Professional Documents
Culture Documents
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Annual Reviews is collaborating with JSTOR to digitize, preserve and extend access to Annual Review of
Ecology, Evolution, and Systematics.
http://www.jstor.org
OF ECOLOGICAL
INDICATORS*
APPLICATION
GeraldJ.Niemi1and MichaelE.McDonald2
'NaturalResourcesResearchInstituteand
of
of Biology,University
Department
Minnesota,Duluth,Minnesota55811;email:gniemi@d.umn.edu
andAssessment
Protection
2U.S.Environmental
Monitoring
Agency,Environmental
email:
20191;
McDonald.Michael@epa.gov
Program,Reston,Virginia
Key Words assessment,condition,monitoring,responses,stressors
0 Abstract Ecologicalindicatorshave widespreadappealto scientists,environmentalmanagers,and the generalpublic. Indicatorshave long been used to detect
changesin nature,but the scientificmaturationin indicatordevelopmentprimarily
has occurredin the past40 years.Currently,indicatorsaremainlyused to assess the
conditionof the environment,as early-warningsignals of ecological problems,and
as barometersfor trendsin ecologicalresources.Use of ecologicalindicatorsrequires
clearlystatedobjectives;the recognitionof spatialandtemporalscales; assessments
of statisticalvariability,precision,andaccuracy;linkageswith specificstressors;and
couplingwitheconomicandsocialindicators.Legislativelymandateduse of ecological
accords.
indicatorsoccursin manycountriesworldwideandis includedin international
As scientificadvancementsandinnovationin the developmentanduse of ecological
indicatorscontinuethroughapplicationsof molecularbiology,computertechnology
and
suchas geographicinformationsystems,datamanagementsuchas bioinformatics,
remotesensing,ourabilityto applyecologicalindicatorsto detectsignalsof environmentalchangewill be substantiallyenhanced.
INTRODUCTION
Humanstryingto understandthe currentconditionor predictthe futurecondition
of ecosystems have often resortedto simple, easily interpretedsurrogates.Often
these surrogateshavebeen indicatorsthatallow humansto isolatekey aspectsof the
environmentfrom an overwhelmingarrayof signals [NationalResearchCouncil
(NRC) 2000]. Early humansused indicatorslike seasonal migratorymovements
of animals or floweringby spring flora to provide insight into changing environmental conditions. The first reference to environmentalindicators is attributed
to Plato, who cited the impacts of human activity on fruit tree harvest (Rapport
1992). Morrison(1986) reviewedthe work of Clements(1920) and noted thatthe
*TheU.S. Government
hastherightto retaina nonexclusive,
royalty-free
licensein andto
anycopyrightcoveringthispaper.
89
90
NIEMI0 MCDONALD
conceptof indicatorsfor plantandanimalcommunitiescan be tracedto the 1600s.
Clements's (1920) work set the scientific stage for using plants as indicatorsof
physical processes, changes to soil conditions, and other factors. In the 1920s,
indicatorswere also being successfullyused to determinechangingenvironmental
conditions, such as waterclarity (Rapport1992) or air qualitywith "thecanaryin
the mine" (Burrell& Siebert 1916), which we continueto use (VanBiema 1995).
One of the more elaborateearly environmentalindicatorswas the saprobiansystem (Kolkwitz & Marsson 1908), which used benthic and planktonicplants as
indicatorspecies for classifying streamdecompositionzones.
The past 40 years have seen a rapidaccelerationof scientificinterestin the developmentandapplicationof ecological indicators.This focus on indicatorsstems
from the need to assess ecological condition in making regulatory,stewardship,
sustainability,or biodiversitydecisions. Forexample,the CleanWaterAct of 1972
requiresthat each state produce a reportevery two years on the condition of all
its waters to the U.S. EnvironmentalProtectionAgency (US EPA) for Congress.
Decisions regardingsustainabilityand biodiversityinvolve both researchandpolicy issues (e.g., Mann & Plummer 1999, Ostrom et al. 1999, Tilman 1999). In
the United States, this researchhas been legislatively mandatedto variousfederal
agencies, in particularto the U.S. Departmentof Agriculturethrough the National ForestManagementAct of 1976, to the U.S. Departmentof Interior(1980,
Parsons2004), andto the US EPA (2002b). These mandateshaveresultedfromthe
increasingconcern for the loss of species, deterioratingwater quality and quantity, sustainabilityof resource use, climate change, and overall condition of the
environment.This interesthas generatedmany new books, articles, and reviews
on ecological indicators(e.g., McKenzieet al. 1992; US EPA2002b,c), as well as
a new journal(Ecological Indicatorsin 2001).
The public has increasinglydemandeda betteraccountingof the conditionor
health of the environmentand whether it is improving or getting worse (Heinz
Center 1999; www.heinzctr.org/ecosystems).Developing scientificallydefensible
indicatorsto establish environmentalbaselines and trendsis a universalneed at
a variety of levels. For instance, federal governmentsin the United States and
Canada(EnvironmentCanadaand US EPA 2003), Europe (www.eionet.eu.int),
and Australia(www.csiro.au/csiro/envind/index.htm)
have developed or are defor
routine
on
veloping programs
reporting ecological indicators.Recent international accords(e.g., RIO Accord) have demandedan accountingand reportingof
indicatorson the stateof the environment.The Montr6alProcess (www.mpci.org)
representing12 countrieswas establishedin 1994 to developand implementinternationallyagreedupon criteriaandindicatorsfor the conservationand sustainable
managementof temperateand boreal forests. In 2003, US EPA (2003a) released
its first state of the environmentreport(www.epa.gov/indicators/roe/index.htm).
As the world human population continues to increase exponentially (Cohen
2003), and with consequent environmentaldemands,the applicationsof indicators to determinestatus and trends in environmentalcondition will continue to
grow.
ECOLOGICAL
INDICATORS
91
DEFINITIONS
Earlyuses of indicatorsprimarilyreflectedenvironmentalconditions,andthe terms
environmentalandecological indicatorshave oftenbeen used interchangeably.Environmentalindicatorsshouldreflectall the elements of the causal chainthatlinks
humanactivitiesto theirultimateenvironmentalimpactsandthe societalresponses
to these impacts(Smeets & Weterings1999). Ecological indicatorsarethen a subset of environmentalindicatorsthatapplyto ecological processes (NRC 2000). For
policy makers,the amountof ecological datais often overwhelming.Environmental indicatorsarean attemptto reducethe informationoverload,isolate key aspects
of the environmentalcondition,documentlarge-scalepatterns,andhelp determine
appropriateactions(Niemeijer2002). An exampleof a large-scale,policy-relevant
environmentalindicatoris the environmentalsustainabilityindex (ESI). The ESI
was developed to allow quantitativeinternationalcomparisonsof environmental
conditions (WorldEconomic Forum 2002). The ESI has five major categories:
environmentalsystems, reducingenvironmentalstresses,reducinghumanvulnerability, social and institutionalcapacity,and global stewardship.In 2001 the ESI
includedinformationon 68 indicatorswithin these categoriesfrom 142 countries
(WorldEconomic Forum2002).
Ecological indicatorsembodyvariousdefinitionsof ecology, such as the "interactionsthatdeterminethe distributionandabundanceof organisms"(Krebs 1978),
or morebroadlythe "structureandfunctionof nature"(Odum1963). Thus,they are
often primarilybiological andrespondto chemical,physical, and otherbiological
(e.g., introducedspecies) phenomena.We have chosen to combine the definitions
of the US EPA(2002b) andthe hierarchyof Noss (1990), andwe defineecological
indicatorsas: measurablecharacteristicsof the structure(e.g., genetic, population,
habitat, and landscape pattern),composition (e.g., genes, species, populations,
communities, and landscape types), or function (e.g., genetic, demographic/life
history,ecosystem, and landscapedisturbanceprocesses) of ecological systems.
Ecological indicatorsare derivedfrom measurementsof the currentcondition
of ecological systems in the field and are either used directly or combined into
one or more summaryvalues (US EPA 2002b). These ecological indicatorscan
be aggregatedinto ecological attributeswith reportingcategories, such as biotic
condition,chemical andphysical characteristics,ecological processes, and disturbance (Harwellet al. 1999, US EPA2002b). Ecosystemdisturbancecan be natural
(e.g., fire, wind, and drought)and partof the functionalattributesof ecosystems
(Noss 1990), or it can be anthropogenic.Ecological indicatorshave been applied
in manyways in the contextof both naturaldisturbancesandanthropogenicstress.
However,the primaryrole of ecological indicatorsis to measurethe responseof the
ecosystem to anthropogenicdisturbances,but not necessarily to identify specific
anthropogenicstress(es) causing impairment(US EPA 2002b). These indicators
have been referredto as "state indicators"in the State of the Lakes Ecosystem
Conference(SOLEC),which is a joint effort of Canadaand the United States to
develop indicatorsfor the GreatLakes (EnvironmentCanadaand US EPA 2003).
92
NIEMI0 MCDONALD
SOLECdefines state indicatorsas response variables(e.g., fish, bird, amphibian
populations)and pressureindicatorsas the stressors(e.g., phosphorusconcentrations, atmosphericdepositionof toxic chemicals, or waterlevel fluctuations).
In this review we focus on ecological indicators,but clearly they can be integratedwith the broaderissues of ecosystem health (Rapportet al. 1998) and
ultimately with economic indicators (Milon & Shogren 1995) to be even more
useful for making policy decisions. There is a continuingdebate on how to accomplishthis integration.A commongoal of linkingeconomic andenvironmental
indicatorsis often based on the concept of sustainability.Forexample,Ekinset al.
(2003) provided a frameworkfor linking economic, social, and environmental
sustainability.Their approachidentifiedhow economic and social options were
constrainedif critical environmentalfunctions were sustained.Lawn (2003) explored the theoreticalfoundationof several indexes of sustainability,including
the Index of SustainableEconomic Welfareand the Genuine ProgressIndicator.
He found that these indexes were theoreticallysound, but more robustvaluation
methodswere necessary.Althoughprogressis being made,thereareno indicators
that link economic, social, or environmentaltrendsin a way thatis meaningfulto
the public.
USE OF ECOLOGICAL
INDICATORS
Ecological indicatorsare primarilyused either to assess the condition of the environment(e.g., as an early-warningsystem) or to diagnose the cause of environmental change (Dale & Beyeler 2001). The widespreaddecline of the peregrine
falcon (Falco peregrinus)in the 1950s is an excellent example of both uses. The
catastrophicdecline of the species served as an early-warningsignal of problems
in the environment,and researchon the cause of the decline led to the diagnosis
of widespreadcontaminationby chlorinatedhydrocarbonssuch as DDT (Ratcliffe
1980). The widespreaddecline of amphibianshas also been viewed as an earlywarningsignal of problemsin the environment,yet furtherresearchhas failed to
identify a specific cause for the decline. Amphibiandeclines are likely due to a
varietyof factors,includinghabitatchange, global climate change, chemical contamination,disease and pathogens,invasive species, and commercialexploitation
(Blaustein & Wake 1995, Semlitsch 2003).
The informationgatheredby ecological indicatorscan also be used to forecast futurechanges in the environment,to identify actions for remediation,or, if
monitoredover time, to identify changes or trendsin indicators(Figure 1). As the
complexity of the system being monitoredincreases (e.g., greaterspatial scales
and levels of biological organization)or as the temporalscale increases,the cost
of gathering,analyzing, and reportingon indicatorsincreases. Complexity also
arises from the need to quantifylinkages between specific stressorsand ecological indicators(Table 1). In the few cases in which such relationshipshave been
determined,these ecological indicatorsare often considereddiagnostic;however,
these linkages have seldom been made (Suter et al. 2002). A major challenge
ECOLOGICAL
INDICATORS
93
Monitoring networks
(decreasecosts4 decreaseinformation)
Landscape
Ecosystem
Community
Population
scale Individual
Tissue
Cell
Spatial
Gene
Condition
Diagnosis
Forecasting
Remediation
Change/Trends
Temporal
scale
Scientific understanding
(increaseinformation
- increasecosts)
94
NIEMIa MCDONALD
TABLE1 Examplesof ecologicalresponsesto naturalandanthropogenic
stress
Stress
Naturaldisturbance
Forestfire
Drought
Herbivory
Wind
Anthropogenicstress
Acidrain
Eutrophication
Introduced
species
Sedimentation
Logging
Heavymetals
Urbanization
Airpollution
Airquality
Ecologicalresponse
Reference
Landscape
pattern
Birdpopulations
andlitter
Vegetation
Foreststands
Turner
et al. 1994
Blakeet al. 1994
McInneset al. 1992
Foster1988
Feathermoss
Aquaticmacrophytes
Birdpopulations
Shrubs
Landscape
pattern
Mosquitogenefrequencies
Birdguilds
Plantspecies
Lichens
Hutchinson
& Scott1988
Kangaset al. 1982
Savidge1984
Johnston2003
Franklin
& Forman1987
Guttman1994
O'Connellet al. 2000
Stolte& Mangis1992
Kinnunen
et al. 2003
ECOLOGICAL
INDICATORS
95
INDICATORSPECIES
Most applicationsof ecological indicatorshave focused at the species level owing to concerns arising from endangeredspecies and species conservationissues
(Fleishmanet al. 2001). For instance,Noss (1990) statedthat "theuse of indicator
species to monitoror assess environmentalconditionsis a firmlyestablishedtradition in ecology, environmentaltoxicology, pollutioncontrol, agriculture,forestry,
and wildlife and range management."The measurementof an indicatorspecies
assumes that a single species representsmany species with similarecological requirements(Landreset al. 1988). Typically,ecological indicatorspecies tend to
be from the macrofloraand macrofauna,especially aquatic macroinvertebrates,
fish, birds, and vascularplants. The primaryreasons for theiruse are: (a) relative
96
NIEMI M MCDONALD
thathavebeenappliedwithindifferentecologicallevelsof
TABLE2 Examplesof indicators
(modifiedfromNoss 1990andUS EPA2003)
organization
Example
Type
References
Compositional Genes
Cell and subcellular
Tissue
Species
Functional
Structural
Integrative
Habitatphysiognomy
Landscapepatterns
Index of biotic integrity
AMOEBA
Multivariate
Species assemblages
Foreststructure
Fragmentation
Fish
Multipletaxa
Biomarkers
Beetles
1991
Lindenmayeret al. 2000
O'Neill et al. 1988
Karr1981
ten Brink et al. 1991
Cormier& Racine 1992
Dufr6ne& Legendre
1997
Paul 2003
INDICATORS
ECOLOGICAL
97
98
NIEMIE MCDONALD
areof public concernin the region.Noss (1999) suggestedthatat least for the first
fourcategoriesumbrellaspecies could be definedthatarethe most sensitiveto the
landscapeattribute.
Birds have been the primaryfocus for most terrestrialapplicationsof indicator
species, but insects hold greatpromisebecause of theirspecies richness,biomass,
and role in ecosystem functioning.McGeoch (1998) recognizedthe potentialapplications of insects as indicatorspecies and defined their use as environmental,
ecological, or biodiversityindicators.Researchershave attemptedto examinevertebrates as possible umbrella species for insects, especially for butterflies.For
example, Rubinoff(2001) analyzedthe use of a bird species, the Californiagnatcatcher(Poliotila californica), as a potentialumbrellaspecies for three species of
butterflies.However, the gnatcatcherwas a poor indicatorprimarilybecause of
its ubiquityin the landscapestudied.Insects and othermicrofaunaoffer excellent
potentialas indicatorspecies. They areof limiteduse in terrestrialsystemsbecause
of the cost of samplingand processingandbecause thereis limited acceptanceby
resourcemanagers,politicians, and the generalpublic.
Researchershave developedotherindexes to providemore holistic approaches
to ecological condition.These indexes range from simple diversityindexes, such
as the Shannonand WienerIndex (Shannon& Weaver 1949), to multimetricindexes (e.g., Karr1981, Kerans& Karr1994, Karr2000, Simon 2003). Multimetric
ecological indicatorsare sets of mathematicallyaggregatedor weighted indicators (US EPA 2000a, Kurtzet al. 2001) that combine attributesof entire biotic
communitiesinto a useful measureof condition (US EPA 2002a). The US EPA
has recently used an index for biotic integrity(IBI) for estuarineinvertebratesas
one of the indicatorsin the assessment of the condition of the nation's estuaries
(US EPA2004). Because of the increasinguse of multimetricandotherindicators,
researchershave developed specific guidelines for evaluatingtheir performance
(US EPA 2000a).
Another aspect of ecological indicatorsis whetherto use them as relative or
absolute measures. As a relative measure, the initial measurementbecomes the
baseline for comparisonof futuremeasures.Most monitoringagencies prefer or
requirea more quantitativebenchmarkfor measuringand regulatingchanges in
ecosystems. These benchmarkor referenceconditionscan be definedas the conditions of ecological resourcesunder minimal contemporaryhumandisturbance
(McDonald et al. 2004). As these conditions are often not available for direct
measurement,models and historicinformationare often invokedas best approximations.However,the selectionof referenceconditionsremainsproblematic(NRC
1990).
In summary,focal species representthose selected as a focus for a specific
investigation.There is no consistent definitionof focal species, but the concept
has been expandedfor use in conservationand management.Focal species have
been used to identify potential indicator species when there is a desire to describe ecological condition or measure the response to a disturbance.Either a
focal species or an indicatorspecies may serve as an umbrellaspecies if the goals
ECOLOGICAL
INDICATORS
99
IN APPLICATIONSOF ECOLOGICAL
COMPLEXITIES
INDICATORS
Monitoring for ecosystem or resource managementoften requiresdata about a
specific site or sites, whereaspublic policy decisions typicallyrequireinformation
across broadergeographicalregions (Olsen et al. 1999). Many of the existing
ecological monitoringprogramsare periodically or continually used at certain
sites, which may lead to a betterunderstandingof the temporalvariabilityat the
site but may not be representativeof a largerarea (Urquhartet al. 1999). Thus,
ecological indicatorsare needed to assess statusand trendsin ecological systems
and to diagnose cause(s) of declining condition across a range of spatial and
temporalscales (Kratzet al. 1995, NRC 2000, Dale & Beyeler 2001, Niemi et al.
2004).
Each ecological indicatorrespondsover differentspatial and temporalscales;
thus,the contextof these scales mustbe explicitly statedfor each ecological indicator.Furthermore,understandingthe responsevariabilityin ecological indicatorsis
essentialfor theireffective use (US EPA2002c). Withoutsuch an understanding,it
is impossibleto differentiatemeasurementerrorfromchangingcondition,or ananthropogenicsignal from backgroundvariation.Workhas begun on understanding
how naturaland anthropogenicvariabilityof indicatorscan affect statusand trend
detection,but it is closely tied to differentstatisticaldesign considerations(Larsen
et al. 1995). Specific monitoringdesigns and indicatorscan be implementedto
detect changes across temporaland spatialscales (US EPA 1997, 2002d).
In general, as one moves up levels of organizationfrom cellular phenomena
to landscape processes, the spatial and temporal scales of applicationincrease
immensely. Similarly, as larger spatial and temporal scales are considered, the
linkage to specific stressorscan be either obscured or refined depending on the
stressor.For example,one of the largestandmost successful monitoringprograms
is the U.S. Geological Survey Breeding Bird Survey (BBS), which has gathered
dataovera 38-yearperiod(1966 to present)(Saueret al. 2003). TheBBS is intended
to indicatebreedingbird species trendsover relativelylargeregional and national
spatial scales. Thus, researchersmust exercise caution in interpretingresults for
specific regions or combining results from differentregions (James et al. 1996).
In contrast,nesting tree swallows (Tachycinetabicolor) are an effective wildlife
100
NIEMIMMCDONALD
indicatorspecies of sedimentchemical contaminantproblems.Because nestlings
are fed flying adult insects, which typically have aquaticearly life histories, the
uptakeof chemicals by nestlings can be relatedto sedimentchemical levels near
the nesting site (Nichols et al. 1995, Jones 2003). Changesin bird trendsfrom the
BBS over large areas are powerfulbecause of the large numberof sample routes
and the a prioriexperimentaldesign, but the causes of changes in species trends
are speculative. In contrast, contaminantuptake in nestling tree swallows and
potentialrisk to wildlife can clearly be connected to food supplies derivedfrom
sediment. Many of the same problems exist for multimetricindexes commonly
used to assess conditionof surfacewatersacross largeregions. These indexes can
distinguishdegradedsites from sites with little or no humanimpact, but they do
not diagnose the causes of impairmentby themselves (US EPA 2003b).
INDICATORSOF
IN ECOLOGICAL
ADVANCEMENTS
BIOLOGICALCOMMUNITIES
Historically,ecological indicatorswere primarilybased on parametersassociated
with individualspecies (e.g., presence)or simple communitymetrics(e.g., species
richnessor diversity).However,manyof these indicatorsdid not fully representthe
entirebiological communityof organismspresent.Hence, Karr(1981) introduced
the IBI using streamfish communities. This index was a numericalsummation
of subsets of the fish communityfrom one area comparedwith a suitable reference area. These reference areas ideally representedareas that were naturalor
undisturbedfrom the same geographic area and with the same general ecological condition. Karr(1981) calculated the IBI using fish community data for a
specific area and subdivided these data into 12 metrics, including the number
of individualsand species found in the sample, the relative abundanceof guilds
(e.g., carnivores),specific species in the sample,andothercategories(e.g., sunfish
species). The IBI was expressedas deviationsfromthe suitablereferenceareasuch
that largervalues representedcommunitiessimilarto the referencearea,whereas
lower values representedareasthatdeviatefromthe reference,potentiallybecause
of stress. The IBI has received considerableattentionand applicationover the
past 20 years, including applicationsto fish (Fausch et al. 1984, Angermeier&
Karr1986, Karret al. 1986, Simon & Emery 1995), macroinvertebrates
(Kerans&
et
al.
communities
al.
et
Klemm
2001,
Karr1994,
(Simon
DeKeyser
2003), plant
et al. 2003), aquaticcommunities(Simon et al. 2000), and birds (O'Connellet al.
2000).
Many other multimetricindexes have evolved over the past 20 years, such
as the Hilsenhoff biotic index (Hilsenhoff 1982) and biological response signatures (Simon 2003). In contrastto multimetricindexes, multivariateindexes
(Reynoldsonet al. 1997, Karr2000) are statisticalanalyses of the biological communityusing a host of multivariatetechniques,such as principalcomponentsanalysis (O'Connoret al. 2000), canonical correspondenceanalysis (Kingston et al.
ECOLOGICAL
INDICATORS
101
CRITICISMSOF INDICATORS
Virtually all attemptsto use ecological indicatorshave been heavily criticized,
and many criticisms are well deserved. For instance, many existing monitoring
programsof indicatorssuffer from two deficiencies: lack of well-articulatedobjectives and neglect of different sources of error(Yoccoz et al. 2001). Indicator
species havebeen especially criticizedin the contextof forestmanagement-related
issues (Landreset al. 1988, Landres1992, Niemi et al. 1997, Rolstad et al. 2002,
Failing & Gregory2003). Many of these criticisms have focused on the lack of:
(a) identificationof the appropriatecontext(spatialandtemporal)for the indicator,
(b) a conceptualframeworkfor what the indicatoris indicating,(c) integrationof
science and values, and (d) validationof the indicator.
Many of these criticismshave led to more focused efforts on individualspecies
and to the developmentof additionalconcepts such as focal species or umbrella
species (Lambeck 1997, Fleischman et al. 2001). Roberge & Angelstam (2004)
recently reviewed the umbrellaspecies concept and concluded that multispecies
strategieswere more compelling. Lawler et al. (2003) evaluatedseveral indicator groups (e.g., birds, fish, mammals, and mussels) to test whether one group
could providehabitatprotectionfor othertaxa in a large areaof the Mid-Atlantic
region of the United States. No single taxonomic indicatorgroup could provide
adequateprotectionfor anothergroup,especially for at-riskspecies withineach of
the groups.The failurewas likely attributableto the narrowgeographicrangesand
restrictedhabitatdistributionof rare species. Hence, informationon rare species
and those thatare at risk was essential,yet gatheringdataon rarespecies is generally difficult,time-consuming,and expensive. In contrastto the indicatorspecies
102
NIEMI0 MCDONALD
approach,Manley et al. (2004) evaluatedan innovative,multispeciesmonitoring
approachthat included all terrestrialvertebratespecies over a large ecoregional
scale (7 million ha). The design of this comprehensiveapproachreducedthe emphasis on indicatorspecies because the spatialcoverage allowed many species to
be adequatelymonitored.A fundamentalproblemwith these approachescontinues to be the inabilityto link species presenceor relativeabundancewith relevant
aspects of habitatquality (VanHome 1983), such as productivity.
Many of these same criticisms apply to indexes (Suter 1993). Indexes have
been viewed as oversimplificationsand generalizationsof biological processes, in
which importantdatacan be lost (May 1985). There are also concernsabouthow
these indexes are calibrated(Seegert 2001) and whether or how they are evaluated across gradients(US EPA 2000a). Despite such criticisms,these indexes can
play an importantmanagementrole by helping characterizeecological condition
(Rakocinskiet al. 1997).
Ecological indicators span broad levels of biological, spatial, and temporal
organizationwithin ecosystems. When establishing a monitoringprogramand
selecting indicators,it is imperativethat researchersarticulatea clear statement
of goals. Once the goals are unambiguouslystated, the scientific soundness and
objectivityof the indicatorbecome a centralissue (Niemeijer2002). Researchers
must recognize these complexities and limitationsin the applicationand use of
ecological indicators(Dale & Beyeler 2001). However,havingeffective indicators
is only one componentof the problem.Sound programdesign and effective data
management,analysis, synthesis, and interpretationare also needed to implement
monitoringand assessment programssuccessfully (NRC 1990). In the past five
years, many publications have provided excellent guidance on how ecological
indicatorscan be improved,includingdocumentsby the NRC (2000) andUS EPA
(2000a, 2002b).
INDICATORS
FUTUREOF ECOLOGICAL
AND CONCLUSIONS
Advances in science and technology at all levels of biological organizationwill
greatly improveour ability to apply ecological indicatorsin the future.Recently
developedtechniquesin molecularbiology such as biomarkershave provenuseful
in rapididentificationof problemsin ecological systems causedby pollutionstress
(e.g., Cormier& Racine 1992, Huggett et al. 1992). For example, Arcand-Hoy
& Metcalfe (1999) found that both fluorescentaromaticcompoundsin bile and
in fish could be used to detect exposureto
hepatic ethoxyresorufin-O-deethylase
the
lowerGreatLakes.Evendon& Depledge
in
polynucleararomatichydrocarbons
(1997) identifiedthe potentialusefulness of genetically susceptiblepopulationsto
environmentalcontaminants.Paerl et al. (2003) have recently used diagnostic
photopigmentsof variousphytoplanktongroupsas ecological indicatorsto detect
changesin nutrients,noxious algal blooms, andoverallwaterquality.Investigators
ECOLOGICAL
INDICATORS
103
104
NIEMIE MCDONALD
has a paucity of informationon which to judge the ecological condition of the
environmentor how the condition might relate to human health or to the economy. Yet, with such information,individualsmake daily and long-termdecisions
on the basis of health indicators(e.g., blood pressure),economic indicators(e.g.,
NASDAQ, Dow Jones IndustrialAverage), and environmentalindicators (e.g.,
weatherforecasts).Despite three decades of discussion of the integrationof economic andecological indicators,therearelimitedapplicationsof integratedanalysis (Milon & Shogren 1995). US EPA's (2003a) state of the environmentreportin
2003 is one of the firststeps in informingthe public of the ecological conditionof
the nation'sresources.Futurereportson integratedand understandablemeasures
will be welcome additionsas indicatorsof environmentalsustainability,but their
acceptanceand impactson policy and public opinion will have to be determined.
Strongpublic interestand legislativemandatesexist at local, state, federal,and
internationallevels to understandthe condition, trends,and cause for change in
our ecosystems. A largearrayof ecological indicatorsareavailablefor application
to environmentalproblems;moreover,the numberof tools and techniques that
are available is rapidly increasing. Therein lies the challenge for the future:to
select appropriatemonitoringdesigns and ecological indicatorsthat will provide
convincingscientificunderpinningsfor managementand policy decisions on realworld problems.
ACKNOWLEDGMENTS
Althoughthe researchdescribedin this articlehas been fundedwholly or in partby
the United StatesEnvironmentalProtectionAgency's Science to Achieve Results
(STAR) programthroughcooperativeagreementR828675-00 to the University
of Minnesota,it has not been subjectedto the agency's requiredpeer and policy
review and thereforedoes not necessarilyreflect the views of the agency, and no
official endorsementshouldbe inferred.We thankJames Cox, RobertHowe, and
LucindaJohnsonfor comments on an earlierversion of this manuscript.This is
publication number 359 of the Center for Waterand the Environment,Natural
Resources ResearchInstitute,Universityof Minnesota,Duluth.
The Annual Reviewof Ecology,Evolution,and Systematicsis online at
http://ecolsys.annualreviews.org
LITERATURE
CITED
tions exposedto copperin vitro.J. Aquat.
Ames BN, Lee F, DurstonW. 1973. An improvedbacterialtestsystemforthedetection Anim.Health1:57-61
of mutagensandcarcino- AndreasenJK, O'Neill RV,Noss R, Slosser
andclassification
for the developNC. 2001. Considerations
gens.Proc.Natl.Acad.Sci. USA70:782-86
of
mentof a terrestrial
index ecologicalinAndersonDP, Dixon OW, BodammerJE,
Lizio EF. 1989. Suppressionof antibody- tegrity.Ecol.Indic.1:21-35
producingcells in rainbowtroutspleensec- AngermeierPL, KarrJR. 1986.Applyingan
ECOLOGICALINDICATORS
index of biotic integritybased on streamfish
communities:considerationsin samplingand
interpretation.North Am. J. Fish. Manage.
6:418-29
Arcand-HoyLD, MetcalfeCD. 1999. Biomarkers of exposure of brown bullheads(Ameiurus nebulosus)to contaminantsin the lower
GreatLakes, North America. Environ.Toxicol. Chem. 18:740-49
Armsworth PR, RoughgardenJE. 2003. The
economic value of ecological stability.Proc.
Natl. Acad. Sci. USA 100(12):7147-51
BairdDJ, BarberI, CalowP. 1990. Clonalvariation in generalresponsesof Daphnia magna
Straus to toxic stress. I: Chroniclife-history
effects. Funct. Ecol. 4:399-408
Blake JG,HanowskiJM,Niemi GJ,Collins PT.
1994. Annual variationin bird populations
of mixed conifer-northernhardwoodforests.
Condor96:381-99
Blaustein AR, Wake DB. 1995. The puzzle of
declining amphibianpopulations. Sci. Am.
272:52-57
BrouwerM, WingeDR, GrayWR. 1989. Structural and functional diversity of coppermetallothioneinfrom the American lobster,
Homarusamericanus.J. Inorg.Biochem.35:
289-303
Browder SF, Johnson DH, Ball IJ. 2002. Assemblages of breedingbirds as indicatorsof
grasslandcondition.Ecol. Indic. 2:257-70
Brown LR. 2003. Plan B: Rescuing a Planet
Under Stress and a Civilizationin Trouble.
New York/London:WW Norton
Burrell GA, Siebert FM. 1916. Gases found
in coal mines. Miner's Circular14. Bureau
Mines, US Dep. Inter.,Washington,DC
CairnsJ Jr,McCormickPV, NiederlehnerBR.
1993. A proposedframeworkfor developing
indicatorsof ecosystem health.Hydrobiologia 263:1-44
Carroll C, Noss RF, Paquet PC. 2001. Carnivores as focal species for conservationplanning in the Rocky Mountain region. Ecol.
Appl. 11:961-80
ClarkDB, ReadJM,ClarkML, CruzAM, Dotti
MF, et al. 2004. Applicationof 1-m and 4-m
resolutionsatellitedatato ecological studies
105
106
NIEMI 0 MCDONALD
ECOLOGICALINDICATORS
KarrJR. 2000. Health, integrity,and biological assessment: the importanceof measuring whole things. In Ecological Integrity:
IntegratingEnvironment,Conservation,and
Health, ed. D Pimentel,L Westra,RF Noss,
pp. 209-26. Washington,DC: Island
Karr JR. 2002. Understanding the consequences of human actions: indicators from
GNP to IBI. In JustEcological Integrity:The
Ethics of MaintainingPlanetary Life, ed. P
Miller, L Westra,pp. 98-110. Lanham,MD:
Rowman& Littlefield
KarrJR, FauschKD, AngermeierPL, YantPR,
Schlosser IJ. 1986. Assessing biological integrity in runningwaters: a method and its
rationale.Spec. Publ. 5, Ill. Nat. Hist. Surv.,
Champaign
KeransBL, KarrJR. 1994. A benthic index of
biotic integrity(B-IBI) for riversof the Tennessee Valley.Ecol. Appl. 4:768-85
Kerr JT, Ostrovsky M. 2003. From space to
species: ecological applicationsfor remote
sensing. TrendsEcol. Evol. 18:299-305
KingstonJC, Birks HJB, UutalaAJ, Cumming
BF, Smol JP. 1992. Assessing trendsin fishery resourcesandlake wateraluminumfrom
paleolimnological analyses of siliceous algae. Can. J. Fish. Aquat.Sci. 49:116-27
Kinnunen H, Holopainen T, KiirenlampiL.
2003. Sources of error in epiphytic lichen
variablesmappedas bioindicators:needs to
modify the Finnish standard.Ecol. Indic.
3:1-11
Klemm DJ, Blocksom KA, Fulk FA, Herlihy
AT,Hughes R, et al. 2003. Developmentand
evaluationof a macroinvertebratebiotic integrity index (MBII) for regionally assessing Mid-Atlantic Highlands streams. Environ. Manage. 31:656-69
Kolkwitz R, Marsson M. 1908. Okologie der
pflanzlichen Saprobien. Ber Dt. Bot. Ges.
26:505-19
KratzTK, MagnusonJJ, Bayley P, Benson BJ,
Barish CW, et al. 1995. Temporaland spatial variabilityas neglected ecosystem properties: lessons learned from 12 American
ecosystems. In Evaluatingand Monitoring
the Health of Large-ScaleEcosystems,ed. D
107
108
NIEMI M MCDONALD
46
O'Connell TJ, Jackson LE, Brooks RP. 2000.
Birdguilds as indicatorsof ecological condition in the CentralAppalachians.Ecol. Appl.
10:1706-21
O'Connor RJ, Walls TE, Hughes RM. 2000.
Using multiple taxonomic groups to index
the ecological condition of lakes. Environ.
Monit.Assess. 61:207-28
Odum EP. 1963. Ecology. New York: Holt,
Rinehart,& Winston
Olsen AR, SedranskJ, EdwardsD, GotwayCA,
Liggett W, et al. 1999. Statisticalissues for
monitoringecological and naturalresources
in the United States.Environ.Monit.Assess.
54:1-45
O'Neill RV, KrummelJR, GardnerRH, Sugihara G, Jackson B, et al. 1988. Indices of
landscapepattern.Landsc.Ecol. 1:153-62
Ostrom E, Burger J, Field CB, Norgarrd
RB, Policansky D. 1999. Revisiting the
ECOLOGICALINDICATORS
commons: local lessons, global challenges.
Science 284:278-82
PaerlHW,ValdesLM, PinckneyJL,PiehlerMF,
Dyble J, MoisanderPH. 2003. Phytoplankton
photopigmentsas indicatorsof estuarineand
coastal eutrophication.BioScience 53:95364
Paine RT. 1969a. A note on trophiccomplexity
and community stability.Am. Nat. 103:9193
Paine RT. 1969b. The Pisaster-Tegulainteraction: prey patches,predatorfood preference,
and intertidalcommunitystructure.Ecology
50:950-61
ParsonsDJ. 2004. Supportingbasic ecological
researchin US nationalparks:challengesand
opportunities.Ecol. Appl. 14:5-13
Paul JF. 2003. Developing and applying an index of environmentalintegrity for the US
Mid-Atlantic region. J. Environ. Manage.
67:175-85
Pdrez-L6pezM, Alonso J, N6via-ValifiasMC,
Melgar MJ. 2003. Assessment of heavy
metal contaminationof seawaterand marine
limpet, Patella vulgata L., from Northwest
Spain. J. Environ.Sci. Health, Part A: Tox.
Hazard.Subst.Environ.Eng. 38:2845-56
PimentelD, WestraL, Noss RF,eds. 2000. Ecological Integrity: Integrating Environment,
Conservation,and Health. Washington,DC:
Island
RakocinskiCF, Brown SS, Gaston GR, Heard
RW,WalkerWW, SummersJK. 1997. Macrobenthic responses to natural and contaminant-relatedgradientsin northernGulf
of Mexico estuaries. Ecol. Appl. 7:127898
RapportDJ. 1992. Evolution of indicators of
ecosystem health.See McKenzieet al. 1992,
1:121-34
RapportD, CostanzaR, EpsteinPR, GaudetC,
Levins R. 1998. EcosystemHealth. Malden,
MA: Blackwell Sci. 72 pp.
RatcliffeDA. 1980. ThePeregrineFalcon. Calton, UK: Poyser
Reynoldson TB, Norris RH, Resh VH, Day
KE, Rosenberg DM. 1997. The reference
condition: a comparisonof multimetricand
109
multivariate approaches to assess waterquality impairmentusing benthic macroinvertebrates.J. North Am. Benthol. Soc. 16:
833-52
Roberge J, Angelstam P. 2004. Usefulness of
the umbrellaspecies concept as a conservation tool. Conserv.Biol. 18:76-85
Rolstad J, GjerdeI, GundersenVS, Setersdal
M. 2002. Use of indicatorspecies to assess
forest continuity: a critique. Conserv.Biol.
16:253-57
Root RB. 1967. The niche exploitationpattern
of the blue-gray gnatcatcher.Ecol. Monogr
37:317-50
Rubinoff D. 2001. Evaluating the California
gnatcatcheras an umbrella species for conservationof southernCaliforniacoastal sage
scrub.Conserv.Biol. 15:1374-83
Rublee PA, KemptonJW, Schaefer EF, Allen
C, Harris J, et al. 2001. Use of molecular
probes to assess geographic distributionof
Pfiesteria species. Environ.Health Perspect.
109:765-67
Rudi K, SkulbergOM, SkulbergR, Jakobsen
KS. 2000. Application of sequence-specific
labeled 16S rRNA gene oligonucleotide
probesfor geneticprofilingof cyanobacterial
abundanceand diversityby arrayhybridization. Appl. Environ.Microbiol.66:4004-11
Sauer JR, Hines JE, Fallon J. 2003. The North
AmericanBreedingBird Survey,Resultsand
Analysis 1966-2002. Version2003.1, USGS
PatuxentWildl. Res. Cent., Laurel,MD
Savidge JA. 1984. Guam: paradise lost for
wildlife. Biol. Conserv.30:305-17
SchillerA, HunsakerCT,KaneMA, Wolfe AK,
Dale VH, et al. 2001. Communicatingecological indicatorsto decision makersandthe
public. Conserv.Ecol. 5:19
Seegert G. 2001. The development, use, and
misuse of biocriteriawith an emphasis on
index of biotic integrity.Environ.Sci. Pol.
3:51-58
SemlitschRD. 2003. AmphibianConservation.
Washington,DC: SmithsonianInst.
ShannonCE,WeaverW. 1949. TheMathematical Theoryof Communication.Urbana:Univ.
Illinois Press
110
NIEMI MMCDONALD
Sierszen ME, Frost TM. 1990. Effects of lake ten Brink BJE, Hosper SH, Colijn F. 1991. A
acidification on zooplankton feeding rates
quantitativemethod for descriptionand assessment of ecosystems: the AMOEBA apand selectivity. Can. J. Fish. Aquat. Sci. 47:
772-79
proach.Mar Pollut.Bull. 23:265-70
SimberloffD. 1998. Flagships, umbrellas,and Tilman D. 1999. Diversity by default.Science
283:495-96
keystones: Is single-species management
Turner
era?
Biol.
Conserv.
in
the
MG, Hargrove WH, Gardner RH,
landscape
pass6
Romme WH. 1994. Effects of fire on land83:247-57
Simon TP, ed. 2003. Biological Response Sigscape heterogeneityin YellowstoneNational
natures: Indicator Patterns Using Aquatic
Park,Wyoming.J. Veg.Sci. 5:731-42
UnderwoodRJ, Roth RR. 2002. Demographic
Communities.Boca Raton,FL: CRC
variablesare poor indicatorsof wood thrush
Simon TP, Emery EB. 1995. Modificationand
assessment of an index of biotic integrity
productivity.Condor 104:92-102
to quantify water resource quality in Great UrquhartNS, Paulsen SG, Larsen DP. 1998.
Rivers.Regul.Rivers:Res. Manage. 11:283Monitoring for policy-relevant regional
trendsover time. Ecol. Appl. 8:246-57
98
Simon TP,JankowskiR, MorrisC. 2000. Mod- US Dep. Inter. 1980. Standardsfor the develification of an index of biotic integrity for
opment of habitat suitabilityindex models.
Ecol. Serv.ManualNo. 103. Div. Ecol. Serv.,
assessing vernal ponds and small palustrine
US Dep. Inter.Fish Wildl.Serv.,Washington,
wetlands using fish, crayfish, and amphibDC
ian assemblagesalong southernLakeMichiEPA. 1997. An ecological assessment
US
3:407Health
Manage.
gan.Aquat.Ecosyst.
of the United States Mid-Atlantic region.
18
SimonTP,StewartPM, RothrockPL. 2001. DeEPA/620/R-97/130,Washington,DC
velopmentof an index of biotic integrityfor US EPA. 1998. Conditionof the Mid-Atlantic
estuaries. EPA 600-R-98-147, Washington,
plant assemblages (P-IBI) in southernLake
DC
Health
Manage.
Michigan. Aquat. Ecosys.
US EPA. 2000a. Evaluationguidelinesfor eco4:293-309
Smeets E, WeteringsR. 1999. Environmental
logical indicators. EPA/620/R-99/005, ResearchTrianglePark,NC.109 pp. httpJi/www.
indicators: typology and overview. Tech.
Environ.
Eur.
25,
ecosystemindicators.org/wg/publication/Jack
Agency, Copenhagen,
Rep.
Den. http://reports.eea.eu.int:80/TEC25/en/ son_KurtzFisher.pdf
US EPA. 2000b. Mid-Atlantic highlands
tech_25_text.pdf
streams assessment. EPA/903/R-00/015,
Stockwell D, Peterson AT. 2003. Comparison
of resolution of methods used in mapping
Washington,DC. 74 pp. http://www.epa.gov/
from
maia/pdf/MAHAStreams.pdf
point-occurrence
biodiversitypatterns
US EPA. 2002a. Environmentalmonitoring
data.Ecol. Indic. 3:213-21
and assessment program research stratStolte KW,MangisDR. 1992. Identificationand
use of plant species as ecological indicators
egy. EPA 620/R-02/002, ResearchTriangle
of air pollution stress in nationalparkunits.
Park,NC. 78 pp. http://www.epa.gov/nheerl/
See McKenzie et al. 1992, 1:373-92
emap/files/emap_research_strategy.pdf
Suter GW II. 1993. A critique of ecosystem US EPA. 2002b. A SAB report: a framework for assessing and reporting on ecohealth concepts and indices. Environ. Toxicol. Chem. 12:1533-39
logical condition. EPA-SAB-EPEC-02-009,
SuterGW II, NortonSB, CormierSM. 2002. A
Washington, DC. 142 pp. http://www.epa.
of
obthe
causes
for
inferring
methodology
gov/sab/pdf/epec02009.pdf
served impairmentsin aquatic ecosystems. US EPA.2002c. Biological indicatorvariability
and streamprogramintegration:a Maryland
Environ.Toxicol.Chem.21:1101-11
ECOLOGICALINDICATORS
111
case study. EPA/903/R-02/008, Washing- Venier LA, Pearce JE, McKee JE, McKenney
DW, Niemi GJ. 2004. Climate and satelliteton, DC. 92 pp. http://www.epa.gov/bio
indicators/pdf/biological-indicator_md_.pdf derived land cover for predicting breeding
US EPA. 2002d. Mid-Atlantic integrated asbird distributionin the GreatLakes basin. J.
sessment (MAIA) estuaries 1997-1998.
Biogeogr 31:315-31
VernerJ. 1984. The guild concept applied to
EPA/620/R-02-003, Washington,DC
US EPA. 2003a. EPA's draft report on the
management of bird populations. Environ.
environment-technical document. EPA/
Manage. 8:1-14
600-R-03-050, Washington, DC. 453 pp. WhittierTR, Paulsen SG, LarsenDP, Peterson
SA, HerlihyAT,KaufmannPR. 2002. Indicahttp://www.epa.gov/indicators
US EPA. 2003b. Developing biological inditorsof ecological stressandtheirextentin the
cators: lessons learned from Mid-Atlantic
populationof northeasternlakes: a regional
streams. EPA/903/R-03/003, Washington,
assessment.BioScience 52:235-47
DC. 52 pp. http://www.epa.gov/bioindica Wolter PT, White MA. 2002. Recent forest cover type transitions and landscape
tors/pdf/MAIAJessonslearnedbiology.pdf
US EPA. 2003c. Response of surface water
structuralchanges in northeastMinnesota.
Landsc.Ecol. 17:133-55
chemistry to Clean Air Act amendmentsof
1990. EPA/620/R-03/001,Washington,DC. WorldEconomic Forum.2002. 2002 Environ74 pp.
mental sustainability index. Annu. Meet.,
US EPA. 2004. Draft national coastal condiGlobal Leaders for TomorrowTask Force,
tion reportII. EPA 620/R-03/002, WashingWorldEcon. Forum,Yale Univ. Cent. Environ. Law Policy, and ColumbiaUniv. Cent.
ton, DC. 362 pp. http://www.epa.gov/owow/
oceans/nccr2/downloads.html
Int. Earth Sci. Inf. Netw., New York http://
Van Biema D. 1995. Prophetof poison. Time
www.ciesin.org/indicators/ESI/downloads.
145:26-33
html
VanHome B. 1983. Density as a misleadingin- Yoccoz NG, Nichols JD, Boulinier T. 2001.
dicatorof habitatquality.J. Wildl.Manage.
Monitoring of biological diversity in space
47:893-901
and time. TrendsEcol. Evol. 16:446-53