve when subjected to environmentalchange These evolutionary changes have the potential toripple throughout communities For example the recently dis-covered rapid evolution of Daphnia in response to pollu-tion in Lake Constance may change phytoplankton dynamicsin ways that are important to community function (Hairstonet al 1999)

Coevolution and ecological dynamics How does the historyof species cooccurrence shape ecological processes commu-nity function and community stability and invasibility Useof molecular markers coupled with new analytical techniques

such as coalescence theory may allow us to estimate the lengthof time that locally and regionally interacting species havecooccurred (Pellmyr et al 1998) These techniques could en-able us to determine whether populations species and com-munities that share a history of cooccurrence exhibit propertiesdifferent from those created largely from recent invasions orrestoration efforts They will help us determine whether long-term coevolution of taxa affects community properties suchas invasibility nutrient cycling and productivity

Scale of evolutionary dynamics How do landscape structureand genetic structure act together in shaping ecologicalprocesses Ecologists are beginning to look hard at how thespatial scale of ongoing evolution and coevolution shapesecological dynamics Species are groups of genetically differ-entiated populations connected by various levels of geneflow Adaptations developed within local communities canspread to populations in other communities thereby alteringlocal ecological dynamics Natural and anthropogenic frag-mentation of populations restricts the exchange of individ-uals among communities both historically and currently Weare just beginning to understand how the landscape structureof fragmentation interacts with the genetic configuration ofspecies and ongoing evolution to shape ecological processesat local regional and continental scales We need a refined the-ory of spatially structured evolutionary dynamics to assess theoverall effects of increased anthropogenic fragmentation oncommunities

Genetic diversity and ecological dynamics How does geneticdiversity within species shape the temporal dynamics of pop-ulations trophic interactions and energy flow in ecosystemsBy incorporating the effects of genetic diversity per se into ourmodels of the temporal dynamics of ecosystem structureand function we may better understand and predict ecolog-ical systems The study of plant hybrid zones has given us aglimpse of the potentially far-reaching effects of the structureof genetic diversity on community dynamics (Whitham et al1999) Particular genotypes or hybrids of some species aremuch more vulnerable to herbivores or are more productivethan other genotypes or either of the parental speciesMoreover some hybrids and genotypes support greater di-versity and biomass of insect herbivores birds and otherspecies thereby becoming local foci of diversity and sup-porting distinct food web interactions

Genomics of ecological dynamics How do genomic con-tent and patterns of gene expression shape species distrib-utions species interactions and responses to changingenvironments We are just starting to evaluate the ways inwhich chromosome number shapes large-scale patterns incommunity structure and dynamics From a small numberof studies we know that chromosome number is tied toecologically important attributes of taxa such as speciesdistributions (eg the proportion of polyploid plant speciesincreases with latitude) invasiveness and susceptibility of


Articles

plants to herbivores (eg moths on saxifrages) (Rosenzweig1995 Thompson et al 1997) Thus the evolutionary ge-nomics of component species may have strong impacts onthe ecological functioning of communities

The role of environmental history and temporalvariability Ecological research needs to build on past ef-forts to integrate processes acting across different time scalesAll environments vary through time at scales from very short(hours days) to long (decades centuries millennia) Becauseof this temporal variability we often cannot understand com-munities without understanding past events or processesWe know that three components of environmental history andtemporal variability appear to be particularly important inshaping community dynamics variability in productivityand the input of resources variability in species compositionabundance and interaction strength and variability in key abi-otic events The impact of these factors can change over dif-ferent time scales shaping species composition abundanceand food web interactions in the short term (eg via recentinvasion) medium term (range expansion and contraction)and long term (historical biogeography)

Historical variability in productivity or resource inputResources vary in availability over time in all communitiesthrough abiotic (eg wet versus dry seasons El Nintildeo versusLa Nintildea years) and biotic influences (mast years in long-lived plants pig wastes flowing into watersheds) (Polis et al1996) Such variation has been characterized by ecologists

working on particular taxa or communities Neverthelesswe are only in the early stages of developing a general bodyof theory on how past periodic or pulsed productivity affectsthe dynamics of populations interactions between resourcesand consumers food webs communities and ecosystems

We need to continue to work toward a synthetic frame-work for explaining how temporally variable productivityinfluences food web processes community dynamics andecosystem function For example masting events and peri-odic emergence of insects (eg cicadas) can greatly alterspecies interactions and community dynamics for years(Jones et al 1998) Long periods between pulses of high pro-ductivity reduce consumer abundance consequently im-peding resource suppression at the next productive periodThis process appears to underlie the evolution of masting inbamboos and trees and synchronous reproduction in insectsand ungulates (Janzen 1976)

In general periodic pulses of productivity (eg goodversus bad years or seasons) may produce resource reservesthat last for varying periods before they are depleted alter-ing food web dynamics long after the productivity pulse(Noy-Meir 1974 Sears et al in press) For example El Nintildeorains (or La Nintildea drought) may stimulate or depress terrestrial productivity thereby allowing subsequently highor low abundances of consumers and altering interactionsbetween resources and consumers (Polis et al 1997) Onsmaller spatial scales the pattern of local competitive outcomes in forests may depend on the pattern of tree fallsold allelopathic footprints and past local patterns of


Articles

Figure 2 The fragmented wet tropics of northern Queensland Australia north of Cairns Similar landscapes nowcharacterize most terrestrial environments worldwide illustrating the need for understanding the dynamics of populationcommunity ecosystem and evolutionary processes in environments whose boundaries are sharp and constantly changingacross large spatial and temporal scales Photo John Thompson

nutrient depletion or regeneration each of which may pro-duce long-lasting changes in local soil chemistry

Historical variability in species composition abundance andinteraction strength We are beginning to disentangle the waysin which past community configurations shape currentprocesses We know for example that the effects of intro-ductions of new species into communities can depend on pastpresence of particular taxa Introduction of cows into inter-mountain western North America led to effects on grasslandand steppe communities that were far different from those re-sulting from similar introductions east of the Rockies Thesediffering impacts came about in part from differences in thehistory of association of these grasslands with bison and theselection these large mammals imposed on plant taxa In ad-dition short-term irruptions of species can produce long-lasting effects The irruption of rinderpest in some Africangrasslands resulted in a single flush of recruitment of acaciasduring this century The acacias are still extant yet current eco-logical conditions or their recent life history statistics (egmeasurement of intrinsic rate of natural increase) do notaccount for their presence

Historical variability in key abiotic events Erratic large-scale disturbance regimes can leave profound ecological lega-cies that may propagate historical echoes Volcanoeshurricanes El Nintildeondashsouthern oscillation events or inter-decadal oscillations all reshape biological communities inways that have long-lasting effects For example interdecadalchanges in anchor ice play a major role in altering the abun-dance composition and interaction dynamics of the entirebottom community in the Antarctic (Dayton 1989)We muststrive to develop a fuller appreciation of when and how thesemajor periodic environmental events are important in shap-ing global and regional patterns in ecological processes

Research frontier 3 Emergent propertiesof complex systemsThe evolutionary and historical determinants of ecologicalprocesses outlined in research frontier 2 are in turn shapedin large part by the laws of physical science Hence thethird frontier is to understand how the joint evolutionaryforces acting on an entire species assemblage result from con-straints imposed by biophysical processes

Phenotypic traits are products of adaptive evolution operating within bounds set by physical and chemical constraints Important progress in determining how theseconstraints affect the morphology and behavior of indi-vidual species has been registered in the subdisciplines ofphysiological and functional ecology Todayrsquos research frontier is the expansion of this viewpoint to collections ofspecies and communities and ecosystems We need to knowhow specific evolutionary responses at the species level produce the collective morphology (physiognomy) and behavior (ecosystem function) of ecosystems One centralquestion is whether first principles of physics chemistry and

evolution by natural selection can successfully predict thecomposition structure and functioning of ecosystems If soevolutionary first principles could enhance ecosystem studyReciprocally advances in studies of matter and energy flowcan provide the information necessary to realistically assessbiophysical constraints on evolutionary change

If we imagine a community as a set of species plotted inmultidimensional phenotypic space then enclosing thisspace is a set of boundaries and constraints imposed by theprinciples of energy and thermodynamics Matter devotedto one use is unavailable for other uses and these con-straints generate evolutionary tradeoffs Selection maxi-mizes fitness within those constraints In the terminology ofdynamic systems theory the result is an attractor a condi-tion (such as a set of phenotypes) toward which a system(such as a community) is drawn over (evolutionary) time

As each organism responds to its environment in a man-ner prescribed by its genome and the specifics of its bio-physical and biotic environment it simultaneously modifiesthat environment The result is a coupled complex dy-namic system of organism and environment wherein nat-ural selection optimizes the fitness of populations amid acontinually changing biotically driven environment If a sin-gle species of plant evolves to have larger leaf area a host ofallometric tradeoffs within that species come into play suchas corresponding increases in sapwood conducting volumeThe cumulative effect of these changes may be an alter-ation in how resources are used within a communityMaterial and energy fluxes for competing species and con-sumers of that plant may in turn also be affected

An example of an attractor resulting from biophysicalproperties is the 075 power law for metabolism in en-dothermic animals which states that whole-body meta-bolic rate scales with body mass to the 075 power Althoughthere has been turnover of species along this curve overtime the physical basis of this power law ensures that thecurve itself has remained fixed Such an evolutionary at-tractor works through a combination of physical principlesand evolutionary optimization Phenotypes that deviatefrom this attractor are at a significant disadvantage comparedwith those near the attractor

The frequency distributions of body mass in mammalswhich demonstrate a unimodal relationship may be anotherexample Brown (1995) suggested that modal mass corre-sponds to an evolutionary optimum (attractor) for a givenbody plan Body mass may be governed by a tradeoff betweenresource harvest ability and bioenergetic patterns deter-mining the rate at which resources can be turned into reproduction Deviations from the optimum may be drivenby niche partitioning Ritchie and Olff (1999) used spatialscaling laws to describe how species of different sizes findfood in patches of varying size and resource concentrationThey derived a mathematical rule for the species that shareresources and used the rule to develop a predictive theoryof the relationship between productivity and diversitySpatial scaling laws they argued provide potentially


Articles

unifying first principles that may explain many importantpatterns in species diversity

Biochemical constraints may also result in an evolution-ary attractor on life history evolution Evidence for such con-straints is accumulating through studies of ecologicalstoichiometry which considers patterns in element con-centration among different species (Sterner 1995) Recentstudies have indicated that rapid growth in biomass re-quires an element composition high in phosphorus (P) aresult of the element composition of anabolic biochemicalmachinerymdashin particular ribosomes (Elser et al 1996)Thus life history evolution is constrained in a phenotypicspace that includes phosphorus as one of its axes Highgrowth rate depends upon high P content Another stoi-chiometric example is the evolution of structural materialsuch as wood or bones in which different biochemical so-lutions to the same or similar structural problems havebeen achieved Hence we see alternative chemical signaturesin living organisms The larger the organism the greater thereliance on carbon (C) (for wood) or calcium (Ca) andphosphorus (for bone) generating an attractor in a phe-notypic space that includes elements such as C Ca or P aswell as organism body size

Specific research questions We need to answer severalquestions about the physical basis of evolutionary attractorsAmong them are these

How are species arranged as collections in phenotypicspace bounded by biophysical constraints

How do adaptive solutions vary as species diversityvaries within those constraints

Can different coherent sets of community members be-come established as multiple stable states or doeseach situation engender a specific ecological and evo-lutionary solution (based on ecological stoichiometryor similar constraints)

What effect does fitness maximization of individualshave on the collective behavior of all species and onthe fluxes of matter and energy within these collec-tions

Can first principles be used to predict the statistical dis-tribution of species traits not just a single optimumCan that distribution be extrapolated to predictionsor estimates of diversity

We cannot begin to cross this frontier successfully unlesswe find the means to perform rigorous experimental testsor find other ways to test hypotheses Multispecies correla-tions of phenotypic traits will be relatively easy to find buttesting among competing hypotheses for these broad-scalepatterns will require considerable ingenuity Experimentaltests of some evolutionary hypotheses are impossible for allbut the smallest (and most quickly reproducing) organ-isms but microbial taxa make excellent experimental toolsfor answering some of these questions For most other taxa

less direct means of testing mechanistic hypotheses will benecessary

New techniques New approaches in the study of structuraland functional genomics can further this highly integratedview of evolution and ecosystems A community is com-posed of suites of genes that are collected into epistatic groupsthese into genomes and finally genomes into communitiesBreakthroughs in gene sequencing and in measuring pat-terns of gene expression should make it increasingly practi-cable for ecologists to study the relationships betweenbiophysical constraints and evolutionary attractors For ex-ample studies have described the patterns of expression of6347 genes in mice raised under caloric restriction (Lee et al1999) Caloric restriction retarded aging by causing a meta-bolic shift toward increased protein turnover and decreasedmacromolecular damage These experiments provide aglimpse of a single genomersquos response to the changed avail-ability of energy They suggest intriguing links between ge-nomics and energy and nitrogen fluxes We need to explorehow speciesrsquo genomes respond to each other as material andenergy availability change Such studies can be viewed asecological community or even ecosystem genomics

Research frontier 4 Ecological topologyFrontiers 1 2 and 3 require that we reevaluate the scale ofecological processes and the analytical methods by which westudy them We need to develop a way of evaluating how ourscales of information (scope and resolution in space andtime) affect our understanding of ecological processesmdashthatis an ecological topology We can think of this ecologicaltopology as a study of the domains of causality For in-stance what are the factors that control rates of primary pro-duction by native plants on ancient lava surfaces in HawaiiThe most important controls may be local (eg competitionamong neighboring plants) regional (eg grazing by wan-dering goats introduced by Europeans centuries earlier andsustained by productivity of the larger surrounding land-scape) or even global (set by the levels of phosphorus in eolian dust from Central Asia mobilized during the lastglacial period) (Jackson et al 1971 Chadwick et al 1999)The appropriate domains of causality in many ecologicalstudies could extend far beyond previously assumed spatialand temporal bounds

We therefore need a unified understanding of the spatialand temporal context of ecological interactions such as themagnitude and importance of cross-habitat fluxes of matterenergy and information sometimes acting over very largescales Even our understanding of fundamental matters suchas the structure of food webs generally lacks a spatiotemporalcontext strongly limiting our ability to explain their originmaintenance or consequences We suggest building an eco-logical topology that addresses these needs Doing so re-quires characterizing patterns of movement of individuals(Holt and Gaines 1992 Holt 1993) and flow of materials


Articles

(Power and Rainey in press) across space and through time(Polis et al 1997) This frontier is consistent with todayrsquos greatneed for scaling up from local processes to much largerones including the planetary scale Global change is one ofthe most pressing issues in ecology today Understanding thetopology of causation of ecological patterns and processescould provide a formal structure for linking studies at thelocal scale to larger ones

We already have some tools and could soon develop oth-ers to investigate and characterize the bounds of suchspatiotemporal domains Genetic and isotopic tracers for fol-lowing the movement of matter and organisms within andacross ecosystem boundaries are becoming more accessibleWe have new satellite and ground-based technologies thatexpand our ability to capture and analyze spatial data Withthese conceptual and technical tools we can address a broadrange of important ecological problems such as those thatfollow

Choosing the right spatial and temporal contextHow do different spatial and temporal domains of causalitycombine to produce local community patterns and processesMoreover how far back in time or out into the spatial land-scape must we probe to understand local phenomena or in-teractions We need to find out just how far we should expandthe domains of our search if we are to interpret ecological pat-terns and processes correctly

Identifying the rules What rules govern the origin andmaintenance of ecological topologies Clearly some generalrules govern the geometries of circulatory systems or thedrainage networks of watersheds This is an area of active

research in scientific fields outside ecology and some of themethods of those fields could inform us about the bounds ofecological systems

Changing the bounds How will ecological communitieschange when their bounds change If there are natural con-straints on the bounds of domains of ecological processes willtruncating or stretching them change system behavior stabilityor sustainability This is a crucial consideration given humandomination and rearrangement of ecosystems For examplethe California water system has been called the most massiverearrangement of nature ever attempted Interbasin watertransfers are now common throughout many arid regions ofthe world depleting or artificially augmenting local ecosys-tems and groundwater systems Release of carboniferous fos-sil fuels back into the active global carbon cycle is anotherexample of human domain stretching along a deep tempo-ral axis Our global homogenization of species could also beconsidered in this light Humans however do not alwaysstretch causality domains Sometimes we greatly contractthem as when we replace wild populations of salmonidswhich harvest huge areas of ocean production with cage-cultured fish that feed only in the mouths of estuaries It isworth exploring whether causality domain theory can enhanceour ability to predict the longer-term consequences of eco-logical distortions of various types and magnitudes

Focus on these ecological topologies will require a changein how ecologists typically think of ecosystems Conventionallythe term ecosystem refers to a particular contiguous unit of thelandscape The system has inputs and outputs and internal dy-namics Most management practices similarly focus on eco-logical processes and patterns defined by such habitat units


Articles

Figure 3 Coalescence of a new complex plant community of native and introduced species at Tejon Pass California Thecommunity includes a mix of annual Mediterranean grass species and a diverse array of native annual wildflowers Beforeinvasion by the introduced grasses the community was probably dominated by perennial tussock-forming grasses alongwith annual grasses and forbs Photo Courtesy of Frank Davis University of California Santa Barbara

Nevertheless many if not most ecological processes can beenvisaged as caused by a set of rules each of which operatesover different spatial and temporal scales Understandingthose rules is among our great current challenges

Meeting the challengeThe challenges ahead will require new ways of working to-gether new tools and greater access to an ever growingnumber of databases

Collaboration and integration No single scientist hasdepth enough across all the physical and biological sciencesto address the questions at the frontiers Therefore we mustfoster research collaborations among scientists who aretrained to understand the language of one another Meetingthat challenge will require renewed effort to make sure thatecologists are well-grounded in earth sciences and mathe-matics and physical scientists solidly grounded in the the-ory of ecology

In addition stronger links between ecology and evolu-tionary biology must be forged across all the frontiersDespite the formation of departments of ecology and evo-lutionary biology in many universities during the last 30years there remains a major gap between these disciplinesRelatively few ecologists have ever had a formal course in evo-lutionary processes and very few have ever had advancedtraining in evolutionary theory and methods Similarly fewevolutionary biologists have much advanced training incurrent ecological theory and methods Incorporating anevolutionary framework into ecological research requires thatwe train scientists to be well versed in both ecological andevolutionary theory

Analogously there is a need to strengthen the links be-tween ecosystem and community ecology Understandinghow biological communities shape and are shaped by phys-ical environmental processes will transpire as the discipli-nary barriers between ecosystem ecology and communityecology break down

The need for stronger links between systematics and ecol-ogy is evident for all taxa but it is especially critical for mi-crobial groups We need to turn out a generation of biologistswho are well grounded both in microbial ecology and mi-crobial systematics These will be scientists capable of grap-pling with the complex structure of microbial taxa in anecological context We are only in the early stages of what willbecome a major field of study how microbial diversity con-tributes to the dynamics of biocomplexity The same needshold for studies of the systematics and ecology of cryptic in-vertebrates and fungi

Analytical tools The mathematical and statistical constructs used in ecological research come mostly from linear theory In contrast many ecological processes are gov-erned by nonlinear dynamics threshold effects and more complex relational structures We must continue to work on

developing ecological theory that builds on the structure ofempirical results in ecology

Access to research and monitoring databasesWe alsoneed ecologists trained to manage large databases who canorganize the storehouse of past ecological data mine it for newresults and make it accessible to others Lack of ready accessto the already available mass of ecological data is a major hur-dle in ecological analysis The experience of dozens of work-ing groups at the National Center for Ecological Analysisand Synthesis in Santa Barbara California who are attemptingto summarize and compare data collected over decades for awide variety of purposes confirms this need Hence the needfor stronger links between empirical ecologists and databasemanagers will continue to grow as new ecological data aregathered for ever-changing questions

The existing storehouse of ecological data includes notonly the results of individual studies but also the manydatabases that come from environmental monitoring effortsMaintenance of these monitoring efforts is important if weare to understand how past environmental events shapecurrent ecological processes Monitoring efforts however areuseful for ecological analyses only if their results are readi-ly accessible to ecologists

Moreover to work with diverse data sets collection meth-ods and data entry methods must be standardized whereverpossible Techniques continue to change and different meth-ods are needed for different ecological situationsNevertheless we need to take a fresh look at some of the com-mon methods used in studies within subdisciplines and at-tempt standardization

ConclusionEcological research is undoubtedly entering a new era build-ing on a firm base of advancements in our ecological knowl-edge achieved in recent years Each of the frontiers of ecologyrequires increased collaboration and technology if we are tounderstand how the complexity of ecological processes con-tinues to reshape the earth Most of the unresolved challengesin ecology arise from incomplete understanding of how bi-ological and physical processes interact over multiple spa-tial and temporal scales These challenges permeate the fourinterrelated research frontiers we have identified Crossingthe frontiers demands strong cross-disciplinary trainingand collaboration coordination of observational and ex-perimental approaches across large scales incorporationof new technologies and greater access to the growing data-base of ecological results Developing the means to cross thefrontiers is essential if the field of ecology is to move forwardas a predictive science

AcknowledgmentsWe are grateful to Scott Collins William Michener andMargaret Palmer program officers for the National ScienceFoundation for initiating this effort to assess the frontiersof ecology We are also grateful to three anonymous


Articles

reviewers for very helpful and thoughtful reviews Thiswork was supported by NSF grant 97-07781 to J N T

References citedBelyea LR Lancaster J 1999Assembly rules within a contingent ecology Oikos

86 402ndash416

Brown JH 1995 Macroecology Chicago University of Chicago PressBrussaard L 1998 Soil fauna guilds functional groups and ecosystem

processes Applied Soil Ecology 9 123ndash135Burke MJW Grime JP 1996 An experimental study of plant community in-

vasibility Ecology 77 776ndash790

Butler MA Losos JB 1997 Testing for unequal amounts of evolution in a con-tinuous character on different branches of a phylogenetic tree usinglinear and squared-change parsimony Evolution 51 1623ndash1635

Chadwick OA Derry LA Vitousek PM Huebert BJ Hedin LO 1999Changing sources of nutrients during four million years of ecosystem de-

velopment Nature 397 491ndash497Dayton P 1989 Interdecadal variation in an Antarctic sponge and its preda-

tors from oceanographic climate shifts Science 245 1484ndash1486de Ruiter PC Neutel A Moore JC 1995 Energetics patterns of interaction

strengths and stability in real ecosystems Science 269 1257ndash1260

Elser JJ Dobberfuhl D MacKay NA Schampel JH 1996 Organism size lifehistory and NP stoichiometry Towards a unified view of cellular andecosystem processes BioScience 46 674ndash684

Futuyma DJ Mitter C 1996 Insectndashplant interactions The evolution ofcomponent communities Philosophical Transactions of the Royal Society

of London B 351 1361ndash1366Gotelli NJ 1999 How do communities come together Science 286

1684ndash1685Grime JP et al 1997 Integrated screening validates primary axes of special-

ization in plants Oikos 79 259ndash281Hairston NG Jr Lampert W Caacuteceres CE Holtmeier CL Weider LJ Gaedke

U Fischer JM Fox JA Post DM 1999 Rapid evolution revealed by dor-mant eggs Nature 401 446

Holt RD 1993 Ecology at the mesoscale The influence of regional processeson local communities Pages 77ndash88 in Ricklefs R Schluter D eds SpeciesDiversity in Ecological Communities Chicago University of Chicago

PressHolt RD Gaines MS 1992 Analysis of adaptation in heterogeneous land-

scapes Implications for the evolution of fundamental niches EvolutionaryEcology 6 433ndash447

Hooper DU Vitousek PM 1997 The effects of plant composition and di-

versity on ecosystem processes Science 277 1302ndash1305Jackson ML Levelt TVM Syers JK Rex RW Clayton RN Sherman GD

Uehara G 1971 Geomorphological relationships of troposphericallyderived quartz in soils of the Hawaiian Islands Soil Science Society ofAmerica Journal 35 515ndash525

Janzen DH 1976 Why bamboos wait so long to flower Annual Review ofEcology and Systematics 7 347ndash91

Jones CG Ostfeld RS Richard MP Schauber EM Wolff JO 1998 Chain re-actions linking acorns to gypsy moth outbreaks and Lyme disease riskScience 279 1023ndash1026

Jordan WR III Gilpin ME Aber JD eds 1987 Restoration Ecology Cambridge(UK) Cambridge University Press

Lee CndashK Klopp RG Weindruch R Prolla TA 1999 Gene expression profileof aging and its retardation by caloric restriction Science 285 2390ndash2393

Levine JM DrsquoAntonio CM 1999 Elton revisited A review of evidence link-ing diversity and invasibility Oikos 87 15ndash26

Naeem S Li S 1997 Biodiversity enhances ecosystem reliability Nature 390507ndash509

NoyndashMeir I 1974 Desert ecosystems Higher trophic levels Annual Reviewof Ecology and Systematics 5 195ndash214

Pellmyr O LeebensndashMack J Thompson JN 1998 Herbivores and molecu-

lar clocks as tools in plant biogeography Biological Journal of theLinnean Society 63 367ndash378

Petchey OL McPhearson PT Casey TM Morin PJ 1999 Environmentalwarming alters food web structure and ecosystem function Nature 40269ndash72

Pimm SL 1989 Theories of predicting success and impact of introducedspecies Pages 351ndash367 in Drake JA DiCastri F Groves RH Kruger FJMooney HA Rejmaacutenek M Williamson MH eds Biological InvasionsA Global Perspective Chichester (UK) John Wiley amp Sons

Polis GA Holt RD Menge BA Winemiller K 1996 Time space and life his-

tory Influences on food webs Pages 435ndash460 in Polis GA WinemillerK eds Food Webs Integration of Patterns and Dynamics New YorkChapman and Hall

Polis GA Anderson W Holt RD 1997 Towards an integration of landscape

ecology and food web ecology The dynamics of spatially subsidized foodwebs Annual Review of Ecology and Systematics 28 289ndash316

Polis GA Hurd SD Jackson CT SanchezndashPintildeero F 1997 El Nintildeo effects onthe dynamics and control of a terrestrial island ecosystem in the Gulf ofCalifornia Ecology 78 1884ndash1897

Power ME Rainey WE In press Food webs and resource sheds Towards spa-tially delimiting trophic interactions In Hutchings MJ John EA edsEcological Consequences of Habitat Heterogeneity London BlackwellScientific

Rejmaacutenek M 1989 Invasibility of plant communities Pages 369ndash388 inDrake JA DiCastri F Groves RH Kruger FJ Mooney HA Rejmaacutenek MWilliamson MH eds Biological Invasions A Global PerspectiveChichester (UK) John Wiley amp Sons

Rejmaacutenek M Richardson DM 1996 What attributes make some plantspecies more invasive Ecology 77 1655ndash1661

Ritchie M Olff H 1999 Spatial scaling laws yield a synthetic theory of bio-diversity Nature 400 557ndash560

Rosenzweig ML 1995 Species diversity in space and time Cambridge (UK)

Cambridge University PressSears AHolt RD Polis GA In press The effects of temporal variability in pro-

ductivity on food webs and communities In Polis GA Power MEHuxelm G eds Food Webs at the Landscape Level Chicago University

of Chicago PressSimberloff D Schmitz DC Brown TC eds 1997 Strangers in Paradise

Impact of Management of Nonindigenous Species in Florida Washington(DC) Island Press

Sterner RW 1995 Elemental stoichiometry of species in ecosystems Pages240ndash252 in Jones CG Lawton JH eds Linking Species and EcosystemsNew York Chapman and Hall

Thompson JN 1998 Rapid evolution as an ecological process Trends inEcology and Evolution 13 329ndash332

______ 1999 The evolution of species interactions Science 284 2116ndash2118Thompson JN Cunningham BM Segraves KA Althoff DM Wagner D

1997 Plant polyploidy and insectplant interactionsAmerican Naturalist150 730ndash743

Tilman D Knops J Wedin D Reich P Ritchie M Siemann E 1997 The in-fluence of functional diversity and composition on ecosystem processesScience 277 1300ndash1302

Vitousek PM Hooper DU 1993 Biological diversity and terrestrial ecosys-

tem biogeochemistry Pages 3ndash14 in Schulze EndashD Mooney HA edsBiodiversity and Ecosystem Function Berlin Springer-Verlag

Wall Freckman D Blackburn TH Brussaard L Hutchings P Palmer MASnelgrove PVR 1997 Linking biodiversity and ecosystem functioning ofsoils and sediments Ambio 26 556ndash562

Webb CO 2000 Exploring the phylogenetic structure of ecological com-munities An example from rain forest trees American Naturalist 156145ndash156

Weiher E Keddy P eds 1999 Ecological Assembly Rules Perspectives

Advances Retreats New York Cambridge University PressWhitham TG Martinsen GD Floate KD Dungey HS Potts BM Keim P 1999

Plant hybrid zones affect biodiversity Tools for a genetic-based under-standing of community structure Ecology 80 416ndash428

Wilson JB 1999 Guilds functional types and ecological groups Oikos 86507ndash522


Articles


Articles

regional species pool This coalescence depends on inter-actions among species availability physical environment evo-lutionary history and temporal sequence of assemblyEcologists have made important strides in understanding theprocess of community coalescence but it will take evengreater integration of approaches for us to be able to con-fidently predict the pathways or endpoints of community as-sembly (Belyea and Lancaster 1999 Gotelli 1999 Weiher andKeddy 1999) So far we cannot predict which species arelikely to invade or to be lost from particular natural com-munities although some patterns are beginning to emerge(Petchey et al 1999) We also have much to learn aboutcommunity responses to perturbations at different stages indevelopment

Much of what we do know comes from a handful of eas-ily studied systems whereas the functioning of communi-ties undoubtedly also depends heavily on little-studiedhidden players such as microbes fungi and soil invertebrates(de Ruiter et al 1995) Indeed many attempts to create orrestore communities (eg freshwater wetlands or saltmarshes) fail for reasons that remain poorly understoodUntil we can fill these fundamental gaps in our knowledgehuman impacts on community patterns will remain hard topredict

Research conducted over the last decade suggests whatkinds of studies are needed to fill the gaps Progress will al-most certainly depend on developing new ways to simplifythe study of complex communities Focusing on functionalgroups or other as yet undeveloped constructs may help Weare also aware that we need to learn far more about the sys-tematics of many smaller or cryptic organisms which mayform links that are crucial to our understanding of com-munity coalescence The difficulty of linking importantecosystem functions to key taxa and interspecific interactionsis a particular challenge

Research in five key areas described below could provideimportant insights into the community assembly process

Functional traits and community composition Aswe study communities that are increasingly a mix of nativeand introduced taxa we need to find out whether coevolvedspecies or local populations differ from coevolutionarily naivepopulations in the strength or nature of their interactions dur-ing or after community assemblyWe need to improve our abil-ity to pinpoint the traits of species that affect the probabilityof invasion or extinction within developing communities(Pimm 1989 Rejmaacutenek and Richardson 1996 Belyea andLancaster 1999) Moreover we need to further our under-standing of how the importance of those traits depends on ex-isting species composition critical thresholds (eg speciesrichness functional group composition) and history of acommunity (Rejmaacutenek 1989 Burke and Grime 1996 Levineand DrsquoAntonio 1999)

Pathways toward community coalescence For anypool of potential community members there are many

possible assembly sequences and endpoints The diversity ofinitial sequences is less problematic if many of the alternativepathways tend to converge on a limited subset of endpointsTo that end we need to find out whether developing com-munities move toward single or multiple points or states (at-tractors) Past some stage of assembly communitycomposition may become canalized but it may be highlysusceptible to perturbations in the early stages of coales-cence Anthropogenic change may limit pathways of com-munity development by changing the physicochemicalenvironment altering biogeochemical cycles or changingthe genetic structure of populations Global homogeniza-tion of the species pool may itself alter the pathways or end-points of community assembly reshaping both short-term andlong-term successional patterns across landscapes

Functional groups Ecologists recognize that the con-cept of functional groups is a valuable tool for simplifyingcommunity complexity to manageable levels Currently mostgroups are defined in a system-specific way (Wilson 1999)based on biochemical morphological or trophic criteria(Vitousek and Hooper 1993 Hooper and Vitousek 1997Naeem and Li 1997 Tilman et al 1997) although some at-tempts have been made to develop general classifications(Grime et al 1997) Each of these attempts has helped us iden-tify the advantages and disadvantages of different approachesTogether they are moving us toward the development ofrobust frameworks that work across multiple taxa and ecological communities Those advances should ultimatelyhelp us to understand how functional group structure shapesassembly dynamics as well as to evaluate how different con-figurations of functional groups affect the development ofcommunities or the dynamics of ecosystems (eg rates andtrajectories of community development with and without nitrogen-fixing species)

Hidden players Community assembly almost certainly de-pends upon cryptic invertebrate microbiological and my-cological groups which include a vast array of free-livingspecies parasites and mutualistic symbionts These groups andinteractions have historically been ignored or unrecognizedbecause of technological limitations and biases in the train-ing of ecologists Many microbial species cannot yet even becultured Nevertheless they may play a keystone role in com-munity development and function and their absence may beresponsible for some failed attempts at community restora-tion (Wall Freckman et al 1997 Brussaard 1998) The im-portance of these hidden players may become apparent onlywhen they become problems as sometimes happens when wealter community structure Lyme disease for example maybe the result of the emergence of a hidden player in responseto changes in community composition and landscape patterns(Jones et al 1998)

Part of our challenge will be to determine how the struc-ture of microbial diversity should be incorporated into dif-ferent kinds of analyses of the local regional and global







Articles












Articles








Articles













Articles















Articles









Articles
















Articles



86 402ndash416




























































Articles







Articles












Articles








Articles













Articles















Articles









Articles
















Articles



86 402ndash416




























































Articles











Articles








Articles













Articles















Articles









Articles
















Articles



86 402ndash416




























































Articles








Articles













Articles















Articles









Articles
















Articles



86 402ndash416




























































Articles












Articles















Articles









Articles
















Articles



86 402ndash416




























































Articles















Articles









Articles
















Articles



86 402ndash416




























































Articles









Articles
















Articles



86 402ndash416




























































Articles















Articles



86 402ndash416




























































Articles



86 402ndash416




























































Articles

Documents

Frontiers of Ecology