#enhydralutris | Explore Tumblr posts and blogs

mcm-curiosity · 4 years ago

Text

Strongly Interacting Species:Conservation Policy,Management, and Ethics

MICHAEL E. SOULÉ, JAMES A. ESTES, BRIAN MILLER, AND DOUGLAS L. HONNOLD

Any legislation or other policy instrument based onempirical science is prone to senescence. Consider theUS Endangered Species Act (ESA) of 1973. This was the firstfederal statute to grant de facto existence rights to species ofplants and animals (Varner 1987) and to impose binding, en-forceable duties on government agencies and private citi-zens to protect imperiled species. The ESA was based on thebest science of the time (Bean and Rowland 1997), and it stillremains in the vanguard of worldwide species protection ef-forts. However, the ESA, like other environmental laws fromthe 1970s, was enacted before conservation biology existed asa discipline, before the field of population viability analysisexisted, before ecologists understood many of the ecosys-tem consequences of species interactions and community dy-namics (Terborgh et al. 1999), and before the spatial andtemporal scale–related complexities of effective protection ofvulnerable species were widely understood (Soulé et al. 2003).Although the ESA was revised and reauthorized in 1988, itdoes not reflect many advances in population biology andcommunity ecology.

Among the scientific anachronisms in this law is the absenceof specific reference to species interactions that contribute sub-stantially to the maintenance of ecological and species diversity.Not only has the understanding of interaction webs ad-vanced (Menge 1995), but it is now widely understood thatthe disappearance of a strongly interactive species can lead toprofound changes in ecosystem composition, structure, anddiversity (Soulé and Terborgh 1999, Terborgh et al. 1999,Oksanen and Oksanen 2000, Schmitz et al. 2000, Soulé et al.2003). For instance, decimation of great whales by industrialwhaling substantially altered krill-consumer dynamics in theSouthern Ocean, and whaling has been proposed as the maincause of a recent megafaunal collapse in the North PacificOcean (Springer et al. 2003). In addition, the disruption offruit dispersal and seed-set patterns following early Holocenemegafaunal extinctions fundamentally altered the speciescomposition of neotropical forests (Janzen and Martin 1982),and the functional dynamics of coastal marine ecosystemsworldwide have been grossly altered by overfishing of largeherbivores and predators (Jackson et al. 2001).The functional extinction of species interactions often oc-curs well before the species themselves have completely dis-appeared. In the oceans, many large, interactive species persistonly as rare adults, or as small or juvenile individuals that donot interact like large adults, qualitatively or quantitatively.

On land, many large animals and other strongly interactivespecies are completely missing from vast areas that they oc-cupied a century or two ago (Laliberte and Ripple 2004). Glob-ally, many, if not most, large-bodied, strongly interactingspecies are increasingly rare, even if they persist in parts of theirformer range. A reasonable hypothesis is that ecosystemsthat have lost one or more strongly interactive species are des-tined to undergo profound degradation and simplification over time.

Nevertheless, most conservation laws, including the ESA,fail to reflect the effects of widespread ecological disappear-ances of strongly interacting species and the resulting ecosys-tem perturbations. For example, the current criteria forrecovery of endangered mammal species under the ESA gen-erally ignore interspecies interactions altogether (Soulé et al.2003), emphasizing short-term, single-species, demographicviability in only a few circumscribed areas. Indeed, manycurrent recovery plans, at least for mammals, call for no in-crease in numbers of individuals, numbers of populations, orgeographic range (Tear et al. 1995; but see USFWS 1998,Jennings 1999). Here we propose that population densities of strongly in-teractive species must not be permitted to fall below thresh-olds for ecological effectiveness, and that the geographicranges of such species should be as large as possible (Conner1988, Soulé et al. 2003). Before this proposal can be imple-mented, however, two issues must be clarified: (1) the defi-nition of strongly interactive species and (2) the achievementof ecologically effective densities of such species.

What are strongly interactive species?

The idea that some species interact more strongly than oth-ers is not new. Paine (1969) first used the term “keystonespecies” for particularly strong interactors: those, for exam-ple, whose activities maintain species and habitat diversity andwhose effects are disproportionate to their abundance (Kotliaret al. 1999). It is worth noting that Paine’s idea, one of the mostinfluential in all of modern ecology, is fundamentally a the-ory of species diversity—that the presence or absence of oneor several key species influences the distribution and abun-dance of many other species. Ecologists recognize, however,that the keystone designation artificially dichotomizes speciesinto groups such as “strongly interactive” (or keystone) and“non-strongly interactive” (Mills et al. 1993). Though such du-alisms have limited utility in science, this particular one is help-ful in education and advocacy.

Species that are relatively interactive have been catego-rized according to the kind of ecological interaction that ismost evident. Among these interactions are habitat enrich-ment, mutualisms, predation,and competition.Species whoseactivities affect and enhance physical or biological habitatstructure have been referred to as “ecological engineers”(Jones et al. 1994). Ecological engineers significantly modifytheir habitat in ways that increase local ��species diversity.Beavers (Castor canadensis), for instance, create wetlands bybuilding dams in streams. Other examples of ecological en- gineering include mound building by termites, burrowing andgrazing by prairie dogs (Cynomysspp.), and habitat conver-sion by elephants (Loxodonta africana) and bison (Bison bi-son) (Naiman et al. 1988, Owen-Smith 1988, Detling 1998,Kotliar et al. 1999).Mutualist species, by virtue of their interactions, can alsomaintain species diversity. An example is the relationshipbetween the whitebark pine (Pinus albicaulus) and Clark’s nut-cracker (Nucifraga columbiana). Clark’s nutcracker is stronglydependent on the seeds of the whitebark pine, and the pinedepends on the nutcracker for the dispersal of its seeds intocaches. These seed caches are also a major food source for bothsmall vertebrates and grizzly bears (Ursus arctos) in theGreater Yellowstone ecosystem (Mattson et al. 1992). Terborgh and colleagues (1999) describe how the loss ofapex mammalian predators can precipitate ecological chainreactions that lead to profound degradation and species loss.Although top-down forcing through three or more trophiclevels has been demonstrated for nonvertebrate taxa (Stronget al. 1996, Terborgh et al. 2001), space constraints and im-mediate policy relevance preclude a detailed review here.Many predator-mediated chains of reaction have been de-scribed or postulated (Estes and Palmisano 1974, Pace et al.1999, Terborgh et al. 2001); some of these (ecological cascades)are summarized in figure 1 and further elaborated in figure2. Figure 2a illustrates the familiar case of gray wolves (Ca-nis lupus) in Yellowstone, representing the scenario in whichthe extirpation of a large carnivore leads to the ecological re-lease of large terrestrial ungulates and other herbivores, caus-ing changes in vegetation structure, species composition,and diversity.

Crooks and Soulé (1999) demonstrated the behavioral re-lease of mesopredators in patches of coastal sage scrub andchaparral in southern California, where the local absence ofcoyotes (Canis latrans) led to an increase in the activity of thehouse cat (Felis catus), in turn causing reductions of native,scrub-requiring bird species (figure 2b). An impressive caseof competitive release (Paine 1966) was described by Henkeand Bryant (1999) and is illustrated in figure 2c. They doc-umented a reduction of rodent diversity from 12 species tojust 1 as a result of coyote removal; the survivor was thecompetitively dominant kangaroo rat,Dipodomys ordii.Thefourth example (figure 2d)—the case of sea otters (Enhydralutris) and kelp forest—is described below.

The fifth example (figure 2e) of a predator-mediated eco-logical cascade is hypothesized to have begun with the deci-mation of the great whales by industrial whaling followingWorld War II. Springer and colleagues (2003) suggest that aseries of ecological extinction events affecting pinnipeds andsea otters in the northern Pacific Ocean and the Bering Seawas initiated when killer whales (Orcinus orca), followingthe effective disappearance of large baleen whales, expandedtheir diets. Though baleen whales are themselves carnivores,they are not considered to be apex predators because of therelatively small size of their prey and because they are preyedon by killer whales.

We know little about the distribution of interaction strengthamong species in most ecosystems. Nor do we know themorphological, physiological, behavioral, or ecological cor-relates of strong interactivity. Paine (1992) showed experi-mentally that the interaction strengths of seemingly similarspecies can vary substantially; he also argued that mammalsare especially strong interactors in many terrestrial ecosystems(Paine 2000). One of us (M. E. S.) surveyed all mammalspecies listed as threatened or endangered under the ESAfor which recovery plans are written (about 44 species orsubspecies). It appears that nearly half of these vulnerablemammals are relatively interactive, according to the criterialisted below, though this estimate may be low because the ques-tion is unstudied for many of the species (Soulé et al. 2003).Sala and Graham (2002) provide the most comprehensiveanalysis to date on species-specific variation in inter actionstrength. They estimate that roughly half of the macroinver-tebrate herbivore species in kelp forest ecosystems are stronginteractors. Based on limited information, therefore, it appearsthat a significant proportion of invertebrate and vertebratespecies are sufficiently interactive to warrant attention if re-covery criteria are an issue. Parenthetically, there are excep-tions to the view that strong interactors are universallybeneficial. Invasive exotic species and some native carnivores,particularly in highly perturbed ecosystems, can exacerbatemanagement problems. For example, coyotes can devastatesmaller, endangered predators such as captive-bred black-footed ferrets (Mustela nigripes), particularly if the coyotes areuncontrolled by wolves and if their prey occur in reduced, dis-turbed, or fragmented habitats (Miller et al. 1996).

The question of how interactivity is dis-tributed in ecosystems has yet another di-mension, namely variability within species.Like all ecological variables, interactionstrength is contingent on place, time, andhistory (Power et al. 1996). Just as it would befutile to assign species-wide, fixed values forage-specific fecundity, population growthrate, coefficients of competition, or othercontext-dependent variables, it would be un-reasonable to assign a fixed value for inter-activity to a widespread species.

Arguably, the related goals of (a) catego-rizing the kinds of interspecific interactionsand (b) assigning species to these categoriestrivialize the variability of species and envi-ronments in space and time. Interactivity isobviously a complex, context-dependent vari-able, and no species trait or feature is uni-versally associated with it across all taxonomicgroups and ecosystems. Nevertheless, the ESAcontains wording that justifies attending tospecies interactions: “The purposes of thisAct [the ESA] are to provide a means wherebythe ecosystems upon which threatened andendangered species depend may be con-served” (16 U.S.C. § 1531[b]). This leaves us with a practicalquestion: How can agencies and managers, in the face of thisuncertainty and variability, determine whether a vulnerablespecies in a particular locality or region is sufficiently inter-active to warrant special consideration with regard to recov-ery goals?

Guidelines for assessing interactivity

Operationally, a given species should receive special attentionfor recovery—beyond mere demographic viability—if itsabsence or unusual rarity causes cascading, dissipative trans-formations in ecosystems, including alterations or simplifi-cations in ecological structure, function, or composition.The following questions may assist in determining whetherthere are grounds to warrant the creation of appropriatemanagement prescriptions and actions that guarantee itsecological effectiveness.

Does the absence or decrease in abundance of the specieslead directly or indirectly to a reduction in local species di-versity? For example, the absence of coyotes from arid ecosys-tems can lead to a reduction in bird species diversity viamesopredator release (Crooks and Soulé 1999) or to a re-duction in rodent species diversity via competitive exclusion(Henke and Bryant 1999), as noted above.

Does the absence, decrease in abundance, or range con-traction of the species directly or indirectly reduce repro-duction or recruitment of other species? For example, thenumber of forest tree species that successfully reproduced on islands in a Venezuelan reservoir lacking large predatorsdropped from about 65 to about 10 because of a super-abundance of herbivores (Terborgh et al. 2001). Likewise, un-gulate herbivory prevented aspen (Populus tremuloides) clonesfrom recruiting sprouts into the overstory after extirpation ofwolves in the northern range of Yellowstone National Park(Romme et al. 1995, Ripple and Larsen 2000, Ripple andBeschta 2004).

Does the absence or decrease in abundance of the specieslead directly or indirectly to a change in habitat structureor composition of ecosystems? For example, excessive elk(Cervus elaphus) herbivory on willow (Salixspp.) in the ab-sence of wolves in Rocky Mountain National Park (Peinettiet al. 2002) and Yellowstone National Park (Ripple and Beschta2004) was apparently the major factor in the disappearanceof beaver and associated wetlands.

Does the absence or decrease in abundance of the specieslead directly or indirectly to a change in productivity or nu-trient dynamics in or between ecosystems? For example,prairie dog colonies shape nutrient cycling, soil chemistry, soilporosity, and the productivity and nutrient content of vege-tation through their burrowing and grazing activities (Whickerand Detling 1993, Kotliar et al. 1999, Miller et al. 2000), andsea otters strongly influence algal productivity (Duggins et al.1989) and food resource availability to herbivores (Konarand Estes 2003).

Does the absence or decrease in abundance of the specieschange an important ecological process in the system? For example, beavers have a profound effect on stream dy-namics, water tables, flooding, and the extent of wetlands(Naiman et al. 1988).

Does the absence or decrease in abundance of the speciesreduce the resilience of the system to disturbances such asfire, drought, flood, or exotic species? For example, the ex-tirpation of the dingo (Canis lupus dingo) in some regions ofAustralia indirectly degrades habitat quality because dingoesprey effectively on exotic rabbits (Oryctolagus cuniculus), redkangaroos (Macropus rufus), and other herbivores (New-some 2001). In addition, dingoes may benefit native fauna,including small, endangered marsupials, by reducing popu-lation densities of the introduced red fox (Vulpes vulpes)(Newsome 2001), a major predator of small animals (O’Neill2002).

These questions cannot eliminate the need for informedjudgment, because interactivity of species is a multidimen-sional continuum, not a simple dichotomy. In addition, theinteraction strength of species is usually not susceptible to rig-orous empirical tests, in part because many appropriate ex-periments would be manipulative (involving the removal ofspecies), long-term, and geographically extensive. With sucha small portion of nature protected, it is difficult to justify experimental removal of a putatively critical species to provea point. There are, however, a number of powerful approachesthat can often be used to make inferences about interactionstrengths. Recovery of ecosystems following the reappearanceof species is one such approach that has been used effec-tively to establish that predators such as gray wolves and seaotters are strongly interactive (see the cases described be-low). Interaction strength has been modeled on the basis ofdemographic and energetic parameters (Williams et al. 2004),even where data are limited. Ecological reconstructions basedon historical records (Jackson et al. 2001), in conjunction withthe comparative approach, provide yet another powerfulmeans of assessing the ecological importance of species.

The estimation of ecologically effective densities

If persistence of species diversity is a management objective,it is essential to consider the densities or population levels thatmaintain interaction effectiveness rather than mere persistenceat minimal numbers. Once it is determined that a species hasrelatively strong interspecies interactions, the proper man-agement of such a species may require the estimation of theminimum threshold of ecological effectiveness. We define anecologically effective density as the population level that pre-vents undesired changes in a defined ecological setting. Asstated above, however, the estimation of effective density isstrongly contextual, depending on locality, season, produc-tivity, and other variables that fluctuate spatially and tem-porally (Estes and Duggins 1995, Soulé et al. 2003). Althougha challenge, this problem may not be more intractable thanthe estimation of population viability. For example, many ofthe relevant parameters in population viability analysis, in-cluding age-specific fecundity and mortality, are similarlysensitive to local conditions. To illustrate some of the factorsthat must be considered in the estimation of ecologically ef-fective densities, we present three examples of strongly in-teractive genera or species: the sea otter, the gray wolf, and theprairie dog.

The sea otter. Abundant sea otter populations inhabitedcoastal waters of the North Pacific Ocean and southernBering Sea throughout most of the Pleistocene and Holocene,but were reduced to a few remnant colonies by the maritimefur trade of the 18th and 19th centuries. Recovery followingthe fur trade was spatially and temporally asynchronous,thus providing contrasts between otherwise similar habitatswith and without sea otters. These contrasts demonstrate astrong limiting influence of sea otters on their most impor-tant prey, kelp-consuming sea urchins (Strongylocentrotusspp.). Thus, lush kelp forests abound where sea otters are abun-dant; where sea otters are absent, the habitat is typically de-forested by hyperabundant sea urchins. Because kelp forestsare highly productive (Duggins et al. 1989), provide habitatfor other coastal species (Dayton 1985), and attenuate watermovements (Jackson and Winant 1983), sea otters exert far-reaching influences on many other species (Estes 1996).Without sea otters, some of these kelp-dependent species decline or disappear, while others, including urchins, eruptto high levels. The ecologically effective population for sea ot-ters, though regionally variable, is always much larger thanminimum viable population sizes based on demography,and in some instances is near the environmental carrying ca-pacity (Estes and Duggins 1995).Geographic variation in the behavior of predators, com-petitors, and prey will also affect the population densitythreshold for ecological effectiveness. For example, the den-sity of sea otters that is effective in suppressing sea urchins dif-fers between sites, because the demography and dispersal ofsea urchins vary geographically. In the Aleutian Islands, whereurchin recruitment is frequent and strong, a higher densityof otters is needed to suppress the urchins and prevent kelpdeforestation than in southeast Alaska, where urchin re-cruitment is weak and episodic, and where just a few ottersare enough to maintain the kelp ecosystem (Estes and Dug-gins 1995). In summary, the estimation of effective densities of sea ot-ters for preventing kelp deforestation depends, among otherthings, on whether the state of the system is kelp dominatedor deforested, on the recruitment dynamics of urchins tothe kelp beds, on whether the substrate is dominated byrocks or mud, and on the mortality rate of otters (see Souléet al. 2003). For these reasons, the ecologically effective den-sities of otters can vary by an order of magnitude, but in allsituations observed, otters eventually attain such densities ifthey are not harassed by human beings or preyed on by killerwhales.

...

[Graphs Removed]

...

The gray wolf. Large areas of the United States, including mostof the East Coast and Midwest, now lack wolves and other largecarnivores, the result of a century of eradication on behalf oflivestock growers, hunters, and other interest groups thatbenefit from the absence of wolf predation on ungulates.Populations of white-tailed deer (Odocoileus virginianus),elk, and moose (Alces alces) have increased both in numbersand in per capita consumption (Soulé et al. 2003), and theseincreases are frequently attributed, at least in part, to the ab-sence of wolves (Messier 1994, Crête 1999). Among the manyharmful consequences of wolf eradication have been in-creased costs for agricultural producers in the Midwest andEast, the widespread degradation of forests and other ecosys-tems, and the decline of many species of plants favored by un-gulates (Rooney et al. 2004). As noted above, aspenrecruitment failed for 80 years in large parts of YellowstoneNational Park, reflecting numerical and behavioral release ofelk subsequent to wolf eradication. Excessive browsing by elkalso affected recruitment of riparian cottonwoods and willows(Beschta 2003), causing the local disappearance of beaverwetlands. These effects are being reversed in Yellowstone Na-tional Park since wolves were reintroduced (starting in 1995),and signs of ecological effectiveness were noted before wolvesreached their current abundance of about 200 (Ripple and Beschta 2004). A similar trend has been observed in GrandTeton National Park, where a decrease in Neotropical migrant bird diversity has been attributed to overbrowsing by moosein riparian willow communities in the absence of wolf pre-dation or sport hunting of moose (Berger et al. 2001). It isnoteworthy that the recovery of willows in northern Yellow-stone National Park is particularly striking in areas where thetopography facilitates capture of elk by wolves (Ripple andBeschta 2003). Several factors affect wolves’ ecologically ef-fective population density. It is lower where hunters can sup-press ungulate numbers; where wolves coexist with otherlarge carnivores, such as bears and cougars; or where deep win-ter snow or periodically severe storms facilitate capture ofprey—for example, El Niño versus La Niña years (Schmitz etal. 2003). We grant that predators do not always control largeherbivores, but given alternative prey, multiple carnivorespecies, or appropriate habitat, wolves can often control su-perabundant ungulates (Soulé et al. 2003).

The prairie dog. A century ago, five species of prairie dog livedin a shifting mosaic of colonies that covered more than40,000,000 hectares (ha) on the Great Plains. By 1960, prairiedog area had declined to about 600,000 ha (Marsh 1984),largely because of poisoning campaigns, land conversion,and the introduction of plague (Yersinia pestis). This is a de-cline of 98 percent, and the remaining colonies are smalland isolated. Prairie dogs are a valuable food for many speciesof predators. In addition, prairie dogs decrease densities ofwoody shrubs and increase densities of grasses and forbs, thuscreating conditions that large grazers prefer. Prairie dog ac-tivities also increase plant productivity, soil nitrogen, nutri-ent cycling, and digestibility of grasses and forbs (Whicker andDetling 1993, Detling 1998). Their burrowing activity changessoil chemistry; increases soil porosity, soil turnover, and the organic content of soil; and enhances the dimensionality ofthe habitat for many other species (Whicker and Detling1993, Outwater 1996). Some species of plants, invertebrates,and vertebrates benefit from prairie dog activities, whileother species benefit from the areas outside of the colony(Kotliar et al. 1999). These effects differ among prairie dogspecies. Furthermore, prairie dogs, unfenced bison, and fireinteracted closely on the midgrass prairies, although thattriad may not have been as tightly associated on the drought-driven shortgrass prairies or the semidesert grasslands andshrublands. Estimating ecologically effective densities of prairie dogsis complicated by the introduction of plague. Plague reducesnumbers and changes the temporal and spatial characteris-tics of the historic “shifting mosaic” between prairie dogcolonies and grasslands. Despite those ambiguities, it is clearthat ecologically effective densities of prairie dogs are farhigher than the densities required for population persistence(Miller et al. 2000). As an example, 762 prairie dogs may berequired to support each female black-footed ferret and heroffspring (Biggins et al. 1993). Thus, conservative recoverygoals that consider only population viability could maintainprairie dogs without providing sufficient resources for ferrets.

...

[Section Removed]

...

Conclusions

Mitigating the current anthropogenic mass extinction will re-quire a scientifically rigorous and ecologically comprehensivegrasp of its drivers. Among these is the increasing rarity of in-teractive species, such as plants that provide critical resources(Terborgh 1986), insect pollinators (Dobson et al. 1999),habitat modifiers (Soulé et al. 2003), coral reef herbivores(Jackson et al. 2001), and carnivores, both marine and ter-restrial (Terborgh et al. 1999). The activities of relatively in-teractive species are disproportionately significant for thesurvival of native species and ecosystems.It is essential, therefore, that conservation practitioners,whether governmental or nongovernmental, adopt an eco-logical view that ensures the persistence of interactive speciesat ecologically effective population densities and maximalspatial occurrence (Soulé et al. 2003). In particular, we believethat natural-resource policymakers and wildlands managersshould determine whether the rarity or absence (Hughes etal. 2000) of a species in a region can be expected to trigger eco-logical degradation, including the disappearances of nativespecies and other elements of biodiversity. Even though interactivity is a quantitative variable, man-agers may be forced to make binary determinations, such aswhether to treat a particular species as strongly interactivewhen formulating recovery goals. Notwithstanding the arbi-trariness of such decisions, a commonsense approach will of-ten suffice. One should assume, for instance, that a substantialreduction of tree species that produce mast or invite cavity for-mation, or of apex predators and many large herbivores—suchas wolves, coyotes, sea otters, killer whales, sharks, predatoryfreshwater fish, and large, predatory or algae-eating reeffish—will trigger cascades of ecological degradation andspecies loss (Terborgh et al. 1999, Jackson et al. 2001, Souléet al. 2003, Springer et al. 2003). Other situations may requireliterature reviews or detailed research to ascertain whether aparticular species in a particular place fulfills any of the cri-teria for relatively strong interactivity given above.A conundrum for managers is that the ecological effec-tiveness of strongly interacting species is not specifically ad-dressed in current laws and policies dealing with biodiversityprotection and management. Until this perspective has beencodified in such laws, conservationists need to consider howbest to provide for such species and the processes they me-diate in accord with the intent of these laws. Population via-bility analyses and conservative recovery goals are aninadequate regulatory context for strongly interacting species.We now understand that the biodiversity of ecosystems willdegrade unless the interactions of species are maintained inas many regions as feasible, particularly those areas within thehistoric range. This more holistic, contemporary view re-quires that strongly interactive species receive special atten-tion to assure that they are well distributed and abundant, aposition consistent with an opinion of the US Ninth CircuitCourt of Appeals (Defenders of Wildlife v.Norton,258 F.3d1136 [2001]). Such a geographic criterion for recovery wouldrequire more than scattered or refugial representation ofsuch species. Ecological function and diversity cannot beconserved in a region by maintaining interactive species in onlya few protected areas. Rather, it is essential that strongly in-teractive species be distributed as broadly as possible and beprotected within well-distributed, secure areas. Applying thisguideline to the wolf in the United States, for example, wouldmean that effective populations should be restored and pro-tected in the Northeast, the Pacific Northwest, the GreatBasin, the Colorado Plateau, the Southwest, and the south-ern Rockies. Moreover, if the current trend of decreasingsport hunting and the spread of chronic wasting disease in deerand elk continue, the pressure to reinstitute natural controlmechanisms will surely increase.The critical roles of interspecies interactions are rarelyconsidered in recovery planning. For example, the US Fish andWildlife Service (USFWS) admits that its goals for wolf re-covery are “somewhat conservative...and should be consideredminimal” (68 Fed. Reg. 15817 [2003]). A recent decision byUSFWS (68 Fed. Reg. 15821 [2003]) states that USFWS is notrequired to restore a species across its available habitat. Thedecision would limit wolf protection to about 5 percent of itshistorical range in the lower 48 states. Similarly, the multistateconservation plan for black-tailed prairie dogs (Luce 2003) sets a 10-year recovery goal for black-tailed prairie dogs(Cynomys ludovicianus) at about 2.5 percent of their histor-ical area, essentially the status quo.

We believe that such conservative recommendations are notbased on current ecological knowledge about the signifi-cance of species interactions. Moreover, minimalist distrib-utional and temporal goals constitute a trivialization of theterm “recovery” as it is used in the ESA. In other words, “re-covery,” at least for mammals, is typically used to mean thepersistence of only a few populations in a limited area for afew generations.Notwithstanding current policies, most natural-resourceand environmental laws require that federal agencies considernew scientific knowledge. Indeed, the ESA’s own mandate isto use “the best scientific and commercial data available” (16U.S.C. § 1533[b][1][A]). Moreover, implementing regulationsfor the National Environmental Protection Act of 1969 requirethat federal agencies disclose and consider “cumulative im-pacts” and the anticipated environmental impacts of proposedfederal agency actions (40 C.F.R. § 1500 et seq. 1995). Any ar-tificially induced reduction in abundance of a strongly in-teractive species, therefore, must be considered in theseenvironmental analyses. In addition, regulations of the Na-tional Forest Management Act of 1976 require that nationalforest plans “provide for the diversity of plant and animal com-munities” and that “such diversity shall be considered through-out the process” (36 C.F.R. § 219.3).Since the recognition of conservation biology as a discipline(Soulé 1985), its practitioners have tacitly assumed that con-servation biologists are “physicians to nature.” Indeed, thereare many parallels between conservation biology and thefields of medicine and public health—disciplines infusedwith morality. Following the Hippocratic principle of doingthe least harm and the most good for patients and the pub-lic, physicians and public health officers are obligated, we think,to consider using new therapies and prophylaxes stemmingfrom peer-reviewed research, even before such practices aregenerally adopted in canonical documents such as textbooks.We propose, therefore, that conservation practitioners, whetherin a public or private (nongovernmental) employ, are simi-larly obligated to apply new biological knowledge in theirwork. Such a doctrine of “best conservation practices basedon the best science” is tantamount to an ethical obligation ofbiologists to adopt a higher standard for management thanis mandated by existing statutes and regulations, if the evidencewarrants it. Environmental codes build the legal and ethicalfoundation of conservation practice, but the best science ofthe day creates the walls and ceiling.In practice, policymakers and managers already haveenough flexibility to implement new knowledge while still ad-hering to relevant statutes and policies, though the exerciseof this doctrine may be inhibited by monetary and politicalconstraints. (Setting relatively stringent recovery objectives forstrongly interactive species, for example, will be opposed byindividuals and organizations who perceive negative conse-quences of such actions.) Even so, ignoring the interspecificinteractions of strongly interactive species will further impairthe diversity and resilience of ecosystems that are alreadyunraveling. In a nation and a world where increasing num-bers of species and ecosystems are shoved toward the brinkof annihilation, it is more important than ever that environ-mental policy and management be buttressed by the bestavailable science.

0 notes