ecosystem management and the conservation of aquatic biodiversity and

VOL. 32, NO. 2WATER RESOURCES BULLETIN

ECOSYSTEM MANAGEMENT AND THE CONSERVATION OFAQUATIC BIODIVERSITY AND ECOLOGICAL INTEGRITy1

Christopher A. Frissell and David Bayles2

ABSTRACT: Ecologically effective ecosystem management willhumans manage the landscape (Warren, 1979; Karr,require the development of a robust logic, rationale, and framework

for addressing the inherent limitations of scientific understanding. 1991, Schlosser, 1991; Roth et all, in press). JudgingIt must incorporate a strategy [’or avoiding irreversible or large- by pervasive and seemingly relentless declines inscale environmental mistakes that arise from social and pnlitica] abundance and natural diversity of many monitoredforces that tend to promote fragmented, uncritical, short-slghted,

groups of aquatic biota throughout the world (e.g.,inflexible, and overly optimistic assessments of resource status,Regier and BaskervilIe, 1986; Williams et al., 1989;management capabilities, and the consequences of decisions andNeh]sen et al., 1991; Allan and Flecker, 1993), we arepolicies. Aquatic resources are vulnerable to tbe effects of human

activities catchment-wide, and many of the landscape changes doing poorly. Ecosystem management seeminglyhumans routinely induce cause irreversible damage (e.g., some implies dramatic improvement in our performance asspecies introductions, extinctions of ecotypes and species) or give

conservators of ecological integrity and biodiversityrise to cumulative, long-term, large-scale biological and cultural(Salwasser, 1992; Montgomery et al., 1995), but until"nsequences (e.g., accelerated erosion and sedimentation, dei’or-humans figure out what ecosystem management is~ation, toxic contamination of sediments). In aquatic ecosystems,

~iotic impoverishment and environmental disruption caused by and learn to implement it successfully for a signifi-past management and natural e.vonts profoundly constrain the abil. cant period of years, how can we be confident it willity el’future management to maintain biodiversity and restore his.

protect and restore our aquatic biota and other watertorical ecosystem functions and values. To provide for rational,resources any better than past ways of environmentaladaptive progress in ecosystem management and to reduce the risk

of irreversible and’unanticipated consequences, managers and sci- management?entists must identify catchments and aquatic networks where eco. Whether people are for or against it, almost every-logical integrity has been least damaged by prior management, and body seems to have a different concept of what ecosys-jointly develop means to ensure their protection as reservoirs of tern management is. We think it is fruitless to arguenatural biodiversity, keystones for regional restoration, manage-ment models, monitoring benchmarks, and resources for ecological ~i.! i~l;.fl S (e.g., Fitzsimmons, 1996), although it can beresearch.

~i~ hbout the conceptual and physical existence of ecosys-

(KEY TERMS: ecosystem management;ecological integrity; aquatic ~’~’!cl~fiite useful to argue about how to most usefullybiodiversity; cumulative effects; conservation reserves; landscape define their structure and boundaries (e.g., Jensen etplanning; watershed analysis.)

al., 1996) and about how they should be managed.Like it or not, the concept of e~osystem management

~" isn’t going to go away anytime soon, because some-

INTRODUCTION thing like it is necessary to address the vast naturaland cultural wreckage of exploitation-focused, single-resource approaches to resource management that

The majority of aquatic organisms have the unfor- has accompanied European colonization of the globe.tunate handicap of living downstream of humans, a There should be little doubt thatthe struggle tobasic fact only the most ideologically motivated can define what ecosystem management is, and how ordeny. As a consequence, the integrity and biodiversity ~vhether it should be implemented on the landscape,of aquatic ecosystems is highly dependent on the way will be critical to the future of aquatic biota and other

1paper No. 95142 of the Waler Resources Bulletin. Disctmsions are open until October 1, 1996.2Respectively, Research Assistant Professor, Flathead Lake Biological Station, The University of Montana, 311 Bio Station Lane, Polson,

Montana 59860-9659; and Senior Program Director, The Pacific Rivers Council, P.O. Box 10798, Eugene, Oregon 97331.

229 WATER RESOURCES BULLETIN

Frissell and Bayles

diminishing water resources. In this paper we argue Baskerville, 1986; Soul~, 1991; Karr, 199.1; Ludwigthat the approaches to ecosystem management pro- et al., 1993).posed to date by government and industry fall far In the following discussion, we examine severeshort of ensuring that future management will halt or commonly proposed approaches to ecosystem man.age-reverse the deterioration of aquatic ecosystems, ment and discuss some of their crucial technical and

In the rush of government agencies to re-define (or, operational shortcomings from the standpoint of con-.more skeptically, re-package) their environmental servation of aquatic ecological integrity and biologicalmanagement programs as ecosystem-friendly endear- values. Then we describe an alternative strategyors, they have failed to acknowledge that human cul- based on establishment of watershed-based conserva-tures have throughout their history deemed their own tion reserves that could potentially reduce many of"state-of-the-art" environmental management prac- the threats to aquatic ecosystems posed by uncertain:rices to be good and proper. At any given moment in ty (and its evil step-sisters, ignorance and hubris) history, there are always people who tout contempo- ecosystem management. If we trul~ aspire to therary management practices as the Panglossian pinna- goals of ecosystem management, we argue (butcher-cle of social and technol.ogical refinement, while the ing a time-honored proverb) that a watershed in theless zealous accept such practices as clearly improved bush can be worth two in the hand.and unquestionably sufficient to maintain desiredresource conditions (i.e:, ecosystem functions).Because our generation believes past generations ofmanagers were wrong in such assumptions, we have TIlE RANGE OF NATURAL VARIABILITY

now invented the term ecosystem management to con-note a new and smarter approach. But what makes us One of the most common precepts invoked to guideso sure we’ve got it right this time (Stanley, 1995)? ecosystem management is the notion that humanAnd what if we don’t? actions should either maintain or return ecosystems

We have heard some scientists and managers claim to within their range of natural or historic variabilitythat now that we have ecosystem management, what- (e.g., FEMAT, 1993; Montgomer.y et al., 1995).ever exactly it may be, we can proceed with large- Although this concept does helpfully point towardscale development of the landscape for human viewing the past as the key to future management, w

’ purposes without fear of ecological retribution. Oth- share the concern of Rhodes et al. (1994) that it hasers, including many of those recently promoting the several major operational and practical limitations.so-called "forest health" agenda in the United States, This concept fails to weigh many of the most funda~seem convinced that because ecosystem management mental environmental realities that constrain ecosys-implicitly incorporates rehabilitation of ecosystems, tern management, either in the technical or policywe must get on with it urgently to correct our past sense.mistakes - we have to re-do management right, and Is the range of variability in ecosystems conditionsright now, everywhere. More modestly, others suggest really what we seek to emulate, or is it more impor-that the concept of ecosystem management at least rant to maintain in a broader sense the full pattern ofopens the door for effective integration of scientific states and successional trajectories (Frissell et al., ink.nowledge into management decisions (Montgomery press)? Strictly speaking, the range of Variability iset al., 1995). defined by extreme states that have occurred due to

However, within the past century or so, rational climatic or geologic events over long time spans.people have advanced remarkably similar arguments Nothing says these extreme states were favorable forand claims for soil conservation, clear-cut logging, water quality or aquatic biodiversity, and in fact suchdams, hatchery fish culture,~ irrigated agriculture, natural-historical extremes were probably no moremaximum sustained yield, multiple use, water quality favorable for these values than present-day extremes.standards, land grant universities, and forest plan- From the point of view of many aquatic species, thening, to name a few once-new concepts. Obviously the range of natural variability at any one site wouldavailability and widespread application of these tech- doubtless include local extirpation. At the scale of anologies, arrangements, a~d institutions did not spare large river basin, management could remain wellus the consequences of aquatic resource degradation, within such natural extremes and we would still faceEach may have in its own way slowed or ameliorated severe degradation of natural resources and possiblesome of the most egregious contemporary examples extinction of species (Rhodes et al., 1994). The missingof environmental destruction (some have fostered element in this concept is the landscape-scale patternmore than their share of damaging side effects), but of occurrence of extreme conditions, and patterns overnone has resulted in truly sustainable resource use space and time of recovery from such stressed states.or maintained .ecological integrity (Regier and How long did ecosystems spend in extreme states vs.

WATER RESOURCES BULLETIN 230

Ecosystem Management and the Conservation of Aquatic Biodive~.~ity and Ecological Integrity

intermediate or mean states? Were extremes chrono- range of variation will likely exclude some physicalogically correlated among adjacent basins, or did . and biotic dimensions important for the maintenanceasynchrony of landscape disturbances provide for of ecological integrity and native species diversity.large-scale refugia for persistence and recolonization To further complicate things, many of the distur-of native species? These are critical questions that are

bance events that dramatically shape terrestrial sys-not well addressed under the concept of range ofnatu-

terns (e.g., fire, windstorms) may have relativelyral variability as it has been framed to date by man-

subdued effects in aquatic ecosystems, whereasagers,

aquatic systems respond more dramatically to pro-We suspect that in most aquatic ecosystems,

cesses such as’floods and acceleration of erosion thatextreme states continue to be determined largely by may have rather subtle or spatially restricted expres-high-magnitude natural.events, whereas most human sion in the terrestrial environment. Such complica-activities predominantly influence ongoing frequent tions in the coupling of terrestrial and aquaticand lower-magnitude processes, although at cumula-

environments mean that extrapolating from one totively vast spatial scales (Frissell et al., 1986; Roth

the other is problematic and fraught with uncertainty.et al. , in press). Repeated, chronic, persistent, or

Each may be driven by different disturbance process-anomalously extensive but sometimes subtle alter- es, even while linkages such as erosion and sedimen-ation of the pattern of lower-magnitude processes,

ration, downstream flow of contaminants, andsuch as the seasonal and diurnal patterns ofriver dis-exchange between sur£ace and ground waters connect

charge, temperature, and sediment mobility, can have the two systems inseparably. Therefore, perceivedmore severe effects on the integrity and resilience of management problems in terrestrial systems, such asaquatic ecosystems and biota than large floods and the depletion of older, larger trees, and proliferationother single-pulse, catastrophic events of much higher of dense younger stands in some western forests thatmagnitude (Yount and Niemi, 1990). Humans, has recently been labeled a "forest.health crisis," dohowever, are more likely to detect, perceive, and not necessarily correspond to the major threats toemphasize the latter category of catastrophic events aquatic systems. Indeed, in the forest health example,as disturbances that exceed the known range of

many of the proposed cures (e.g., salvage logging and~atural-historical conditions. The consequence is that massive thinning programs, continuing existing live-.t~e vast array of human activities that cause subtle stock grazing policies) pose far greater threats to fish

but creeping and pervasive effects is de-emphasizedpopulations and aquatic ecosystem integrity than do

or ignored by managers and regulators, and most fires and other natural events that might (or mightplanning and protection measures focus primarily on not) be associated with the "undesired" changes inhuman activities known to directly trigger massive,

forest structure (Henjum et al., 1994; Rhodes et al.,unprecedented events of episodic proportions (e.g., 1994). For aquatic systems in the west, the manage-massive industrial discharges, or collapse of a mine ment crisis arises from the cumulative and persistentdrainage retention structure). No environmental

effects of thousands of miles of roads, thousands ofimpact, statement we have seen in the Pacific North-

’dams, and a century Of logging, grazing, mining, crop-west has evaluated (or otherwise disclosed) the chron- land farming, channelization, and irrigation diversionic, cumulative effects of human activities that can (Frissell, 1993a; Wissmar et al., 1994; Rhodes et al.,cause small increases in the rate of local extirpation

1994).of breeding groups and simt~ltaneous decreases in theFinally, for many kinds of ecosystems (e.g., low-

rate of recolonization, which coupled can clearly pro-elevation alluvial fans in tbe Great Basini forested

duce strikingly rapid declines in species with popula- floodplain rivers of New England and the midwest,tion ecology typical of that in Pacific salmon (Frissell, the cedar forests of Florida, grasslands of the Great1993a) and many other formerly abundant aquatic

Plains and Columbia Plateau), we have few or noorganisms,unaltered representative sites and sparse historical

The concept of range of natural variability also suf- records to reconstruct what natural-historical condi-fers from its failure to provide defensible criteria

tions looked like and how they were maintained.about which factors’ ranges should be measured. Pro- These ecosystems have been so starkly and extensive-ponents of the concept assume that a finite set of vari- .ly modified (and so sparsely studied, relative to theirables can be used to define the range of ecosystem scale) that we cannot presently determine how theybehaviors, when ecological science strongly indicates varied over time and space before destruction of abo-~any diverse factors can control and limit biota and riginal cultures and colonization by European man.atural resource productivity, often in complex, inter-

acting, surprising, and species-specific and time-variant ways. Any simple index for measuring the


Frisscll and Bayles

MIMICKING NATURAL DISTURBANCE natural disturbances, such as major landslides, capREGIMES: TIlE GHOSTS OF IMPACTS have similar effects. The result is that the impacts

PAST AND FUTURE future disturbances are contingent on the legacy b.past disturbances - the Ghost of Impacts Past. The

Another frequently cited guiding principle for same magnitude or pattern of" disturbance may have

ecosystem management is the notion that human dramatically different effects in a catchment that hasactions should attempt to mimic natural or historical experienced Such prior disturbances than in an iden,disturbance.regimes (thus presumably ~:emaining tical catchment that has not had such a history. The

within the natural range of variation) (e.g., FEMAT, Ghost of Impacts Past determines the response of an

1993). Even on its face, this concept faces logical trou- ecosystem to any particular disturbance regime, even

ble as a rationale for management. If natural distur- though its presence isn’t always obvious. If you don’t

bance regimes are the best way to maintain or restore believe this ghost exists, we are certain you will not

desired ecosystem values, then it seems nature should see it. You will nevertheless suffer its consequences,

be able to accomplish this task very well without and you will be left (as have many in the past) with-

human intervention. It is difficultto imagine how pro- out a defensible explanation for why your manage-

gramming of additional artificial disturbances, such ment objectives were not achieved.

as more road construction and logging, can be neces- Presently, however, planning for ecosystem man-sary to return an ecosystem to its natural disturbance agement remains largely focused on defining how andregime or tO somehow improve or optimize that where traditional resource extraction activities andecosystem’s operation. The principle exception might environmental disruption can be continued withoutbe where a human intervention, creating a relatively irreversible net harm to water quality, bi0diversity,

small disturbance, is necessary to undo a prior alter- and other ecosystem values (Frissell et al., 1992;

ation that otherwise would persistently impact the Grumbine, 1994). This emphasis presumes there

ecosystem - such as the removal of a dam, or an exists some ecological space in which such disruptive

unstable road network. Strictly speaking, under this activities can be pursued with no ~onsequential illprinciple thesole task of management should be the effects. It assumes that we have the capability toreversal of artificial legacies to allow restoration of identify such "free space" and to implement activities

natural, self-sustalning ecosystem processes, that will not violate it. It assumes (without checking.

In actual application, this principle is not so strict- that watershed ecosystems retain inherent resilience

ly applied. It becomes a credo for shaping manage- that allows them to recover from continued humanment actions such that they more nearly resemble the disturbances (Frissell, 1993b; Rhodes et al., 1994).quality, spatial distribution, and temporal pattern of However, because of the persistence of many kinds ofnatural disturbance processes. Butin this sense, the impacts in aquatic systems and because of the exten-

concept is haunted by at least two very consequential sive nature of human activities in most catchments,problems. Due to their shadowy, nearly phantom-like any inherent ecological resilience or resistance these

nature, we caricature these problems as ghosts that aquatic ecosystems may once have had is likely corn-

haunt the concept of ecosystem management, promised by the Ghost of Impacts Past. Ecosystem

Most ecosystem management plans that embrace management must be more than the search for thethe natural disturbance regime concept assume that last free lunch. For watersheds and aquatic ecosys-

we can simply start managing this way today, and all terns, we have ample evidence that if there ever was a

our management problems should vanish. The legacy, free lunch, we already ate it.of past disturbances, both natural and human, is tac- Even more problematic is the Ghost of Impacts

itly denied. However, it is, imperative to account for Future. Natural disturbance events, both small andspecific historical even~s, and their long-term lega- large, will continue, by definition in an unmanaged

cies, in any attempt to consider how far an ecosystem and unpredictable fashion. We think of these events

has deviated from its natural-historical disturbance as "wild disturbances," in the same sense that natu-regime and what actions may be necessary to return rally produced fish are wild fish. Their behavior is not

it to some semblance of its former domain of behavior, under human control and cannot always be anticipat-

A simple example is the case of aggradation of coarse .ed. That is their nature. Even if the probability ofsediments in streams following extensive human dis- occurrence, magnitude, and effects of such events can

turbance of their catchments. The effects of increased be predicted, their timing cannot be. The result is

sediment yield on channel morphology and stability that even the best-laid management programs based

can persist for many decades (perhaps centuries) after on disturbance regimes can go badly awry. In fact, thethe causal disturbance of the slopes of the catchment more meticulously a management program is

(Hagans et al., 1986; Ziemer et al., 1991). Large-scale designed around a particular expected or desired dis-turbance regime and sequence of actions, the more


Ecosystem Management and ’the Conservation of Aquatic Biodiversity and Ecological Integrity

likely it is to fail because of unanticipated naturalmost people, and not very rapidly. Data, for instance,vents. The Ghost of Impacts Future prevents us fromon the mechanics of debris flows in headwater chan-.ontrolling disturbance regimes, and it is little morenels have little intrinsic meaning or obvious relevancethan hubris to assume that we manage disturbanceto most engineers, farmers, or silviculturists. Theregimesin the sense that they can be programmedtragic risk is that by promising a technical solution toand optimized for specific objectives. While humansenvironmental problems, watershed analysis will pro-can and do influence disturbance regimes (often invide an excuse that allows managers to avoid facing

unanticipated ways), it seems to us disturbance the full array of political, philosophical, and adminis-regimes ultimately manage themselves,trative dimensions of management reform. Themantra of managers will remain, "Let the scientiststake care of it."

WATERSHED ANALYSIS AND MANAGEMENT Perhaps more importantly, better scientific data: may not unambiguously point the way for an honestly

repentant manager either. Fundamental uncertain-Another important component of ecosystem man, ties, the Ghost of Impacts Future among them, willagement is watershed analysis (FEMAT, 1993; Mont- remain. In fact, after a really good watershed analy-gomery et ai., 1995) and relate~l methods that attemptsis, critical uncertainties will probably appear to theto evaluate and eventually prescribe management alert decision maker to loom larger than they everactions based on cause-effect analysis and simulationhave before, simply because the complexities of ecolo-models. Watershed analysis is a set of technical toolsgy and history will be a little less blurred by the lensthat, unfettered by bureaucratic encumberments, of ideology. When was the last time science or newholds promise for assisting in the retrospective analy- technology simplified your life? It happens now andsis of qatchment change and aquatic ecosystem then, but the opposite seems much more the rulel

response. In this sense, it can be a way of getting a fix As McNab (1983) pointed out to wildlife managerson the Ghost of Impacts Past and reducing the likeli- and scientists, one fundamental problem is that natu-hood of management failure from this cause. But ral resource managers tend to resist close working~here are some serious limitations to what we canrelationships with researchers. Managers feel morepect to achieve from watershed analysis, and we comfortable in the political arena when they portray~re concerned it is increasingly being oversold as a the ecological assumptions underlying their programspanacea for an accumulating burden of management as proven facts rather than as tentative hypotheses.

problems that are as much political, ideological, and Good scientists are inherently skeptical and thereforeadministrative in origin as scientific and technical, if often seem more a nuisance than a help to managers.not more so.To acknowledge uncertainty in the principles guidingAlthough properly focused scientific analysis might a management program is to accept that failure or

help clarify the inadequacy of management premisessuccess of the program is itself a test of the underly-and the causes of management failure (Underwood, ing ecological assumptions. This requires managers

1995), watershed analysis as it presently exists (e.g., and scientists to work together to establish and moni-as portrayed in FEMAT, 1993, and subsequent federaltot criteria for evaluation, which in most cases willdocuments) is not designed to accomplish this task. Inrequire explicit experimental designs incorporatingfact, perhaps the principle flaw of watershed analysisunmanipulated control systems (McNab, 1983; McA1-as a management tool is that it does not provide alister and Peterman, 1992). Unfortunately, unexpect-clear vehicle or protocol to link technical analysis and ed results can embarrass managers, especially the

policies and decisions. The arguments of Montgomery. more intrepid or audacious ("ecosystem") managerset al. (1995) suffer from what we would characterize who tend to take the lead in the development of newas an overly optimistic assumption that better techni- programs. Unless the largely uncertain and experi-cal analysis will in some unspecified way lead to bet- mental basis of all ecosystem management programster management plans, decisions, and outcomes, is squarely faced, watershed analysis and similarWhile agreement on scientific facts may help reduce assessment procedures conducted by researchers aresome of the uncertainty and illusion in management, .unlikely to themselves markedly change or improvethis view denies the overriding importance ofideologi- management. Only the most egregious mistakes ofcal, philosophical, and political perspective in man- past management will be exposed, and the virtually

~’ement. Facts gain their meaning through the lens uniform response of managers to retrospective analy-theory and world views (C. E. Warren, unpublished sis is that the lessons of the past are largely irrele-

,aanuscript, Department of Fisheries and Wildlife, rant because "We don’t do that anymore" (they pointOregon State University, Corvallis), and better data out that now we use Best Management Practices, orare unlikely to change world views - at least not for buffer strips, or standards guaranteed to produce


Frisscll and Bayles

sediment~free roads, or some other change in manage- This difference in’biogeographic context creates

ment style they presume to lessen environmental profound implications for ecosystem management and

impact), the conservation of" aquatic biodiversity (Zwick, 1992,

Another serious problem with watershed analysis Frissell, 1993b; Doppelt et al., 1993). Unstated

is that many of" the necessary scientific tools for relat- assumptions of past approaches to modeling and man-

ing changes in physical systems to biological respons- agement of" biological populations in aquatic ecosys-

es are weak or nonexistent. On the research side, terns include the following: (1) disturbances are

some proponents of watershed analysis perpetuate isolated and independent in their effects, and the

the illusion that models or simple relationships exist ecosystem as a whole remains functionally intact; (2)

that allow prediction of changes in particular fish biotic recovery at each disturbed site proceeds lade-

populations based on changes in its physical habitat, pendently and relatively rapidly, also independent of

In fact such capabilities are crude, and may in the the site’s context in the ecosystem; (3) a steady, virtu-

best circumstances extend only to the general popula- ally unlimited supply of organisms is available to col-

tion trend or time to extinction that might be likely onize disturbed habitat patches as they recover

with a given physical scenario and good biological physically; and (4) biota and riverine habitats are

information on current population status (Rieman et largely homogeneous in distribution so that habitatal., 1993). Existing models do a relatively poor jobof and fish populations are readily replaceable, genericpredicting the general range of fish biomass likely techniques of habitat modification are widely applica-under a givdn set of physical conditions (Fausch et ble, and the risk of" failure or unintended side effectsal., 1988; Hall, 1988), let alone the far more delicate of management actions is minimal (Frissell, 1993b).task of’ predicting the abundance or harvest of indi- These assumptions may be at least partly valid in avidual species and’populations (e.g., Hecky et al., landscape that is relatively free of recent, large-scale1984). And we have even less experience with other human alteration or catastrophic natural disturba, nce

kinds of organisms. General indices of ecological (Figure la).integrity, developed for multi-species assemblages o£ However, in a landscape that has been highlyaquatic organisms or for habitat factors in specific altered in a relatively short period of time, much dif-geographic areas, do often have predictable correla- ferent biogeographic dynamics may prevail (Frissell,rive relationships with environmental stressors at 1993b). In this context, an aquatic habitat mosaicleast at coarser scales (e.g., Karr, 1991; Roth et al., in that is inherently heterogeneous becomes more highlypress), but because the underlying causal mecha- fragmented and, from the standpoint of sensitivenisms of these relationships are not fully understood, species, more patchy (Figure lb). Most present pro-many managers and some scientists continue to reject duction, abundance, and diversity of sensitive biotathem. : may be supported by the small proportion of the over-

all habitat mosaic that remains relatively undisrupt-ed. Fragmented and isolated populations sufferelevated vulnerability to extinction through further

THE ECOLOGICAL AND SPATIAL CONTEXT habitat alteration or demographic or genetically-OF WATERSHED CHANGE AND mediated reproductive failure (Zwick, 1992; Rieman

BIOTIC RESPONSES et al., 1993; Bradford et al., 1993). The present distri-bution and life history patterns of such populations,

The lack of success in mechanistically linking bio- largely governed by the availability of habitat refugia

logical response models to physical driving models and the specific historical pattern of habitat alteri

stems at least partly from a failure to pay appropriate ation, determine their ability to respond to future

attention to the geographical and ecological context in changes in habitat. Biological responses are thus his-

which models are derived and applied. This means torically and spatially constrained, determined by the

not only that biological responses are likely to be proximity and preadaptation of potential colonists for

regionally and locally variable depending on habitat local conditions, the sequence of events and conditions

conditions but also that the response in a specific that has occurred in key patches, and the specific

habitat unit may strongly depend on its spatial rela- local vagaries of juxtaposition of habitat patch types

tionship to other habitat patches in the ecosystem (~chlosser, 1991). Biotic recovery in such circum-

(Sheldon, 1988; Schlosser, 1991). For example, histori- stances may lag far behind apparent physical recov-

cal responses of fish populations and other biota may ery of local habitat patches (Zwick, 1992; Frissell

not reliably reflect future responses because the larg- 1993b; Doppelt et al., 1993). As a result, many appar-

er-scale context or habitat and metapopulation mosaic ently suitable habitat patches across the landscape

at the catchment level is changing (Figure 1). will remain unoccupied, leading the biologically naive


Ecosystem M~anagement and the Conservation ol’Aquatic Biodivc~. ity and Ecological Integrity

a bPatch Disturbance Refuge

Model Model

Figure 1. The Changing Biogeographic Context of Aquatic Ecosystem Management. In catchment a, degraded aquatic habitats(shaded) constitute isolated patches within a matrix of high-quality, richly-inhabited areas. Abundant, well-cbnnected populationssupply a steady supply of colonists (arrows represent colonizatinn vectors) to re-establish populations in disturbed areas. catchment b, high-quality habitats are isolated remnants in a matrix of disturbed and degraded habitat. Fragmented habitatislands serve as refugia for sensitive species and provide weak and localized or unidirectional sources of colonists to the degradedand relatively hostile matrix. Many refugia are sufficiently distant from others that little suqccssful exchange of individualsbetween populations occurs, but some migration still occurs between patches that are closely spaced.

to wrongly assume that habitat factors are not thealtered landscapes (Zwick, 1992; Frissell, 1993b;, Rie-cause of population declines,man et al., 1993). In Other words, continuity through

Unfortunately, for aquatic’ ecosystems in North time and space, of both particular habitats and partic-America (and most of the rest of the world), the frag-

ular populations, is increasingly important in frag-mentation scenario is probably a more realistic repre-mented and human-altered landscapes. If watershedsentation of the ecological status of most sensitiveanalysis is to become truly effective as a set of tools

species. Spatially informed and taxon-specific models,for biological conservation, it will have to be consider-

as yet largely undeveloped and untested, will be nec-ably .expanded to explicitly account for these kinds ofessary to understand and predict biotic responses inbiophysical and biogeographic relationships.

this kind of landscape. Such models, assuming theyAlthough the preceding discussion has stressed thecan be someday successfully developed, will be corn-

individuality of watershed responses to human distur-plex and highly site-specific in many of their predic-bance, there is a level of general understanding thattions. One of the few general rules of thumb thatcan be gained from retrospective analysis of physical

emerges from preliminary work in this vein is thatand biological histories. However, we suggest this

aintaining existing undisturbed habitat patchesunderstanding is best gained by a design that~pecially large or complex patches, or those withincludes comparative analysis of multiple watershedsnigh density or diversity of sensitive species) is criti-over time, in which some watersheds are heavilycal for maintaining native species biodiversity inimpacted by human activities and ~others are less or


Frisscll and Bayles

differently altered, or affected at different times and predictable, optimiz6d, and socially desirable outputsplaces. Spatially and temporally consistent patterns (Grumbine, 1994; Stanley, 1995; Frissell et al., ivin the response of key assemblages and populations press). Moreover, despite our well-demonstratecan reveal the presence of general, predictable effects inability to prescribe and forge institutional arrange-(McAllister and Peterman, 1992)- e.g., that increased ments capable of successfully implementing thesediment reduces survival of fall-spawning salmonids, principles and practice of integrated ecosystem man-or that summer water withdrawals reduce survival of agement over a sustained time frame and at suffi-specific species and age classes. We believe much ciently large spatial scales, would-be ecosystemmore powerful understanding of physical-biological managers have neglected to acknowledge and critical-interactions and their ecosystem management impli- ly analyze past institutional and policy failurescations can generally be gained from the integrated, (Grumbine, 1994; Underwood, 1995; Stanley, 1995).comparative analysis of multiple watersheds across a They say we need ecosystem management becauseregional landscape than from intensive, reductionistic public opinion has changed, neglecting the obviousanalysis of a single catchment. Such investigations point that public opinion has been shaped by themight better be called watersheds synthesis than glowing promises of past managers and by their clearwatershed analysis, and spectacular failure to deliver on such promises

(Frissell et al., in press).These fundamental limitations on our ability to

anticipate and optimize environmental outcomes inECOSYSTEM MANAGEMENT IN THE FACE OF ecosystem management are particularly striking in

UNCERTAINTY, IGNORANCE, AND RISK aquatic ecosystems, which are strongly linked and yetspatially and temporally removed from many of the

We do not intend to discourage or disdain the fragmented institutions, human activities, and natu-development of new and more sustainable approaches ral events that affect them. Like Bella and OVertonfor environmental management, or in the terms of (1972), Regier and Baskerville (1986), Ludwig et al.Regier and Baskerville (1986), sustainable redevelop- (1993), Stanley (1995), and many other eminent scien-ment. Our fundamental point is that we need new fists, we emphasize that no foreseeable science ormanagement, but to get it we must change our expec- management will eliminate the fundamental chal.rations of management. If ecosystem management is lenges and risks posed by uncertainties about futuresold with the promise of no net environmental ecosystem response to human actions, and humanimpacts, jobs for everybody, and restoration for every response to ecosystem changes..habitat and species, nothing has really changedexcept the jargon. Should we instead.choose to frameecosystem management as a consciously experimental

THE NEED FOR WATERSHED RESERVESendeavor with a largely uncertain outcome - anacknowledgment that we have been playing a losing ’ IN ECOSYSTEM MANAGEMENT

game and if we are not extremely careful with ourremaining natural resources, we stand to suffer envi- The concept of definition and establishment ofronmental and eventually economic check-mate - large-watershed reserves can provide several crucialthen perhaps we can indeed move forward to a new functions that are lacking in other, less spatially-perspective that provides a clear (if slim) chance for explicit approaches to eeosystem management. Anlong-term maintenance and restoration of our envi- ecosystem management plan without reserves is aronment and its aquatic resources, plan that fails to address what we now know about

Most philosophies and approaches for ecosystem ecosystems, the state of the environment, and ourmanagement put forward to date are limited (perhaps management capabilities (Stanley, 1995). No plan candoomed) by a failure to acknowledge and rationally eschew or gloss over these issues and still claim toaddress the overriding problems of uncertainty and provide a valid map to recovery or maintenance ofbio-ignorance about the mechanisms by which complex diversity, ecological integrity, and other ecosystemecosystems respond to human actions. They lack services. In a sense, we are arguing that the worldhumility and historical perspective about science and ’does not face an ecosystem problem; it faces a man-about our past failures in management. They still agement problem.implicitly subscribe to the scientifically discredited What would a watershed reserve system look like,illusion that humans are fully in control of an ecosys- and how would it help us cope with or avoid manage-temic machine and can foresee and manipulate all the ment problems? Such reserves would constitute a net-possible consequences of particular actions while work of the best-remaining examples of relativelydeliberately altering the ecosystem to produce only unaltered ecosystems and aquatic communities; in

WATER RESOURCES BULLETIN 236 ~

Ecosystem Management and the Conservation 0fAquatic Biodive~ity and Ecological Iotegrity

extensively-altered landscapes, these would need tobe supplemented or replaced by the least-disruptedecosystems that retain much of their ecological valueand hold good promise for relatively rapid and cost-efficlentrestoration. In Figure2wepresentabrief ~ M~ ]example, but we refer readers to several receiptsources for deeper discussion of these issues than weare able to provide here (e.g., Reeves and Sedell, 1992;Frissell, 1993b; Frissell et al., 1993; Doppelt et al, ,~.1993; Noss and Cooperrider, 1994). Such a reservenetwork should ideally encompass a regionally repre-sentative range of terrestrial and aquatic ecosystemtypes and natural successional conditions, and incor-porate areas that have especially high ecologicalintegrity or natural diversity, high incidence of rare orseriously declining aquatic and riparian species andassemblages, and relatively unimpaired natural-his-torical catchment-wide biophysical processes and dis-turbance regimes (Moyle and Sato, 1991; for e×amp]essee Henjum et al., 1994; Frissell et al., 1995).

Figure 2. Example of a Recommended Design ~’or an AquaticDiversity Reserve Network for the Swan River Basin, an Area of2,070 km2 in Northwest Montana, USA (a~ter Frissell et al., 1995).The figure shews critical watersheds (black tone) and river-lakecorridors and wetland complexes (line shaded). Critical watershedscontain relatively well-distributed populations or native fishes,~estricted distribution of non-native fishes, and limited fish stock.

,g history. Watersheds selected by biological criteria turned out to~e among those least-impacted by land use activities in the basin.

0


Frissell and Bayles

The reserve approach acknowledges there is much that of densen et al. (1996), could provide a usefuluncertainty about the success of future management framework for planning and evaluation of ecosystemactions and ensures that some large ecosystems will management activities in the landscape outside ofnot be directly exposed to new management manipu- special reserve areas as well.lations that are bound to have unanticipated and The concept of conservation reserves, particularlyunforeseeable consequences (Bella and Overton, 1972; encompassing whole catchments, offers a badly need-Ehrenfeld, 1991; Henjum et al., 1994; Stanley, ’1995). ed logic to more effectively and clearly link watershedWatershed reserves offer a fundamentally Conserva- analysis, adaptive monitoring, and decisions aboutrive hedge against uncertainties about the outcome of the planning and scheduling of human activities, withpast and future management in four ways. First, they the goal of reducing long-term uncertainties in ecosys-ensure we won’t make the same mistakes everywhere, tern management while simultaneously minimizingSecond, a network of such reserves could provide nec- its irreversible consequences to aqLmtic ecosystemessary and appropriately-scaled scientific controls for integrity and biodiversity. Contrary to Fitzsimmonsthe landscape-level experiment; that ecosystem man- (1996), who seems to assume that any landscape-wideagement constitutes. For example, such watersheds strategy for conservation would necessarily be imple-can be absolutely indispensable in distinguishing the mented by gestapo-like government control, weeffects of climate change from those of direct land- believe that conservation of natural resources and thescape alteration in ecosystem research and monitor- stewardship ethic it entails are not intrinsicallying. Third, from the aquatic point of" view, watershed threatening to human liberty or economies. Althoughreserves or aquatic diversity areas represent the best we of’ten forget to think about it (or choose to ignoreremaining places to focus restoration resources for the . it), at many levels the conservation of naturalnear-term, where the likelihood of physical and bio- resources in some way serves the interest of everylogical success is greatest, and where the greatest person. We know that other cultures have sustainablyshare of threatened biotic resources can be protected inhabited ecosystems for many generations withoutwith the limited resources that are available (Moyle evolving Biodiversity Police, and it is obvious thatand Sato, 1991; Frissell, 1993b; Doppelt et al., 1993). such tactics are not long tolerated in most societies.Finally, the less-disrupted land-aquatic ecosystems For years we have been discussing the concepts dis-within watershed reserves can serve as our best cussed in this paper in public forums, and most citi-remaining living models for the development of truly zens (but not all managers) roundly accept them restorative ecosystem management on more severely just plain common sense, a good, conservative, andaltered parts of the landscape (Frissell, 1993b; Eber- pragmatic basis to begin discussing cooperative man-sole et al., in press), agement across the landscape. Serious challenges

Perhaps the best (however imperfect) example remain, of course, in visualizing, formulating, imple-the formulation and attempted implementation (vir- menting, and evaluating actual landscape plans basedtually aborted by Congress in 1995) of a similar strat- on these principles, and this is where we should beegy at a regional scale we know of is the President’s investing the bulk of our creative energy and dwin-Forest Plan for the national forests in the range of the dling resources.northern spotted owl (FEMAT, 1993). Exciting propos- Ultimately, if ecosystem management attains itsals exist for bi-national redevelopment efforts in the goal of widespread ecological sustainability, thereGreat Lakes region (Regier and Baskerville, 1986; could be reduced need for maintaining discrete biodi-Steedman and Regier, 1987), and some large-scale versify reserves and for spatially focusing restorationconservation programs in progress in developing activities as hedges against the loss of ecosystemiccountries. Similarly spatially-stratified and conserva- functions and biodiversity in the remainder of thetive management strategies,, based on river-floodplain landscape. Whether this opportunity will come to passvalley segments and ,smaller-scale land-aquatic units depends critically on our actions today, but remainswithin watersheds (Warren, 1979; Frissell et al., 1986; for our grandchildren to see.Jensen et al., 1996), need to be developed for large-river ecosystems (Sparks, 1995) and for landscapeswhere the spatial extent of prior human development ACKNOWLEDGEMENTSmay preclude the establishment of functional whole-

We thank The Pacific Rivers Council and its funders for continu-watershed reserves (Moyle and Sato, 1991; Doppelt ing support that allowed development o[~this paper.et al., 1993; Frissell, 1993b; Moyle and Yoshiyama,1994). Smaller-scale refinement of the concept of spa-tial stratification of risk-taking in managementactions, coupled with ecological classifications such as


Ecosystem Management and the Conservation of Aquatic Biodiversity and Ecological Int%~’ity

LITERATURE CITED Hagans, D. K., W. E. Weaver, and M. A. Madej, 1986. Long-Termand Off-Site Effects of Logging and Erosion in the RedwoodAllan, J. D. and A. S. Flecker, 1993. Biodiversity Conservatiee inCreek Basin, Northern California. In papers presented at theRunning Waters. BioScience 43:32-43.American Geophysical Union Meeting on Cumulative Effects.Bella, D. A. and W. S. Overton, 1972. Environmental Planning andNational Council for Air and Stream Improvement TechnicalEcological Possibilities. Journal of the Sanitary EngineeringBulletin 490:38-66, New York, New York.Division, American Society of Civil Engineers 98(SA3):579-592.

tIeeky, R. E., R. W. Ncwbury, R. A. Bodaly, K. Patalas, and D. W.Bradford, D. F., F. Tabatabai, and D. M. Graber, 1993. Isolation ofRosenberg, 1984. Environmental Impact Prediction and Assess-Remaining Populations of the Native Frog, Rana muscosa, byment: The Southern Indian Lake Experience. Canadian JournalIntroduced Fishes in Sequoia and Kings Canyon Nationalof Fisheries and Aquatic Sciences 41:720-731.Parks, California. Conservation Biology 7:882-888.

Henjnm, M. G., J. R. Karl D. L. Bottom, D. A. Perry, J. C. Bednarz,Doppelt, B., M. Scurlock, C: Frisseil, and J. Karr, 1993. EnteringS.G. Wright, S. A. Beckwitt, and E. Beckwitt, 1994. Interimthe Watershed: A New Approach to Save America’s RiverProtection for Late-Successional Forests, Fisheries, and Water-Ecosystems. Island Press, Covelo, California.sheds, National Forests East of the Cascade Crest, Oregon andEbersole, J. L., W. J. Liss, and C. A. Frissell (in press). RestorationWashington. Eastside Forests Scientific Society Panel. Theof Stream Habitats in Managed Landscapes in- the WesternWildlife Society Technical Review 94-2, Bethesda, Maryland.USA: Restoration as Re-Expression of Habitat Capacity. Envi-

Jensen, Mark E., Patrick Bourgeren, Richard Everett, and Irisronmental Management.Goodman, 1966. Ecosystem Management: A Landscape EcologyEhrenfeld, D. W., 1991. The Management of Diversity: A Conserva-Perspective. Water Resources Bulletin 32(2):203-216.ties Paradox. In: Ecology, Economics, Ethics: the Broken Circle,

Karr, J. R., 1991. Ecoh)gical Integrity: A Long-Neglected Aspect ofF. H. Bormann and S. R. Kellert (Editors). Yale UniversityWater Resource Management. Ecological Applications 1:66-84.Press, New Haven, Connecticut, pp. 26-39.

Ludwig, D., R. Hilborn, and C. Waltcrs, 1993. Uncertainty,Fausch, K. D., C. L. Hawkes, and M. G. Parsons, 1988. Models thatResource Exploitation, and C( nservatmn: Lessons from Histms,.Predict Standing Crop of Stream Fish from Habitat Varia})]es:Science 260:17,36.1950-85. USDA Forest Service General TechnicAl Report PNW-

McA]listcr, M. K. and R. J. Peterman, 1992. Experimental DesignGTR-213.in the Management of Fisheries: A Review. North AmericanFEMAT, 1993. Forest Ecosystem Management: An Ecological, Eco-Journal of Fisheries Management 12:1-18.nomic, and Social Assessment. Report of the Forest Ecosystem

NcNab, J., 1983. Wildlife Management as Scientific Experimenta.~Management Assessment Team (a federal agency consortium),

ties. Wildlife Society Bulletin 11:397.401.Portland, Oregon.Montgomery, D. R., G. E. Grant, and K Sullivan, 1995. WatershedFitzsimmons, Allan K., 1996. Sound Policy or Smoke and Mirrors:

Analysis as a Framewm-k for Implementing Ecosystem Manage-Does Ecosystem Management Make Sense? Water Resourcesmeat. Water Resources Bulletin 31:369-386.Bulletin, 32(2):217-227.

Moyle, P. B. and G. M. Sate, 1991. On the Design of Preserves toFrissell, C. A., 1993a. Topology of Extinction and Endangerment ofProtect Native Fishes. In: Battle Against Extinction: NativeNative Fishes in the Pacific Northwest and California (U.S.A.).Fish Management in the American West, W. L: Minckley and

Conservation Biology 7:342-354.d.E. Deacan (Editors). University of Arizona Press, Tucson, Ari-Frissell, C. A., 1993b. A New Strategy for Watershed Restorationzona, pp. 155-170.and Recovery of Pacific Salmon on the Pacific Northwest. Report

Merle, P. B. and R. M. Yoshiyama, 1994. Protection of Biodiversityprepared for The Pacific Rivers Council, Eugene, Oregon.in California: A Five-Tiered Approach. Fisheries 19(2):6-18.Frisse]l, C. A., J. Doskocil, J. T. Gangemi, and J. A. Stanford, 1995.

Nchlscn, W., J. A. Lichatowich, and J. E. Williams, 1991. Pacific .Identifying Priority Areas for Protection and Restoration ofSalmon at the Crossroads: Stocks at Risk from California, Ore-Aquatic Biodiversity: A Case Study in the Swan River Basin,gon, Idaho, and Washington. Fisheries 16(2):4-21.Montana, USA. Flathead Lake Biological Station, The Universi-

Noss, R. F. and A. Y. Coopcrrider, 1994. Saving Nature’s Legacy:ty of Montana, Open File Report 136-95, Polson, Montana.Protecting and Restoring Biodiversity. Island Press, Washing-Frissell, C. A., W. J. Liss, and D. Bayles, 1993. An Integrated, Bio-ton, D.C.

physical Strategy for Ecological Restoration of Large Water- Reeves, G. H. and J. R. Sedell, 1992. An Ecosystem Approach to thesheds. In: Changing Roles in Water Resources Management and

Conservation and Management of Freshwater Habitat forPolicy, N. E. Spangenberg and D. F. Ports (Editors). AmericanAnadromous Salmonids in the Pacific Northwest. TransactionsWater Resources Association, Herndon, Virginia, pp. 449-456.of the 57th North American Wildlife and Natural ResourcesFrissell, C. A., W. J. Liss, R. E. Gresswell, R. K. Nawa, and J.L.Conference, pp. 408-415.Ebersole (in press). A Resource in Crisis: Changing the Measure

Regier, H. A. and G. L. Baskcrvi[le, 1986. Sustainable Redevelop-of Salmon Management. In: Pacific Salmon and their Ecosys-meat of Regional Ecosystems Degraded by Exploitive Develop-terns: Status and Future Options, D. J. Stouder, P. A. Bisson,meat. Dr: Sustainable Devclepment of the Biosphere. W. C.and R. J. Naiman (Editors). Chapman and Hall, New York, NewClarke and R. E. Munn (Editors). Cambridge University Press,York.Cambridge, Un ted Kingdom pp. 75-101.Frissell, C. A., W. J. Liss, C. E. Warren, and M. D. Hurley, 1986. A

Rieman, B., D. Lee, J. McIntyre, K. Overton, and R. Thurow, 1993.Herarchical Approach to Stream Habitat Classification: Viewing

Consideration of Extinction Risks for Salmonids. USDA ForestStreams in a Watershed Context. Environmental ManagementService, Fish Habitat Relationships Tcchn.ical Bulletin 14, Six

10:199-214.Rivers National Forest, Eureka, California.Frissell, C. A., R. K. Nawa, and R. Noss, 1992. Is There Any Con-

Rhodes, J. J., D. A. McCullough, and F. A. Espinosa, Jr., 1994. Aservation Biology in "New Perspectives?": A Response to Sa]-Coarse Screening Process for Evaluation of the Effects of Landwasser. Conservation Biology 6:461-464.Management Activities nn Salmon Spawning and Rearing Habi-rumblne, R. E., 1994. What is Ecosystem Management? Censer.tat in ESA Consultations. Columbia River Inter-Tribal Fishration Biology 8:27-38.Commission Technical Report 94-4, Portland, Oregon.Hall, C. A. S., 1988. An Assessment of Several of the Historically

Roth, N. E., J. D. Allan, and D. L. Erickson (in press). LnndscapeMost Influential Theoretical Models Used in Ecology and of theInfluences On Stream Biotic Integrity Assessed at Multiple Spa-

Data Used in Their Support. Eoological Modeling 43:5-31.tial Scales. LandscapeEcology.


Frissell and Bayles

Salwasser, H., 1992. From New Perspectives to Ecosystem Manage-ment: Response to Frisscll et al., and Lawrence and Murphy.Conservation Biology 6:469-472.

Schlosser, I. J., 1991. Stream Fish Ecology: A Landscape Perspec-tive. BioScience 41:704-712.

Sheldon, A. L., 1988. Conservation of Stream Fishes: Patterns ofDiversity, Rarity, and Risk. Conservation Biolegy 2:149-156.

Soul6, M. E., 1991. Conservation: Tactics for a Constant Crisis. Sci-ence 253:744-750.

Sparks, R. E., 1995. Need I’or Ecosystem Management of LargeRivers and Their Floodplains. BioScience 45:168-182.

Stanley, T. R., 1995. Ecosystem Management and the Arrogance ofHuman ism. Conservation Biology 9:255-262.

Steedman, R. J. and H. A. Regier, 1987. Ecosystem Science for theGreat Lakes: Perspectives on Degradative and RehabilitativeTransformations Canadian Journal of Fisheries and AquaticSciences 44(Supplement 2):95-103.

Underwood, A. J., 1995. Ecological Research and (and researchinto) Environmental Management. Ecological Applications5:232-247.

Warren, C. E., 1979. Toward Classification and Rationale fi~r Water-shed Management and Stream Protection. U. S. Envil"onmentalProtection Agency Ecnlogical Researcb Series EPA-600/3-79-059.

Williams, J. E. et al., 1989. Fishes of Nerth America Endaagcrcd,Threatened, or of Special Concern. Fisheries 14(6):2-20.

Wissmar, R. C., J. E. Smith, B. A. McIntosh, H. W. Li, G. It. Reeves,and J. R Sedcll, 1994. A History of Resource Use and Distur-bance in Rivcrine Basins of Eastern Oregon and Washington(early 1800s-1900s). Northwes~ Science 68:1-35.

Yount, J. D. and G. J. Niemi, 1990. Recovery of Lotic Communitiesfrom Disturbance - A Narrative Review of Case Studies. Envi-ronmental Management 14:547-569.

Ziemer, R. R., J Lewis, R. M. Rice, and T. E. Lisle, 1991. Modelingthe Cumulative Watershed Effects of Forest ManagementStrategies. Journal of Envlmnmental Quality 20:36-42.

Zwick, P., 1992. Stream Habitat Fragmentation - A Threat to Bie-diversity. Biodivcrsity and Conservation 1:80:97.


ecosystem management and the conservation of aquatic biodiversity and

Documents