|
Items in cart [1] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 91
Characterizing Ecosystem Responses to Stress
MARK A. HARWELL
CHRISTINE C. HARWELL
DAVID A. WEINSTEIN
JOHN R. KELLY
Cornell University
Environmental risk assessment and management involve the use of
methodologies to assess risks to the health of biological systems, especially
the stresses from human activities. The use of appropriate ecological
indicators to measure environmental effects of these stresses can allow a
realistic evaluation of risks.
ECOLOGICAL ll:lSK ASSESSMENT
Effective protection and management of environmental systems re-
quires an adequate understanding of stress ecology. Three facets are
central to the science of stress ecology: 1) how various components of
ecosystems are exposed to stress; 2) how ecosystems respond to these
stresses; and 3) how ecosystems recover from or adapt to stress. When
there is a solid understanding of these relationships, as well as the inherent
uncertainties in predicting stress response, then risks to ecological systems
can be properly balanced with risks and benefits to other systems of human
concern, such as economic or societal systems. Instances of unprotected or
unexpected adverse effects on the environment from a particular human
activity will continue to occur, along with instances of expensive over pro-
tection from effects of other human activities. Risk assessment based on
ecological science is essential to minimize these problems.
In principle, ecological risk assessment is intended to illustrate and
accommodate differences in stress/response relationships and to provide a
basis for balancing environmental concerns with other economic and soci-
etal issues through the associated process of risk management. However,
the current reality of risk assessment and risk management is quite remote
from this ideal, for a number of important reasons:
91
OCR for page 92
92
ECOLOGICAL RISKS
1. A satisfactory basis does not exist for making cross-comparisons
among alternative risks and values; for example, what value is placed on a
single human life or on an endangered species? How do you compare the
loss of an endangered whale species versus an endangered liverwort species?
Ecological and societal values typically cannot be reduced solely to monetary
units; consequently, those factors in an environmental management decision
that are easier to quantify, such as economic costs of compliance with
regulations, often dominate in debates about the effects of human activities
on the environment.
2. Even if there were a common method of valuation, there would be
great difficulty in establishing a common level of acceptability of risk across
different problems. Emotional and subjective factors play a major role
in defining social acceptability. Many factors contribute to this disparity,
including very different perceptions about voluntary versus involuntary risks,
and society's inability to handle infrequent but catastrophic events on the
same basis as frequent but relatively low-consequence events.
For example, the routine emissions from a nuclear power plant have
few, if any, demonstrated adverse impacts on the health of the environment
or the general public; nearby residents could receive a much greater dose
of radiation from the thorium-daughter products in soils and rocks of
the neighborhood, including radon in their homes, or from flying a few
times across the country at high altitudes, than from a properly operated
nuclear facility. In contrast, a fossil-fuel plant continuously emits a host of
compounds that are known to cause adverse ecological and human health
effects, including acid precipitation, greenhouse-induced global climate
changes, and long-lived alpha-emitting radionuclides.
On the other hand, an accident involving the fossil-fuel plant could at
most lead to local-scale fires and injuries, whereas an accident at a nuclear
facility might cause extensive ecological and human health consequences,
as witnessed by increased cancer risks for tens of thousands of inhabitants
in Eastern Europe, and has the potential to eradicate a whole lifestyle for
cultures in Northern Scandinavia based on herding reindeer populations,
all from the single event at Chernobyl.
3. The effectiveness of ecological risk assessment is seriously limited
by the inability of scientists accurately to predict ecological responses to
stress. There are many reasons for this limitation, including:
.
the considerable variety of ecosystems and potential types of human
disturbances to those ecosystems;
· the wide range of spatial, temporal, and organizational scales in-
herent in any ecosystem;
· the lack of an adequate baseline database for comparison of dis-
turbed and undisturbed ecosystems;
OCR for page 93
HUMAN EFFECTS ON THE TERRESTRIAL ENVIRONMENT
and
93
fundamental limitations in ecological theory and understanding;
environmental variability and other irreducible forms of uncertainty
associated with stress/response/recovery predictions.
ECOLOGICAL RISK MANAGEMENT: REGUI^TORY ENDPOINTS
Each state or nation has its own traditions and systems of environmen-
tal values and regulations. In the United States, the U.S. Environmental
Protection Agency (EPA) is charged with the responsibility of protecting
and managing many aspects of the environment. ~ date, the predominant
focus at EPA has been on environmental threats to human health, rather
than on human actions affecting the health of the environment. This reality
may be shifting however, as EPA:s interest grows in using ecological risk
assessment as a tool for objective environmental decision making (Chapter
6, this volume).
A variety of legislative actions provides the framework for the role
of EPIC The language of the laws written by the U.S. Congress typically
contains both broad statements as to the law's general purpose and narrower
statements or sections in the law which detail the particular activities to
be regulated by a government agency such as EPA, sometimes including
detailed instructions as to how that regulation is to occur. The exact
wording of the legislation, often clarified by a study of the legislation's
history, indicates to EPA the directions to follow in formulating regulations
and mechanisms to enforce the law. Thus, regulations developed by EPA
are one way to translate both general and specific legislative directions
provided by Congress (or directions provided by the President's executive
orders or by various courts' judicial interpretations) into regulatory actions
and requirements. There are sometimes certain issues or phrases in a law
that are key to deciding which regulatory actions are to be taken by the
government agency; these are often called regulatory endpoints, defined as
those regulatory norms that translate fundamental legislative purposes into
regulatory decisions or actions (C. Harwell, 1989; Limburg et al., 1986)
(Figure 1~.
The regulatory endpoints for environmental protection can be very
specific, such as the requirement that sulfur dioxide (SO2) and ozone
(03) concentrations in urban areas should not exceed particular numerical
amounts; that lead (Pb) levels in gasoline should be below specified con-
centrations; or that fecal conform counts in effluents from sewage facilities
should remain under a certain value. In part, the degree of specificity in
regulations often reflects the level of certainty about causal relationships or
the simplicity and consensus within society about the endpoint of concern.
Other environmental regulatory endpoints, however, are quite generic;
OCR for page 94
94
ECOLOGICAL RISKS
Congress
/ Regulatory
Endpoints
courts /
\ generic
specific
regulatory ~ ~
agencies - ; <`
onitoringt
Characterizing
Ecosystem
Responses
To Stress
ecological
understanding /
~ / Ecological
/ Endpoints
human health
species-level
community -level
\:osystem-le:
)~>
/'
\ /issues of
\ concern to
humans
I:
Indicators of
Ecological Effects
intrinsic importance
early warning
sensitive
process/s~uctural
select subset
of ecological
endpoints to
monitor
FIGURE 1 Environmental decision making process. Regulatory endpoints are specified
by legislation, courts, or regulations written by agencies. These must be translated into
ecologically meaningful endpoints. focusing on those nroDe~ie~ of Vim. of Urn
to humans. For each selected endpoint, one or more specific indicator is appropriate to
measure or monitor for changes in that endpoint. The suite of selected indicatom provides
the basis for evaluating ecological responses to stress and characterizing ecosystem health.
they require EPA to take action on an area of legislative concern, though
the specific action to be taken is not designated. For example, EPA is
required by several laws to develop regulations that will accomplish the
general, overall purpose of environmental and human health protection by:
.
maintaining and propagating a "balanced indigenous population"
in estuarine ecosystems exposed to less-than-secondarily treated municipal
effluents Howell, 1984a);
preventing point-source discharges to marine ecosystems that cause
"unreasonable degradation of the marine environment" (Harwell, 1984b);
· minimizing 'Significant adverse impacts";
· protecting '`areas of biological concern";
· preventing ``irreparable harm" to the environment;
· maintaining Biological integrity"; or
· not allowing actions that result in "accumulation of toxic materials
in the human food chain."
The remainder of this chapter focuses on such generic regulatory endpoints,
because these tend to be associated with regulation based on ecological
responses to human activity, as opposed to the highly specific regulatory
endpoints, which tend to focus on technological capabilities or on chemical
concentrations.
OCR for page 95
HUMAN EFFECTS ON THE TERRESTRIAL ENVIRONMENT
95
Generic environmental regulatory endpoints typically do not incorpo-
rate language that has clear, intrinsic ecological meaning. For instance,
maintaining a "balanced indigenous population" as required by Section
301(h) of the Federal Water Pollution Control Act is actually interpreted
in EPAs regulations to relate to maintaining a biological community that
is similar to other communities in the region surrounding a local area
of disturbance. Thus, the regulatory phrasing is not what was literally
specified by the legislation, though it may be what was intended: EPA:s
regulatory interpretation of the law refers to a community, rather than
a single "population"; the biota of concern are not necessarily "indige-
nous" to the area; and since natural populations continuously fluctuate
and experience dynamic interactions, it is not clear that "balanced" has
any biological meaning at all. Likewise, it is not immediately evident what
constitutes "degradation" of an ecosystem, or what "biological integrity"
means. Terms like these were often selected by legislators precisely because
the words are subject to alternate interpretations or because they allow flex-
ibility and discretion on the part of the regulatory agency. Nevertheless,
there is a common theme in generic regulatory endpoints, i.e., they call for
some measure of maintenance of the health of the ecosystem. The intent
is not to preserve all ecosystems in their pristine state, uninfluenced by
human activities; such an endpoint is simply not possible for an industrial
nation with a quarter-billion human inhabitants. But, on the other hand,
the intent is to prevent serious adverse impacts on ecosystems from human
activities that are so extensive that the environment is perceived to be
excessively degraded or irreversibly disturbed- that is to say, unhealthy.
Unfortunately, unlike measures of adverse human health effects (such
as mortality, induced cancers, respiratory illnesses, or chromosomal aber-
rations), there are no readily comparable, integrative, simple measures or
indices of adverse effects on ecosystem health caused by stress. Attempts
at drawing an analogy between ecological health and human health (e.g.,
Rapport et al., 1985) have been unsatisfactory, in part because exposure of
ecosystems to stress is very complex, with differential exposure to different
parts of the ecosystem; in part because ecosystems are both more diverse
and more complex than the human metabolic system; and in part because
ecosystems are much less internally integrated, i.e., they have a far less
coordinated and controlled response to stress, and fewer mechanisms for
compensation and homeostasis. If ecosystems truly were superorganisms,
then ecological stress/response/recovery predictions in principle could be as
reliable as human health predictions; but the reality is that the science of
ecology is not in a position now to meet the needs of ecological risk as-
sessment. Nevertheless, we believe that reasonable environmental decision
making can be accomplished in the presence of uncertainties (Ha~well et
al., 1986; Harwell and Howell, 1989~.
OCR for page 96
96
ECOLOGICAL RISKS
ECOLOGICAL ENDPOINTS
The single most useful criterion to apply to measure ecosystem health
is the requirement of relevance to issues of concern to humans. That is,
a change in an ecosystem is only considered relevant if it relates directly
or indirectly to something affecting humans. By focusing on such human-
centered ecological endpoints, a structured way of evaluating ecological
effects can be developed, and a framework can be created for incorporat-
ing non-ecological issues into environmental decision malting (Figure 1~.
While regulatory endpoints are specific goals or standards stated in laws or
regulations, ecological endpoints are selected characteristics of ecosystems
at various levels, the examination of which can allow evaluation of societally
important environmental issues. These issues may have been covered by
existing regulatory endpoints or may yet remain to be regulated.
Ecological endpoints are categorized vis-a-vis issues of human concern
(Table 1~. The first item, human health effects, actually dominates environ-
mental regulation/protection in the United States. Our concern here, how-
ever, is limited to ecosystems as vectors for human exposure to potentially
harmful substances; we do not consider human effects themselves as part of
ecological endpoints. For example, a bathing beach contaminated by high
fecal coliform counts involves a serious risk to human health. Similarly,
radiocesium deposited in fallout on tundra ecosystems and subsequently
biomagnified to dangerous levels for human consumption of reindeer is a
serious concern. These examples are demonstrably ecological endpoints,
even if ecologically no adverse reactions occur. Even though the radioce-
sium is unlikely to affect the biota population levels, or productivity, or
nutrient cycling rates, its presence in potential human food-chain pathways
is prima facie an ecological endpoint.
All of the other categories of ecological endpoints in Able 1 are not
directly related to human health issues, but are associated with ecological
responses and recovery. These can be separated into species-, community-,
or ecosystem-level endpoints.
Species-level Endpoints: Primary
The simplest, primary ecological endpoints concern direct effects on
particular species that have a direct interest to humans. Such interest can
involve the economic value of the species, such as Douglas fir, salmon,
or oysters. Similarly, direct importance can accrue to recreational species,
such as the striped bass of the eastern United States, or the variety of
deer and gamebird populations throughout North America. Other species
do not have a specific, direct economic value, but are of special concern
OCR for page 97
HUM 4N EFFECTS ON THE I:E:RRESTRIAL ENVIRONMENT
TABLE 1 Ecological endpoints: issues of concern to humans.
l
e HUMAN HEALTH EFFECTS
~ vector for exposure to humans
· SPECIES-LEVEL ENDPOINTS
· direct interest
· econonnc, aesd~edc, recreational, nuisance, endangered
Dies
· indirect interest (secondary endpoints)
· hi- species effects Radon , conned don , pollinado n)
· habitat role
· ecological role
· Atrophic relationship
· functional relationship
· cndcal species
.
.
- COMMUNITY-LEVEL ENDPOINTS
· food-web structure
· species diversity
· biotic diversity
ECOSYSTEM-LEVEL ENDPOINTS
· ecologically important process
· economically unportant process
· water quality
· habitat quality
97
because of aesthetics or other human values, e.g., dolphins, eagles, wild
horses, and grizzly bears.
Many species on earth are endangered or threatened with extinc-
tion; unprecedented losses of species are underway, especially in tropical
biomes, from massive deforestation. A select few endangered species also
hold particular recognition, usually because of aesthetic values rather than
because they have a particular ecological value; after all, an endangered
species typically is too rare to play a major ecological role. Examples are
many species of birds (e.g., least tern, California condor); large predators
(e.g., Florida panther, Peregrine falcon); or large herbivores (e.g., white
rhinoceros, American bison).
Finally, many species are of direct concern to humans because of a
negative role. Such nuisance species include disease-vectors (e.g., certain
species of mosquitoes); exotic plants that outcompete native vegetation
(e.g., kudzu, Casunna); and noxious species, such as blue-green algal
blooms.
Species-level Endpoints: Secondary
Indirect effects on species, mediated by effects on other components
of the ecosystem, must also be considered. If the ecosystem component
OCR for page 98
98
ECOLOG CAL RISKS
is another species which is not directly important to humans, it then
becomes a secondary ecological endpoint. Its relationship to the species of
primary concern can involve several different mechanisms. For instance, bi-
specific interactions can be important, i.e., where the primary and secondary
species are closely linked through some interrelationships. One example is
the predator/prey relationship: if society is concerned with the population
levels of blue whales near Antarctica, then there must be a concern for
the dynamics of the krill population. In this case, adverse impacts on the
density of krill would constitute an ecological endpoint, even though the
krill population itself might have little direct concern for humans.
Other bi-specific relationships include pollination (e.g., concern for a
species of fruit tree might translate into concern for the insects necessary
for fertilization); mutualism (where two species mutually benefit from the
presence of each other); and competition. The latter has been shown
through theoretical studies to be potentially quite important; for example,
studies using computer simulation models of forested ecosystems have
shown that direct effects of air pollutants on one species of trees may allow
another species to become dominant, as it outcompetes the other by being
less sensitive to the pollution (Weinstein et al., in prep.~. Also, host/parasite
interactions may prove important, such as when air pollutants impact forest
tree species indirectly through enhancing the prospects for pest or disease
outbreaks (Bedford, 1987~.
Other secondary endpoints may be found in ejects mediated by habi-
tat alterations, such as changes in the physical structure of the environment
that alter the vertical or horizontal heterogeneity of the ecosystem. This is
particularly important when a single species dominates the environmental
structure for other species in the community, such as mangrove forests,
many coniferous forests, seagrass beds? and agroecosystems. Thus, par-
ticular concern for habitat-mediated stresses occurs for near-monoculture-
dominated ecosystems. For example, mangrove trees provide a complex
physical substrate for other plant and animal species, allowing the differ-
entiation and development of a series of diverse ecological communities
in particular niches within that ecosystem. Loss of the mangroves would
consequently result in loss of habitat for many other species that might
not be directly affected by the stress that destroys the mangroves, but that
might have a particular importance for humans.
Another important mechanism for habitat-mediated indirect effects
is the amelioration of physicochemical conditions by biota or other com-
ponents of the ecosystem. For example, the presence of tree and shrub
species in semi-arid environments can induce evapotranspiration, which in
turn can increase precipitation regionally; hence, the biota act as an en-
hancing pump in the hydrological cycle upon which virtually all terrestrial
ecosystems depend. Adverse impacts on this role can indirectly result in
OCR for page 99
HUA!4N EFFECTS ON THE TERRESTRIAL ENVIRONMENT
99
other ecological effects, as starkly demonstrated by the current positive
feedback from desertification underway in sub-Saharan Africa; indeed, a
major component of the drought and famines recently experienced is the
human impacts on the environment through biomass harvesting for energy
and food. Thus, the ecological effect of concern here (tree and shrub pro-
ductivity) is of both direct importance (with respect to the economic value
of this resource) and of indirect, habitat-mediated importance (with respect
to inducing local- to synoptic-scale reductions in precipitation). Evaluation
of this species-level effect becomes an evaluation of an ecological endpoint,
determining the ability of that environment to support human life.
Species-level Endpoints: Ecological Role
Another species-level, indirect effect relates to the ecological role the
affected species plays in the community, such as in maintaining the trophic
structure of the community. Such critical species have been identified
in several ecosystems; a well-known example is provided in Paine's work
(1974, 1980) on keystone species, i.e., particular predatory invertebrates
whose presence or absence is the determinant of the presence of other
species in intertidal ecosystems. In other ecosystems, a few lower trophic-
level species control the food availability for an entire suite of trophic levels
directly or indirectly consuming that energy resource base.
Consequently, a species-level ecological endpoint might be the seagrass
species Thalassia, which supports a diverse ecosystem through reliance
on it as a detritus base and relatively stable substrate; another might
be the wolves that control populations of herbivores through predation.
Other examples include the alligator-controlled ecosystems in the Flonda
Everglades, known as "alligator holes," in which the alligator determines
both the habitat structure and the trophic structure of the ecosystem.
Community-level Endpoints
Another type of ecological endpoint involves community-level issues.
At this level of endpoint is the overall trophic structure per se, not as a
mechanism for supporting a particular species of concern but as a charac-
teristic of direct importance. Humans have come to value the diversity of
ecosystems as having intrinsic worth, and consequently any change in species
diversity constitutes an ecological change of concern. This value is reflected
in some regulatory endpoints that specify maintenance of species diversity
as an endpoint and a measure of overall ecological health. For instance,
Section 403(c) of the Federal Water Pollution Control Act specifically calls
OCR for page 100
100
ECOLOGICAL RISKS
for consideration of changes in species diversity (Hanvell, 1984b). Simi-
larly, Section 301(h) of the Federal Water Pollution Control Act indirectly
calls for maintenance of species diversity through its "balanced indigenous
nnn~lntion" endnnint ns interpreted hv regulations and litigation (Harwell.
rare --—A---- i- -A --me -c~-
9Ma).
The issue of species diversity applies to ecosystem-scale biological units;
but the idea also extends across landscape and regional units incorporating
many different ecosystems and ecosystem types. Stresses that extend across
this spatial scale can result in loss of species, or at least regional-scale loss
of species. Consequently, a broader community-level concern arises, often
termed a concern for biotic diversity. Thus, the elimination of large numbers
of species in the tropics, primarily through human destruction of forests for
biomass and for agricultural uses, is of immense importance because of the
overall reduction in biotic diversity that is occurring essentially on a global
scale.
Ecosystem-level Endpoints
The final level of ecological endpoints involves direct or indirect effects
at the ecosystem level. Here the concern is for maintenance of processes
that are of particular importance. Often, the roles that ecosystems perform
in ameliorating environmental extremes are greatly underappreciated; but
clearly, many instances exist of ecosystem processes providing tremendous
economic or other societal benefit to humans. Examples include maintain-
ing the biogeochemical cycles in wetlands to decrease the high amounts
of nutrients in wastewater; maintaining a forest for water retention and
floodcontrol; and maintaining dune ecosystems to protect coastal areas
from storms.
Other important ecosystem-level endpoints relate to how changes in
ecosystem processes cause other changes of concern to humans. In general,
changes in biotic populations may not be important, as redundancy or
other compensatory mechanisms may prevent adverse changes in ecosystem
processes. But the converse is not true, and changes in ecosystem processes
almost invariably result in changes in biological constituents.
Some ecological endpoints that need assessment are explicitly recog-
nized in regulations, such as endpoints relating to water quality of surface
waters and endpoints relating to habitat quality. There are no simple
measures of such concepts as water or habitat quality, but often physical
parameters (e.g., soil structure), chemical parameters (e.g., dissolved oxy-
gen levels), or biological parameters (e.g., available habitat for waterfowl,
or forage base for deer populations) can be used as surrogates for these
ecosystem-level concerns.
OCR for page 101
HUAL4N EFFECTS ON THE TERRESTRIAL ENVIRONMENT
DEFINING ECOSYSTEM RESPONSES TO STRESS
Frequency and Novelty of Stress
101
Qualitatively different responses by ecosystems to stress will depend
on the frequency of occurrence and novelty of the disturbance (or closely
analogous disturbance) in the evolutionary history of the ecosystem. Thus,
the same disturbance can have dramatically different consequences on
different ecosystems, e.g., fire affecting grassland ecosystems versus tropical
rain forests. In the former case, fire is a natural part of the long-term
biogeochemical cycles of the ecosystem, necessary to rejuvenate a biotic
community that is adapted to survival or redevelopment after fire. In the
case of a tropical rain forest, a fire would lead to extreme disruption of
the physical habitat and nutrient reservoirs, and reestablishment of the
biotic community would take a very long time, if it were possible at all.
Similarly, a particular ecosystem will likely respond differently to different
disturbances; for example, the grassland may do well in the presence of
fire, but be devastated by overgrazing.
Time Scale of Stress
An ecosystem's response to stress must also be characterized in rela-
tion to the particular time scale of occurrence. Acute disturbances (i.e.,
those involving abrupt, large-magnitude changes in some characteristic of
an ecosystem) can be exemplified by the removal of live biomass (Grime,
1979), or by the removal of total biotic material, including previously liv-
ing material such as litter and detritus (Reiners, 1983~. Large-magnitude
changes typically involve alterations of species composition, as in the con-
version of forests in Vietnam to grass and bamboo ecosystems following the
spraying of defoliants (Ischirley, 1969~. Elimination of sensitive species,
reduction in pools of organic matter, and decreases in diversity have been
observed by numerous researchers to occur simultaneously following acute
disturbance (Weinstein and Bunce, 1981; Freedman and Hutchinson, 1980;
Woodwell, 1970; and Gordon and Gorham, 1963~.
Ecosystems often are adapted to cope with many types of natural dis-
turbances, especially chronic stresses that are predictable (e.g., intertidal
ecosystems adapted to diurnal and monthly Cycles) or periodic (e.g., grass-
land ecosystems adapted to fire). When these stresses are small relative
to the scale of the ecosystem, and when they have occurred commonly
during historical development of the ecosystem, the disturbance often will
be absorbed within the system structure, adding more heterogeneity but
not changing the basic ecosystem functioning. For example, in northern
temperate forests a response to wind-induced treefalls can be a formation
OCR for page 105
HZJM;4N EFFECTS ON THE TERRESTRIAL ENVIRONMENT
TABLE 2 Indicators of ecological effects.
· PURPOSES FOR INDICATORS
· intrinsic importance - key: indicator is endpoint
- e.g., economic species
· early warning indicator - key: rapid indication of potential effect
- use when endpoint is slow or delayed in response
- Animal time lag in response to stress; rapid response rate
- signal-t~noise low; discrimination low
- screening tool; accept false positives
· sensitive indicator - key: reliability in predicting actual response
- use when endpoint is relatively insensitive
- stress specificity
- signal-t~noise high
- minimize false positives
· processlfunctional indicator - key: endpoint is process
- monitoring other than biota; e.g., decomposition rates
- complement structural indicators
· CRl l ELLA FOR SELECI ING INDICATORS
· signal-to-noise ratio
- sensitivity to stress
- intrinsic stochasticity
· rapid response
- early exposure; e.g., low atrophic level
- quick dynamics; e.g., short life span, short life cycle phase
· reliabilitylspecifcily of response
· easeleconomy of monitoring
- field earning
- lab identification
- pre-existing data base; e.g., fisheries catch data
- easy process test; e.g., decon~osition, chlorophyll
· relevance to endpoint
- addresses "so what?" question
· monitoring feedback to regulation
- Captive management
.,
105
must be with respect to a particular stress, or combination of stresses, since
the nature of the different ecosystems may evoke different responses from
different stresses. In this light, one cannot accurately characterize a type
of ecosystem as being intrinsically resistant; further, the level of resistance
may be determined from measuring one indicator of the ecosystem, but
differ in measuring another. How an ecosystem responds to perturbation
and how readily it recovers, i.e., the stability of an ecosystem, like the
ecosystem itself, can only be defined operationally.
Sensitivity
A similar concept to resistance is the idea of sensitivity. A sensitive
ecosystem is one that responds readily to a particular stress; an insensitive
ecosystem may be oblivious to the stress. Sensitivity is not identical to
OCR for page 106
106
ECOLOGICAL RISKS
~0
`-_
O
o
m
~v,
lo
sp. A
o OK ~,~ O
By, 0
I'" ,~0
o ...,,.
TV
~ O
O x.
o
0:'
NV ..'
'I...,, ,,'
1~_ 0
Sp.C
1
Legend:
O Species
Ecosystem 3<
Properties
community structure
physical structure
processes
Ecological
Endpoints
Ecological
Indicators
O (folded)
(shaded)
FIGURE 2 Relanons~p among ecosystem properties ecological endpoints, and ecological
indicators. Ecosystems are characterized by a variety of properties that exist across many
space and time scales; included are species- and communi~-level properties, physical
structure, and ecological processes. Stress on an ecosystem can change some or all of these
properties. Ecological endpoints (shown in bolded figures) are those ecosystem properties
(species, community, structural, or process) for which changes would have importance to
humans and thus would represent changes in ecosystem health. Each ecological endpoint is
measured or monitored by ecological indicators (shown as shaded figures). Sometimes, the
endpoint itself is its own indicator (as in the Sample here for the process, community, and
structural endpoints, and for species A and C). Other endpoints are measured indirectly,
as for species B and C, and their indicatom are other species, community, structural, or
process properties of the ecosystem.
resistance, although both measure how much an ecosystem is affected by a
disturbance. But sensitivity also has a temporal component, and a system
that responds more rapidly than another, or to lower levels of disturbance,
is considered to be more sensitive.
OCR for page 107
HUAf4N EFFECTS ON THE TERRESTRIAL ENVIRONMENT
Recovery and Resilience
107
Another concept of importance is recovery, i.e., how the ecosystem
responds following removal of the stress. Again, there are two components,
one related to how rapidly the ecosystem recovers, the other to how
effectively the ecosystem recovers. The temporal aspect is characterized
as the ecosystems' resilience, which is defined as the inverse of the length
of time required for an ecosystem to return to near-normal. Note that
one cannot define this as a complete return to a pre-perturbed state,
because natural heterogeneity might preclude ever attaining that precise
state; however, the time required for an ecosystem state to return to a
point within a specified range of its pre-stress state could be a measure of
resilience (Harwell et al., 1981~.
One complicating factor sometimes considered is that the non-perturbed
ecosystem may well not be at steady-state, even in the absence of human
interference. Properties of the ecosystem may change over time. For ex-
ample, diversity of a forest ecosystem will increase during the early stages
of ecosystem development, decline in the middle stages of succession, and
increase again during the later stages (Woodwell, 1970~. In this case, the
ecosystem is characterized by a moving set of values describing the trajectory
of the undisturbed ecosystem. How resilient a non-steady-state ecosystem
is to disturbance reflects the mechanism of homeorhesis of the ecosystem,
i.e., feedbacks that tend to direct the ecosystem's state along a specific
time sequence. The analog for the steady-state ecosystem responding to
disturbance is homeostasis, a more commonly known term because of its
applicability to physiological control of individuals, such as in maintaining a
human's body temperature or blood pH. The human analog to homeorhesis
is the developmental sequence and timing associated with embryology and
with maturation of the individual into an adult.
Beyond the resilience issues are questions as to whether the ecosystem
effectively ever will return to its pre-perturbed state or trajectory. It
is possible that some complex ecosystems, when subjected to particular
disturbances, will become irreversibly transformed into another system,
having different components, steady-states, and dynamics; this is a well-
known characteristic of many ecosystems. For example, deforestation in
the coastal hills of Venezuela has changed soil structure, seed sources, and
the local physical environment sufficiently that forests have not returned
even after the areas were abandoned by humans. This phenomenon repeats
the irreversible loss of the great forests in Britain during neolithic times,
as humans cleared land for agricultural production and energy resources.
Perhaps the above examples merely reflect an exceedingly long time period
of recovery, and the ecosystem will eventually recover. Yet for practical
purposes, these examples of ecosystem change are permanent.
OCR for page 108
108
ECOLOGICAL RISKS
Recovery of ecosystems is in part dependent upon the characteristics of
the stress (i.e., the disturbance regime, including such factors as the nature
of the stress, its frequency, duration, and intensity), and in part the history
of the ecosystem (e.g., the level of preadaptation to disturbance, past history
of disturbance, and susceptibility of organisms within the ecosystem). An
ecosystem that has been subjected to repeated disturbances may tend to
deteriorate over time because of loss of nutrient reserves or substrate.
Examples of such deterioration can be found in the forests of the San
Bernadino Mountains of California following periodic ozone exposures
(Miller, 1973), and in salt marshes exposed to a series of oil spills (Baker,
1973~.
Recovery from repeated stress may be rapid if most of the important
species within the ecosystem complete their life cycles within the interim
between disturbance events (Noble and Slatyer, 1980~. Alternatively, for
single disturbances of longer duration, recovery will be promoted if the
important organisms within the ecosystem are capable of outlasting the
toxicant by remaining in a latent or resting stage. For example, poor
recovery has been noted in grassland systems exposed to oil- because oil
degradation proceeds slowly, and the actively growing portions of the
grasses become directly exposed to the toxicant during their growth periods
(Hutchinson and Freedman, 1978~.
Characterizing recovery of ecosystems has the same problems as char-
acterizing the ecosystem response to stress, specifically which indicator to
examine. Is an ecosystem recovered when its pools of nutrients are back to
the pre-stressed state; or when a specific species has reestablished its pop-
ulation at a particular density; or when the residues of a toxic chemical in
sediments or in biological tissues have decreased to below some threshold?
Just as an ecosystem functions and responds to stress at widely differing
rates, hierarchical levels, and spatial extents, it also recovers differentially.
There are substantial difficulties added in establishing an appropriate base-
line for comparison with the stressed ecosystem, especially since when
evaluating homeorhesis, one must not only have an adequate existing base-
line but also a representation of what the ecosystem dynamics would have
been had the ecosystem not been disturbed. Also, natural heterogeneity
and fluctuations again raise the issue of detecting signals from among the
noise of natural variations.
Summary
In summary, the health of an ecosystem is much too complex a concept
to be quantified by a single measure. The multiscaled and multilayered
nature of ecosystems establishes an almost infinite variety of ways of char-
acterizing the ecosystem's state and relationship to some baseline condition.
OCR for page 109
HUMAN EFFECTS ON THE TERRESTRIAL ENVIRONMENT
109
Simple schemes to overcome this intrinsic complexity of ecosystems are by
necessity simplistic and cannot be trusted. This is a true, but unfortunate
reality to which society must accommodate; it means there can be no sim-
ple, generic answers to the complex problems of environmental protection
and decision making. However, an operationally selected suite of indica-
tors, chosen in the context of the ecosystem of interest and the regulatory
and ecological endpoints of concern, can offer a reasonable and realistic
approach to evaluating ecosystem response and recovery to stress (Figure
2~.
SELECTION OF INI)ICATORS
Given that specific ecological endpoints need to be evaluated for a
particular ecosystem and stress, the next step is to identify what indicators
should be measured to detect potential changes in the ecosystem (Figure
14. Again there are innumerable components of the ecosystem that could
be evaluated, but some careful thought can reduce these to a manageable
set of indicators selected to optimize the detection of potential or actual
changes in the selected ecological endpoint of concern (Table 2~. The first
approach to this process is to focus on the purposes of indicators.
Purposes of Indicators
Intrinsic Importance
Some indicators have intrinsic importance, such as when populations of
direct human interest are measured directly. An example is the valuable
striped bass population of the Hudson River and other estuaries of the
east coast of the United States. Through the regulatory and, especially,
the litigation process, measurements of this species have developed into
the central concern for major human disturbances to the Hudson, such as
involving thermal power plant siting, the management of the PCB-laden
sediments in the river, and the recently resolved Westway Project in New
York City (Limburg et al., 1986a, b). Striped bass populations became a
primary indicator of ecological effects, especially through evaluations of
population levels, age structures, recruitment rates, mortality rates, and
migratory patterns. Many other examples of the endpoint itself being
the indicator include: deer population levels, breeding success in bald
eagles, productivity of Douglas fir stands, and harvested yields of shrimp.
The common theme for this intrinsic importance criterion is some direct,
usually economic, value of the species or processes.
Early Warning
But a big problem can develop if too much reliance is placed on just
monitoring economically important species as the indicator of effects on
OCR for page 110
110
ECOLOGICAL RISKS
ecological endpoints of concern. In many cases, by the time an effect
shows up on the indicator, it is too late for effective management or
mitigation. Thus, another category of indicators is warranted, i.e., earAf'
warning indicators.
The key characteristic of an early warning indicator is for it to respond
rapidly to a stress. Often this criterion means the indicator needs to
be exposed to the stress early in the introduction of the stress into the
ecosystem. Further, the indicator needs to respond rapidly once exposed.
Thus, there is both a time lag and a rate of response factor involved here.
Since the key issue is the rapid indication of a potential effect, the
early warning indicator is a red flag hoisted to signal the need for closer
examination of a potential problem. Consequently, the discrimination of
the indicator can be rather low, i.e., it need not provide all the information
needed to evaluate effects on the ecological endpoints of concern, and
tight, causal relationships between the stress and the triggering of the early
warning indicator are not required. Hence, this functions as a screening
tool, where false positives are acceptable at a relatively high rate (i.e.,
having the flag go up even though further evaluation demonstrates no
ecological effects of concern). Conversely, early warning indicators need
to minimize false negatives; thus, they need to avoid missing a warning for
a problem which is real. One way of enhancing this protective aspect is
to incorporate more than one early warning indicator in an environmental
protection scheme.
Reliabili~/Sensitivity
Another goal is for it to be a reliable indicator, with high capability in
characterizing an adverse effect on an ecological endpoint of concern. Note
that this category of indicators is focused on actual ecological effects rather
than on potential ecological effects. Thus, the key issue is not the rapidity
of response, but the reliability for characterizing changes in ecological
endpoints. This type of indicator does require strong evidence of causal
relationships with the stress, and the response should be relevant to the
state of the ecosystem.
This type of indicator is used when the ecological endpoint itself is
relatively insensitive to the stress, or when it is difficult to separate stress-
induced changes from the normal variation that occurs over time and/or
space. Stress specificity is essential here if the indicator can demonstrate
causality needed to Justin specific management or protection policies. Also,
a criterion for this category of indicators is to minimize false positives,
since incorrectly predicting unacceptable adverse impacts could lead to
uneconomical overregulation.
Long-term indicators might be necessary to reflect alterations at large
OCR for page 111
HUA~N EFFECTS ON THE TERRESTRIAL ENVIRONMENT
111
spatial/temporal scales, alterations that might not become evident by only
examining short-term indicators. As an example, remote sensing of land-
use patterns can illustrate loss of estuarine habitats in coastal regions of
southwest Florida, alterations that might not be as clearly apparent by
monitoring, say, the population levels of tarpon.
Process/Function
Indicators may be chosen to represent alterations in ecological func-
tions and processes. Such process indicators may be the ecological endpoints
themselves, but it is more likely that process indicators represent the poten-
tial for changes in other ecological endpoints of more immediate concern
to humans. Note that process indicators are not excluded from also being
early-warning indicators, or reliable indicators of change or of state.
Much has been made of the relative value of structural indicators (i.e.,
biotic indicators of population and community structures) as opposed to
functional indicators (i.e., of ecosystem processes; see Kelly et al., 1987
and Kelly, 1989~. Some authors have suggested that structural indicators,
involving effects on biotic populations, are more sensitive and better early-
warning indicators than functional indicators (cf., Schindler et al., 1985~.
Several reasons are offered for this generalization:
· ecological effects are first manifest as effects on individual organ-
isms and subsequently on populations; thus, functional responses would
imply prior associated changes in biotic populations performing those func-
tions;
· there is often functional redundancy in ecosystems, so that effects
on specific biota may not translate into functional effects; and
· recovery of biotic structure of an ecosystem often lags behind
recovery of functional attributes.
However, there are instances where functional indicators respond at least
as rapidly and sensitively to stress as structural indicators (Kelly et al.,
1987~. The point is not to prefer functional over structural indicators or
vice versa, but rather, carefully to select functional indicators that can
significantly enhance our ability to evaluate ecological responses to stress.
Criteria for Selection of Indicators
Sensitivity
Criteria can be listed for selecting a particular indicator to measure a
specified ecological endpoint. One factor is the sensitivity of the indicator
to stress, i.e., how large is the response of the indicator to a unit of stress.
This measure of indicator resistance is important with respect to the normal
OCR for page 112
112
ECOLOGICAL RISKS
variation that the indicator experiences over time and space in the absence
of stress. These two factors, sensitivity and variation, combine to form the
major determinants of the signal-to-noise ratio for the indicator. A high
signal-to-noise ratio is required for sensitive, stress-specific indicators; a
low signal-to-noise ratio is acceptable for screening indicators, especially
involving inexpensive or easily measured variables.
As an example of this issue, consider the purported impacts on the
striped bass population in the Hudson River ecosystem in comparison to
the natural variability of that species. Having density-independent mecha-
nisms as the primary control for these populations means a poor signal-to-
noise relationship, and experts were able to argue effectively on both sides
of the controversy concerning the presence or absence of demonstrated
effects, compensatory mechanisms, and other issues. By contrast, consider
the data on CO2 concentrations in the atmosphere at the Mauna Loa ob-
servatory in Hawaii (Keeling et al., 1982~. The annual cycle in CO2 levels,
related to seasonal turning on and off the primary production potential of
the Northern Hemisphere, is clearly discernable; and it is superimposed
over a rather constant, inexorable rise in the annually averaged CO2 levels,
reflecting effects from human inputs of CO2 and human-caused destruction
of primary production. Here the signal-to-noise ratios of both the long-term
trend and the annual cycle are quite good, and the indicator is convincing.
Rapidity of Response
A second criterion relates to the rapidity of response of the indicator,
especially with no time lag and a high rate of signal processing, as discussed
previously. Early exposure is important; consequently, for some stresses,
especially those transmitted through food webs, the rapidly responding
indicator is likely to occur at lower trophic levels. Quick response also
implies quick population dynamics, such as having a short life span or at
least a short duration for one phase of the life cycle; for example, changes
in phytoplankton are likely to occur much more rapidly than changes in
whale populations.
Specificity
Another criterion is the specific of the response indicator. High
specificity may be critical to establishing causal relationships and, hence,
appropriate management decisions. Conversely, broader response charac-
teristics (i.e., low specificity) may be much more appropriate for screening
indicators.
OCR for page 113
HUMAN EFFECTS ON THE TERRESTRIAL ENVIRONMENT
Ease/Economy of Monitoring
113
The criterion of the ease and/or economy of monitoring has historically
been of special importance. In one sense, seeking the most economical
indicator is of major benefit, in that a larger data base is likely to be
developed. For example, historical records such as fisheries' catch data
or board-feet of lumber harvested provide a major means for compar-
ing current environmental conditions with those in existence prior to the
development of stresses from industrial society. On the other hand, we
seem to become too enamored with the ease or historical precedent for
indicators; and we may find ourselves focusing efforts on large amounts of
data with great precision, but with poor accuracy and little relationship to
ecological endpoints of concern. This problem applies to laboratory testing
(e.g., bioassays on easily maintained but ecologically insignificant species);
field sampling (e.g., counting the 95th species in benthic samples even
though looking at the top dozen or so will provide virtually all the relevant
information); and pre-existing databases (e.g., fishery catches, where the
endpoint is a poor indicator of anthropogenic stress on the environment
because of poor signal-to-noise ratios).
Relevance
A final criterion is the degree of relevance of the indicator to the
ecological endpoint of concern. Clearly, if the indicator itself is identical
to the endpoint (e.g., the population levels of an endangered species), the
relevancy is maximal. Otherwise, the more closely linked the indicator is to
the ecological endpoints of concern, the less difficult it is to answer the "so
what?" question that often haunts demonstrations of environmental change.
Process indicators tend automatically to be considered more relevant than,
for example, sensitive species indicators, since loss of a species sensitive
to stress, but otherwise not of particular note for humans or ecosystems,
raises the 'So what?" question (cf., Kelly et al., 1987~.
SUMA/L\RY
Ecosystems are complex and varied, multiscaled and multitiered, and
subject to continuing change and adaptation. Consequently, a sophisticated
approach is needed to characterize ecological effects from human activities,
relying on a suite of ecological response/recovery indicators that reflect
the status of the variety of facets about the ecosystem, or endpoints, of
concern to humans. Focusing on these suites of indicators and endpoints
can provide a systematic framework for incorporating scientific knowledge
and understanding into a broader process of ecological risk assessment. It
is unreasonable and futile to expect that a simple, generically applicable,
OCR for page 114
114
ECOLOGICAL RISKS
single measure of ecosystem health can ever be realized. However, such a
scheme is not required, and an ecological risk assessment methodology for
enhancing environmental decision making is a reachable goal for ecological
science.
Acknowledgment
This report is ERC-153a, Ecosystems Research Center, Cornell Uni-
versity. The ERC was established under a cooperative agreement between
the U.S. Environmental Protection Agengy (EPA) and Cornell. This chap-
ter represents the views of the authors, and not necessarily the views of the
EPIC
REFERENCES
Amundson, R.G., and LH. Weinstein. 1980. Effects of airborne fluoride on forest ecosystems.
Pp. 63-78 in the Proceedings of the Symposium on the Erects of Air Pollutants on
Mediterranean and Temperate Forest Ecosystems. U.S.D.A. Forest Service, Pacific SW
Forest Experiment Station, General Technical Report PSW43, Berkeley, California.
Baker, J.M. 1973. Recovery of salt marsh vegetation from successive oil spillages. Environ-
mental Pollution 4:2~-230.
Bedford, B.L., ed. 1987. Modification of plant-pest interactions by air pollutants. ERC-117.
Ecosystems Research Center, Cornell University, Ithaca, NY.
Bormann, F.H., and G.E. Likens. 1979. Pattern and process in a forested ecosystem. New
York: Springer-Verlag.
Freedman, B., and T.C. Hutchinson. 1980. Long-term effects of smelter pollution at Sudbury,
Ontario, on forest community composition. Canadian Journal of Botany 58:2123-2140.
Gordon, A.G., and E. Gorham. 1963. Ecological effects of air pollution from an iron-sintering
plant at Wawa, Ontario. Canadian Journal of Botany 41:1063-1078.
Gnme, J.P. 1979. Plant Strategies and Vegetation Processes. Chichester, U.K: Wiley and
Song
Maxwell, CC. 1984a. Regulatory Framework of the Federal Water Pollution Control Act,
Section 301(h). ERG-29. Ecosystems Research Center, Cornell University, Ithaca, NY.
Maxwell, C.C. 1984b. Analysis of Federal Water Pollution Control Act, Section 403: Ocean
Discharge Criteria. ERG-26. Ecosystems Research Center, Cornell University, Ithaca,
NY.
Harwell, C.C. 1989. Regulatory framework for ecotoxicology. Pp. 497-516 in Ecotoxicology:
Problems and Approaches, S.A. Levin, M.A. Howell, J.R. Kelly, and K. Kimball, eds.
New York: Springer-Verlag.
Maxwell, M.A., W.P. Cropper, and H.L~ Ragsdale. 1978. Nutrient cycling and stability: A
reevaluation. Ecology 58:660-666.
Harwell, M.A., W.P. Cropper, Jr., and H.L" Ragsdale. 1981. Analyses of transient character-
istics of a nutrient cycling model. Ecological Modelling 12:105-131.
Maxwell, M.A., T.W. Duke, and J.R. Kelly. 1985. Evaluation of proposed environmental
scoring methodology. ERC-106. Ecosystems Research Center, Cornell University,
Ithaca, NY.
Maxwell, M.A., and T.C. Hutchinson, with W.P. Cropper, Jr., C. C. Maxwell, and H.D.
Grover. 1989. Environmental Consequences of Nuclear War. Volume II. Ecological
and Agricultural Effects. Second Edition. Chichester, U.K.: John Wiley & Sons.
Maxwell, M.^, and C.C. Howell. 1989. Environmental decision making in the presence
of uncertainty. Pp. 517-540 in Ecotoxicology: Problems and Approaches, S.A. Levin,
M.A. Howell, J.R. Kelly, and K. Kimball, eds. New York: Springer-Verlag.
OCR for page 115
HUMAN EFFECTS ON THE TERRESTRIAL ENVIRONMENT
115
Harwell, M.A., CC Howell, and J.R. Kelly. 1986. Regulatory endpoints, ecological
uncertainties, and environmental decision making. Oceans 86 Proceedings 3:993-998.
Hutchinson, T.C, and W. Freedman. 1978. Effects of experimental spills on subarctic boreal
forest vegetation near Norman Wells, N.W.T., Canada. Canadian Journal of Botany
56:2424-2433.
Keeling, CD., R.B. Bacastow, and T.P. Whorf. 198Z Measurements of the concentration of
carbon dioxide at Mauna Loa, Hawaii. Carbon Dioxide Review 1982:377-382.
Kelly, J.R., T.W. Duke, M.A. Howell, and CC Harwell. 1987. An ecosystem perspective on
potential impacts of drilling fluid discharges on seagrasses. Environmental Management:
11~4~:537-562.
Kelly, J.R. 1989. Ecotoxicology beyond sensitivity: A case study involving 'unreasonableness'
of environmental change. Pp. 473496 in Ecotoxicology: Problems and Approaches,
S.A. Levin, M.A. Harwell, J.R. Kelly, and K. Kimball, eds. New York: Springer-Verlag.
Levin, S.A., AD. Kimball, W.H. McDowell, and S.H. Kimball, eds. 1984. New perspectives
in ecotoxicology. Environmental Management 8~5~:375442.
Limburg, K.E., S.A. Levin, and C C. Howell. 1986a. Ecology and estuarine impact
assessment: Tons learned from the Hudson River and other estuarine experiences.
J. Environmental Management 22:255-280.
Limburg, K.E., M.A. Moran, and W.H. McDowell. 1986b. The Hudson River Ecosystem.
New York: Springer-Verlag.
Miller, P.L~ 1973. Oxidant-induced community change in a mixed conifer forest. Pp. 101-117
in Air Pollution Damage to Vegetation, J.A. Naegle, ed. Adv. Chem. Series No. 122,
American Chemical Society, Washington D.C.
O'Neill, R.V., D.L. DeAngelis, J.B. Waide, and T.F.H. Allen. 1986. A hierarchical concept of
ecosystems. Monographs in Population Biology 22. New Jersey: Princeton University
Press.
Paine, R.T. 1974. Intertidal community structure: Experimental studies on the relationship
between a dominant competitor and its principal predator. Oecologia 15:93-120.
Paine, R.T. 1980. Food webs: Linkage, interaction strength, and community infrastructure.
J. Animal. Ecol. 49:667-685.
Rapport, D.J., H.A. Regier, and 12G Hutchinson. 1985. Ecosystem behavior under stress.
Amer. Nat. 125~5~:617-640.
Reiners, W.A. 1983. Disturbance and basic properties of ecosystem energetics. Pp. 83-98
in Disturbance and Ecosystems, H.A. Mooney, and M. Godron, eds. New York:
Springer-Verlag.
Schindler, D.W., K.H. Mills, D.F. Malley, D.L. Findlay, J.A. Shearer, I.J. Davies, M.A.
Turner, G.A. Linsey, and D.R. Cruikshank. 1985. Long-term ecosystem stress: The
effects of years of experimental acidification on a small lake. Science 228:1395-1401.
Smith, W.H. 1980. Air pollution: A 20th century allogenic influence on forest ecosystems.
Pp. 79-87 in the Proceedings of the Symposium on the Effects of Air Pollutants on
Mediterranean and Temperate Forest Ecosystems. U.S.D.A. Forest Sentence, Pacific SW
Forest Experiment Station, General Technical Report PSW-43, Berkeley, California.
Tschirley, F.H. 1969. Defoliation in Vietnam. Science 167:779-786.
Webster, J.R., J.B. Waide, and B.C. Patten. 1975. Nutrient recycling and the stability of
ecosystems. In Mineral Cycling in Southeastern Ecosystems, F.G. Howell, J.B. Gentry,
and M.H. Smith, eds. ERDA CONF-740513, NTIS, Springfield, VA.
Weinstein, D.A., M.A. Harwell, and L. Buttel. (in prep.) A sensitivity analysis of a forest
simulator Methodology and application. ERC-147. Ecosystems Research Center,
Cornell University, Ithaca, NY.
Weinstein, L.H., and H.\UF. Bunce. 1981. Impact of emissions from an alumina reduction
smelter on the forests at Kitimat, B.C.: A synoptic review. Pp. 2-16 in the Proceedings
of the 74th Annual Meeting of the Air Pollution Control Association, Philadelphia,
Pennsylvania, June 21-26, 1981.
Woodwell, G.M. 1970. Effects of pollution on the structure and physiology of ecosystems.
Science 168:429433.
Representative terms from entire chapter:
regulatory endpoints
