Society for Conservation Biology

Indicators for Monitoring Biodiversity: A Hierarchical Approach

Author(s): Reed F. Noss
Source: Conservation Biology, Vol. 4, No. 4 (Dec., 1990), pp. 355-364
Published by: Blackwell Publishing for Society for Conservation Biology
Stable URL: http://www.jstor.org/stable/2385928 .
Accessed: 19/09/2011 23:10

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

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 [email protected].

Blackwell Publishing and Society for Conservation Biology are collaborating with JSTOR to digitize, preserve
and extend access to Conservation Biology.

http://www.jstor.org
Essay

Indicators for Monitoring Biodiversity:

A Hierarchical Approach

REED F. NOSS*
U.S. Environmental Protection Agency
Environmental Research Laboratory
Corvallis, OR 97333, U.S.A.

Abstract: Biodiversity is presently a minor consideration in Resumen: La biodiversidad es basta abora una consid-
environmentalpolicy. It has been regarded as too broad and eracion menor en lapolitica ambiental. Se ha visto como un
vague a concept to be applied to real-world regulatory and concepto demasiado amplio y vago para ser aplicado en las
managementproblems. Thisproblem can be corrected if bio- regulaciones y el maneyo de los problemas del mundo real.
diversity is recognized as an end in itself and if measurable Este problema se puede corregir si la biodiversidad es re-
indicators can be selected to assess the status of biodiversity conocida como un fin por si misma y si se pueden seleccio-
over time. Biodiversity, as presently understood, encom- nar indicadores cuantificables para determinar el estado de
passes multiple levels of biological organization. In tbis pa- la biodiversidad a trave'sdel tiempo. La biodiversidad, como
per, I expand the three primary attributes of biodiversity se entiende actualmente comprende muiltiples niveles de or-
recognized by Jerry Franklin - composition, structure, and ganizacion biologica En esta disertacion, extiendo los tres
function - into a nested hierarchy that incorporates ele- atributos primarios de la biodiversidad reconocidos por
ments of each attribute at four levels of organization: re- Jerry Franklin - composicion, estructura y funcion -
gional landscape, community-ecosystem, population- dentro de unajerarquia que encaja a incorpora los elemen-
species, and genetic. Indicators of each attribute in terrestrial tos de cada uno de los atributos en cuatro niveles de orga-
ecosystems, at the four levels of organization, are identified nizacion: paisaje regional, ecosistemas de las comunidades,
for environmental monitoring purposes. Projects to monitor poblacion de especies y gene'tica Los indicadores de cada
biodiversity will benefit from a direct linkage to long-term atributo en los ecosistemas terrestres,en los cuatro niveles de
ecological research and a commitment to test hypotheses organizacion, son identificados para propositos de moni-
relevant to biodiversity conservation. A general guideline is toreo ambiental. Los proyectos para el monitoreo de la bio-
to proceed from the top down, beginning with a coarse-scale diversidad se beneficiarian de una union directa con la in-
inventory of landscape pattern, vegetation, habitat structure, vestigacion ecol6gica a largo plazo y de un compromiso
and species distributions, then overlaying data on stress lev- para probar bipotesis relevantes a la conservacion de la bio-
els to identify biologically significant areas at high risk of diversidad. Un lineamiento general es proceder de arriba
impoverishment Intensive research and monitoring can be para abajo, empezando con una escala-burda de inventario
directed to high-risk ecosystems and elements of biodiversity, de lospatrones delpaisaje, de la vegetacion, de la estructura
while less intensive monitoring is directed to the total land- del bacbitaty de la distribucion de ls especies, despue'ssuper-
scape (or samples thereof). In any monitoring program, par- poner los datos sobre niveles de presion para identificar las
ticular attention should be paid to specifying the questions areas da alto riezgo y de empobrecimiento. La investigacion
that monitoring is intended to answer and validating the intensivay el monitorio peude ser dirigido a los ecosistemas
relationships between indicators and the components of bio- de alto riezgoy a los elementos de la biodiversidad, mientras
diversity they represent que un monitoreo me'nos intenso sepuede dirigir al total del
paisaje (o a muestras del mismo). En cualquierprograma de
monitoreo, se debe de poner atencion especial al estar espe-
Paper submitted August 22, 1989; revised manuscript accepted No- cificando laspreguntas que el monitoreopretende resolvery
vember 29, 1989. al estar validando las relaciones entre los indicadores y los
* Current address is 925 N W 31st Street, Corvallis, OR 47330 componentes de la biodiversidad que representen.

355

Conservation Biology
Volume 4, No. 4, December 1990
356 for Biodiversity
Indicators Noss

Introduction cies, ecosystems, or any other group of things in a de-

fined area. Knowing that one community contains 500
Biological diversity (biodiversity) means different species and another contains 50 species does not tell us
things to different people. To a systematist, it might be much about their relative importance for conservation
the list of species in some taxon or group of taxa. A purposes. Ecologists usually define "diversity" in a way
geneticist may consider allelic diversity and heterozy- that takes into consideration the relative frequency or
gosity to be the most important expressions of biodi- abundance of each species or other entity, in addition to
versity, whereas a community ecologist is more inter- the number of entities in the collection. Several differ-
ested in the variety and distribution of species or ent indices, initially derived from information theory,
vegetation types. To a wildlife manager, managing for combine richness with a measure of evenness of relative
biodiversity may mean interspersing habitats to maxi- abundances (e.g., Shannon & Weaver 1949; Simpson
mize edge effects, thereby building populations of pop- 1949). Unfortunately, the number of indices and inter-
ular game species. Some nonbiologists have complained pretations proliferated to the point where species diver-
that biodiversity is just another "smokescreen" or envi- sity was in danger of becoming a "nonconcept" (Hurl-
ronmentalist ploy to lock up land as wilderness. No bert 1971). Diversity indices lose information (such as
wonder agencies are having difficulty defining and im- species identity), are heavily dependent on sample size,
plementing this new buzzword in a way that satisfies and generally have fallen out of favor in the scientific
policy-makers, scientists, and public user groups alike. community. As Pielou (1975:165) noted, "a communi-
Conservation biologists now recognize the biodiver- ty's diversity index is merely a single descriptive statis-
sity issue as involving more than just species diversity or tic, only one of the many needed to summarize its char-
endangered species. The issue is grounded in a concern acteristics, and by itself, not very informative." Despite
about biological impoverishment at multiple levels of such warnings, diversity indices still are used in mislead-
organization. Increasingly, the American public sees bio- ing ways in some environmental assessments (Noss &
diversity as an environmental end point with intrinsic Harris 1986).
value that ought to be protected (Nash 1989). The Agencies prefer to promulgate and enforce regula-
heightened interest in biodiversity presents an oppor- tions based on quantitative criteria, even though quali-
tunity to address environmental problems holistically, tative changes in community structure are often the
rather than in the traditional and fragmentary species- best indicators of ecological disruption. When a natural
by-species, stress-by-stress fashion. One way to escape landscape is fragmented, for example, overall commu-
the vagueness associated with the biodiversity issue is to nity diversity may stay the same or even increase, yet
identify measurable attributes or indicators of biodiver- the integrity of the community has been compromised
sity for use in environmental inventory, monitoring, and with an invasion of weedy species and the loss of species
assessment programs. The purpose of this paper is to unable to persist in small, isolated patches of habitat
provide a general characterization of biodiversity and to (Noss 1983). Qualitative changes at local and regional
suggest a set of indicators and guidelines by which bio- scales correspond to a homogenization of floras and fau-
diveristy can be inventoried and monitored over time. I nas. As a biogeographical region progressively loses its
emphasize terrestrial systems, but many of the guide- character, global biodiversity is diminished (Mooney
lines apply to aquatic and marine realms. 1988).

A HierarchicalCharacterizationof Biodiversity
Definingand Characterizing
Biodiversity
A definition of biodiversity that is altogether simple,
WhatIt Is and WhatIt Is Not
comprehensive, and fully operational (i.e., responsive to
A widely cited definition of biological diversity is "the real-life management and regulatory questions) is un-
variety and variability among living organisms and the likely to be found. More useful than a definition, per-
ecological complexes in which they occur" (OTA haps, would be a characterization of biodiversity that
1987). The OTA document described diversity at three identifies the major components at several levels of or-
fundamental levels: ecosystem diversity, species diver- ganization. This would provide a conceptual framework
sity, and genetic diversity. These three kinds of biodi- for identifying specific, measurable indicators to moni-
versity were noted earlier by Norse et al. (1986). Un- tor change and assess the overall status of biodiversity.
fortunately, most definitions of biodiversity, including Franklin et al. (1981) recognized three primary at-
OTA's, fail to mention processes, such as interspecific tributes of ecosystems: composition, structure, and
interactions, natural disturbances, and nutrient cycles. function. The three attributes determine, and in fact
Although ecological processes are as much abiotic as constitute, the biodiversity of an area. Composition has
biotic, they are crucial to maintaining biodiversity. to do with the identity and variety of elements in a
Biodiversity is not simply the number of genes, spe- collection, and includes species lists and measures of

Conservation Biology
Volume 4, No. 4, December 1990
Noss Indicatorsfor Biodiversity 357

species diversity and genetic diversity. Structure is the sessment be limited to higher levels (e.g., remote sens-
physical organization or pattern of a system, from hab- ing of regional landscape structure). Lower levels in a
itat complexity as measured within communities to the hierarchy contain the details (e.g., species identities and
pattern of patches and other elements at a landscape abundances) of interest to conservationists, and the
scale. Function involves ecological and evolutionary mechanistic basis for many higher-order patterns.
processes, including gene flow, disturbances, and nutri- The hierarchy concept suggests that biodiversity be
ent cycling. Franklin (1988) noted that the growing monitored at multiple levels of organization, and at mul-
concern over compositional diversity has not been ac- tiple spatial and temporal scales. No single level of or-
companied by an adequate awareness of structural and ganization (e.g., gene, population, community) is funda-
functional diversity. Hence, structural simplification of mental, and different levels of resolution are appropriate
ecosystems and disruption of fundamental ecological for different questions. Big questions require answers
processes may not be fully appreciated. Here, I elabo- from several scales. If we are interested in the effects of
rate Franklin's three attributes of biodiversity into a climate change on biodiversity, for instance, we may
nested hierarchy (Fig. 1). Because the compositional, want to consider (1) the climatic factors controlling
structural, and functional aspects of nature are interde- major vegetation ecotones and patterns of species rich-
pendent, the three spheres are interconnected and ness across continents; (2) the availability of suitable
bounded by a larger sphere of concern (i.e., Earth). habitats and landscape linkages for species migration;
Hierarchy theory suggests that higher levels of orga- (3) the climatic controls on regional and local distur-
nization incorporate and constrain the behavior of bance regimes; (4) the physiological tolerances, auteco-
lower levels (Allen & Starr 1982; O'Neill et al. 1986). If logical requirements, and dispersal capacities of individ-
a big ball (e.g., the biosphere) rolls downhill, the little ual species; and (5) the genetically controlled variation
balls inside it will roll downhill, also. Hence, global within and between populations of a species in response
problems such as greenhouse warming and strato- to climatic variables. "Big picture" research on global
spheric ozone depletion impose fundamental con- phenomena is complemented by intensive studies of the
straints on efforts to preserve particular natural areas or life histories of organisms in local environments.
endangered species. The importance of higher-order Another value of the hierarchy concept for assessing
constraints should not suggest that monitoring and as- biodiversity is the recognition that effects of environ-
mental stresses will be expressed in different ways at
different levels of biological organization. Effects at one
level can be expected to reverberate through other lev-
els, often in unpredicatable ways. Tree species, for ex-
ample, are known to be differentially susceptible to air
pollution, with some (e.g., Pinusponderosa) highly sen-
sitive to photochemical oxidants such as ozone (Miller
1973). Different genotypes within tree species vary in
their tolerance of air pollution. A decline in a tree pop-
ulation due to air pollution would alter the genetic com-
position of that population, and reduce genetic varia-
\~~~~~~~~-
0 tion, as pollution-intolerant genotypes are selected out
genetic
processes (Scholz 1981). If a declining tree species is replaced by
demogrophic
species that are either more or less pyrogenic, or oth-
erwise regulatory of disturbance dynamics, changes in
\ \
interspecific
I~~~~~~~~nteract,ons,
//
biodiversity could be dramatic as the system shifts
abruptly to a new stable state.
landscope processes
\ \ ~~~~~and
disturbances, //

SelectingBiodiversityIndicators
FUCINAL
WhyIndicators?
Figure 1. Compositional, structural, and functional Indicators are measurable surrogates for environmental
biodiversity, shown as interconnected spheres, each end points such as biodiversity that are assumed to be of
encompassing multiple levels of organization. This value to the public. Ideally, an indicator should be (1)
conceptual framework may facilitate selection of sufficiently sensitive to provide an early warning of
indicators that represent the many aspects of biodi- change; (2) distributed over a broad geographical area,
versity that warrant attention in environmental or otherwise widely applicable; (3) capable of providing
monitoring and assessment programs. a continuous assessment over a wide range of stress; (4)

Conservation Biology
Volume 4, No. 4, December 1990
358 for Biodiversity
Indicators Noss

relatively independent of sample size; (5) easy and cost- itself, rather than as an index of air quality, water quality,
effective to measure, collect, assay, and/or calculate; (6) or some other anthropocentric measure of environmen-
able to differentiate between natural cycles or trends tal health. (2) Selection of indicators depends on for-
and those induced by anthropogenic stress; and (7) rel- mulating specific questions relevant to management or
evant to ecologically significant phenomena (Cook policy that are to be answered through the monitoring
1976; Sheehan 1984; Munn 1988). Because no single process. (3) Indicators for the level of organization one
indicator will possess all of these desirable properties, a wishes to monitor can be selected from levels at, above,
set of complementary indicators is required. or below that level. Thus, if one is monitoring a popu-
The use of indicator species to monitor or assess en- lation, indicators might be selected from the landscape
vironmental conditions is a firmly established tradition level (e.g., habitat corridors that are necessary to allow
in ecology, environmental toxicology, pollution control, dispersal), the population level (e.g., population size,
agriculture, forestry, and wildlife and range manage- fecundity, survivorship, age and sex ratios), the level of
ment (Thomas 1972; Ott 1978; Cairns et al. 1979). But individuals (e.g., physiological parameters), and the ge-
this tradition has encountered many conceptual and netic level (e.g., heterozygosity). (4) The indicators in
procedural problems. In toxicity testing, for example, Table 1 are general categories, most of which cut across
the usual assumption that responses at higher levels of ecosystem types. In application, many indicators will be
biological organization can be predicted by single- specific to ecosystems. Coarse woody debris, for exam-
species toxicity tests is not supportable (Cairns 1983). ple, is a structural element critical to biodiversity in
Landres et al. (1988) pointed out a number of difficul- many old-growth forests, such as in the Pacific North-
ties with using indicator species to assess population west (Franklin et al. 1981), but may not be important in
trends of other species and to evaluate overall wildlife more open-structured habitats, including forest types
habitat quality, and noted that the ecological criteria subject to frequent fire.
used to select indicators are often ambiguous and falli-
ble. RegionalLandscape
Recent criticisms of the use and even the concept of
indicator species are valid. Indicator species often have The term "regional landscape" (Noss 1983) emphasizes
told us little about overall environmental trends, and the spatial complexity of regions. "Landscape"refers to
may even have deluded us into thinking that all is well "a mosaic of heterogeneous land forms, vegetation
with an environment simply because an indicator is types, and land uses" (Urban et al. 1987). The spatial
thriving. These criticisms apply, however, to a much scale of a regional landscape might vary from the size of
more restricted application of the indicator concept a national forest or park and its surroundings up to the
than is suggested here. The final recommendation of size of a physiographic region or biogeographic prov-
Landres et al. (1988) is to use indicators as part of a ince (say, from 102 to 107 kM2).
comprehensive strategy of risk analysis that focuses on The relevance of landscape structure to biodiversity
key habitats (including corridors, mosaics, and other is now well accepted, thanks to the voluminous litera-
landscape structures) as well as species. Such a strategy ture on habitat fragmentation (e.g., Burgess & Sharpe
might include monitoring indicators of compositional, 1981; Harris 1984; Wilcove et al. 1986). Landscape fea-
structural, and functional biodiversity at multiple levels tures such as patch size, heterogeneity, perimeter-area
of organization. ratio, and connectivity can be major controllers of spe-
cies composition and abundance, and of population vi-
ability for sensitive species (Noss & Harris 1986). Re-
An IndicatorSelectionMatrix
lated features of landscape composition (i.e., the
Table 1 is a compilation of terrestrial biodiversity indi- identity and proportions of particular habitats) are also
cators and inventory and monitoring tools, arranged in a critical. The "functional combination" of habitats in the
four-level hierarchy. As with most categorizations, some landscape mosaic is vital to animals that utilize multiple
boxes in Table 1 overlap, and distinctions are somewhat habitat types and includes ecotones and species assem-
arbitrary. The table may be useful as a framework for blages that change gradually along environmental gradi-
selecting indicators for a biodiversity monitoring proj- ents; such gradient-associated assemblages are often
ect, or more immediately, as a checklist of biodiversity rich in species but are not considered in conventional
attributes to consider in preparing or reviewing envi- vegetation analysis and community-level conservation
ronmental impact statements or other assessments. (Noss 1987)
Four points about choosing indicators deserve em- The indicators listed for the regional landscape level
phasis. (1) The question "What are we monitoring or in Table 1 are drawn mostly from the literature of land-
assessing, and why?" is fundamental to selecting appro- scape ecology and disturbance ecology. General refer-
priate indicators. I assume that the purpose is to assess ences include Risser et al. (1984), Pickett and White
biodiversity comprehensively and as an end point in (1985), and Forman and Godron (1986). O'Neill et al.

Conservation Biology
Volume 4, No. 4, December 1990
Noss Indicatorsfor Biodiversity 359

Table 1. Indicatorvariablesfor inventorying,monitoring,and assessing terrestrialbiodiversityat four levels of organization,including

compositional,structural,and functionalcomponents;includes a samplingof inventoryand monitoringtools and techniques.
Indicators
Inventory and
Composition Structure Function monitoring tools

Regional Identity, distribution, Heterogeneity; connectivity; Disturbance processes (areal Aerial photographs (satellite
Landscape richness, and proportions spatial linkage; patchiness; extent, frequency or and conventional aircraft)
of patch (habitat) types porosity; contrast; grain return interval, rotation and other remote sensing
and multipatch landscape size; fragmentation; period, predictability, data; Geographic
types; collective patterns configuration; intensity, severity, Information System (GIS)
of species distributions juxtaposition; patch size seasonality); nutrient technology; time series
(richness, endemism) frequency distribution; cycling rates; energy flow analysis; spatial statistics;
perimeter-area ratio; rates; patch persistence mathematical indices (of
pattern of habitat layer and turnover rates; rates pattern, heterogeneity,
distribution of erosion and connectivity, layering,
geomorphic and diversity, edge,
hydrologic processes; morphology,
human land-use trends autocorrelation, fractal
dimension)
Community- Identity, relative abundance, Substrate and soil variables; Biomass and resource Aerial photographs and
Ecosystem frequency, richness, slope and aspect; productivity; herbivory, other remote sensing data;
evenness, and diversity of vegetation biomass and parasitism, and predation ground-level photo
species and guilds; physiognomy; foliage rates; colonization and stations; time series
proportions of endemic, density and layering; local extinction rates; analysis; physical habitat
exotic, threatened, and horizontal patchiness; patch dynamics (fine-scale measures and resource
endangered species; canopy openness and gap disturbance processes), inventories; habitat
dominance-diversity proportions; abundance, nutrient cycling rates; suitability indices (HSI,
curves; life-form density, and distribution human intrusion rates and multispecies);
proportions; similarity of key physical features intensities observations, censuses and
coefficients; C4:C3 plant (e.g., cliffs, outcrops, inventories, captures, and
species ratios sinks) and structural other sampling
elements (snags, down methodologies;
logs); water and resource mathematical indices (e.g.,
(e.g., mast) availability; of diversity, heterogeneity,
snow cover layering dispersion, biotic
integrity)
Population- Absolute or relative Dispersion Demographic processes Censuses (observations,
Species abundance; frequency; (microdistribution); range (fertility, recruitment rate, counts, captures, signs,
importance or cover (macrodistribution); survivorship, mortality); radio-tracking); remote
value; biomass; density population structure (sex metapopulation dynamics; sensing; habitat suitability
ratio, age ratio); habitat population genetics (see index (HSI);
variables (see below); population species-habitat modeling;
community-ecosystem fluctuations; physiology; population viability
structure, above); life history; phenology; analysis
within-individual growth rate (of
morphological variability individuals); acclimation;
adaptation
Genetic Allelic diversity; presence of Census and effective Inbreeding depression; Electrophoresis; karyotypic
particular rare alleles, population size; outbreeding rate; rate of analysis; DNA sequencing;
deleterious recessives, or heterozygosity; genetic drift; gene flow; offspring-parent
karyotypic variants chromosomal or mutation rate; selection regression; sib analysis;
phenotypic polymorphism; intensity morphological analysis
generation overlap;
heritability

( 1988) developed and tested three indices of landscape regimes. Statistical techniques applicable to landscape
pattern, derived from information theory and fractal ge- pattern analysis were summarized by Risser et al.
ometry and found them to capture major features of (1984) and Forman and Godron (1986).
landscapes. Landscape structure can be inventoried and Monitoring landscape composition requires more in-
monitored primarily through aerial photography and tensive ground-truthing than monitoring structure, as
satellite imagery, and the data organized and displayed the dominant species composition of patch types (and,
with a Geographical Information System (GIS). Time perhaps, several vertical layers) must be identified.
series analysis of remote sensing data and indices of Landscapefunction can be monitored through attention
landscape pattern is a powerful monitoring technique. to disturbance-recovery processes and to rates of bio-
Monitoring the positions of ecotones at various spatial geochemical, hydrologic, and energy flows. For certain
scales may be particularly useful to track vegetation re- ecosystems, such as longleaf pine-wiregrass communi-
sponse to climate change and disruptions of disturbance ties in the southeastern United States, a disturbance

Conservation Biology
Volume 4, No. 4, December 1990
360 for Biodiversity
Indicators Noss

measure as simple as fire frequency and seasonality may cling rates) that may be appropriate to monitor for spe-
be one of the best indicators of biodiversity. If fires cialized purposes.
occur too infrequently, or outside of the growing sea- Tools and techniques for monitoring biodiversity at
son, hardwood trees and shrubs invade, floristic diver- the community-ecosystem level of organization are
sity may decline, and key species may be eliminated nearly as diverse as the taxa and systems of concern.
(Noss 1988). In many landscapes, human land-use indi- Plant ecology texts (e.g., Greig-Smith 1964; Mueller-
cators (both structural and functional: e.g., deforestation Dombois & Ellenberg 1974) contain much information
rate, road density, fragmentation or edge index, grazing on community-level sampling methodology. A tremen-
and agricultural intensity, rate of housing development) dous literature exists on bird census techniques, the
and the protection status of managed lands may be the most complete single reference being Ralph and Scott
most critical variables for tracking the status of biodi- (1981). Bird community data can be readily applied
versity. to environmental evaluations (e.g., Graber & Graber
In addition to strictly landscape-level variables, col- 1976). Long-term bird surveys across the United States,
lective properties of species distributions can be inven- such as the U.S.Fish and Wildlife Service's Breeding Bird
toried at the regional landscape scale. Terborgh and Survey (BBS) program (Robbins et al. 1986), are used to
Winter (1983), for example, mapped the distribution of monitor temporal trends in species populations, but
endemic land birds in Columbia and Ecuador and iden- could be interpreted to monitor guilds or the entire
tified areas of maximum geographic overlap where pro- avian community of a defined area. Small mammal, rep-
tection efforts should be directed. Scott et al. (1990) tile, and amphibian monitoring are discussed in several
developed a methodology to identify centers of species papers in Szaro et al. (1988). Useful summaries of wild-
richness and endemism, and vegetative diversity, at a life habitat inventory and monitoring are in Thomas
scale of 1:100,000 to 1:500,000, and to identify gaps in (1979), Verner et al. (1986), and Cooperrider et al.
the distribution of protected areas. In most cases, re- (1986). Karr'sindex of biotic integrity (IBI; Karr et al.
peating extensive inventories of species distributions 1986), which collapses data on community composition
would be impractical for monitoring purposes. Periodic into a quantitative measure, has been applied with suc-
inventories of vegetation from remove sensing, how- cess to aquatic communities, and terrestrial applications
ever, can effectively monitor the availability of habitats are possible (J. R. Karr,personal communication).
over broad geographic areas. Inferences about species
distributions can be drawn from such inventories. Population-Species
Monitoring at the species level might target all popula-
Community-Ecosystem
tions of a species across its range, a metapopulation
A community comprises the populations of some or all (populations of a species connected by dispersal), or a
species coexisting at a site. The term "ecosystem" in- single, disjunct population. The population-species level
cludes abiotic aspects of the environment with which is where most biodiversity monitoring has been fo-
the biotic community is interdependent. In contrast to cused. Although the indicator species approach has
the higher level of regional landscape, the community- been criticized for its questionable assumptions, meth-
ecosystem level is relatively homogenous when viewed, odological deficiencies, and sometimes biased applica-
say, at the scale of a conventional aerial photograph. tion, single species will continue to be important foci of
Thus, monitoring at this level or organization must rely inventory, monitoring, and assessment efforts, for two
more upon ground-level surveys and measurements basic reasons: (1) species are often more tangible and
than on remote sensing (although the latter is still useful easy to study than communities, landscapes, or genes;
for some habitat components). (2) laws such as the U.S. Endangered Species Act (ESA)
Indicator variables for the community-ecosystem mandate attention to species but not to other levels of
level (Table 1) include many from community ecology, organization (except that the ESAis supposed to "pro-
such as species richness and diversity, dominance- vide a means whereby the ecosystems upon which en-
diversity curves, life-form and guild proportions, and dangered species and threatened species depend may
other compositional measures. Structural indicators in- be conserved").
clude many of the habitat variables measured in ecology Noss (1990) lists five categories of species that may
and wildlife biology. Ideally, both biotic and habitat in- warrant special conservation effort, including intensive
dicators should be measured at the community-eco- monitoring (1) ecological indicators: species that signal
system and population-species levels of organization the effects of perturbations on a number of other spe-
(Schamberger 1988). The functional indicators in Table cies with similar habitat requirements; (2) keystones:
1 include biotic variables from community ecology pivotal species upon which the diversity of a large part
(e.g., predation rates) and biotic-abiotic variables from of a community depends; (3) umbrellas: species with
ecosystem ecology (e.g., disturbance and nutrient cy- large area requirements, which if given sufficient pro-

Conservation Biology
Volume 4, No. 4, December 1990
Noss for Biodiversity 361
Indicators

tected habitat area, will bring many other species under at the genetic level usually is restricted to zoo popula-
protection; (4) flagships: popular, charismatic species tions of rare species, or species of commercial impor-
that serve as symbols and rallying points for major con- tance such as certain trees. Lande and Barrowclough
servation initiatives; and (5) vulnerables: species that (1987) discussed techniques available to directly mea-
are rare, genetically impoverished, of low fecundity, de- sure and monitor genetic variation, and much of the
pendent on patchy or unpredictable resources, ex- genetic portion of Table 1 is adapted from their paper.
tremely variable in population density, persecuted, or Although some indices of morphological variability may
otherwise prone to extinction in human-dominated be good indicators of genetic stress (Leary & Allendorf
landscapes (see Terborgh & Winter -1980; Karr 1982; 1989), variation in morphology can be confounded by
Soule 1983, 1987; Pimm et al. 1988; Simberloff 1988). phenotypic effects. Electrophoresis of tissue samples is
Not all of these categories need to be monitored in any the preferred technique for monitoring heterozygosity
given case. It may be that adequate attention to catego- and enzyme variability (allozymes), probably the most
ries 2-5 would obviate the need to identify and monitor common measures of genetic variation. Heritability
putative ecological indicator species (D. S. Wilcove, per- studies (e.g., offspring-parent regression, sib analysis)
sonal communication). can be used to determine the level of genetic variation
For species at risk, intensive monitoring may be di- for quantitative traits. Chromosomal polymorphisms
rected at multiple population-level indicators, as well as can be monitored by karyotypic analysis, and the use of
appropriate indicators at other levels - the genetic restriction endonucleases to cut DNA allows direct as-
level, for example (Table 1). Measurements of morpho- sessment of genetic variation (Mlot 1989). The severity
logical characters are often useful. The regression of of inbreeding depression can be evaluated from pedi-
weight on size for amphibians and reptiles, for instance, grees (which, however, are seldom available for wild
provides an index of the general health of a population populations).
(Davis 1989). Within-individual morphological variabil-
ity (e.g., fluctuating asymmetry in structures of bilater-
ally symmetrical organisms) can be a sensitive indicator Implementation
or environmental and genetic stress; composite indices
that include information from several morphological Monitoring has not been a glamorous activity in science,
characters are particularly useful (Leary & Allendorf in part because it has been perceived as blind data-
1989). Growth indicators (e.g., tree dbh) and reproduc- gathering (which, in some cases, it has been). The kinds
tive output (e.g., number of fruits, germination rates) of questions that a scientist asks when initiating a re-
are common monitoring targets for plants. search project - about causes and effects, probabilities,
Often, monitoring at the population-species level is interactions, and alternative hypotheses - are not com-
directed not at the population itself, but at habitat vari- monly asked by workers initiating a monitoring project.
ables determined or assumed to be important to the In most agencies, monitoring and research projects are
species. Habitat suitability indicators can be monitored uncoordinated and are carried out by separate branches.
by a variety of techniques, including remote sensing of Explicit hypothesis-testing only rarely has been a part of
cover types required by a species (Cooperrider et al. monitoring studies, hence the insufficient concern for
1986). It has sometimes been assumed that monitoring experimental design and statistical analysis (Hinds
habitat variables obviates the need to monitor popula- 1984). Perhaps monitoring will be most successful
tions; however the presence of suitable habitat is no when it is perceived (and actually qualifies) as scientific
guarantee that the species of interest is present. Popu- research and is designed to test specific hypotheses that
lations may vary tremendously in density due to biotic are relevant to policy and management questions. In this
factors, while habitat carrying capacity remains roughly context, monitoring is a necessary link in the "adaptive
constant (Schamberger 1988). Conversely, inferences management" cycle that continuously refines regula-
based solely on biotic variables such as population den- tions or management practices on the basis of data de-
sity can be misleading. Among vertebrates, for example, rived from monitoring and analyzed with an emphasis
concentrations of socially surbordinate individuals may on predicting impacts (Holling 1978).
occur in areas of marginal habitat (Van Horne 1983). As an illustration of how a biodiversity monitoring
Monitoring both habitat and population variables seems project might be implemented, imagine that a hypothet-
to be essential in most cases. ical agency wants to assess status and trends in biolog-
ical diversity in the Pacific Northwest. This grandiose
Genetic project might be carried out in ten steps:
1. Wbat and wby? It is first necessary to establish
In wild populations, demography is usually of more im- goals and objectives, and the "sub-end points" of bio-
mediate significance to population viability than is pop- diversity that the agency wishes to assess (and main-
ulation genetics (Lande 1988). Due to cost, monitoring tain). This is more a matter of policy-making than of

Conservation Biology
Volume 4, No. 4, December 1990
362 Indicatorsfor Biodiversity Noss

science. Goals for the Pacific Northwest might include: lands, for example, encompass a wide variety of silvicul-
no net loss of forest cover or wetlands; recovery of old- tural treatments.
growth coniferous forests to twice the present acreage; 8. Design and implement a sampling scbeme. Apply-
recovery of native grasslands and shrub-steppe from ing principles of experimental design, select monitoring
overgrazed condition; maintaining viable populations of sites for identified questions and objectives. A design
all native species; and eradicating troublesome exotic might include intensive sampling of high-risk ecosys-
species from federal lands. Sub-end points would cor- tems and species (identified in Step 4) and less intensive
respond to these goals and encompass the health and sampling of general control and treatment areas identi-
viability of all elements of biodiversity identified to be of fied in Step 7 (but with sampling points and plots se-
concern. lected randomly within treatments). All treatments and
2. Gatber and integrate existing data Existing bio- controls should be replicated. Randomized systematic
diversity-related data bases in state naturalheritage pro- sampling of the total regional landscape (stratified by
grams, agency files, and from other sources would be ecosystem type, if desired) would provide background
collected, digitized, and overlayed in a GIS. These data monitoring and may serve to identify unforeseen
would be mapped for the region as a whole at a scale of stresses. Biology should drive the statistical design, how-
1:100,000 to 1:500,000 (see Scott et al. 1990). ever, rather than letting the design assume a life of its
3. Establish "baseline" conditions. From current own.
data, determine the extent, distribution, and condition 9. Validate relationsbips between indicators and
of existing ecosystem (vegetation) types and the prob- sub-endpoints. Detailed, ongoing research is needed to
able distribution of species of concern. Also, map the verify how well the selected indicators correspond to
distribution and intensity of identified stressors (e.g., the biodiversity sub-end points of concern. For exam-
tropospheric ozone, habitat fragmentation, road density, ple, does a particular fragmentation or edge index (such
grazing intensity). as perimeter-area ratio or patch size frequency distribu-
4. Identify "bot spots" and ecosystems at high risk. tion) really correspond to the intensity of abiotic and
Proceeding from the previous two steps, delineate areas biotic edge effects or the disruption of dispersal be-
of concentrated biodiversity (e.g., centers of species tween patches in the landscape? Does the relationship
richness and endemism) and ecosystems and geograph- between indicator and sub-end point hold for the entire
ical areas at high risk of impoverishment due to anthro- range of conditions encountered?
pogenic stresses. Such areas warrant more intensive 10. Analyze trends and recommend management
monitoring. In the Pacific Northwest, one prominent actions. Temporal series of measurements must be an-
center of endemism is the Klamath-Siskiyoubioregion; alyzed in a statistically rigorous way and the results syn-
an ecosystem type at high risk is old-growth forest (of all thesized into an assessment that is relevant to policy-
species associations). makers. If the assessment can be translated into positive
5. Formulate specific questions to be answered by changes in planning assumptions, management direc-
monitoring. These questions will be guided by the sub- tion and practices, laws and regulations, or environmen-
end points, goals, and objectives identified in Step 1. tal policy, the monitoring project will have proved itself
Questions might include: Is the ratio of native to exotic a powerful tool for conservation.
range grasses increasing or decreasing? Is the average
patch size of managed forests increasing or decreasing?
Are populations of neotropical migrant birds stable?Are Acknowledgments
listed endangered species recovering? It will help if pol-
icy-makers can specify thresholds at which changes in Allen Cooperrider, Sandra Henderson, Bob Hughes,
management practices or regulations will be imple- David Wilcove, and two anonymous referees provided
mented. many helpful comments on an earlier draft of this manu-
6. Select indicators. Identify indicators of structural, script.
functional, and compositional biodiversity at several lev-
els of organization that correspond to identified sub- Literature Cited
end points (Step 1) and questions (Step 5). Indicators
Allen, T. F. H., and T. B. Starr.1982. Hierarchy:perspectives for
can be chosen from the "laundrylist" in Table 1, based
ecological complexity. University of Chicago Press, Chicago,
on the criteria for ideal indicators reviewed above (un- Illinois.
der "Why Indicators?").
7. Identify control areas and treatments. For each Burgess, R. L., and D. M. Sharpe, editors. 1981. Forest island
major ecosystem type, identify control areas (e.g., des- dynamics in man-dominated landscapes. Springer-Verlag,New
ignated Wilderness and Research Natural Areas) and ar- York.
eas subjected to different kinds and intensities of stress
and management practices. Public and private forest Cairns,J. 1983. Are single species toxicity tests alone adequate

Conservation Biology
Volume 4, No. 4, December 1990
Noss for Biodiversity 363
Indicators

for estimating environmental hazard?Hydrobiologia 100:47- size, genetic variation, and their use in population manage-
57. ment. Pages 87-123 in M. E. Soule, editor. Viable populations
for conservation. Cambridge University Press, Cambridge, En-
Cairns,J., G. P. Patil, and W. E. Waters, editors. 1979. Environ- gland.
mental biomonitoring, assessment, prediction, and manage-
ment - certain case studies and related quantitative issues. Landres, P. B., J. Verner, and J. W. Thomas. 1988. Ecological
International Co-operative Publishing House, Fairland,Mary- uses of vertebrate indicator species: a critique. Conservation
land. Biology 2:316-328.

Cook, S. E. K 1976. Quest for an index of community structure Leary,R. F., and F. W. Allendorf. 1989. Fluctuating asymmetry
sensitive to water pollution. Environmental Pollution 11:269- as an indicator of stress: implications for conservation biology.
288. Trends in Ecology and Evolution 4:214-217.

Cooperrider, A. Y., R.J. Boyd, and H. R. Stuart, editors. 1986. Miller, P. L. 1973. Oxidant-induced community change in a
Inventory and monitoring of wildlife habitat. USDI Bureau of mixed conifer forest. Pages 101-117 in J. Naegle, editor. Air
Land Management, Washington, D.C. pollution damage to vegetation. Advances in Chemistry, Series
122. American Chemical Society, Washington, D.C.
Davis, G.E. 1989. Design of a long-term ecological monitoring
program for Channel Islands National Park,California.Natural Mlot, C. 1989. Blueprint for conserving plant diversity. Bio-
Areas Journal 9:80-89. Science 39:364-368.

Forman, R. T. T., and M. Godron. 1986. Landscape ecology. Mooney, H. A. 1988. Lessons from Mediterranean-climate re-
John Wiley and Sons, New York. gions. Pages 157-165 in E. 0. Wilson, editor. Biodiversity. Na-
tional Academy Press, Washington, D.C.
Franklin,J. F. 1988. Structural and functional diversity in tem-
perate forests. Pages 166-175 in E. 0. Wilson, editor. Biodi- Mueller-Dombois, D., and H. Ellenberg. 1974. Aims and meth-
versity. National Academy Press, Washington, D.C. ods of vegetation science. John Wiley and Sons, New York.
Franklin,J. F., K Cromack, W. Denison, et al. 1981. Ecological Munn, R. E. 1988. The design of integrated monitoring systems
characteristics of old-growth Douglas-fir forests. USDAForest to provide early indications of environmental/ecological
Service General Technical Report PNW-118. Pacific North- changes. Environmental Monitoring and Assessment 11:203-
west Forest and Range Experiment Station, Portland, Oregon. 217.
Graber, J. W., and R. R. Graber. 1976. Environmental evalua- Nash, R. F. 1989. The rights of nature: a history of environ-
tions using birds and their habitats. Biological Notes No. 97, mental ethics. University of Wisconsin Press, Madison, Wis-
Illinois Natural History Survey, Urbana, Illinois. consin.

Greig-Smith,P. 1964. Quantitative plant ecology, 2nd edition Norse, E.A., K L. Rosenbaum, D. S. Wilcove, et al. 1986. Con-
Butterworth, London, U.K serving biological diversity in our national forests. The Wil-
derness Society, Washington, D.C.
Harris,L.D. 1984. The fragmented forest: island biogeography
theory and the preservation of biotic diversity. University of
Noss, R. F. 1983. A regional landscape approach to maintain
Chicago Press, Chicago, Illinois.
diversity. BioScience 33:700-706.
Hinds, W. T. 1984. Towards monitoring of long-term trends in
terrestrial ecosystems. EnvironmentalConservation 11:11-18. Noss, R. F. 1987. From plant communities to landscapes in
conservation inventories: a look at The Nature Conservancy
Holling, C. S., editor. 1978. Adaptive environmental assess- (USA). Biological Conservation 41:11-37.
ment and management. Volume 3. International Series on Ap-
plied Systems Analysis. John Wiley and Sons, New York. Noss, R. F. 1988. The longleaf pine landscape of the Southeast:
almost gone and almost forgotten. Endangered Species Update
Hurlbert, S. H. 1971. The nonconcept of species diversity: a 5(5):1-8.
critique and some alternative parameters. Ecology 52:577-
586. Noss, R. F. 1990. From endangered species to biodiversity. In
K Kohm, editor. Balancing on the brink: A retrospective on
Karr,J. R. 1982. Population viability and extinction in the avi- the Endangered Species Act. Island Press, Washington, D.C. (In
fauna of a tropical land bridge island. Ecology 63:1975-1978. press).

Karr,J. R., K D. Fausch, P. L. Angermeier, P. R. Yant, and I.J. Noss, R. F., and L.D. Harris. 1986. Nodes, networks, and MUMs:
Schlosser. 1986. Assessing biological integrity in running wa- preserving diversity at all scales. Environmental Management
ters: a method and its rationale. Illinois Natural History Survey 10:299-309.
Special Publication No. 5. Champaign, Illinois.
Office of Technology Assessment (OTA). 1987. Technologies
Lande, R. 1988. Genetics and demography in biological con- to maintain biological diversity. U.S. Government Printing Of-
servation. Science 241:1455-1460. fice, Washington, D.C.

Lande, R., and G. F. Barrowclough. 1987. Effective population O'Neill, R V., D. L DeAngelis, J. B. Waide, and T. F. H. llen.

Conservation Biology
Volume 4, No. 4, December 1990
364 for Biodiversity
Indicators Noss

1986. A hierarchical concept of ecosystems. Princeton Univer- Simberloff, D. 1988. The contribution of population and com-
sity Press, Princeton, New Jersey. munity biology to conservation science. Annual Review of
Ecology and Systematics 19:473-511.
O'Neill, R.V., J. R. Krummel, R. H. Gardner, G. Sugihara, B.
Jackson, D. L.DeAngelis, B. T. Milne, M. G. Turner, B. Zygmunt, Simpson, E. H. 1949. Measurement of diversity. Nature
S. W. Christensen, V. H. Dale, and R L.Graham.Indices of land- 163:688.
scape pattern. Landscape Ecology 1:153-162.
Soule, M. E. 1983. What do we really know about extinction?
Ott, W. R. 1978. Environmental indices: theory and practice. Pages 111-124 in C. M. Schonewald-Cox, S. M. Chambers, B.
Ann Arbor Science, Ann Arbor, Michigan. MacBryde, and L. Thomas, editors. Genetics and conservation:
a reference for managing wild animal and plant populations.
Pickett, S. T. A., and P. S. White. 1985. The ecology of natural Benjamin/Cummings, Menlo Park, California.
disturbance and patch dynamics. Academic Press, Orlando,
Florida. Soule, M. E., editor. 1987. Viable populations for conservation.
Cambridge University Press, Cambridge, England.
Pielou, E. C. 1975. Ecological diversity. Wiley-Interscience,
New York.
Szaro, R. C., K E. Severson, and D. R. Patton, editors. 1988.
Management of amphibians, reptiles, and small mammals in
Pimm, S. L., H. L.Jones, and J. Diamond. 1988. On the risk of
North America. USDAForest Service General Technical Report
extinction. American Naturalist 132:757-785.
RM-166, Fort Collins, Colorado.
Ralph, C.J., and J. M. Scott. 1981. Estimating numbers of ter-
restrial birds. Studies in Avian Biology No. 6. Allen Press, Terborgh, J., and B. Winter. 1980. Some causes of extinction.
Lawrence, Kansas. Pages 119-133 in M. E. Soule and B. A. Wilcox, editors. Con-
servation biology: an evolutionary-ecological perspective. Sin-
Risser, P. G., J. R. Karr, and R. T. T. Forman. 1984. Landscape auer Associates, Sunderland, Massachusetts.
ecology: directions and approaches. Illinois Natural History
Survey Special Publication No. 2, Champaign, Illinois. Terborgh, J., and B. Winter. 1983. A method for siting parks
and reserves with special reference to Columbia and Ecuador.
Robbins, C. S., D. Bystrak,and P. H. Geissler. 1986. The Breed- Biological Conservation 27:45-58.
ing Bird Survey: its first fifteen years, 1965-1979. USDI Fish
and Wildlife Service, Washington, D.C. Thomas, J. W., editor. 1979. Wildlife habitats in managed for-
ests: the Blue Mountains of Oregon and Washington. USDA
Schamberger, M. 1988. Monitoring wildlife habitat: a critique Forest Service Agricultural Handbook No. 553, Washington,
of approaches. Statistical Journal of the United Nations ECE D.C.
5:303-313.
Thomas, W. A., editor. 1972. Indicators of environmental qual-
Scholz, F. 1981. Genecological aspects of air pollution effects ity. Plenum Press, New York.
on northern forests. Silva Fennica 15:384-391.
Urban, D. L., R. V. O'Neill, and H. H. Shugart. 1987. Landscape
Scott, J. M., B. Csuti, K Smith, J. E. Estes, and S. Caicco. 1990. ecology. BioScience 37:119-127.
Gap analysis of species richness and vegetation cover: an in-
tegrated conservation strategy for the preservation of biolog-
Van Home, B. 1983. Density as a misleading indicator of hab-
ical diversity. In K. Kohm, editor. Balancing on the brink: A
itat quality. Journal of Wildlife Management 47:893-901.
retrospective on the Endangered Species Act. Island Press,
Washington, D.C. (In press).
Verner, J., M. L. Morrison, and C.J. Ralph, editors. 1986. Wild-
Shannon, C. E., and W. Weaver. 1949. The mathematical the- life 2000: Modeling habitat relationships of terrestrial verte-
ory of communication. University of Illinois Press, Urbana. brates. University of Wisconsin Press, Madison, Wisconsin.

Sheehan, P. J. 1984. Effects on community and ecosystem Wilcove, D. S., C. H. McLeilan,and A. P. Dobson. 1986. Habitat
structure and dynamics. Pages 51-99 in P.J. Sheehan, D. R. fragmentation in the temperate zone. Pages 237-256 in M. E.
Miller, G. C. Butler, and P. Boudreau, editors. Effects of pollut- Soule, editor. Conservation biology: the science of scarcity and
ants at the ecosystem level. John Wiley and Sons, New York. diversity. Sinauer Associates, Sunderland, Massachusetts.

Conservation Biology
Volume 4, No. 4, December 1990